Split (1)

train · 521k rows

text stringlengths 254 1.16M
--- title: Induced Inflammatory and Oxidative Markers in Cerebral Microvasculature by Mentally Depressive Stress authors: - Yuequan Zhu - Yazeed Haddad - Ho Jun Yun - Xiaokun Geng - Yuchuan Ding journal: Mediators of Inflammation year: 2023 pmcid: PMC9966573 doi: 10.1155/2023/4206316 license: CC BY 4.0 --- # Induced Inflammatory and Oxidative Markers in Cerebral Microvasculature by Mentally Depressive Stress ## Abstract ### Background Cerebrovascular disease (CVD) is recognized as the leading cause of permanent disability worldwide. Depressive disorders are associated with increased incidence of CVD. The goal of this study was to establish a chronic restraint stress (CRS) model for mice and examine the effect of stress on cerebrovascular inflammation and oxidative stress responses. ### Methods A total of forty 6-week-old male C57BL/6J mice were randomly divided into the CRS and control groups. In the CRS group ($$n = 20$$), mice were placed in a well-ventilated Plexiglas tube for 6 hours per day for 28 consecutive days. On day 29, open field tests (OFT) and sucrose preference tests (SPT) were performed to assess depressive-like behaviors for the two groups ($$n = 10$$/group). Macrophage infiltration into the brain tissue upon stress was analyzed by measuring expression of macrophage marker (CD68) with immunofluorescence in both the CRS and control groups ($$n = 10$$/group). Cerebral microvasculature was isolated from the CRS and controls ($$n = 10$$/group). mRNA and protein expressions of tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), interleukin-6 (IL-6), vascular cell adhesion molecule-1 (VCAM-1), and macrophage chemoattractant protein-1 (MCP-1) in the brain vessels were measured by real-time PCR and Western blot ($$n = 10$$/group). Reactive oxygen species (ROS), hydrogen peroxide (H2O2), and nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) activities were quantified by ELISA to study the oxidative profile of the brain vessels ($$n = 10$$/group). Additionally, mRNA and protein expressions of NOX subunits (gp91phox, p47phox, p67phox, and p22phox) in the cerebrovascular endothelium were analyzed by real-time PCR and Western blot ($$n = 10$$/group). ### Results CRS decreased the total distances ($p \leq 0.05$) and the time spent in the center zone in OFT ($p \leq 0.001$) and sucrose preference test ratio in SPT ($p \leq 0.01$). Positive ratio of CD68+ was increased with CRS in the entire region of the brain ($p \leq 0.001$), reflecting increased macrophage infiltration. CRS increased the expression of inflammatory factors and oxidative stress in the cerebral microvasculature, including TNF-α ($p \leq 0.001$), IL-1β ($p \leq 0.05$), IL-6 ($p \leq 0.05$), VCAM-1 ($p \leq 0.01$), MCP-1 ($p \leq 0.01$), ROS ($p \leq 0.001$), and H2O2 ($p \leq 0.001$). NADPH oxidase (NOX) was activated by CRS ($p \leq 0.01$), and mRNA and protein expressions of NOX subunits (gp91phox, p47phox, p67phox, and p22phox) in brain microvasculature were found to be increased. ### Conclusions To our knowledge, this is the first study to demonstrate that CRS induces depressive stress and causes inflammatory and oxidative stress responses in the brain microvasculature. ## 1. Introduction Cerebrovascular diseases (CVD) are conditions that narrow blood flow and the blood vessels in the brain, with stroke being the most important and devastating clinical manifestation of CVD [1]. CVD is the second most common cause of death and one of the top five causes of morbidity worldwide, causing a huge burden on patients and their families [2, 3]. Common risk factors for CVD include family history, hyperlipidemia, hypertension, and hyperglycemia [4]. Additionally, psychological stress has been recognized as a risk factor, which includes work-related stress, psychological distress, depression, anxiety, and negative personality traits, namely, anger or hostility [5–7]. It has been found that approximately $40\%$ of CVD cases are associated with psychological distress [8]. Psychological factors are related to increased incidence of death due to CVD [9]. A correlation has been found between mental stress and the incidence of myocardial infarction, stroke, and coronary heart disease (CHD) [10]. Numerous clinical guidelines explain that psychosocial support plays a vital role in the rehabilitation for patients suffering from cardiovascular disease [11–13]. Most studies on psychological stress-induced vascular diseases primarily focus on cardiovascular diseases; consequently, the effects and pathophysiology of psychological stress on CVD remains largely unknown [10]. Inflammation and oxidative stress play a vital role in the progression of CVD [14–17]. Psychological stress is found to increase inflammation [18] and oxidative stress responses [19], additionally it appears that cerebrovascular inflammation and oxidative stress may be the foundation of mental stress-induced CVD. Nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (NOX) is a multisubunit enzyme complex which utilizes NADPH to produce superoxide anions and reactive oxygen species (ROS) in the brain [20, 21]. NOX complex contained membrane subunits (gp91phox, p22phox) and cytosolic subunits (p47phox, p67phox), and major NOX subunits have been found in the brain. Upregulation of these subunits is associated with increased NOX activity [20, 22]. ROS from NOX is the basis of early and late inflammatory responses in atherosclerosis of the aorta, and inhibition of NOX activity is found to reduce the progression of atherosclerotic lesions [23, 24]. The primary objective of this study was to study the implications of mental stress on CVD mediated by depressive stress-modulated vascular inflammation and oxidative injuries [25]. This study analyzed the effects of depressive stress on vascular inflammation and oxidative injury by applying a well-established depressive-like model in wild-type C57BL/6J mice. The results of this study could improve the understanding of the risk profile in cerebrovascular disease with an emphasis on psychological stressors. ## 2.1. Subjects The animal research protocol in this study was consistent with the NIH Guide for the Care and Use of Laboratory Animals and approved by the Animal Care and Use Committee of Capital Medical University in Beijing, China. A total of forty 6-week-old adult male C57BL/6J mice (18-22 g) from the Vital River Laboratory Animal Technology Co. Ltd. (Beijing, China) were randomly grouped into the chronic restraint stress (CRS) ($$n = 10$$/group × 2) and control groups ($$n = 10$$/group × 2). All mice were housed in an environment with a 12-hour dark/light cycle, controlled temperature (22 ± 2°C), and humidity. Mice were provided with unlimited access to water and food. Figure 1(a) illustrates the timeline along which CRS, validation using the open field tests (OFT), sucrose preference test (SPT), and tissue collection were performed. ## 2.2. Chronic Restraint Stress (CRS) CRS was performed as described in the preceding study [25]. Twenty mice in the CRS groups were placed in well-ventilated Plexiglas tubes (inner diameter, 6 cm) without food and water for 6 hours per day (from 09:00 to 15:00 every day) for 28 consecutive days. The control animals were handled for 5 minutes in the same manner as the CRS group and kept in their cages without food or water. OFT and SPT were then performed starting from day 29. ## 2.3. Open Field Tests (OFT) Ten mice were randomly selected from each group and examined for OFT. OFT has been reported as the most sensitive test for environmental factors [26]. They were conducted to measure the depressive changes 24 hours after the last stress session (day 29). Mice were placed in an open-field apparatus (50 × 50 × 50 cm) to measure anxiety-like behaviors. Their behaviors were monitored for 10 minutes, using a digital camera. The images were captured by an IBM computer with SMART 3.0 animal behavior recording and analysis system (Panlab, Spain). The running paths, total distances traveled, and the time spent in the center zone (12.5 × 12.5 cm) were calculated. ## 2.4. Sucrose Preference Test (SPT) Ten mice were randomly selected from both groups and tested for SPT. SPT was performed as described in the previous study [25]. Mice were acclimated to $1\%$ sucrose solution (w/v) for 24 hours with two bottles of $1\%$ sucrose solution in each cage on day 30. On the second day, a bottle of $1\%$ sucrose solution and a bottle of water were provided for another 24 hours. The sucrose and water bottles were then placed in randomly assigned sides of the cage. Following a 12-hour period of food and water deprivation, mice were given the sucrose solution and water for 24 hours. Consumption of water and sucrose during the last 24-hour period was measured. Sucrose preference ratio was calculated according to the following formula: [1]Sucrose preference ratio %=sucrose intake g/sucrose intakeg+water g×$100\%$. ## 2.5. Immunofluorescence of CD68 After the behavioral tests, ten mice from each group were selected randomly and a total of twenty mice were used for immunofluorescence assay to study the cellular expression of CD68 [27]. The brain tissues were dehydrated, embedded in paraffin, and prepared in 5 μm sections. The sections were deparaffinized with xylene, rehydrated, and washed with phosphate-buffered saline (PBS) for 5 minutes three times. Antigen retrieval was achieved, using a 10 mM citrate buffer (pH = 6.0) for 15 minutes at 96°C. All sections were blocked in $2\%$ milk for 1 hour and probed with the anti-CD68 antibodies (1: 1000, ab237968, Abcam) overnight at 4°C. After washing, the sections were incubated for 2 hours at room temperature and exposed to anti-rat/anti-rabbit IgG secondary antibody (ZB-2301, 1: 500, Beijing ZhongShan Inc.). Cellular CD68 expression was analyzed using a computer-assisted microscope (Neurolucid™, MicrobrightWled, USA) and an image analysis system. Quantitative analysis was conducted by randomly counting CD68-immunolabeled cells under a light microscope at 400x magnification throughout 24 nonoverlapping regions of the brain tissue (0.025 mm2 each) [26]. The positive cell ratio was calculated for all representative images. ## 2.6. Isolation of Microvasculature of the Entire Brain After the behavioral tests, ten mice from each group were sacrificed and microvasculature of the whole brain was isolated as performed in the previous studies [28, 29]. The brain was removed after cardiac perfusion, homogenized in a 3 mL ice-cold sucrose buffer (0.32 mol/L sucrose, 3 mmol/L HEPES, and pH 7.4), and centrifuged at 4°C for 10 minutes at 1000 g. The supernatant was discarded, and the pellet was resuspended once again in 3 mL of cold sucrose buffer on ice, homogenized, and centrifuged at 4°C for 10 minutes at 1000 g. The sediment was resuspended in a sucrose buffer and centrifuged twice for 30 seconds at 100 g. The pellets were, then, pooled and washed twice with a sucrose buffer and once with phosphate-buffered saline+$0.1\%$ bovine plasma albumin at 200 g. The final pellet was suspended in 1.0 mL of phosphate-buffered saline+$0.1\%$ bovine plasma albumin and centrifuged at 14,000 g, and the precipitate was stored at −70°C. ## 2.7. mRNA Expression in Microvasculature The expression of mRNA of the isolated cerebral vessels was detected, using PCR as previously described [29]. Isolated vessels were homogenized ($$n = 10$$/group), and RNA was isolated with the TRIzol reagent (Invitrogen). Total RNA was converted into cDNA with the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantification of gene expression was performed by the Prism 7500 real-time PCR system (Applied Biosystems). All reactions were performed under the following conditions: 95°C for 15 minutes, 40 cycles of 95°C for 10 seconds, and 60°C for 32 seconds. β-Actin was used as the control gene, and all data were recorded as fold differences from β-actin level. The primers for mouse, including TNF-α, IL-6, IL-1β, VCAM-1, MCP-1, gp91phox, p47phox, p67phox, p22phox, and β-actin, were recorded as shown in Table 1. ## 2.8. Protein Expression in Microvasculature Western blotting was used to detect protein expression in the cerebrovascular endothelium as performed in the past study [29]. Isolated brain vessels were homogenized and processed, and protein concentrations were measured with a bicinchoninic acid protein assay kit (Pierce Biotechnology, Inc.). Primary antibodies to tumor necrosis factor-α (TNF-α) (1: 1000, 11948, Cell Signaling Technology), interleukin-1β (IL-1β) (1: 5000, ab9722, Abcam), interleukin-6 (IL-6) (1: 200, ab7737, Abcam), vascular cell adhesion molecule-1 (VCAM-1) (1: 2000, ab134047, Abcam), macrophage chemoattractant protein-1 (MCP-1) (1: 1000, ab25124, Abcam), gp91phox (1: 1000, ab129068, Abcam), p47phox (1: 1000, sc-17845, Santa Cruz, Inc.), p67phox (1: 1000, sc-374510, Santa Cruz, Inc.), p22phox (1: 1000, sc-271968, Santa Cruz, Inc.), and β-actin (1: 5000, A5060, Sigma-Aldrich) were incubated on the membrane at 4°C overnight. The membranes were washed three times with PBS for 10 minutes each and reincubated with secondary antibody (1: 10,000, Invitrogen) for 1 hour at room temperature. Protein expressions were analyzed with an enhanced chemiluminescence kit (Millipore). Quantification of relative target protein expression was performed by ImageJ 1.42 (National Institutes of Health). ## 2.9. Expression of ROS and H2O2 in Microvasculature by ELISA Mouse enzyme-linked immunosorbent assay (ELISA) kit was used in accordance with the manufacturer's instructions to quantify the expression of reactive oxygen species (ROS) (ml009876-1, MLBIO) and hydrogen peroxide (H2O2) (ml076343, MLBIO) in the isolated brain microvasculature [29]. Isolated microvasculature was homogenized and washed with saline, and the supernatant was collected. ## 2.10. NADPH Oxidase (NOX) Activity NOX activity was assessed as described previously [30]. The brain microvasculature samples containing phenylmethylsulfonyl fluoride and protease inhibitor cocktail (Thermo Fisher Scientific; 20 μL) were added to a 96-well luminescence plate containing 6.25 μmol/L of lucigenin. The reaction was initiated by an addition of nicotinamide adenine dinucleotide phosphate (100 μmol/L). NOX activity was measured by the change in luminescence recorded by the DTX-880 Multimode Detector. ## 2.11. Statistical Analysis Data were expressed as mean values ± SE. The differences between the mean values of two groups were calculated by Student's t-test, followed by post hoc comparisons, using the unpaired comparison test. All analyses were performed by GraphPad Prism v5.0 (GraphPad Software, San Diego, CA). In all cases, $p \leq 0.05$ was considered statistically significant. ## 3.1. CRS Induced Depressive-Like Behaviors Images of running paths of open field test (OFT) after CRS were shown in Figure 1(b). Ten mice were randomly selected from each group, and a total of twenty mice were examined for OFT. CRS decreased the total running distances to 3561 ± 704 cm as compared to the control (4280 ± 785 cm, $p \leq 0.05$; Figure 1(c)). The time spent in the center zone was decreased to 48 ± 15 seconds (8.6 ± $2.5\%$) versus 88 ± 23 seconds from the control group in 10-minute tests (14.8 ± $3.9\%$, $p \leq 0.001$; Figures 1(d) and 1(e)). The sucrose preference test ratio was decreased from 84 ± $1.9\%$ to 74 ± $6.8\%$ ($p \leq 0.01$, Figure 1(f)). These results indicated that CRS induces depressive-like behaviors. ## 3.2. CRS Increased Infiltration of CD68+-Positive Cells Ten mice of each group were randomly chosen, and a total of twenty mice were examined for immunofluorescence of CD68. To delineate the influence of CRS on neuroinflammation, the expression of CD68, or a macrophage marker, was examined with immunofluorescence. As compared to the control group, larger numbers of macrophages were observed in the brain, indicating increased macrophage infiltration into the brain tissue after CRS. The CD68+-positive ratio was increased to 53 ± $9.6\%$ (versus 9.3 ± $2.6\%$ from the control group) ($p \leq 0.001$, Figure 2). ## 3.3. CRS and Microvascular Inflammation Ten mice of each group were examined for microvasculature isolation. The effects of CRS on inflammatory factors in the brain microvasculature were studied with mRNA and protein levels 24 hours after the last session of CRS. mRNA and protein expressions of TNF-α were increased to 2.1 ± 0.2-fold ($p \leq 0.001$, Figure 3(a)) and 4.7 ± 0.1-fold ($p \leq 0.001$, Figure 3(d)), respectively, in the CRS group. Similarly, mRNA expression of IL-1β and IL-6 was upregulated to 2.4 ± 0.7-fold ($p \leq 0.05$, Figure 3(b)) and 3 ± 1-fold ($p \leq 0.05$, Figure 3(c)) after CRS and the protein expression was increased to 2.8 ± 0.3-fold ($p \leq 0.01$, Figure 3(e)) and 8.2 ± 2.2-fold ($p \leq 0.01$, Figure 3(f)), respectively. As shown in Figures 3(g) and 3(i), there was a clear increase in the expression level of VCAM-1 upon CRS. VCAM-1 mRNA expression was increased 3.6 ± 0.6 times ($p \leq 0.01$, Figure 3(g)) and a similar trend was seen in the protein expression ($p \leq 0.01$, Figure 3(i)). MCP-1, an important proinflammatory cytokine in atherosclerosis, showed increased expression of mRNA and protein (5.2 ± 1.4 times, $p \leq 0.01$ and 6.7 ± 1 times, $p \leq 0.001$), respectively, by CRS (Figures 3(h) and 3(j)). ## 3.4. Increased Oxidative Stress and NOX Activity in the Brain Microvasculature by CRS When the same brain tissues were further studied, production of ROS ($p \leq 0.001$, Figure 4(a)) and H2O2 ($p \leq 0.001$, Figure 4(b)) were noted to be significantly increased by stress. The activity of NADPH oxidase (NOX), which facilitates the production of ROS and H2O2, was detected. Increased activity of NOX by CRS is a potential explanation of ROS and H2O2 production ($p \leq 0.01$, Figure 4(c)). ## 3.5. Effects of CRS on Expression of NOX Subunits in Microvasculature As above, the effect of CRS on NOX subunits (gp91phox, p47phox, p67phox, and p22phox) in the brain microvasculature was analyzed. The mRNA expression of gp91phox, p47phox, p67phox, and p22phox increased 2.9 ± 0.4-fold ($p \leq 0.001$, Figure 5(a)), 3 ± 0.5-fold ($p \leq 0.001$, Figure 5(b)), 3.7 ± 1.2-fold ($p \leq 0.001$, Figure 5(e)), and 3.8 ± 1.5-fold ($p \leq 0.001$, Figure 5(f)), respectively. Western blot analyses demonstrated the same trend for these subunits. Comparing to the control group, protein expression of gp91phox, p47phox, p67phox, and p22phox increased 1.4 ± 0.4 ($p \leq 0.05$, Figure 5(c)), 1.8 ± 0.3 ($p \leq 0.05$, Figure 5(d)), 1.5 ± 0.3 ($p \leq 0.05$, Figure 5(g)), and 1.6 ± 0.2 times ($p \leq 0.01$, Figure 5(h)), respectively. ## 4. Discussion Previous studies regarding the effects of psychological stress on vascular disease primarily focused on the cardiovascular system with use of high-fat diet or apolipoprotein E -/- (ApoE-/-) mice [31, 32]. There are limited studies available addressing the effects of psychological stress on cerebrovascular diseases. The purpose of this study was to investigate the effects of psychological stress on cerebrovascular vessels in wild-type mice with normal diet, using a well-established psychological stress model, namely, CRS. It has been acknowledged that mental stress induces microglia activation, especially in the hippocampus and prefrontal cortex [33]. Activated microglia and neuroinflammation have been noted from repeated social defeat stress (RSDS) [1]. CD68 is a transmembrane glycoprotein highly expressed in activated and phagocytic microglia [34]. Increased expression of CD68 reflects the activation of microglia [35, 36]. Stress from chronic social defeat [37] or chronic mild stress [38] are found to increase CD68+ microglia and promote their activation. A recent study also reports that CD68 is enriched in the plaque shoulder and necrotic core of atherosclerotic lesions from the human carotid artery [39]. Animal experiments have proved excess inflammatory responses and oxidative injuries in cardiovascular systems from mental stress. Endothelial inflammation is noted to be the initial step in atherosclerosis [40, 41], and cerebrovascular inflammation has been found to play an important role in intracranial atherosclerosis [29, 42, 43] and CVD [44, 45]. Studies have confirmed the involvement of inflammatory mediators in the early proatherogenic processes, such as upregulation of adhesion molecules and highly activated inflammatory responses on endothelial cells [46]. TNF-α, IL-6, IL-1β, VCAM-1, and MCP-1 play critical roles in regulating vascular inflammation, and increased numbers of these inflammatory factors in brain vessels indicate vascular dysfunction [29, 47]. Mental stress exacerbates vascular inflammation (i.e., TNF-α, CRP, MCP-1, and ICAM-1) and decreases endothelial nitric oxide syntheses [48]. Stress-induced vascular inflammation ultimately results in plaque destabilization in atherosclerosis [49]. Stress-induced hyperactivation of hypothalamic-pituitary-adrenal (HPA) and sympathetic-adrenal-medullary (SAM) axes can influence the vascular endothelium function as well as the recruitment of circulating monocytes and their conversion to foam cells [50, 51]. Lehmann et al. [ 52] describes different gene expressions of the brain endothelial cells (bECs) upon mental stress (i.e., unpredictable chronic mild stress (UCMS)); the results show that stress could cause gene expression changes in inflammation, oxidative stress, growth factor signaling, and wound healing in bECs. Although the findings have not been validated by molecular biology studies, the study utilized the CRS model, one of the most effective and commonly used depressive model [25, 53, 54], and demonstrated increased mRNA and protein expressions of inflammatory factors in the brain vessels, including TNF-α, IL-1β, IL-6, VCAM-1, and MCP-1. A few mechanisms of disease have been proposed to explain the impact of mental stress on cerebrovascular inflammation and oxidative injuries. One study suggests that CRS may decrease the size of cerebral vessels and increases blood-brain barrier (BBB) permeability, which eventually promotes leakage of plasma immunoglobulins across the BBB into perivascular and parenchymal spaces [55]. We have reported that repeated social defeat stress could induce microglia activation [1], which subsequently causes cerebrovascular inflammation [56, 57]. In addition, mental stress is found to elevate levels of circulating proinflammatory cytokines and glucocorticoids, which potentially impacts the brain vessels by altering the central inflammatory state via glucocorticoid feedback to the brain [52]. As this study shows, oxidative stress occurs with cerebrovascular inflammation. Oxidative injuries have been shown to play an important role in pathophysiology of CVD as it contributes to endothelial cell dysfunction, monocyte/macrophage recruitment and activation, and ultimately vascular inflammation [58, 59]. Brooks et al. have found increased ROS and reduced NO bioavailability in the MCA of mice with mental stress, and mental stress (UCMS) could impair endothelial-dependent dilation (EDD) and exaggerate vascular constriction [60]. Although there are many potential sources of ROS, NOX has been considered the main contributor. NOX is a multisubunit enzyme complex formed by cytosolic subunits (p47phox and p67phox) translocated to membrane subunits (p22phox and gp91phox) to function as a catalytic unit. NOX activation has been found in many diseases, such as neurodegenerative diseases, traumatic brain injury (TBI), and stroke. Additionally, inhibition of NOX reduces the development of atherosclerotic plaques in the aorta [23, 61] and improves clinical outcomes from ischemic brain injuries [62], which supports the fact that NOX activation is strongly involved in the pathogenesis of CVD. The findings of this study also demonstrate that NOX activation is involved in the pathogenesis of intracranial atherosclerosis. Mental stress can induce NOX activation in the endothelium of thoracic aorta [63] and increase the expression of NOX-4 in carotid [64, 65]. This study shows the NOX activation in the brain vessels and the expression of NOX subunits in mRNA and protein levels upon mental stress. This finding suggested that NOX activation may play a vital role in vascular inflammation. Dyslipidemia induced by a high-fat diet promotes atherosclerotic lesions. In fact, ApoE-/- mice has been widely used for spontaneous atherosclerosis since ApoE-/- mice accumulate high levels of cholesterol in the blood and thus develop atherosclerosis [66]. This study demonstrates mental stress could not only increase the expression of inflammatory factors but also induce oxidative stress responses and NOX activation in cerebrovascular endothelial cells of wild-type C57BL/6J mice with normal diet. These findings suggest that psychological events may play greater roles than lipid disorders or genetic defects in cerebrovascular diseases than thought previously. ## 5. Conclusion CRS-induced depressive stress instigates the infiltration of macrophage, cerebrovascular inflammation, and oxidative stress responses in the brain. Activation of NOX and the expression of its subunits increase after mental stress. It is very likely that cerebrovascular inflammation and oxidative injuries from psychological stress without high-fat diet serve as an important risk factor for cerebrovascular diseases. ## Data Availability The data used to support the findings in this study are available from the corresponding author upon reasonable request. ## Conflicts of Interest The authors declare that they have no conflicts of interest. ## References 1. Zhu Y., Geng X., Stone C., Guo S., Syed S., Ding Y.. Forkhead box 1(FoxO1) mediates psychological stress-induced neuroinflammation. (2022) 44 483-495. DOI: 10.1080/01616412.2021.2022913 2. Mirzaei M., Truswell A. S., Arnett K., Page A., Taylor R., Leeder S. R.. Cerebrovascular disease in 48 countries: secular trends in mortality 1950-2005. (2012) 83 138-145. DOI: 10.1136/jnnp-2011-300408 3. Leng T., Xiong Z. G.. Treatment for ischemic stroke: from thrombolysis to thrombectomy and remaining challenges. (2019) 5 8-11. DOI: 10.4103/bc.bc_36_18 4. Zhao M., Guan L., Collet J. P., Wang Y.. Relationship between ischemic stroke locations, etiology subtypes, neurological outcomes, and autonomic cardiac function. (2020) 42 630-639. DOI: 10.1080/01616412.2020.1782103 5. Ai H., Fang W., Hu H., Hu X., Lu W.. Antidiabetic drug metformin ameliorates depressive-like behavior in mice with chronic restraint stress via activation of AMP-activated protein kinase. (2020) 11 31-43. DOI: 10.14336/AD.2019.0403 6. Hackett R. A., Steptoe A.. Type 2 diabetes mellitus and psychological stress -- a modifiable risk factor. (2017) 13 547-560. DOI: 10.1038/nrendo.2017.64 7. Al-Kuraishy H. M., Abdulhadi M. H., Hussien N. R., Al-Niemi M. S., Rasheed H. A., Al-Gareeb A. I.. Involvement of orexinergic system in psychiatric and neurodegenerative disorders: a scoping review. (2020) 6 70-80. DOI: 10.4103/bc.bc_42_19 8. Black P. H. G. L., Garbutt L. D.. Stress, inflammation and cardiovascular disease. (2002) 52 1-23. DOI: 10.1016/S0022-3999(01)00302-6 9. Mark Hamer M. K., Emmanuel Stamatakis G.. Psychological distress as a risk factor for death from cerebrovascular disease. (2012) 184 1461-1466. DOI: 10.1503/cmaj.111719 10. Xue Y. T., Tan Q. W., Li P.. Investigating the role of acute mental stress on endothelial dysfunction: a systematic review and meta-analysis. (2015) 104 310-319. DOI: 10.1007/s00392-014-0782-3 11. Kivimaki M., Steptoe A.. Effects of stress on the development and progression of cardiovascular disease. (2018) 15 215-229. DOI: 10.1038/nrcardio.2017.189 12. Pesah E., Supervia M., Turk-Adawi K., Grace S. L.. A review of cardiac rehabilitation delivery around the world. (2017) 60 267-280. DOI: 10.1016/j.pcad.2017.08.007 13. Liu M., Liu J., Zhang L.. An evidence of brain-heart disorder: mental stress-induced myocardial ischemia regulated by inflammatory cytokines. (2020) 42 670-675. DOI: 10.1080/01616412.2020.1783879 14. Nguyen H., Aum D., Mashkouri S.. Growth factor therapy sequesters inflammation in affording neuroprotection in cerebrovascular diseases. (2016) 16 915-926. DOI: 10.1080/14737175.2016.1184086 15. Jin R., Liu L., Zhang S., Nanda A., Li G.. Role of inflammation and its mediators in acute ischemic stroke. (2013) 6 834-851. DOI: 10.1007/s12265-013-9508-6 16. Sivandzade F., Bhalerao A., Cucullo L.. Cerebrovascular and neurological disorders: protective role of NRF2. (2019) 20 p. 3433. DOI: 10.3390/ijms20143433 17. Fan X., Elkin K., Shi Y.. Schisandrin B improves cerebral ischemia and reduces reperfusion injury in rats through TLR4/NF-. (2020) 42 693-702. DOI: 10.1080/01616412.2020.1782079 18. Tetsuka S.. Depression and dementia in older adults: a neuropsychological review. (2021) 12 1920-1934. DOI: 10.14336/AD.2021.0526 19. Ghezzi P. F. L., Boraschi D., Cuadrado A.. Oxidative stress and inflammation induced by environmental and psychological stressors: a biomarker perspective. (2018) 28 852-872. DOI: 10.1089/ars.2017.7147 20. Rastogi R., Geng X., Li F., Ding Y.. NOX activation by subunit interaction and underlying mechanisms in disease. (2016) 10 p. 301. DOI: 10.3389/fncel.2016.00301 21. Li W., Yang S.. Targeting oxidative stress for the treatment of ischemic stroke: upstream and downstream therapeutic strategies. (2016) 2 153-163. DOI: 10.4103/2394-8108.195279 22. Shen J., Rastogi R., Geng X., Ding Y.. Nicotinamide adenine dinucleotide phosphate oxidase activation and neuronal death after ischemic stroke. (2019) 14 948-953. DOI: 10.4103/1673-5374.250568 23. Manea S. A., Vlad M. L., Fenyo I. M.. Pharmacological inhibition of histone deacetylase reduces NADPH oxidase expression, oxidative stress and the progression of atherosclerotic lesions in hypercholesterolemic apolipoprotein E-deficient mice; potential implications for human atherosclerosis. (2020) 28, article 101338. DOI: 10.1016/j.redox.2019.101338 24. Buchmann G. K., Schurmann C., Warwick T.. Deletion of NoxO1 limits atherosclerosis development in female mice. (2020) 37, article 101713. DOI: 10.1016/j.redox.2020.101713 25. Zhu Y., Klomparens E. A., Guo S., Geng X.. Neuroinflammation caused by mental stress: the effect of chronic restraint stress and acute repeated social defeat stress in mice. (2019) 41 762-769. DOI: 10.1080/01616412.2019.1615670 26. Sturman O., Germain P. L., Bohacek J.. Exploratory rearing: a context- and stress-sensitive behavior recorded in the open-field test. (2018) 21 443-452. DOI: 10.1080/10253890.2018.1438405 27. Jiang Q., Geng X., Warren J.. Hypoxia Inducible Factor-1. (2020) 448 126-139. DOI: 10.1016/j.neuroscience.2020.09.036 28. Shen J., Rastogi R., Guan L.. Omega-3 fatty acid supplement reduces activation of NADPH oxidase in intracranial atherosclerosis stenosis. (2018) 40 499-507. DOI: 10.1080/01616412.2018.1451290 29. Guan L., Geng X., Stone C.. PM2.5 exposure induces systemic inflammation and oxidative stress in an intracranial atherosclerosis rat model. (2019) 34 530-538. DOI: 10.1002/tox.22707 30. Guo S., Cosky E., Li F.. An inhibitory and beneficial effect of chlorpromazine and promethazine (C + P) on hyperglycolysis through HIF-1. (2021) 1763, article 147463. DOI: 10.1016/j.brainres.2021.147463 31. Paine N. J., Ring C., Aldred S., Bosch J. A., Wadley A. J., Veldhuijzen van Zanten J. J. C. S.. Eccentric-exercise induced inflammation attenuates the vascular responses to mental stress. (2013) 30 133-142. DOI: 10.1016/j.bbi.2013.01.082 32. Paine N. J., Ring C., Bosch J. A., Drayson M. T., Aldred S., Veldhuijzen van Zanten J. J. C. S.. Vaccine-induced inflammation attenuates the vascular responses to mental stress. (2014) 93 340-348. DOI: 10.1016/j.ijpsycho.2014.06.016 33. Calcia M. A., Bonsall D. R., Bloomfield P. S., Selvaraj S., Barichello T., Howes O. D.. Stress and neuroinflammation: a systematic review of the effects of stress on microglia and the implications for mental illness. (2016) 233 1637-1650. DOI: 10.1007/s00213-016-4218-9 34. Xiao K., Luo Y., Liang X.. Beneficial effects of running exercise on hippocampal microglia and neuroinflammation in chronic unpredictable stress-induced depression model rats. (2021) 11 p. 461. DOI: 10.1038/s41398-021-01571-9 35. Hendrickx D. A. E., van Eden C. G., Schuurman K. G., Hamann J., Huitinga I.. Staining of HLA-DR, Iba1 and CD68 in human microglia reveals partially overlapping expression depending on cellular morphology and pathology. (2017) 309 12-22. DOI: 10.1016/j.jneuroim.2017.04.007 36. Xu P., Hong Y., Xie Y.. TREM-1 exacerbates neuroinflammatory injury via NLRP3 inflammasome-mediated pyroptosis in experimental subarachnoid hemorrhage. (2021) 12 643-659. DOI: 10.1007/s12975-020-00840-x 37. Lehmann M. L., Cooper H. A., Maric D., Herkenham M.. Social defeat induces depressive-like states and microglial activation without involvement of peripheral macrophages. (2016) 13 p. 224. DOI: 10.1186/s12974-016-0672-x 38. Feng X., Fan Y., Chung C. Y.. Mefenamic acid can attenuate depressive symptoms by suppressing microglia activation induced upon chronic stress. (2020) 1740, article 146846. DOI: 10.1016/j.brainres.2020.146846 39. Rykaczewska U., Zhao Q., Saliba-Gustafsson P.. Plaque evaluation by ultrasound and transcriptomics reveals BCLAF1 as a regulator of smooth muscle cell lipid transdifferentiation in atherosclerosis. (2022) 42 659-676. DOI: 10.1161/ATVBAHA.121.317018 40. Yamauchi T. Y. T., Sasaki T., Matsumoto S., Takahashi M., Suka M., Yanagisawa H.. Cerebrovascular/cardiovascular diseases and mental disorders due to overwork and work-related stress among local public employees in Japan. (2018) 56 85-91. DOI: 10.2486/indhealth.2017-0131 41. Steptoe A., Kivimaki M.. Stress and cardiovascular disease: an update on current knowledge. (2013) 34 337-354. DOI: 10.1146/annurev-publhealth-031912-114452 42. Shen J., Hafeez A., Stevenson J.. Omega-3 fatty acid supplement prevents development of intracranial atherosclerosis. (2016) 334 226-235. DOI: 10.1016/j.neuroscience.2016.08.013 43. Kelley R. E., Bir S. C.. Carotid atherosclerotic disease: a systematic review of pathogenesis and management. (2022) 8 127-136. DOI: 10.4103/bc.bc_36_22 44. Mun K. T., Hinman J. D.. Inflammation and the link to vascular brain health: timing is brain. (2022) 53 427-436. DOI: 10.1161/STROKEAHA.121.032613 45. Wolf V. L., Ergul A.. Progress and challenges in preclinical stroke recovery research. (2021) 7 230-240. DOI: 10.4103/bc.bc_33_21 46. Ridker P. M.. Anticytokine agents: targeting interleukin signaling pathways for the treatment of atherothrombosis. (2019) 124 437-450. DOI: 10.1161/CIRCRESAHA.118.313129 47. Murphy J. M., Jeong K., Rodriguez Y. A. R., Kim J. H., Ahn E. Y. E., Lim S. T. S.. FAK and Pyk2 activity promote TNF-. (2019) 9 p. 7617. DOI: 10.1038/s41598-019-44098-2 48. Lu X. T., Liu Y. F., Zhao L.. Chronic psychological stress induces vascular inflammation in rabbits. (2013) 16 87-98. DOI: 10.3109/10253890.2012.676696 49. Hinterdobler J., Schott S., Jin H.. Acute mental stress drives vascular inflammation and promotes plaque destabilization in mouse atherosclerosis. (2021) 42 4077-4088. DOI: 10.1093/eurheartj/ehab371 50. Gu H. F., Tang C. K., Yang Y. Z.. Psychological stress, immune response, and atherosclerosis. (2012) 223 69-77. DOI: 10.1016/j.atherosclerosis.2012.01.021 51. Huang C. J., Webb H. E., Zourdos M. C., Acevedo E. O.. Cardiovascular reactivity, stress, and physical activity. (2013) 4 p. 314. DOI: 10.3389/fphys.2013.00314 52. Lehmann M. L., Poffenberger C. N., Elkahloun A. G., Herkenham M.. Analysis of cerebrovascular dysfunction caused by chronic social defeat in mice. (2020) 88 735-747. DOI: 10.1016/j.bbi.2020.05.030 53. Schoenfeld T. J., McCausland H. C., Morris H. D., Padmanaban V., Cameron H. A.. Stress and loss of adult neurogenesis differentially reduce hippocampal volume. (2017) 82 914-923. DOI: 10.1016/j.biopsych.2017.05.013 54. van der Kooij M. A., Fantin M., Kraev I.. Impaired hippocampal neuroligin-2 function by chronic stress or synthetic peptide treatment is linked to social deficits and increased aggression. (2014) 39 1148-1158. DOI: 10.1038/npp.2013.315 55. Lee S., Kang B. M., Kim J. H.. Real-time _in vivo_ two-photon imaging study reveals decreased cerebro- vascular volume and increased blood-brain barrier permeability in chronically stressed mice. (2018) 8. DOI: 10.1038/s41598-018-30875-y 56. Zhou T., Liu Y., Yang Z.. IL-17 signaling induces iNOS+ microglia activation in retinal vascular diseases. (2021) 69 2644-2657. DOI: 10.1002/glia.24063 57. Park E. S., Kim S., Yao D. C.. Soluble endoglin stimulates inflammatory and angiogenic responses in microglia that are associated with endothelial dysfunction. (2022) 23 p. 1225. DOI: 10.3390/ijms23031225 58. El Hadri K., Smith R., Duplus E., El Amri C.. Inflammation, oxidative stress, senescence in atherosclerosis: thioredoxine-1 as an emerging therapeutic target. (2022) 23 p. 77. DOI: 10.3390/ijms23010077 59. Hao Y., Xiong R., Gong X.. Memantine, NMDA receptor antagonist, attenuates ox-LDL-induced inflammation and oxidative stress via activation of BDNF/TrkB signaling pathway in HUVECs. (2021) 44 659-670. DOI: 10.1007/s10753-020-01365-z 60. Brooks S., Branyan K. W., DeVallance E.. Psychological stress-induced cerebrovascular dysfunction: the role of metabolic syndrome and exercise. (2018) 103 761-776. DOI: 10.1113/EP086892 61. Di Marco E. G. S., Chew P., Koulis C.. Pharmacological inhibition of NOX reduces atherosclerotic lesions, vascular ROS and immune-inflammatory responses in diabetic Apoe. (2014) 57 633-642. DOI: 10.1007/s00125-013-3118-3 62. Kim J. Y., Park J., Lee J. E., Yenari M. A.. NOX inhibitors - a promising avenue for ischemic stroke. (2017) 26 195-205. DOI: 10.5607/en.2017.26.4.195 63. Marques B. V. D., Higashi C. M., Novi D. R. B. D. S.. Intrauterine and lactation exposure to fluoxetine blunted in the offspring the aortic adaptive response induced by acute restraint stress. (2017) 813 147-152. DOI: 10.1016/j.ejphar.2017.08.018 64. Moreira J. D., Pernomian L., Gomes M. S.. Acute restraint stress increases carotid reactivity in type-I diabetic rats by enhancing Nox4/NADPH oxidase functionality. (2015) 765 503-516. DOI: 10.1016/j.ejphar.2015.09.020 65. Coco H., Pernomian L., Pereira P. C.. Chronic restraint stress increases angiotensin II potency in the rat carotid: role of cyclooxygenases and reactive oxygen species. (2016) 69 52-65. DOI: 10.1111/jphp.12659 66. Gaudreault N., Kumar N., Posada J. M.. ApoE suppresses atherosclerosis by reducing lipid accumulation in circulating monocytes and the expression of inflammatory molecules on monocytes and vascular endothelium. (2012) 32 264-272. DOI: 10.1161/ATVBAHA.111.238964
--- title: Integrated Serum Metabolome and Gut Microbiome to Decipher Chicken Amino Acid Improvements Induced by Medium-Chain Monoglycerides authors: - Tao Liu - Shengyue Ruan - Qiufen Mo - Minjie Zhao - Jing Wang - Zhangying Ye - Li Chen - Fengqin Feng journal: Metabolites year: 2023 pmcid: PMC9966585 doi: 10.3390/metabo13020208 license: CC BY 4.0 --- # Integrated Serum Metabolome and Gut Microbiome to Decipher Chicken Amino Acid Improvements Induced by Medium-Chain Monoglycerides ## Abstract Chicken muscle yield and amino acid composition improvements with medium-chain monoglyceride (MG) supplementation were reported by previous studies, but the underlying mechanism was uncertain. This study aimed to decipher chicken amino acid improvements induced by medium-chain monoglycerides in the views of metabolomics, gene expression, and the gut microbiome. Newly hatched chicks (12,000 chicks) were weighed and randomly divided into two flocks, each with six replicates (1000 chicks per replicate), and fed a basal diet (the control group, CON) or a basal diet enriched with 300 mg/kg MG (the treated group, MG). Results demonstrated that MGs significantly increased the chicken flavor and essential and total amino acids. The serum amino acids and derivatives (betaine, l-leucine, l-glutamine, 1-methylhistide), as well as amino acid metabolism pathways in chickens, were enhanced by MG supplementation. Gene expression analysis exhibited that dietary MGs could improve muscle protein synthesis and cell growth via the mTOR/S6K1 pathway. Dietary MGs enhanced the cecal amino acid metabolism by selectively increasing the proportion of genera Lachnospiraceae_NK4A136_group and Bacteroides. Conclusively, the present study demonstrated that dietary MGs improved chicken amino acid composition via increasing both gut amino acid utilization and muscle amino acid deposition. ## 1. Introduction In the last two decades, metabolomics has been widely used to study low molecular weight metabolites (<1000 daltons), providing us with information about metabolite profiles and integrated metabolic pathways of farmed animals in response to nutritional intervention [1]. Alterations in the circulating metabolites profile can partially reflect the influences of nutritional treatments on energy and nutrient metabolism, among them, some key metabolites are identified to be closely related to animal performance and meat quality. For instance, l-glutamine is the richest amino acid in both the bloodstream and the body’s free amino acid pool. Broilers fed diets containing l-glutamine present improved productive performance and better meat quality [2,3,4]. As an essential amino acid, higher plasma leucine can significantly stimulate muscle protein synthesis by enhancing translation initiation factor activation in neonatal pigs [5,6]. Higher levels of betaine exert a positive influence on both animal performance (average daily gain and feed nutrients utilization) and carcass yield [7]. Currently, mass spectrometry (MS)-based and non-MS-based techniques such as nuclear magnetic resonance (NMR) are the two main types of platforms that have been widely used for metabolomic studies [8]. MS-based metabolomics is widely used in tissues, biofluids, or cells due to the lipophilicity and polarity of metabolites. Medium-chain monoglycerides (MGs) are a group of saturated 8–12 carbon fatty acids monoglycerides containing glycerol monolaurate (GML), glycerol monodecanoate (GMD), and glycerol monocaprylin (GMC), showing broad antibacterial spectrum and strong synergistic antimicrobial activity in vivo and in vitro [9,10,11,12]. Recently, numerous studies demonstrated that dietary supplementation with MGs modulates the structure and function of gut microbiota and exhibits a close relationship with chicken production and quality. Junhong declares that feed additive GML can promote Lachnospiraceae, Christensenellaceae, and Ruminococcaceae colonization in the chicken cecum, thereby increasing the short-chain fatty acids (SCFAs) content and improving the chicken body weight [13]. Kong et al. [ 14] state that dietary GML selectively increases the abundance of Lachnospiraceae, Faecalibacterium, and Bacteroides, which is found to be positively related to a lowered feed conversion rate (FCR) in broiler chickens [15]. Chickens fed diets containing a GML and GMD mixture are selectively enriched with an unclassified genus of the Lachnospiraceae family, Bifidobacteriaceae, and Bacteroides [16], along with improved average body weight, muscle amino acid, and carcass yield [17]. Intestinal microecology has proven to be the main functional target of dietary GML ameliorating metabolic syndrome and obesity in mice, as GML-mediated metabolic improvements are all abolished after the addition of antibiotics [18,19]. In addition, 16S rRNA gene sequencing combined with MS-based metabolomics reveals that dietary butyrate glycerides modulate lipid metabolism and energy homeostasis in broilers through increasing Bifidobacterium and bacterial metabolites [20]. Dietary lauric acid modulates gut microbiota (Faecalibacterium, Ruminococcaceae_UCG-014) and serum metabolites (lysophosphatidylcholines and phosphatidylcholines) to enhance immune functions, suppress inflammation, and modulate lipid metabolism of broilers [21]. These findings offer us an effective way to directly investigate the relationship between dietary intervention and meat quality. Moreover, our previous study reported that dietary MG-induced muscle amino acid changes were closely related to the increased relative abundance of an unclassified genus of Lachnospiraceae, Bifidobacteriaceae, Bacteroides, and bacterial amino acid metabolism gene in experimentally reared broilers [16,17]. In another study, the content and metabolism pathways of muscle amino acids were significantly improved with dietary MGs in large-scaled broilers [9]. However, the influences and linkages of related serum metabolism and key gut microbiota that were responsible for muscle amino acid and yield were not uncovered. The present study was conducted to integrally reveal chicken amino acid improvements induced by medium-chain monoglycerides using combined approaches of 16S rRNA sequencing, quantitative PCR assays, and MS-based metabolomics in large-scaled broilers. ## 2.1. Animal Management and Diets Newly hatched chicks (male Chinese indigenous, yellow-feathered broiler breeds) were weighed and randomly divided into two flocks, each with six replicates (1000 chicks per replicate). Chicks were raised on the ground and fed and drunk ad libitum on a modern farm with 24 h constant lighting for a 70-day experiment. The broiler chicks were fed a basal diet (the control group, CON) or a basal diet enriched with 300 mg/kg MG (the treated group, MG). The basal diet was formulated referring to previous work in Table S1 [17]. MG is a mixture of GML and GMD that is produced by Hangzhou Longyu Biotechnology Co., Ltd. (Zhejiang, China). Feed consumption and the number and corresponding body weight of dead birds were recorded daily for each replicate to calculate and adjust the feed conversion rate of broilers throughout the entire experiment. All of the chickens were weighed with replicates at 0 and 70 days of age. Body weight (BW), average daily gain (ADG), average feed intake (FI), and feed conversion rate (FCR) were calculated subsequently. ## 2.2. Sample Collection After fasting for 12 h, two randomly selected chickens out of a replicate were sacrificed in the morning at 71 days of age. Blood was drawn from the wing veins and the serum was obtained using centrifugation (2000× g, 15 min, 4 °C). Pectoralis major and cecal digesta was promptly isolated and frozen [22]. All samples were kept at −80 °C before analysis. ## 2.3. Muscle Amino Acid Determination The method of muscle amino acid measurement refers to a previous study [17]. The dried muscle samples were dissolved in 6M HCl and digested at 150 °C for 2 h under a pure nitrogen atmosphere. Then, the digested samples were derived with reagent (ethanol: water: triethylamine: phenyl-isothiocyanate, 7:1:1:1, v/v/v/v). The amino acid composition was measured using HPLC (Waters e2695; Waters Corporation, Milford, MA, USA) equipped with an Ultimate AQ-C18 column. ## 2.4.1. Extraction of Serum Metabolites Serum (100 μL) from each chicken ($$n = 6$$) was added into 2.0 mL tubes containing a cold methanol/acetonitrile solution (1:1, v/v, 400 μL). After 60 s vortex, all tubes were left standing at −20 °C for 1 h; thereafter, the supernatant was separated (15,000× g, 4 °C, 20 min) and freeze-dried. The dried samples were separately redissolved in an acetonitrile/water solution (1:1, v/v, 100 μL) before analysis. ## 2.4.2. Mass Spectrometry Analysis The serum metabolites profile was separated using an ACQUITY UPLC HSS T3 column (1.8 µm, 2.1 mm × 100 mm, Waters, Milford, MA, USA) and an ACQUITY UPLC BEH Amide column (1.7 µm, 2.1 mm × 100 mm, Waters, Milford, MA, USA) on liquid chromatography (1290 Infinity UHPLC, Agilent, Santa Clara, CA, USA). The mobile phase consisted of A (25 mM ammonium hydroxide and ammonium acetate in water) and B ($100\%$ acetonitrile). The gradient elution procedure was: 0–1 min, $95\%$ B; 1–14 min, B linearly reduced to $65\%$; 14–16 min, B linearly reduced to $40\%$; 16–18 min, $40\%$ B; 18–18.1 min, B linearly increased to $95\%$; 18.1–23 min, $95\%$ B. The mass spectrometric data were acquired with a time-of-flight mass spectrometer (Triple TOF $\frac{5600}{6600}$, AB SCIEX, Framingham, MA, USA). The Electrospray ionization source conditions and operating conditions referred to a previous study [9]. The raw MS data processing referred to previous works [9,23]. Principal component analysis (PCA) and orthogonal partial least-squares discriminant analysis (OPLS-DA) were performed to reveal the metabolic alterations between groups. The VIP value of each metabolite was used to reflect its contribution to the classification. The metabolites with VIP values larger than 1 were further applied to measure their significance using the Student’s t-test. The significantly different metabolites were blasted against with the KEGG database to retrieve the KOs and corresponding pathways. The significantly different pathways were screened based on Fisher’s exact test. ## 2.5. Quantitative PCR Analysis The mRNAs of pectoralis major were extracted using reagent Kits following the instructions (Invitrogen, Waltham, MA, USA). The total RNA of each sample was reverse-transcribed using the PrimeScriptTMRT Kit (Takara, Kyoto, Japan). Gene expression was analyzed in the Roche Light Cycler 480 system (Indianapolis, IN, USA) using Takara TB Green Premix Ex TaqTM (Kyoto, Japan). The reference gene was Glyceraldehyde-3-phosphate dehydrogenase, and the results were calculated relative to the CON group using the 2–ΔΔCt method [24]. Primer sequences of the genes were listed in Table S2. ## 2.6. Gut Microbiome The DNA of cecal digesta was extracted using QIAGEN Kits referring to instructions (Venlo, The Netherlands). After concentration determination, the DNA content of each sample was diluted to 1 ng/μL. Then, the V3−V4 region of the 16S rRNA gene was amplified with PCR using primers 338F and 806R. The amplicons were purified and quantified before they were sequenced (HiSeq2500 platform, San Diego, CA, USA). Clean reads were obtained after assembly (FLASH software v1.2.11) and quality control (QIIME v1.9.1) of raw paired-end reads. Then the operational taxonomic units (sequences ≥ $97\%$ similarity) were picked, and taxonomic information was annotated using the UPARSE v 7.0.1009 and Silva$\frac{128}{16}$S_bacteria database [25]. Microbial composition, α-diversity (Shannon, Simpson, Chao, and Ace), and β-diversity (principal coordinates analysis (PCoA) based on unweighted unifrac) were analyzed. Predictive bacterial function profiling was performed using PICRUSt (http://picrust.github.io/picrust). The changes in Orthologue (KOs) and pathways were identified using the KEGG database. The characteristic taxa and function were revealed with the LEfSe (http://huttenhower.sph.harvard.edu/galaxy/) algorithm and showed with STAMP (version 2.1.3) [16]. The raw data have been uploaded to NCBI (PRJNA638502). ## 2.7. Statistical Analysis The student’s t-test and Wilcoxon rank-sum test were used to screen for significant differences. Differences were expressed as * $p \leq 0.05$, $p \leq 0.01$, and * $p \leq 0.001.$ A p-value greater than 0.05 but less than 0.1 was discussed as tendencies. ## 3.1. Chicken Productive Performance Compared to the CON group, the BW, FI, and ADG were 74.26 g, 110.11 g, and 1.05 g higher in the MG group (Table 1), respectively, while the FCR decreased by $1.08\%$. ## 3.2. Dietary MG Alters Muscle Amino Acids Content The muscle glutamic acid, proline, serine, leucine, glycine, phenylalanine, alanine, tyrosine, methionine, and threonine content were increased with MG supplementation ($p \leq 0.05$, Table S3). The dietary MG increased ($p \leq 0.05$) the total and essential amino acids by $7.14\%$ and $7.40\%$, respectively. ## 3.3. Chicken Serum Metabolome A total of 132 serum metabolites were identified out of 9157 acquired ion peaks. Derivatives [37], choline [13], pyridines and derivatives [10], purines and derivations [8], carbohydrates and conjugates [7], and dipeptides were the major chicken serum metabolites (6, Figure 1A). The principal component analysis (PCA) scores plot showed obvious clustering and drift according to MG treatment with a slight overlap (Figure 1B). The OPLS-DA scores plot showed intensive clustering in the MG group and a distinct shift away from the CON group (Figure 1C), indicating that the chicken serum metabolic profiles differed a lot. Moreover, the R2 and Q2 values in the permutation test were 0.999 and 0.694, respectively, showing that the model had credible cumulative interpretation and predictive ability. Twelve differential serum metabolites were selected using a VIP value over 1 and a p-value less than 0.1 in this model including four down-regulated metabolites and eight up-regulated metabolites that were responsible for the serum profile differences (Figure 1D). A total of thirteen metabolic pathways were notably influenced by MG addition (Figure 2A), including five lipid metabolism pathways (linoleic acid metabolism, α-linolenic acid metabolism, glycerophospholipid metabolism, fatty acid biosynthesis, arachidonic acid metabolism), three amino acid metabolism pathways (d-glutamine and d-glutamate metabolism, arginine biosynthesis, valine, leucine, and isoleucine biosynthesis), one energy metabolism pathway (nitrogen metabolism) and another four metabolism pathways. Pathways involving in the chicken serum metabolite differences by dietary MG were summarized and sketched in Figure 2B according to the KEGG database, mainly involving amino acid metabolism pathways (valine, leucine and isoleucine biosynthesis, glycine, serine and threonine metabolism, histidine metabolism, valine, leucine, and isoleucine degradation, d-glutamine and d-glutamate metabolism) and lipid metabolism pathways (fatty acid degradation, sphingolipid metabolism, fatty acid elongation), as well as seven key differential serum metabolites related to MG supplementation (betaine, l-leucine, l-glutamine, 1-methylhistide, sphingomyelin (d18:$\frac{1}{18}$:0), and PC (16:$\frac{0}{16}$:0)). ## 3.4. Dietary MG Affects the mRNA Expression of Muscle Growth Regulation The relative expression of S6K1, MEF2C, and MEF2D increased by $83.52\%$, $336.48\%$, and $353.54\%$ in the MG group, respectively ($p \leq 0.05$, Figure 3B,J,K). ## 3.5.1. Changes in Chicken Microbial Diversity and Structure The current study obtained 1119 OTUs, and 119 and 113 OTUs only existed in the CON and MG groups (Figure 4A), respectively. The dietary MG did not exert changes on the ACE, Chao, Simpson, and Shannon index (Figure 4B). However, the principal coordinates analysis (PCoA) plot showed intensive clusters in the MG group and distinct drift away from the CON group without any overlaps. Moreover, the distance between the MG and CON group in the PC1 coordinates represented significant differences with an explanation of $22.57\%$ observed total variance ($p \leq 0.001$, Figure 4C), suggesting a distinct variation of the gut microbiota community. Taxonomic profiling demonstrated that cecal flora was dominated by Bacteroidetes, followed by Firmicutes, Tenericutes, Proteobacteria, Actinobacteria, and Patescibacteria (Figure 4D), and the relative content of Patescibacteria showed significant differences (Figure 4F). Bacteroidaceae, Ruminococcaceae, Rikenellaceae, Lachnospiraceae, Muribaculaceae, Tannerellaceae, Prevotellaceae, an unclassified family of Bacteroidales order, and Clostridiales_vadinBB60_group were the main families of phyla Bacteroidetes and Firmicutes (Figure 4E). At the family level (Figure 4E), increased Bacteroidaceae and decreased Rikenellaceae were found in the MG group. Dietary MG increased ($p \leq 0.05$) the abundance of Bacteroides, a no rank genus of the Ruminococcaceae family, Lachnospiraceae_NK4A136_group, Oscillospira, and an unclassified genus of Burkholdderiaceae and reduced ($p \leq 0.05$) the content of Rikenellaceae_RC9_gut_group, Alloprevotella, an unclassified genus of Barnesiellaceae, Butyricicoccus, a no rank genus of the Saccharimonadales order, Phascolarctobacterium, Olsenella, Prevotellaceae_NK3B31_group, an unclassified genus of Barnesiellaceae, Oribacterium, Collinsella, [Eubacterium]_hallii_group, Anaeroplasma, CHKCI002, CAG_56, Faecalicoccus, Lachnospiraceae_FCS020_group, Candidatus_Saccharimonas, a no rank genus of Victivallaceae, and Succinatimonas (Figure 5A). LEfSe revealed 27 specific bacterial genera (Figure 5B). Rikenellaceae_RC9_gut_group, Alloprevotella, a no rank genus of Victivallaceae, Succinatimonas, an unclassified genus of Barnesiellaceae, Lachnospiraceae_FCS020_group, CAG_56, a no rank genus of the Saccharimonadales order, Phascolarctobacterium, Prevotellaceae_NK3B31_group, Candidatus_Saccharimonas, [Ruminococcus]_gauvreauii_group, Collinsella, Butyricicoccus, Faecalicoccus, Oribacterium, CHKCI002, [Eubacterium]_hallii_group, Anaeroplasma, an unclassified genus of Barnesiellaceae, and Olsenella were rich in the CON group. ( Figure 5B). Bacteroides, an unclassified genus of Burkholdderiaceae, a no rank genus of the Ruminococcaceae family, Lachnospiraceae_NK4A136_group, Enorma, and Oscillospira were abundant in the MG group (Figure 5B). ## 3.5.2. Changes in Chicken Gut Microbiota Community and Function The predictive functions of the microbial genes belonging to carbohydrate and amino acid metabolism were significantly changed with dietary MG (Figure 5C). The dietary MG significantly increased the gene abundances of four carbohydrate metabolism pathways (amino sugar and nucleotide sugar metabolism ($$p \leq 0.015$$), the pentose phosphate pathway ($$p \leq 0.0003$$), pentose and glucuronate interconversions ($$p \leq 0.016$$), and C5-branched dibasic acid metabolism ($$p \leq 0.0002$$)) and three amino acid metabolism pathways (tyrosine metabolism ($$p \leq 0.0006$$), glutathione metabolism ($$p \leq 0.0009$$), and glycine, serine, and threonine metabolism ($$p \leq 0.018$$)). The dietary MG also depleted the microbial gene abundance of carbohydrate metabolism (starch and sucrose metabolism ($$p \leq 0.0005$$), citrate cycle (TCA cycle, ($$p \leq 0.012$$), and galactose metabolism ($$p \leq 0.0001$$)) and amino acid metabolism (lysine biosynthesis ($$p \leq 0.016$$) and phenylalanine, tyrosine, and tryptophan biosynthesis ($$p \leq 0.0001$$)). ## 3.5.3. Correlation of Chicken Gut Microbiota Composition with Amino Acid Content Spearman’s correlation analysis was conducted to reveal the relationship between the chicken gut microbiota composition (at the genus level) and amino acid content (Figure 5D). A total of 16 genera out of the top 50 were closely related to chicken amino acids. Among them, 11 genera were positively ($p \leq 0.05$) related to chicken amino acid content, but 4 genera showed a notable negative relationship ($p \leq 0.05$). Moreover, Lachnospiraceae_NK4A136_group, the significantly increased genera in the MG group, were positively ($p \leq 0.05$) related to chicken amino acid content. The increased Bacteroides represented a notably positive relationship with muscle amino acid, but no significant differences were observed. However, the decreased Phascolarctobacterium were negatively ($p \leq 0.05$) correlated with chicken amino acid content. The results suggested the alterations in chicken amino acids were closely correlated to the changes in gut microbiota composition with thee dietary MG. ## 4. Discussion Muscle protein is the main component in raw poultry meat (18.4–$23.4\%$), playing a key role in poultry meat quality evaluation [26]. Amino acids are the basic building blocks of muscle proteins, and their composition and relative quantity are important indexes for meat quality evaluation [17]. Dietary MGs significantly increased the content of 10 muscle amino acids including both flavor taste and essential amino acids in the current study, enhancing the umami and sweet taste, as well as the nutritional value of chicken. These results were consistent with a previous study where muscle aspartic acid, serine, proline, glutamic acid, and arginine content were increased in chickens fed 300 mg/kg MG [17]. The inclusion of 300 mg/kg GML was also reported to increase chicken muscle serine, glycine, arginine, histidine, tyrosine, threonine, methionine, phenylalanine, and lysine content [27]. Consistent with previous studies where both the body weight and carcass yield were improved with MG supplementation [17,27], the dietary MG notably increased the chicken body weight and daily weight gain in the current study. Similarly, a couple of recently published studies reported that dietary medium-chain monoglycerides increased chicken body weight in both yellow-feathered and white-feathered broilers [11,13,28,29]. These findings indicated that the increased muscle amino acid content induced with MG supplementation did not only improve the chicken meat quality, but also the main contributions to improved chicken muscle mass and body weight. In this research, betaine, l-leucine, l-glutamine, 1-methylhistide, sphingomyelin (d18:$\frac{1}{18}$:0), and PC (16:$\frac{0}{16}$:0) were screened as chicken serum metabolites in response to dietary MG supplementation. Betaine is a trimethyl derivative of glycine and has amino acid properties, and it is widely distributed in most organisms and could participate in protein metabolism by donating its methyl group [7,30,31]. Betaine was reported to be involved in regulating chicken growth, nutrient metabolism, and antioxidant balance, which was recognized as a “carcass modifier” due to its lipotropic and growth-promoting effects [32]. Chen et al. [ 30,32] stated that dietary betaine improved chicken productive performance and lean meat yield. The higher serum betaine level provides us with biochemical evidence of dietary MGs improving the chicken growth performance and carcass yield. l-Glutamine is the richest amino acid in both the bloodstream and the body’s free amino acid pool [2,3], which plays a vital role in body ammonia balance due to its two ammonia groups that can accept excess ammonia and release it when needed to form biologically important molecules such as amino acids, nucleotides, protein, etc. Broilers receiving diets containing l-glutamine showed better performance and meat quality [2,3,4]. The up-regulated serum l-glutamine in the current study indicated that MG supplementation may improve the muscle amino acid content and meat quality of broilers by boosting the amino acid utilization and deposition [9]. Escobar et al. [ 5] and Duan et al. [ 6] stated that increased plasma leucine could stimulate muscle protein synthesis by enhancing translation initiation factor activation in neonatal pigs [5,6], suggesting that dietary MGs may promote the muscle protein deposition of broilers through up-regulating the serum l-leucine level. In addition, 1-Methylhistide was the end-product of histidine metabolism and the important precursor for the synthesis of anserine, which exerts a significant impact on the nutritional values, antioxidant capacity, and the umami taste of meat [33,34]. In the present study, the up-regulated serum 1-methylhistide in the MG group reflected the higher level of muscle anserine that has been confirmed in our previous study [9]. Similar to Wu et al. [ 21], the content changes of serum phosphatidylcholine (including PC (16:$\frac{0}{16}$:0), SOPC, sphingomyelin (d18:$\frac{1}{18}$:0), and S-LYSO-PC, P-LYSO-PC) implied that inclusion of MGs affected chicken lipid metabolism, but their effects on muscle fat content and mass were limited [9,17]. The metabolic pathway analysis offered an integrated and direct connection of metabolites by the reconstruction of a biochemical reaction network [35,36]. The dietary MG supplementation enhanced the chicken amino acids and lipid metabolism pathways which was reflected in the KEGG enrichment analysis. Moreover, serum metabolites are the end-products of physiological processes caused by body states or by exposure to environmental factors or drugs, which can provide an integrated view of the biochemical environment of both body fluid and tissue. Therefore, the enhanced amino acid metabolic pathways and related metabolites revealed the chicken body’s amino acid improvement in the MG-treated group. Our previous study showed that both the composition and metabolism pathways of muscle amino acids in broilers were altered in response to MG supplementation [9]. Results obtained with serum untargeted metabolomics in the current study reminded us that the muscle protein synthesis in the pectoralis major may be affected by MG supplementation. Protein synthesis and proteolysis were regulated by the mTOR signaling pathway in avain [37,38,39], in which the phosphorylation of 4E binding protein (4EBP1) and 70 kDa ribosomal protein S6 kinase (S6K1) by mTOR activation initiated mRNA translation and protein synthesis. Higher expression of S6K1 in the MG group indicated that the dietary MG promoted muscle protein synthesis via the mTOR/S6K1 pathway in the present study, resulting in improved skeletal muscle mass and carcass yield [17]. Moreover, the up-regulated serum leucine was recognized as the activator of the mTOR/S6K1 pathway, as the up-regulated effects of branched-chain amino acids on the mTOR/S6K1 pathway have been well demonstrated [40,41,42]. The unchanged relative expression of AKT, FOXO4, FOXO1, and MURF1 demonstrated that the dietary MG exerted no effects on muscle protein proteolysis [39,43]. Myocyte enhancer factor 2 family (MEF2) including MEF2A, MEF2B, MEF2C, and MEF2D, play multiple roles (regulating the differentiation, maintenance, and regeneration of muscle cells) in muscle cells to regulate myogenesis and morphogenesis [44,45]. Wen et al. [ 46] demonstrated that dietary methionine improved breast muscle growth and carcass yield of commercial broilers with increased mRNA levels of MEF2A and MEF2B. Chen et al. [ 30] stated that betaine (up-regulated in serum in the present study) supplementation enhanced muscle growth with increased MEF2B expression in breast muscle. Therefore, the significantly higher expression of MEF2C and MEF2D indicated that MG supplementation may also promote muscle growth and performance by improving the growth and development of muscle cells. Numerous studies proved that intestinal microecology is the main functional target of MG supplementation exerting benefit effects on animals [16,18,19]. Similarly, dietary supplementation of MGs altered the cecal microbiota profile of broilers with increases in the family Bacteroidaceae and decreases in the family Rikenellaceae in this study. Similar to our previous study, broilers receiving the 300–600 mg/kg MG-supplemented diet showed a distinct variation of gut microbiota structure and a notably higher family of Bacteroidaceae in cecum compared with the CON group [16]. The supplementation of graded levels of single GML (450 and 600 mg/kg) in diets also exerted significant alterations on both the structure and composition of chicken gut microbiota [27]. Kong et al. [ 14] reported notable different gut microbiota profiles and improved microbial diversity in broilers fed diets containing 300, 600, 900, or 1200 mg/kg GML at both 7 and 14 days of age. Lan et al. [ 13] stated that the dietary supplementation of GML significantly modulated the gut microbiota community of broilers at both 28 and 56 days of age. Specifically, at the genus level, dietary MGs increased the content of chicken cecal Bacteroides, a no rank genus of the Ruminococcaceae family, and Lachnospiraceae_NK4A136_group. Similar to our previous study, the relative abundance of an unclassified genus of the Lachnospiraceae family, Bacteroides, and Bifidobacteriaceae in cage-reared chickens was increased [16]. Dietary supplementation of GML selectively increased the proportion of an unclassified genus of the Lachnospiraceae family and Bifidobacteriaceae in broilers [27]. It has been reported that the abundance of Lachnospiraceae_FE2018_group and Bacteroides in chicks was increased with the addition of graded levels of GML (300, 600, 900, and 1200 mg/kg) [14]. Similarly, supplementation of lauric acid to the basal diet increased the colonization of Bacteroides and Lactobacillus in lipopolysaccharide-challenged broilers [47]. In summary, these findings revealed that dietary supplementation of medium-chain monoglycerides increased the colonization of Lachnospiraceae and Bacteroides, which belong to the Firmicutes and Bacteroidetes phylum, respectively. It has been reported that Bacteroides play a vital role in the breakdown of complex molecules, especially in the utilization of nitrogenous substances by the host and the gut microbiota [48]. The increased Bacteroides were probably responsible for improved protein digestibility in this study, which was increased from 51.40 to $54.49\%$ after MG supplementation (data not shown). Lachnospiraceae can utilize complex plant-derived carbohydrates, in particular, they readily degrade less recalcitrant indigestible polysaccharides and starch to release sugars for both the gut microbiota and host [49]. Apajalahti et al. [ 50] and Singh et al. [ 15] stated that a higher relative abundance of Lachnospiraceae was closely related to the increased body weight of commercial broilers chickens, suggesting increased Lachnospiraceae_NK4A136_group may contribute to the improved productive performance in the present study. Additionally, similar to a previous study [16], the increased Lachnospiraceae_NK4A136_group and Bacteroides were positively correlated with muscle amino acid. Moreover, the bacterial gene abundance of the carbohydrate and amino acid metabolism pathways by PICRUSt function prediction was significantly increased with MG supplementation. Therefore, the increase of feed protein and carbohydrate utilization efficiency in the gut by selectively increasing the proportion of Lachnospiraceae_NK4A136_group and Bacteroides may partially explain the chicken amino acid improvements. ## 5. Conclusions The present study demonstrated that the enhancement of feed protein digestion and absorption with 300 mg/kg MG supplementation mainly started from the gut microbiota modulation (Lachnospiraceae_NK4A136_group and Bacteroides). Moreover, we observed remarkably increased serum amino acids and derivatives (betaine, l-leucine, l-glutamine, 1-methylhistide), as well as enhanced amino acid pathways in the serum after MG supplementation. Coincidentally, the chicken amino acid composition and gene expression of chicken protein synthesis were improved after MG treatment. Conclusively, the present study partially explained that the dietary MG improved the chicken amino acid composition by increasing amino acid utilization in the gut microbiota, serum, and muscle. The current study offered us a new approach to control chicken quality in future poultry production. ## References 1. Zampiga M., Flees J., Meluzzi A., Dridi S., Sirri F.. Application of omics technologies for a deeper insight into quali-quantitative production traits in broiler chickens: A review. J. Anim. Sci. Biotechnol. (2018) 9. DOI: 10.1186/s40104-018-0278-5 2. Moghaddam H.N., Alizadeh-Ghamsari A.H.. Improved performance and small intestinal development of broiler chickens by dietary L-glutamine supplementation. J. Appl. Anim. Res. (2013) 41 1-7. DOI: 10.1080/09712119.2012.738214 3. Bartell S.M., Batal A.B.. The effect of supplemental glutamine on growth performance, development of the gastrointestinal tract, and humoral immune response of broilers. Poult. Sci. (2007) 86 1940-1947. DOI: 10.1093/ps/86.9.1940 4. Dai S., Wang L., Wen A., Wang L., Jin G.. Dietary glutamine supplementation improves growth performance, meat quality and colour stability of broilers under heat stress. Br. Poult. Sci. (2009) 50 333-340. DOI: 10.1080/00071660902806947 5. Escobar J., Frank J.W., Suryawan A., Nguyen H.V., Kimball S.R., Jefferson L.S., Davis T.A.. Physiological rise in plasma leucine stimulates muscle protein synthesis in neonatal pigs by enhancing translation initiation factor activation. Am. J. Physiol.-Endocrinol. Metab. (2005) 288 E914-E921. DOI: 10.1152/ajpendo.00510.2004 6. Duan Y., Guo Q., Wen C., Wang W., Li Y., Tan B., Li F., Yin Y.. Free amino acid profile and expression of genes implicated in protein metabolism in skeletal muscle of growing pigs fed low-protein diets supplemented with branched-chain amino acids. J. Agric. Food Chem. (2016) 64 9390-9400. DOI: 10.1021/acs.jafc.6b03966 7. Eklund M., Bauer E., Wamatu J., Mosenthin R.. Potential nutritional and physiological functions of betaine in livestock. Nutr. Res. Rev. (2005) 18 31-48. DOI: 10.1079/NRR200493 8. Muroya S., Ueda S., Komatsu T., Miyakawa T., Ertbjerg P.. Meatabolomics: Muscle and meat metabolomics in domestic animals. Metabolites (2020) 10. DOI: 10.3390/metabo10050188 9. Liu T., Mo Q., Wei J., Zhao M., Tang J., Feng F.. Mass spectrometry-based metabolomics to reveal chicken meat improvements by medium-chain monoglycerides supplementation: Taste, fresh meat quality, and composition. Food Chem. (2021) 365. DOI: 10.1016/j.foodchem.2021.130303 10. Batovska D.I., Todorova I.T., Tsvetkova I.V., Najdenski H.M.. Antibacterial study of the medium chain fatty acids and their 1-monoglycerides: Individual effects and synergistic relationships. Pol. J. Microbiol. (2009) 58 43-47. PMID: 19469285 11. Fortuoso B.F., dos Reis J.H., Gebert R.R., Barreta M., Griss L.G., Casagrande R.A., de Cristo T.G., Santiani F., Campigotto G., Rampazzo L.. Glycerol monolaurate in the diet of broiler chickens replacing conventional antimicrobials: Impact on health, performance and meat quality. Microb. Pathog. (2019) 129 161-167. DOI: 10.1016/j.micpath.2019.02.005 12. Li Y., Zhang H., Yang L., Zhang L., Wang T.. Effect of medium-chain triglycerides on growth performance, nutrient digestibility, plasma metabolites and antioxidant capacity in weanling pigs. Anim. Nutr. (2015) 1 12-18. DOI: 10.1016/j.aninu.2015.02.001 13. Lan J., Chen G., Cao G., Tang J., Li Q., Zhang B., Yang C.. Effects of α-glyceryl monolaurate on growth, immune function, volatile fatty acids, and gut microbiota in broiler chickens. Poult. Sci. (2021) 100. DOI: 10.1016/j.psj.2020.11.052 14. Kong L., Wang Z., Xiao C., Zhu Q., Song Z.. Glycerol monolaurate ameliorated intestinal barrier and immunity in broilers by regulating intestinal inflammation, antioxidant balance, and intestinal microbiota. Front. Immunol. (2021) 12. DOI: 10.3389/fimmu.2021.713485 15. Singh K.M., Shah T., Deshpande S., Jakhesara S.J., Koringa P.G., Rank D.N., Joshi C.G.. High through put 16S rRNA gene-based pyrosequencing analysis of the fecal microbiota of high FCR and low FCR broiler growers. Mol. Biol. Rep. (2012) 39 10595-10602. DOI: 10.1007/s11033-012-1947-7 16. Liu T., Guo L., Zhangying Y., Ruan S., Liu W., Zhang X., Feng F.. Dietary medium-chain 1-monoglycerides modulates the community and function of cecal microbiota of broilers. J. Sci. Food Agric. (2021) 102 2242-2252. DOI: 10.1002/jsfa.11562 17. Liu T., Li C., Zhong H., Feng F.. Dietary medium chain α-monoglycerides increases body weight, feed intake, and carcass yield in broilers with muscle composition alteration. Poult. Sci. (2020) 100 186-195. DOI: 10.1016/j.psj.2020.09.056 18. Zhao M., Cai H., Jiang Z., Li Y., Zhong H., Zhang H., Feng F.. Glycerol-monolaurate-mediated attenuation of metabolic syndrome is associated with the modulation of gut microbiota in high-fat-diet-fed mice. Mol. Nutr. Food Res. (2019) 63. DOI: 10.1002/mnfr.201801417 19. Zhao M., Jiang Z., Cai H., Li Y., Mo Q., Deng L., Zhong H., Liu T., Zhang H., Kang J.X.. Modulation of the gut microbiota during high-dose glycerol monolaurate-mediated amelioration of obesity in mice fed a high-fat diet. mBio (2020) 11. DOI: 10.1128/mBio.00190-20 20. Yang X., Yin F., Yang Y., Lepp D., Yu H., Ruan Z., Yang C., Yin Y., Hou Y., Leeson S.. Dietary butyrate glycerides modulate intestinal microbiota composition and serum metabolites in broilers. Sci. Rep. (2018) 8. DOI: 10.1038/s41598-018-22565-6 21. Wu Y., Zhang H., Zhang R., Cao G., Li Q., Zhang B., Wang Y., Yang C.. Serum metabolome and gut microbiome alterations in broiler chickens supplemented with lauric acid. Poult. Sci. (2021) 100. DOI: 10.1016/j.psj.2021.101315 22. Shi K., Zhao Q., Shao M., Duan Y., Li D., Lu Y., Tang Y., Feng C.. Untargeted metabolomics reveals the effect of selective breeding on the quality of chicken meat. Metabolites (2022) 12. DOI: 10.3390/metabo12050367 23. Zhang F., Zhao Y., Wang H., Nan X., Wang Y., Guo Y., Xiong B.. Alterations in the milk metabolome of dairy cows supplemented with different levels of calcium propionate in early lactation. Metabolites (2022) 12. DOI: 10.3390/metabo12080699 24. Cui Y., Liu P., Yu S., He J., Afedo S., Zou S., Zhang Q., Liu J., Song L., Xu Y.. Expression analysis of molecular chaperones Hsp70 and Hsp90 on development and metabolism of different organs and testis in cattle (Cattle & Yak). Metabolites (2022) 12. PMID: 36422254 25. Cai L., Li M., Zhou S., Zhu X., Zhang X., Xu Q.. Trans-species fecal transplant revealed the role of the gut microbiome as a contributor to energy metabolism and development of skeletal muscle. Metabolites (2022) 12. DOI: 10.3390/metabo12080769 26. Soriano-Santos J.. Chemical composition and nutritional content of raw poultry meat. Handb. Poult. Sci. Technol. (2010) 1 467-491 27. Liu T., Tang J., Feng F.. Glycerol monolaurate improves performance, intestinal development, and muscle amino acids in yellow-feathered broilers via manipulating gut microbiota. Appl. Microbiol. Biotechnol. (2020) 104 10279-10291. DOI: 10.1007/s00253-020-10919-y 28. Fortuoso B.F., Galli G.M., Griss L.G., Armanini E.H., Silva A.D., Fracasso M., Mostardeiro V., Morsch V.M., Lopes L.Q., Santos R.C.. Effects of glycerol monolaurate on growth and physiology of chicks consuming diet containing fumonisin. Microb. Pathog. (2020) 147. DOI: 10.1016/j.micpath.2020.104261 29. Saleh A.A., El-Gharabawy B., Hassan A., Badawi N., Eid Y., Selim S., Shukry M., Dawood M.. Effect of dietary inclusion of alpha-monolaurin on the growth performance, lipid peroxidation, and immunity response in broilers. Sustainability (2021) 13. DOI: 10.3390/su13095231 30. Chen R., Zhuang S., Chen Y., Cheng Y., Wen C., Zhou Y.. Betaine improves the growth performance and muscle growth of partridge shank broiler chickens via altering myogenic gene expression and insulin-like growth factor-1 signaling pathway. Poult. Sci. (2018) 97 4297-4305. DOI: 10.3382/ps/pey303 31. Alirezaei M., Gheisari H., Ranjbar V., Hajibemani A.. Betaine: A promising antioxidant agent for enhancement of broiler meat quality. Br. Poult. Sci. (2012) 53 699-707. DOI: 10.1080/00071668.2012.728283 32. Chen R., Wen C., Gu Y., Wang C., Chen Y., Zhuang S., Zhou Y.. Dietary betaine supplementation improves meat quality of transported broilers through altering muscle anaerobic glycolysis and antioxidant capacity. J. Sci. Food Agric. (2020) 100 2656-2663. DOI: 10.1002/jsfa.10296 33. Liu C., Pan D., Ye Y., Cao J.. 1H NMR and multivariate data analysis of the relationship between the age and quality of duck meat. Food Chem. (2013) 141 1281-1286. DOI: 10.1016/j.foodchem.2013.03.102 34. Uenoyama R., Miyazaki M., Miyazaki T., Shigeno Y., Tokairin Y., Konno H., Yamashita T.. LC-ESI-MS/MS quantification of carnosine, anserine, and balenine in meat samples. J. Chromatogr. B (2019) 1132. DOI: 10.1016/j.jchromb.2019.121826 35. Klamt S., Stelling J.. Two approaches for metabolic pathway analysis?. Trends Biotechnol. (2003) 21 64-69. DOI: 10.1016/S0167-7799(02)00034-3 36. Xiao Z., Ge C., Zhou G., Zhang W., Liao G.. 1H NMR-based metabolic characterization of Chinese Wuding chicken meat. Food Chem. (2019) 274 574-582. DOI: 10.1016/j.foodchem.2018.09.008 37. Wang X., Proud C.G.. The mTOR pathway in the control of protein synthesis. Physiology (2006) 21 362-369. DOI: 10.1152/physiol.00024.2006 38. Tesseraud S., Everaert N., Ezzine S.B.-O., Collin A., Métayer-Coustard S., Berri C.. Manipulating tissue metabolism by amino acids. World’s Poult. Sci. J. (2011) 67 243-252. DOI: 10.1017/S0043933911000274 39. Goodman C.A., Mayhew D.L., Hornberger T.A.. Recent progress toward understanding the molecular mechanisms that regulate skeletal muscle mass. Cell. Signal. (2011) 23 1896-1906. DOI: 10.1016/j.cellsig.2011.07.013 40. Tesseraud S., Bigot K., Taouis M.. Amino acid availability regulates S6K1 and protein synthesis in avian insulin-insensitive QM7 myoblasts. FEBS Lett. (2003) 540 176-180. DOI: 10.1016/S0014-5793(03)00260-6 41. Yoshizawa F.. Regulation of protein synthesis by branched-chain amino acids in vivo. Biochem. Biophys. Res. Commun. (2004) 313 417-422. DOI: 10.1016/j.bbrc.2003.07.013 42. Kimball S.R., Jefferson L.S.. Signaling pathways and molecular mechanisms through which branched-chain amino acids mediate translational control of protein synthesis. J. Nutr. (2006) 136 227S-231S. DOI: 10.1093/jn/136.1.227S 43. Tesseraud S., Bouvarel I., Collin A., Audouin E., Crochet S., Seiliez I., Leterrier C.. Daily variations in dietary lysine content alter the expression of genes related to proteolysis in chicken pectoralis major muscle. J. Nutr. (2009) 139 38-43. DOI: 10.3945/jn.108.095752 44. McKinsey T.A., Zhang C.L., Olson E.N.. MEF2: A calcium-dependent regulator of cell division, differentiation and death. Trends Biochem. Sci. (2002) 27 40-47. DOI: 10.1016/S0968-0004(01)02031-X 45. Black B.L., Olson E.N.. Transcriptional control of muscle development by myocyte enhancer factor-2 (MEF2) proteins. Annu. Rev. Cell Dev. Biol. (1998) 14 167-196. DOI: 10.1146/annurev.cellbio.14.1.167 46. Wen C., Jiang X., Ding L., Wang T., Zhou Y.. Effects of dietary methionine on breast muscle growth, myogenic gene expression and IGF-I signaling in fast-and slow-growing broilers. Sci. Rep. (2017) 7. DOI: 10.1038/s41598-017-02142-z 47. Wu Y., Li Q., Liu J., Liu Y., Xu Y., Zhang R., Yu Y., Wang Y., Yang C.. Integrating serum metabolome and gut microbiome to evaluate the benefits of lauric acid on lipopolysaccharide- challenged broilers. Front. Immunol. (2021) 12. DOI: 10.3389/fimmu.2021.759323 48. Sakamoto M., Lapidus A.L., Han J., Trong S., Haynes M., Reddy T., Mikhailova N., Huntemann M., Pati A., Ivanova N.N.. High quality draft genome sequence of Bacteroides barnesiae type strain BL2. Stand. Genom. Sci. (2015) 10. DOI: 10.1186/s40793-015-0045-6 49. Meehan C.J., Beiko R.G.. A phylogenomic view of ecological specialization in the Lachnospiraceae, a family of digestive tract-associated bacteria. Genome Biol. Evol. (2014) 6 703-713. DOI: 10.1093/gbe/evu050 50. Apajalahti J., Vienola K.. Interaction between chicken intestinal microbiota and protein digestion. Anim. Feed. Sci. Technol. (2016) 221 323-330. DOI: 10.1016/j.anifeedsci.2016.05.004
--- title: 'The Association between Vegan, Vegetarian, and Omnivore Diet Quality and Depressive Symptoms in Adults: A Cross-Sectional Study' authors: - Hayley Walsh - Megan Lee - Talitha Best journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9966591 doi: 10.3390/ijerph20043258 license: CC BY 4.0 --- # The Association between Vegan, Vegetarian, and Omnivore Diet Quality and Depressive Symptoms in Adults: A Cross-Sectional Study ## Abstract Dietary patterns and depressive symptoms are associated in cross-sectional and prospective-designed research. However, limited research has considered depression risk related to meat-based and plant-based dietary patterns. This study explores the association between diet quality and depressive symptoms across omnivore, vegan, and vegetarian dietary patterns. A cross-sectional online survey utilised the Dietary Screening Tool (DST) and the Centre for Epidemiological Studies of Depression Scale (CESD-20) to measure diet quality and depressive symptoms, respectively. A total of 496 participants identified as either omnivores ($$n = 129$$), vegetarians ($$n = 151$$), or vegans ($$n = 216$$). ANOVA with Bonferroni post hoc corrections indicates that dietary quality was significantly different between groups F[2, 493] = 23.61, $p \leq 0.001$ for omnivores and vegetarians and omnivores and vegans. Diet quality was highest in the vegan sample, followed by vegetarian and omnivore patterns. The results show a significant, moderately negative relationship between higher diet quality and lower depressive symptoms (r = −0.385, $p \leq 0.001$) across groups. Hierarchical regression showed that diet quality accounted for $13\%$ of the variability in depressive symptoms for the omnivore sample, $6\%$ for vegetarians, and $8\%$ for vegans. This study suggests that diet quality in a meat-based or plant-based diet could be a modifiable lifestyle factor with the potential to reduce the risk of depressive symptoms. The study indicates a greater protective role of a high-quality plant-based diet and lower depressive symptoms. Further intervention research is needed to understand the bi-directional relationship between diet quality and depressive symptoms across dietary patterns. ## 1. Introduction The cost of depression to the global economy is an estimated US$1 trillion annually, impacting $5\%$ of the global adult population [1] and $10\%$ of Australians [2]. This chronic disorder is characterised by mood dysregulation, sad, empty, or irritable emotions, as well as negative self-appraisal and withdrawal/isolating behaviours [3]. Depression is responsible for 50–$70\%$ of suicides globally and is the second leading cause of death for 15–29 year-olds [4]. The average treatment response rate for depression is 20–$30\%$ [5], prompting research to consider other modifiable lifestyle factors, such as diet, to address depressive symptoms [6]. The emerging field of nutritional psychiatry is the nexus between nutrition and psychology, focusing on the role of dietary patterns in mental health conditions such as depression [6,7,8]. Dietary patterns, defined as “the quantity, variety, or combination of different foods and beverages in a diet and the frequency with which they are habitually consumed” [9], are theorised to impact mood due to the differing nutrient profiles and biological mechanisms [10]. Dietary patterns are categorised as either ‘healthy’—rich in fresh vegetables, fruits, seeds, nuts, whole grains, legumes, and water or ‘unhealthy’—high in refined, sugary, and ultra-processed foods [11]. The most commonly researched healthy dietary pattern—the Mediterranean dietary pattern—consists of high consumption of fruit, vegetables, nuts, and olive oil, moderate consumption of oily fish, and limited intake of red meat and highly processed foods [12]. Adherence to the Mediterranean diet has been associated with a lower risk of depression onset [13,14], whilst consumption of a Western dietary pattern typically includes high consumption of processed foods, meat, dairy, and alcohol [15] and is associated with increased risk of depression [16]. Many observational, longitudinal, and intervention studies exploring diet and depression primarily focus on healthy (Mediterranean, anti-inflammatory) and unhealthy (Western) dietary patterns [6,17,18,19,20,21,22,23]. Four randomised control trials assessed dietary change from unhealthy (Western) to healthy (Mediterranean) dietary patterns and depression; two assessed the general population [14,18,24] and two assessed young adults [19,20]. All four studies found that the depressive symptoms of the participants significantly improved after the healthy dietary intervention compared with the control group. To date, associations between dietary patterns and depression outcomes are diverse and exist across the lifespan, including childhood through to older adulthood [25]. For example, narrative systematic review findings of 20 studies in children and young adults show that high diet quality is associated with lower levels of depression. Conversely, low-quality diets are associated with higher levels of depression [26]. Additionally, a systematic review and meta-analysis of 18 studies on dietary patterns and depression risk in older adults found that high diet quality was associated with lower depression risk (OR, 0.85; $95\%$CI, 0.78–0.92) [27]. Other systematic reviews and meta-analyses in the general population have found similar findings [25,28,29]. However, high levels of heterogeneity and risk of bias are inherent across these findings [30,31]. More research is needed in young adult populations to support awareness of dietary patterns and depressive symptoms. Dietary patterns can be further categorised into omnivore and/or plant-based. The typical omnivore diet includes no restrictions on animal products and is generally high in arachidonic acid, a fatty acid found in meat and linked to lower mood [32,33]. Plant-based dietary patterns are characterised by their emphasis on fruits, vegetables, whole grains, soy foods, nuts, and seeds; whilst a vegan diet excludes all animal products, a vegetarian diet may include dairy and eggs [34]. Importantly, low-quality foods such as those high in sugar, saturated fats, and refined grain consumption are also consumed in plant-based dietary patterns. Therefore, both meat and plant-based diets have the potential to be high or low in diet quality [35]. The relationship between dietary patterns, predominately plant-based vegetarian and vegan patterns, and depression is equivocal. Some research suggests that vegetarians and vegans have increased depressive symptoms compared to their omnivore counterparts [36]. For example, meat abstinence is associated with depressive symptoms and, when compared to other dietary patterns, meat abstainers exhibit more significant symptoms of depression than meat eaters [37,38,39]. Conversely, other studies demonstrate lower symptoms of depression in plant-based diet samples than in omnivore samples [33,40,41,42]. Systematic reviews report inconsistent findings between a plant-based dietary pattern and depressive symptoms [43,44]. These inconsistent findings suggest that it may not be plant-based dietary patterns that are linked to depression but rather the quality of the plant-based diet. Additionally, the association between plant-based diets and depression may be confounded by other factors that impact mental health, such as food restriction and food group exclusion, despite the absence of meat [45]. Therefore, more research is needed to understand the relationship between plant-based dietary patterns and depressive symptoms. A recent Australian study surveyed 219 vegans and vegetarians aged 18 to 44 years and considered a plant-based diet quality measurement. Results showed that a high-quality plant-based diet was associated with reduced depressive symptoms, and a low-quality plant-based diet was associated with increased depressive symptoms [42]. Whilst this finding aligns with high and low-quality diet impacts on health in the general population [17], the study focused solely on the relationship between a plant-based diet quality measure and depressive symptoms. As such, further diet quality comparison with omnivore diets is needed. The primary objective of this research is to extend the previous study and compare diet quality in vegan, vegetarian, and omnivore populations and the association with depressive symptoms. It is hypothesised that high diet quality will be associated with decreased depressive symptoms and that there will be a difference in diet quality and depressive symptoms between the vegan, vegetarian, and omnivore populations. ## 2. Material and Methods A total of 581 participants started the survey. The complete data response rate was $85\%$ ($$n = 496$$) for the primary outcome variables (diet quality and depressive symptoms). A G Power analysis indicated that 219 participants would be adequate for a moderate effect (0.15, α = 0.05, β = 0.80). Participants were recruited online via social media sites Facebook, Twitter, or SONA (Bond University student research portal). Inclusion criteria required English-speaking participants between 18 and 44 years of age to have access to an internet-enabled device. Participants were informed that no remuneration was provided for participation, excluding students who accessed the survey through the SONA platform for course credit. The research team first piloted the survey to ensure usability. Informed consent was granted after participants read the information sheet and commenced the survey. Due to a large influx of vegan participants, an attempt to acquire representation from omnivores and vegetarians was made through purposive sampling. Data collection occurred between November 2021 and January 2022. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE-nut) checklist ensured a common reporting standard Supplementary Material [46]. The Bond University Human Research Ethics Committee approved the study (#ML01980). Participants completed an online survey hosted on Qualtrics, which asked for demographic measures of gender (male, female), age (continuous), marital status (partnered, not partnered), and education (high school, university degree, trade certificate). Self-reported height and weight were collected to calculate the participants’ Body Mass Index (BMI). Participants self-reported dietary patterns (omnivore, vegan, or vegetarian). ## 2.1. Depression (CESD-20) The CESD-20 [47] is a 20-item measurement of symptoms of depression. The scale indicates the experience of depressive symptoms in the general population, not in a clinical population. Depressive symptoms experienced within the previous seven days were calculated on a four-point Likert scale: rarely [0] to most of the time [3]. An example question included: “I felt I was just as good as other people”. Scores ranged between 0 and 60; the higher the score, the greater the experience of depressive symptoms. A score of 16 or above exceeded the criterion cut-off score and indicated the experience of depressive symptoms. The CESD-20 has been successfully employed on younger populations, older populations, and populations with health comorbidities [48]. The CESD-20 demonstrated high validity, reliability [49], and internal consistency (α > 0.9). The tool is concurrently valid with the Beck Depression scale, classified as the gold standard scale for depression measurement [48]. ## 2.2. Dietary Screening Tool (DST) The DST determines diet quality and nutritional risk scores by capturing the frequency of commonly eaten foods over a seven-day period [50]. The original American scale consists of 37 items and is structured according to the 2005 Dietary Guidelines for Americans [50]. The DST version used in this study was condensed to 21 items, with adaptations for an Australian population, which have been used in previous research [14]. For example, the fast-food question was amended to include Australian fast-food chains: “How often do you eat McDonald’s, Kentucky Fried Chicken, Pizza Hut, or Hungry Jacks?” Each food item response was assigned a number between 0 and 8. The total sum of all scores determined diet quality, with 0 indicating low diet quality and 105 signifying high diet quality. The DST had a high test-retest reliability coefficient of 0.83 ($p \leq 0.001$) and high validity and was validated in an Australian population [50,51]. ## 2.3. International Physical Activity Questionnaire (IPAQ) The IPAQ [52] determined physical activity levels over a seven-day period, capturing activity as vigorous, moderate, walking, or sitting across 7 items, for example: “During the last 7 days, on how many days did you do vigorous physical activities like heavy lifting, digging, aerobics, or fast bicycling?” For the purpose of this study, total active hours per week were calculated by tallying vigorous, moderate, and walking activities. The IPAQ had a high test-retest reliability coefficient of 0.80 ($p \leq 0.001$) and high validity [53]. ## 2.4. The Social Connectedness Scale—Revised The Social Connectedness Scale [54] determines the degree to which participants have felt connections to others in social settings. Participants provide a response to eight items on a six-point Likert scale, ranging from strongly disagree [1] to strongly agree [6]. An example question includes: “I feel so distant from people”. Negatively worded items were reverse-coded. Scores were summed to provide total scores (range 0–48), with higher scores indicating a strong sense of social connectedness, high reliability (internal consistency a >0.92), and validity [55]. ## 2.5. Statistical Analysis Data were analysed using SPSS Statistics software version 28 (IBM SPSS Statistics, Chicago, FL, USA). Steps were taken to ensure data integrity; missing values in the CESD-20 and DST were imputed with mean estimates based on the dietary pattern category. Frequency descriptives were calculated for all variables prior to ANOVA and chi-square statistics were conducted to compare group means. The final ANOVA analysis included a Bonferroni correction post hoc in determining between-group differences, segmented by dietary pattern, for all variables. Correlational analyses were used to determine the association between depressive symptoms (DV), diet quality (IV), and other covarying factors such as age, gender, BMI, social connectedness, physical activity levels, marital status, and education level. Pearson’s correlation was used to assess the significance of continuous variables and Spearman’s correlation for categorical variables. Subsequently, a hierarchical multiple linear regression was conducted for all significant continuous predictors of depressive symptoms (CESD-20); BMI and physical activity levels for model one and DST for model two. ## 3. Results The demographic and lifestyle characteristics of 496 participants are detailed in Table 1, split by dietary pattern and total population. Overall, the participants’ mean age was 30.95 years (SD = 7.47), with $76\%$ identifying as female. Most were partnered ($66\%$) and had a university degree ($75\%$). The overall population was split by dietary pattern: omnivore ($$n = 129$$), vegetarian ($$n = 151$$), and vegan ($$n = 216$$). The mean depressive symptom score (CESD-20) for participants following an omnivore dietary pattern was 16.27 (SD = 10.98). The mean omnivore score is above the cut-off criterion score of 16, indicating possible experiences of depressive symptoms for this dietary pattern. Depressive symptom scores for those following a vegetarian dietary pattern ($M = 12.99$, SD = 9.78) and vegan dietary pattern ($M = 11.08$, SD = 9.83) were below the criterion cut-offs as were the total population scores ($M = 13.01$, SD = 10.32). A one-way ANOVA found a significant difference in diet quality, F[2, 493] = 23.61, $p \leq 0.001$, and depressive symptoms, F[2, 493] = 10.61, $p \leq 0.001$ between the three dietary patterns (omnivore, vegetarian, and vegan). A post hoc analysis with Bonferroni correction determined that vegans ($M = 76.55$, SD = 10.44, $p \leq 0.001$) and vegetarians ($M = 73.00$, SD = 12.00, $p \leq 0.001$) had greater dietary quality than omnivores ($M = 67.03$, SD = 15.66). In the overall sample, Pearson’s correlations on continuous variables and Spearman’s correlations on categorical variables (Table 2) show a significant moderate negative relationship between diet quality and depressive symptoms (r = −0.385, $p \leq 0.001$), irrespective of dietary type. For dietary type, the significant moderately negative relationship between diet quality and depressive symptoms remained, see Table 3, omnivore (r = −0.440, $p \leq 0.001$), vegetarian (r = −0.302, $p \leq 0.001$), and vegan (r = −0.300, $p \leq 0.001$). A two-stage hierarchical multiple linear regression was conducted to estimate the proportion of variance in depressive symptoms that can be accounted for by BMI, physical activity levels, and diet quality across omnivore, vegetarian, and vegan dietary patterns. Standardised (β) regression coefficients and squared semi-partial correlations (sr2) for each predictor at each step are reported in Table 4. Prior to interpreting the results, the assumptions of linearity, homoscedasticity, and multicollinearity were met. On the first step of the hierarchical multiple linear regression, model one, BMI and physical activity levels collectively accounted for $9\%$ of the variability in depressive symptoms in omnivores, R2 = 0.094, F[2, 126] = 6.57, $$p \leq 0.002.$$ Model one accounted for $8\%$ of depressive symptoms in vegetarians, R2 = 0.078, F[2, 148] = 6.25, $$p \leq 0.002$$; however, no significant results were determined for vegans ($$p \leq 0.079$$). In the second step of the model, diet quality was added to the regression equation and accounted for an additional $13\%$ of the variability in depression symptoms for an omnivore diet, R2 = 0.229, F[1, 125] = 21.8, $p \leq 0.001.$ Model two accounted for an additional $6\%$ of the variability in depression symptoms for a vegetarian diet R2 = 0.134, F[1, 147] = 9.45, $p \leq 0.001$, and $8\%$ for a vegan diet, R2 = 0.100, F[1, 212] = 18.06, $p \leq 0.001.$ ## 4. Discussion The current study explored the relationship between diet quality and symptoms of depression in self-reported vegan, vegetarian, and omnivore dietary samples. Across the whole sample of dietary patterns, a high-quality diet was associated with lower depressive symptoms. These findings align with recent research that showed a high-quality plant-based diet may protect against the onset or severity of depression [42] and that improved dietary quality, in general, is related to lowering the symptoms of depression [17,56,57,58]. Notably, the relationship between dietary quality and reported depressive symptoms was irrespective of dietary patterns and reflects the broad findings from intervention studies. For example, four Australian randomised controlled trials show that by increasing dietary quality, symptoms of depression decreased over a 12-week time frame [14,18,19,20]. Further, systematic review and meta-analysis findings show that adhering to a high-quality diet protects against depressive symptoms [17]. Whilst the relationships between high dietary quality and lower depression appear robust, the type of dietary pattern has not been considered, and the neurophysiological mechanisms of effect for differential dietary patterns and relationship with depression are yet to be clearly understood. Our results show that across and between dietary patterns, diet quality relates to depressive symptoms, irrespective of dietary pattern. Probable explanations for the relationship between diet quality and depressive symptoms rely on biological mechanisms underpinning the biochemical effects of specific food components in a low-quality diet. For example, a low-quality diet is typically characterised by ultra-processed foods [59] or those constituting high sugar or fat content [28]. Consumption of high-sugar and high-fat foods is associated with heightened markers of inflammation and inflammatory disease [60], which in turn are associated with a higher risk of depression [61,62]. Some low-quality foods, including sugar-sweetened beverages, have a high glycaemic index that also contributes to inflammation [63]. In individuals with depression, higher inflammation markers are present compared to healthy control groups, suggesting that biological mechanisms of inflammation underpin depression [64]. Similarly, higher oxidative stress and inflammation levels have been identified in the brains of those with depression than in those without depression [65]. To date, dietary interventions such as the Mediterranean diet (high in plant foods) are known to reduce inflammation and inflammatory markers [66,67] and reduce depressive symptoms [13,18]. Similarly, diets high in fruit and vegetables, which are staples of a vegan or vegetarian diet, are rich in antioxidants, notably polyphenols, which are negatively correlated with depression [68,69] and depression severity [65]. Conversely, nutritional deficiencies in a plant-based diet could be involved in the increase of depressive symptoms. One study showed that $52\%$ of vegans and $7\%$ of vegetarians were deficient in vitamin B12 [70], a vitamin generally gained through red meat consumption and thought to help combat depressive symptoms [38]. Similarly, omega-3 polyunsaturated fatty acids play a vital role in brain function and are linked with mood outcomes [33]. Given that the most bioavailable source of Omega-3 polyunsaturated fatty acids is oily fish, the intake of fish is reduced in some plant-based diets, and as such omega-3 deficiency may play a role in decreased mood [44]. When split by dietary pattern, the current findings reveal that the vegan sample reported the highest diet quality score, followed by the vegetarian and omnivore samples. These findings are consistent with previous research that reported that both vegan and vegetarian dieters scored higher on the Healthy Eating Index 2010 (HEI-2010) than omnivores (Clarys, et al. [ 71]). Results are likely attributed to the nutritional value of foods consumed in each dietary pattern. For example, a vegan diet typically includes more whole foods, such as vegetables and legumes. In contrast, an omnivore diet typically includes more ultra-processed and refined foods such as cake, pastries, and chocolate [72]. When exceeding recommended amounts, sugar, sodium, and fat levels are associated with a low-quality diet and are considerably higher in self-nominated omnivore foods [73]. Conversely, a more recent study showed that both vegetarian and vegan dietary patterns of a low-quality diet were associated with high levels of depression compared to a high-quality vegan or vegetarian diet [42]. These studies suggest that diet quality may be associated with depressive symptoms, irrespective of the dietary pattern consumed. In this study, the omnivore diet quality score was significantly lower than both the vegetarian and vegan groups and may result from the frequency of consumption of red and processed meat and lower consumption of fruit and vegetables compared to their vegan and vegetarian counterparts. Red meat is commonly touted as a valuable contributor to the recommended dietary intake levels of protein, iron, vitamin B12, and zinc [74], all of which contribute to a healthy diet. However, excess consumption can result in the digestion of too many saturated fats, which is associated with a low-quality diet and a higher risk of depressive symptoms [75]. It has been reported that a typical omnivore consumes approximately $58\%$ and $81\%$ more meat than the recommended intake for women and men, which correspondingly [76] links excess red meat consumption with lower diet quality in comparison to plant-based diets. In the first study to look at meat consumption in depression Jacka, Pasco, Williams, Mann, Hodge, Brazionis and Berk [77] found that when consumption rates fell below or exceeded the recommended daily intake (28–57 g per day), red meat consumption was linked to an increased risk of depressive disorder prevalence. These findings highlight that both abstinence from red meat and overconsumption may have adverse implications for mood. Depressive symptoms were highest in the omnivore group in this study, followed by the vegetarian and vegan groups. The correlation between diet quality and depressive symptoms between groups followed a similar trend, as the strongest relationship was evident in the omnivore sample, followed by the vegetarian and then vegan samples. Results show a significant between-group difference when comparing the omnivore and vegan sample and the omnivore and vegetarian sample; however, no difference was found between the vegan and vegetarian sample. Literature on plant-based diets and depressive symptoms is ambiguous. Some studies found that plant-based diets were associated with a risk of depressive symptoms [42,70]. Other studies only concluded gender differences; male vegetarians demonstrated higher depression than male omnivores, but females did not [38]. Interestingly, a recent meta-analysis of over 170,000 participants concluded that people who eat meat, predominantly omnivores, had lower depression than plant-based diet samples [78]. ## Strengths, Limitations, and Future Direction The study is novel to the field, providing a unique insight into the relationship between diet quality and depressive symptoms across vegan, vegetarian, and omnivore dietary patterns. Key strengths of the study include the adequate power of the sample size and high completion rate ($79\%$), which indicates a highly motivated sample. A sensitivity analysis was conducted to explore the impact of the three DST items that scored favourably on the omnivore dietary pattern, determining no disadvantage to the plant-based diet sample and further endorsing the study’s findings. Conversely, the study’s highly motivated participants are also a limitation as the sample may not reflect the general Australian population who have a predominantly Western dietary pattern. Further, in this sample, $61\%$ of participants were within the self-reported healthy BMI of 25 compared to the national average of $32\%$ [79]. Data collection for this study occurred during the COVID-19 worldwide pandemic. As such, it would be expected that data may not reflect habitual dietary patterns due to the global increase in the consumption of high-energy-density snack foods and emotional eating [80]. In addition, global depression rates had also increased during the pandemic and may have impacted the results of this study [81]. Further, selection bias may be a concern as participation in the study may be driven by following a plant-based diet or experiencing depressive symptoms. This may explain why omnivore participants reported higher levels of depressive symptoms. Some of the recruitment occurred on a plant-based social media site and a podcast in which a study reflecting poorly on plant-based diets was discussed. Therefore, a proportion of the vegan and vegetarian responses to the CESD-20 depression scale in this study may have been influenced by social desirability. Therefore, depressive symptoms being highest in the omnivore diet, in comparison to the plant-based diet samples, may be influenced by this social desirability bias. Being a morally motivated minority, vegan groups are often stigmatised and the target of discrimination [82] and may inadvertently mask depressive symptoms to maintain socially desirable attributes of group identity. Therefore, comparisons between meat-based and plant-based diet followers and depressive symptoms within this study should be interpreted with due prudence. ## 5. Conclusions This cross-sectional study demonstrates an association between high-quality dietary omnivore, vegan, and vegetarian diets and lower depressive symptoms. It also indicates that vegan and vegetarian dietary patterns were associated with higher diet quality. Regardless of the dietary pattern (meat or plant-based), the importance of these findings shows that more frequent intake of fruits, vegetables, nuts, seeds, legumes, whole grains, and water and reduced ultra-processed, refined, and sugary foods, are associated with lower depressive symptoms. ## References 1. World Health Statistics 2. National Health Survey: First Results, 2017–2018 3. 3. American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders5th ed.American Psychiatric AssociationWashington, DC, USA2013. Diagnostic and Statistical Manual of Mental Disorders (2013.0) 4. 4. World Health Organization Depression and Other Common Mental Disorders: Global Health EstimatesWorld Health OrganizationGeneva, Switzerland2017. Depression and Other Common Mental Disorders: Global Health Estimates (2017.0) 5. Cohen S.E., Zantvoord J.B., Wezenberg B.N., Bockting C.L.H., van Wingen G.A.. Magnetic resonance imaging for individual prediction of treatment response in major depressive disorder: A systematic review and meta-analysis. Transl. Psychiatry (2021.0) 11 168. DOI: 10.1038/s41398-021-01286-x 6. Jacka F.N.. Nutritional psychiatry: Where to next?. EBioMedicine (2017.0) 17 24-29. DOI: 10.1016/j.ebiom.2017.02.020 7. Adan R.A.H., van der Beek E.M., Buitelaar J.K., Cryan J.F., Hebebrand J., Higgs S., Schellekens H., Dickson S.L.. Nutritional psychiatry: Towards improving mental health by what you eat. Eur. Neuropsychopharmacol. (2019.0) 29 1321-1332. DOI: 10.1016/j.euroneuro.2019.10.011 8. Marx W., Moseley G., Berk M., Jacka F.N.. Nutritional psychiatry: The present state of the evidence. Proc. Nutr. Soc. (2017.0) 76 427-436. DOI: 10.1017/S0029665117002026 9. Sánchez-Villegas A., Martínez-Lapiscina E.H., Sánchez-Villegas A., Sánchez-Tainta A.. A healthy diet for your heart and your brain. The Prevention of Cardiovascular Disease through the Mediterranean Diet (2018.0) 169-197 10. Berk M., Jacka F.N.. Diet and depression, from confirmation to implementation. JAMA (2019.0) 321 842-843. DOI: 10.1001/jama.2019.0273 11. Khosravi M., Sotoudeh G., Majdzadeh R., Nejati S., Darabi S., Raisi F., Esmaillzadeh A., Sorayani M.. Healthy and unhealthy dietary patterns are related to depression: A case-control study. Psychiatry Investig. (2015.0) 12 434-442. DOI: 10.4306/pi.2015.12.4.434 12. Altun A., Brown H., Szoeke C., Goodwill A.M.. The Mediterranean dietary pattern and depression risk: A systematic review. Neurol. Psychiatry Brain Res. (2019.0) 33 1-10. DOI: 10.1016/j.npbr.2019.05.007 13. Yin W., Lof M., Chen R., Hultman C.M., Fang F., Sandin S.. Mediterranean diet and depression: A population-based cohort study. Int. J. Behav. Nutr. Phys. Act. (2021.0) 18 1-10. DOI: 10.1186/s12966-021-01227-3 14. Jacka F.N., Neil A., Opie R.S., Itsiopoulos C., Cotton S., Mohebbi M., Castle D., Dash S., Mihalopoulos C., Chatterton M.. A randomised controlled trial of dietary improvement for adults with major depression (the SMILES trial). BMC Med. (2017.0) 15. DOI: 10.1186/s12916-017-0791-y 15. Jacka F.N., Pasco J.A., Mykletun A., Williams L.J., Hodge A.M., O’Reilly S.L., Nicholson G.C., Kotowicz M.A., Berk M.. Association of Western and traditional diets with depression and anxiety in women. Am. J. Psychiatry (2010.0) 167 305-311. DOI: 10.1176/appi.ajp.2009.09060881 16. Matison A.P., Mather K.A., Flood V.M., Reppermund S.. Associations between nutrition and the incidence of depression in middle-aged and older adults: A systematic review and meta-analysis of prospective observational population-based studies. Ageing Res. Rev. (2021.0) 70 101403. DOI: 10.1016/j.arr.2021.101403 17. Lassale C., Batty G.D., Baghdadli A., Jacka F.N., Sánchez-Villegas A., Kivimäki M., Akbaraly T.N.. Healthy dietary indices and risk of depressive outcomes: A systematic review and meta-analysis of observational studies. Mol. Psychiatry (2018.0) 1 1-22. DOI: 10.1038/s41380-018-0237-8 18. Parletta N., Zarnowiecki D., Cho J., Wilson A., Bogomolova S., Villani A., Itsiopoulos C., Niyonsenga T., Blunden S., Meyer B.. A Mediterranean-style dietary intervention supplemented with fish oil improves diet quality and mental health in people with depression: A randomized controlled trial (HELFIMED). Nutr. Neurosci. (2017.0) 22 474-487. DOI: 10.1080/1028415X.2017.1411320 19. Francis H.M., Stevenson R.J., Chambers J.R., Gupta D., Newey B., Lim C.K.. A brief diet intervention can reduce symptoms of depression in young adults—A randomised controlled trial. PLoS ONE (2019.0) 14. DOI: 10.1371/journal.pone.0222768 20. Bayes J., Schloss J., Sibbritt D.. A randomised controlled trial assessing the effect of a Mediterranean diet on the symptoms of depression in young men (the ‘AMMEND’ study): A study protocol. Br. J. Nutr. (2021.0) 126 730-737. DOI: 10.1017/S0007114520004699 21. Lee M., Bradbury J., Yoxall J., Sargeant S.. Is dietary quality associated with depression? An analysis of the Australian longitudinal study of women’s health data. Br. J. Nutr. (2022.0) 1-8. DOI: 10.1017/S0007114522002410 22. Lee M.F., Bradbury J.F., Yoxall J., Sargeant S.. It’s about what you’ve assigned to the salad: Focus group discussions on the relationship between food and mood. Int. J. Environ. Res. Public Health (2023.0) 20. DOI: 10.3390/ijerph20021476 23. Sanchez-Villegas A., Henriquez-Sanchez P., Ruiz-Canela M., Lahortiga F., Molero P., Toledo E.A., Martinez-Gonzalez M.A.. A longitudinal analysis of diet quality scores and the risk of incident depression in the SUN Project. BMC Med. (2015.0) 13. DOI: 10.1186/s12916-015-0428-y 24. Lee M.F., Angus D., Walsh H., Sargeant S.. “Maybe it’s not just the food?” A food and mood focus group study. Int. J. Environ. Res. Public Health (2023.0) 20. DOI: 10.3390/ijerph20032011 25. Molendijk M.L., Molero P., Ortuño Sánchez-Pedreño F., Van der Does W., Angel Martínez-González M.. Diet quality and depression risk: A systematic review and dose-response meta-analysis of prospective studies. J. Affect. Disord. (2018.0) 226 346-354. DOI: 10.1016/j.jad.2017.09.022 26. Khalid S., Williams C.M., Reynolds S.A.. Is there an association between diet and depression in children and adolescents? A systematic review. Br. J. Nutr. (2016.0) 116 2097-2108. DOI: 10.1017/S0007114516004359 27. Wu P.-Y., Chen K.-M., Belcastro F.. Dietary patterns and depression risk in older adults: Systematic review and meta-analysis. Nutr. Rev. (2020.0) 79 976-987. DOI: 10.1093/nutrit/nuaa118 28. Quirk S.E., Williams L.J., O’Neil A., Pasco J.A., Jacka F.N., Housden S., Berk M., Brennan S.L.. The association between diet quality, dietary patterns and depression in adults: A systematic review. BMC Psychiatry (2013.0) 13. DOI: 10.1186/1471-244X-13-175 29. Li Y., Lv M.R., Wei Y.J., Sun L., Zhang J.X., Zhang H.G., Li B.. Dietary patterns and depression risk: A meta-analysis. Psychiatry Res. (2017.0) 253 373-382. DOI: 10.1016/j.psychres.2017.04.020 30. Xu Y., Zeng L., Zou K., Shan S., Wang X., Xiong J., Zhao L., Zhang L., Cheng G.. Role of dietary factors in the prevention and treatment for depression: An umbrella review of meta-analyses of prospective studies. Transl. Psychiatry (2021.0) 11 478. DOI: 10.1038/s41398-021-01590-6 31. Gianfredi V., Dinu M., Nucci D., Eussen S., Amerio A., Schram M.T., Schaper N., Odone A.. Association between dietary patterns and depression: An umbrella review of meta-analyses of observational studies and intervention trials. Nutr. Rev. (2022.0) 81 346-359. DOI: 10.1093/nutrit/nuac058 32. Martínez-González M.A., Sánchez-Villegas A., Martínez-González M.A., Sánchez-Villegas A.. Food patterns and the prevention of depression. Proc. Nutr. Soc. (2016.0) 75 139-146. DOI: 10.1017/S0029665116000045 33. Beezhold B., Johnston C.. Restriction of meat, fish, and poultry in omnivores improves mood: A pilot randomized controlled trial. Nutr. J. (2012.0) 11 9. DOI: 10.1186/1475-2891-11-9 34. Dreher M.L.. Dietary Patterns and Whole Plant Foods in Aging and Disease (2018.0) 35. Baden M.Y., Liu G., Satija A., Li Y., Sun Q., Fung T.T., Rimm E.B., Willett W.C., Hu F.B., Bhupathiraju S.N.. Changes in plant-based diet quality and total and cause-specific mortality. Circulation (2019.0) 140 979-991. DOI: 10.1161/CIRCULATIONAHA.119.041014 36. Fazelian S., Sadeghi E., Firouzi S., Haghighatdoost F.. Adherence to the vegetarian diet may increase the risk of depression: A systematic review and meta-analysis of observational studies. Nutr. Rev. (2022.0) 80 242-254. DOI: 10.1093/nutrit/nuab013 37. Forestell C.A., Nezlek J.B.. Vegetarianism, depression, and the five factor model of personality. Ecol. Food Nutr. (2018.0) 57 246-259. DOI: 10.1080/03670244.2018.1455675 38. Li X.-D., Cao H.-J., Xie S.-Y., Li K.-C., Tao F.-B., Yang L.-S., Zhang J.-Q., Bao Y.-S.. Adhering to a vegetarian diet may create a greater risk of depressive symptoms in the elderly male Chinese population. J. Affect. Disord. (2019.0) 243 182-187. DOI: 10.1016/j.jad.2018.09.033 39. Michalak J., Zhang X.C., Jacobi F.. Vegetarian diet and mental disorders: Results from a representative community survey. Int. J. Behav. Nutr. Phys. Act. (2012.0) 9 67. DOI: 10.1186/1479-5868-9-67 40. Jin Y., Kandula N.R., Kanaya A.M., Talegawkar S.A.. Vegetarian diet is inversely associated with prevalence of depression in middle-older aged South Asians in the United States. Ethn. Health (2021.0) 26 504-511. DOI: 10.1080/13557858.2019.1606166 41. Yu-Chih S., Chang C.-E., Ming-Nan L., Chin-Lon L.. Vegetarian diet Is associated with lower risk of depression in Taiwan. Nutrients (2021.0) 13. DOI: 10.3390/nu13041059 42. Lee M.F., Eather R., Best T.. Plant-based dietary quality and depressive symptoms in Australian vegans and vegetarians: A cross-sectional study. BMJ Nutr. Prev. Health (2021.0) 4 e000332. DOI: 10.1136/bmjnph-2021-000332 43. Iguacel I., Huybrechts I., Moreno L.A., Michels N.. Vegetarianism and veganism compared with mental health and cognitive outcomes: A systematic review and meta-analysis. Nutr. Rev. (2020.0) 79 361-381. DOI: 10.1093/nutrit/nuaa030 44. Askari M., Daneshzad E., Darooghegi Mofrad M., Bellissimo N., Suitor K., Azadbakht L.. Vegetarian diet and the risk of depression, anxiety, and stress symptoms: A systematic review and meta-analysis of observational studies. Crit. Rev. Food Sci. Nutr. (2020.0) 62 261-271. DOI: 10.1080/10408398.2020.1814991 45. Matta J., Czernichow S., Kesse-Guyot E., Hoertel N., Limosin F., Goldberg M., Zins M., Lemonge C.. Depressive symptoms and vegetarian diets: Results from the Constances Cohort. Nutrients (2018.0) 10 1695. DOI: 10.3390/nu10111695 46. Lachat C., Hawwash D., Ocké M.C., Berg C., Forsum E., Hörnell A., Larsson C., Sonestedt E., Wirfält E., Åkesson A.. STrengthening the Reporting of OBservational studies in Epidemiology-Nutritional epidemiology (STROBE-nut): An extension of the STROBE statement. PLoS Med. (2016.0) 13. DOI: 10.1371/journal.pmed.1002036 47. Radloff L.S.. The CES-D Scale: A self-report depression scale for research in the general population. Appl. Psychol. Meas. (1977.0) 1 385-401. DOI: 10.1177/014662167700100306 48. Cosco T.D., Prina M., Stubbs B., Wu Y.-T.. Reliability and validity of the center for epidemiologic studies depression scale in a population-based cohort of middle-aged U.S. adults. J. Nurs. Meas. (2017.0) 25 476-485. DOI: 10.1891/1061-3749.25.3.476 49. Jiang L., Wang Y., Zhang Y., Li R., Wu H., Li C., Wu Y., Tao Q.. The reliability and alidity of the Center for Epidemiologic Studies Depression Scale (CES-D) for Chinese university students. Front. Psychiatry (2019.0) 10 315. DOI: 10.3389/fpsyt.2019.00315 50. Bailey R.L., Miller P.E., Mitchell D.C., Hartman T.J., Lawrence F.R., Sempos C.T., Smiciklas-Wright H.. Dietary screening tool identifies nutritional risk in older adults. Am. J. Clin. Nutr. (2009.0) 90 177-183. DOI: 10.3945/ajcn.2008.27268 51. Ventura Marra M., Thuppal S.V., Johnson E.J., Bailey R.L.. Validation of a Dietary Screening Tool in a middle-aged appalachian population. Nutrients (2018.0) 10. DOI: 10.3390/nu10030345 52. Craig C.L., Marshall A.L., Sjöström M., Bauman A.E., Booth M.L., Ainsworth B.E., Pratt M., Ekelund U., Yngve A., Sallis J.F.. International physical activity questionnaire: 12-country reliability and validity. Med. Sci. Sport. Exerc. (2003.0) 35 1381-1395. DOI: 10.1249/01.MSS.0000078924.61453.FB 53. Hallal P.C., Victora C.G.. Reliability and validity of the international physical activity questionnaire (IPAQ). Med. Sci. Sport. Exerc. (2004.0) 36 556. DOI: 10.1249/01.MSS.0000117161.66394.07 54. Lee R.M., Robbins S.B.. Measuring belongingness: The social connectedness and the social assurance scales. J. Couns. Psychol. (1995.0) 42 232. DOI: 10.1037/0022-0167.42.2.232 55. Carpenter C.J., Boster F.J., Kotowski M., Day J.P.. Evidence for the validity of a social connectedness scale: Connectors amass bridging social capital online and offline. Commun. Q. (2015.0) 63 119-134. DOI: 10.1080/01463373.2015.1012217 56. Nanri A., Nagai C., Kochi T., Eguchi M., Kabe I., Mizoue T.. Diet quality and depressive symptoms among workers. Clin. Nutr. (2020.0) 39 1951-1957. DOI: 10.1016/j.clnu.2019.08.021 57. Collins S., Dash S., Allender S., Jacka F., Hoare E.. Diet and mental health during emerging adulthood: A systematic review. Emerg. Adulthood (2020.0) 10 645-659. DOI: 10.1177/2167696820943028 58. Recchia D., Baghdadli A., Lassale C., Brunner E., Verdier J., Kivimäki M., Akbaraly T.N.. Associations between long-term adherence to healthy diet and recurrent depressive symptoms in Whitehall II Study. Eur. J. Clin. Nutr. (2019.0) 59 1031-1041. DOI: 10.1007/s00394-019-01964-z 59. Ares G., Vidal L., Allegue G., Giménez A., Bandeira E., Moratorio X., Molina V., Curutchet M.R.. Consumers’ conceptualization of ultra-processed foods. Appetite (2016.0) 105 611-617. DOI: 10.1016/j.appet.2016.06.028 60. Elizabeth L., Machado P., Zinöcker M., Baker P., Lawrence M.. Ultra-processed foods and health outcomes: A narrative review. Nutrients (2020.0) 12 1955. DOI: 10.3390/nu12071955 61. Berk M., Williams L.J., Jacka F.N., O’Neil A., Pasco J.A., Moylan S., Allen N.B., Stuart A.L., Hayley A.C., Byrne M.L.. So depression is an inflammatory disease, but where does the inflammation come from?. BMC Med. (2013.0) 11. DOI: 10.1186/1741-7015-11-200 62. Kivimäki M., Shipley M.J., Batty G.D., Hamer M., Akbaraly T.N., Kumari M., Jokela M., Virtanen M., Lowe G.D., Ebmeier K.P.. Long-term inflammation increases risk of common mental disorder: A cohort study. Mol. Psychiatry (2014.0) 19 149-150. DOI: 10.1038/mp.2013.35 63. Bullmore E.T.. The Inflamed Mind: A Radical New Approach to Depression (2019.0) 64. Haapakoski R., Mathieu J., Ebmeier K.P., Alenius H., Kivimäki M.. Cumulative meta-analysis of interleukins 6 and 1β, tumour necrosis factor α and C-reactive protein in patients with major depressive disorder. Brain Behav. Immun. (2015.0) 49 206-215. DOI: 10.1016/j.bbi.2015.06.001 65. Prohan M., Amani R., Nematpour S., Jomehzadeh N., Haghighizadeh M.H.. Total antioxidant capacity of diet and serum, dietary antioxidant vitamins intake, and serum hs-CRP levels in relation to depression scales in university male students. Redox Rep. Commun. Free Radic. Res. (2014.0) 19 133-139. DOI: 10.1179/1351000214Y.0000000085 66. González R., Ballester I., López-Posadas R., Suárez M.D., Zarzuelo A., Martínez-Augustin O., Medina F.S.D.. Effects of flavonoids and other polyphenols on inflammation. Crit. Rev. Food Sci. Nutr. (2011.0) 51 331-362. DOI: 10.1080/10408390903584094 67. Elliott M., Chithan K.. The Impact of Plant Flavonoids on Mammalian Biology: Implications for Immunity, Inflammation and Cancer (2017.0) 619-652 68. Bayes J., Schloss J., Sibbritt D.. Effects of polyphenols in a mediterranean diet on symptoms of depression: A systematic literature review. Adv. Nutr. (2019.0) 11 602-615. DOI: 10.1093/advances/nmz117 69. Daneshzad E., Keshavarz S.-A., Qorbani M., Larijani B., Bellissimo N., Azadbakht L.. Association of dietary acid load and plant-based diet index with sleep, stress, anxiety and depression in diabetic women. Br. J. Nutr. (2020.0) 123 901-912. DOI: 10.1017/S0007114519003179 70. Hibbeln J.R., Northstone K., Evans J., Golding J.. Vegetarian diets and depressive symptoms among men. J. Affect. Disord. (2018.0) 225 13-17. DOI: 10.1016/j.jad.2017.07.051 71. Clarys P., Deliens T., Huybrechts I., Deriemaeker P., Vanaelst B., De Keyzer W., Hebbelinck M., Mullie P.. Comparison of nutritional quality of the vegan, vegetarian, semi-vegetarian, pesco-vegetarian and omnivorous diet. Nutrients (2014.0) 6 1318-1332. DOI: 10.3390/nu6031318 72. Larsson C.L., Johansson G.K.. Dietary intake and nutritional status of young vegans and omnivores in Sweden. Am. J. Clin. Nutr. (2002.0) 76 100-106. DOI: 10.1093/ajcn/76.1.100 73. Larsson C.L., Klock K.S., Nordrehaug Astrøm A., Haugejorden O., Johansson G.. Lifestyle-related characteristics of young low-meat consumers and omnivores in Sweden and Norway. J. Adolesc. Health (2002.0) 31 190-198. DOI: 10.1016/S1054-139X(02)00344-0 74. McAfee A.J., McSorley E.M., Cuskelly G.J., Moss B.W., Wallace J.M.W., Bonham M.P., Fearon A.M.. Red meat consumption: An overview of the risks and benefits. Meat Sci. (2010.0) 84 1-13. DOI: 10.1016/j.meatsci.2009.08.029 75. Nucci D., Fatigoni C., Amerio A., Odone A., Gianfredi V.. Red and processed meat consumption and risk of depression: A systematic review and meta-analysis. Int. J. Environ. Res. Public Health (2020.0) 17. DOI: 10.3390/ijerph17186686 76. Carvalho A.M.d., César C.L.G., Fisberg R.M., Marchioni D.M.L.. Excessive meat consumption in Brazil: Diet quality and environmental impacts. Public Health Nutr. (2013.0) 16 1893-1899. DOI: 10.1017/S1368980012003916 77. Jacka F.N., Pasco J.A., Williams L.J., Mann N., Hodge A., Brazionis L., Berk M.. Red meat consumption and mood and anxiety disorders. Psychother. Psychosom. (2012.0) 81 196-198. DOI: 10.1159/000334910 78. Dobersek U., Teel K., Altmeyer S., Adkins J., Wy G., Peak J.. Meat and mental health: A meta-analysis of meat consumption, depression, and anxiety. Crit. Rev. Food Sci. Nutr. (2021.0) 1-18. DOI: 10.1080/10408398.2021.1974336 79. Overweight and Obesity 80. Coulthard H., Sharps M., Cunliffe L., van den Tol A.. Eating in the lockdown during the COVID-19 pandemic; self-reported changes in eating behaviour, and associations with BMI, eating style, coping and health anxiety. Appetite (2021.0) 161 105082. DOI: 10.1016/j.appet.2020.105082 81. Stanton R., To Q.G., Khalesi S., Williams S.L., Alley S.J., Thwaite T.L., Fenning A.S., Vandelanotte C.. Depression, anxiety and stress during COVID-19: Associations with changes in physical activity, sleep, tobacco and alcohol use in Australian adults. Int. J. Environ. Res. Public Health (2020.0) 17 4065. DOI: 10.3390/ijerph17114065 82. Minson J.A., Monin B.. Do-gooder derogation: Disparaging morally motivated minorities to defuse anticipated reproach. Soc. Psychol. Personal. Sci. (2012.0) 3 200-207. DOI: 10.1177/1948550611415695
--- title: Seasonal Changes in the Structure and Function of Gut Microbiota in the Muskrat (Ondatra zibethicus) authors: - Fengcheng Song - Yishu Xu - Peng Peng - Hongxu Li - Ranxi Zheng - Haolin Zhang - Yingying Han - Qiang Weng - Zhengrong Yuan journal: Metabolites year: 2023 pmcid: PMC9966595 doi: 10.3390/metabo13020248 license: CC BY 4.0 --- # Seasonal Changes in the Structure and Function of Gut Microbiota in the Muskrat (Ondatra zibethicus) ## Abstract The gut microbiota plays a crucial role in the nutrition, metabolism, and immune function of the host animal. The muskrat (Ondatra zibethicus) is a typical seasonal breeding animal. The present study performed a metagenomic analysis of cecum contents from muskrats in the breeding and non-breeding seasons. The results indicated that the breeding muskrats and non-breeding muskrats differed in gut microbiota structure and function. During the breeding season, the relative abundance of phylum Bacteroidetes, genus Prevotella, and genus Alistipes increased, while the relative abundance of phylum Firmicutes and phylum Actinobacteria decreased. The muskrat gut microbiota was enriched in the metabolism-related pathways, especially amino acid and vitamin metabolism, and genetically related metabolites in the breeding season. We presumed that the muskrat gut microbiota might seasonally change to secure reproductive activity and satisfy the metabolic demands of different seasons. This study could explore potential mechanisms by which gut microbiota affects reproduction. Moreover, this study may provide a new theoretical basis for the management of muskrat captive breeding. ## 1. Introduction The gut microbiota refers to the diverse microorganisms present in the digestive system of animals, which plays a vital role in animal metabolism, immunity, and reproduction [1,2,3]. Sometimes it is called a “forgotten organ” [4]. For the past few years, depending on the rapid development of bioinformatics, especially the rise of metagenomics, the functions of gut microbiota are being gradually understood [5,6]. The most important function of the intestinal microbiota is the nutritional function, providing energy to the host. Up to $35\%$ of the digestive and metabolic enzymes in mammals are secreted by gut microbiota [7]. The gut microbiota is not static. It is diverse and unstable and is highly susceptible to external environmental influences, such as food [8,9], age [10], disease [11], and living areas [12,13]. The gut microbiota of the same species can vary greatly at different times and in different environments, which can help the host to adapt to its surroundings by influencing host energy metabolism or other aspects. Recently, numerous studies have revealed that animal reproduction is closely linked to gut microbiota. Clostridium scindens American Type Culture Collection 35,704 can convert glucocorticoids to androgens via side-chain cleavage [14]. The gut microbiota can also be involved in gut metabolism and deglucuronidation of dihydrotestosterone (DHT) and testosterone (T) and results in higher DHT levels in the colon of young healthy mice than in germ-free mice [15]. Accordingly, studies for the composition and function of the gut microbiota might be essential for further research on animal reproduction. Seasonal breeding is a phenomenon in which some animals mate only at certain times of the year. Seasonal breeding activity may be influenced mainly by photoperiodism [16]. Photoperiodic changes are sensed by the pineal gland in the brain. It secretes melatonin at night to regulate the secretion of the Gonadotropin-releasing hormone and the Luteinizing hormone, leading to seasonal changes in reproductive activity [17]. Meanwhile, several studies have confirmed that many other factors, such as estrogen, thyroid hormone, kisspeptin, and living environment, affect seasonal breeding animals [18,19,20,21]. According to a recent study, the gut microbiota may regulate seasonal breeding and behavior based on photoperiodic timing in rodents via the Hypothalamic Pituitary Gonadal (HPG) axis, melatonin, and the Kisspeptin/G-protein coupled receptor 54 (GPR54) system in the hypothalamus [22]. This report has aroused our interest in the further investigation of the association between gut microbiota and seasonal breeding. The muskrat (Ondatra zibethicus) is a seasonal breeding animal. It has a breeding season from March to October and a non-breeding season from November to February [23]. During the breeding season, muskrats secrete a substance called musk, which is highly valued for its medicinal properties [24,25]. Most of the previous studies on muskrats have focused on the differences between the reproductive organs or secretory glands of muskrats during the breeding and non-breeding seasons [23,26,27]. Nevertheless, the variation in the composition and function of the gut microbiota of muskrats in different seasons is unknown, and the correlation between gut microbiota and seasonal reproductive phenomena in muskrats has not been explored. The present study utilized high-throughput sequencing by metagenome sequencing to analyze the variation in the structure and function of the muskrat gut microbiota during different seasons. α and β diversity analysis, LEfSe analysis, Metastats analysis, and functional difference analysis on sequencing results were performed to explain differences in gut microbiota. This study aimed to investigate the seasonal variation of composition and function in the gut microbiota of muskrats, further study the secrets of seasonal breeding in muskrats and provide a new theoretical basis for the progress of captive breeding of muskrats. ## 2.1. Sample Collection Adult male muskrats were obtained in January 2022 (the non-breeding season, $$n = 5$$) and May 2022 (the breeding season, $$n = 5$$) from a muskrat breeding base in Xinji, Hebei, China. Based on the references given by the breeding base, the nutritional compositions of the diets of muskrats for each season are as follows. Breeding season: total energy 19.84 MJ/kg, dry matter 36 g, crude protein 7.2 g, crude fiber 5.4 g, and crude fat 1.4 g. Non-breeding season: total energy 15.95 MJ/kg, dry matter 30 g, crude protein 3.6 g, crude fiber 9.0 g, and crude fat 1.1 g. Additionally, vitamins A, D, and E were added to the diet during the breeding season for better reproduction of muskrats [28,29]. Animals were executed after being anesthetized (CO2 inhalation) [30], and the animals were dissected using antiseptic equipment to collect cecum contents. As soon as the cecum contents were collected, they were frozen in liquid nitrogen. All animal experimental procedures were approved by the Institutional Review Board (or Ethics Committee) of Beijing Forestry University (protocol code EAWC_BJFU_2021004). ## 2.2. Metagenomic Sequencing and Data Processing Using the TIANamp Stool DNA Kit (TIANGEN BIOTECH, Beijing, China), the total genomic DNA was extracted from cecum contents. The libraries were constructed through the Enzymic Universal DNAseq Library Prep Kit (Kaitai-Bio, Zhejiang, China), and then the libraries were tested for quality control. The qualifying libraries were sequenced on the Illumina NovaSeq PE150 platforms (Illumina, San Diego, CA, USA). To obtain clean data for further analysis, the raw data were pre-processed to remove low-quality and host sequence contamination. The clean metagenome data assembly was performed using Megahit 1.2.9 [31]. In the gene assembly, we keep the sequences (contigs) with lengths longer than 500 bp [32,33,34] for subsequent analysis. Gene prediction was performed on contigs using MetaGeneMark v.3.38 (Georgia Institute of Technology, Atlanta, GA, USA) [35,36,37], clustering with identity $95\%$, converge $90\%$ [38,39], and statistical Unigene abundance information by Salmon 1.8.0 [40]. ## 2.3. Analysis of Species Composition and Function Gene prediction information was compared with the Non-Redundant Protein Sequence Database (NR), evolutionary genealogy of genes: Non-supervised Orthologous (eggNOG) database [41,42] and Kyoto Encyclopedia of Genes and Genomes (KEGG) PATHWAY database [43,44] using DIAMOND 2.0.7 (Max Planck Institute for Biology, Tübingen, Germany)(blastp, evalue ≤ 1 × 10−5) [45,46]. Species-specific taxonomic information at all levels (phylum, class, order, family, genus, and species) was obtained by MEGAN 6.21.2 (Tübingen University, Tübingen, Germany) [47,48]. The Bray–Curtis distance used for β-diversity was calculated and visualized with principal coordinate analysis (PCoA). Biomarkers with significant variation between groups were ascertained using linear discriminant analysis effect size (LEfSe) [49]. The images were drawn through R 4.1.1 or the online site ImageGP [50]. Analysis of functional pathways with significant differences was conducted using STAMP 2.1.3 [51] and the Metastats method [52]. ## 3.1. Metagenomic Sequencing and Gene Prediction We performed sequencing on muskrat cecum contents samples and obtained a total of 117.50 gigabases (Gbps) of raw data. The raw data were quality controlled to obtain 115.52 Gbps of clean data. The quality control efficiency of the clean data was $98.29\%$ (Table S1). A total of 1,944,649 contigs with the longest length of 585,984 bp were obtained (Table S2). A total of 7,414,843 open reading frames and a total length of 4411.7 Mbp Unigene were obtained for gene prediction (Table S3). The gene information was used to construct a core-pan gene rarefaction curve to evaluate the sample sequencing depth (Figure 1A,B). The curve tends to increase and flatten out with increasing sample size, proving that our sequencing results have largely covered all species. The sample size was reasonable for our study. Meanwhile, the correlation heat map was constructed, and the richness varied greatly between groups, which further proved the reliability of the experiment (Figure 1C). ## 3.2. Gut Microbiota Composition and Difference The following analyses were performed using the abundance information obtained from the species annotations. The analysis of the top 10 microbiota in relative abundance at the phylum level and genus level were plotted separately (Figure 2). At the phylum level, the dominant phylum was Bacteroidetes and Firmicutes in both the breeding and non-breeding seasons, and the mean abundance of Firmicutes was higher in the non-breeding season ($41.30\%$) than in the breeding season ($23.94\%$), while the mean abundance of Bacteroidetes was lower ($21.71\%$) than in the breeding season ($36.19\%$). From the genus perspective, only $24.76\%$ of sequences were unclassified. Prevotella ($6.63\%$), Muribaculaceae_noname ($6.15\%$), Bacteroides ($5.20\%$), and Bacteria_noname ($5.19\%$) had high relative abundance (>$5\%$) during both the breeding and non-breeding seasons. The detailed abundance table was shown in Table S4. α-Diversity analysis showed significantly higher gut microbiota richness in the non-breeding season muskrats than in the breeding season (Figure 3A). PCoA at the genus level can be used to reveal the variation in the gut microbiota of muskrats between the breeding and non-breeding seasons (Figure 3B). Samples from different seasons were well clustered together, with PCo1 explaining $68.34\%$ of the total variation in samples and PCo2 explaining $11.03\%$ of the total variation in samples, indicating that the gut microbiota of muskrats varied at the genus level in different seasons. Prevotella, Bacteroides, and Alistipes were considerably higher in the breeding season than in the non-breeding season, whilst Flavonifractor, Oscillibacter, and Colidextribacter were significantly higher in the non-breeding season than in the breeding season (Figure 3C). LEfSe demonstrated significantly different biomarkers between groups (Figure 3D,E). The non-breeding season microbiota substantially enriched in phylum Firmicutes and phylum Actinobacteria. In contrast, phylum Tenericutes and order family Acetobacteraceae enriched in the breeding season. These results further validated the difference in gut microbiota between the breeding season and the non-breeding season. ## 3.3. Functional Analysis of Gut Microbiota in Muskrat To investigate the metabolic-related changes in the gut microbiota during the breeding and non-breeding seasons, the KEGG pathway (Figure 4A) and eggNOG (Figure 5A) analyses were performed using DIAMOND. KEGG pathways were mainly enriched in metabolism (Figure 4A), including carbohydrate metabolism ($12.65\%$), amino acid metabolism ($9.26\%$), metabolism of cofactors and vitamins ($6.90\%$), energy metabolism ($6.40\%$), and glycan biosynthesis and metabolism ($4.88\%$). Other pathways were genetic information processing ($18.95\%$), environmental information processing ($13.24\%$), cellular processes ($9.66\%$), human diseases ($6.70\%$), and organismal systems ($3.12\%$) (Table S5). Difference analysis of the KEGG level2 pathways ($p \leq 0.05$) showed that there are 10 differentially significant pathways enriched in the non-breeding season, such as the carbohydrate metabolism, cell motility, signal transduction, drug resistance: antimicrobial, endocrine and metabolic disease, infectious disease: parasitic, cellular community-prokaryotes, infectious disease: viral, membrane transport, and signaling molecules and interaction. There are 12 differentially significant pathways enriched in the breeding season, such as the circulatory system, development and regeneration, environmental adaptation, folding, sorting and degradation, neurodegenerative disease, immune disease, glycan biosynthesis and metabolism, lipid metabolism, metabolism of cofactors and vitamins, metabolism of terpenoids and polyketides, translation and transport, and catabolism (Figure 4B). PCoA of specific pathways showed a clear separation between the breeding and non-breeding samples (Figure 4C). The heat map showed the clustering of metabolic pathways with significant differences in relative abundance in the top 30 ($p \leq 0.01$). The breeding season was enriched with many functional pathways related to the amino acid, cofactor, and vitamin metabolism, such as biotin metabolism, D-Glutamine and D-glutamate metabolism, folate biosynthesis, and nicotinate and nicotinamide metabolism (Figure 4D). The plot of all significantly different KEGG pathway analyses was shown in Figure S1. The 25 functions of the eggNOG database were annotated (Figure 5A). The major functions (annotated Unigene > 100,000) were translation, ribosome structure, and biogenesis (J), carbohydrate transport and metabolism (G), cell wall/membrane/envelope biogenesis (M), replication, recombination, and repair (L), amino acid transport and metabolism (E), transcription (K), general function prediction only (R), signal transduction mechanism (T), energy production and conversion (C), coenzyme transport and metabolism (H), defense mechanism (V), posttranslational modification, protein turnover, and chaperones (O), inorganic ion transport and metabolism (P), nucleotide transport and metabolism (F), cell cycle control, cell division, and chromosome partitioning (D), lipid transport and metabolism (L), and function unknown (S). The top 10 significantly different functions obtained using the Metastats method were shown in Figure 5B. During the breeding season, several metabolism-related functions were enriched: coenzyme transport and metabolism (H), lipid transport and metabolism (I), and inorganic ion transport and metabolism (P). The functional analysis illustrated that the gut microbiota of breeding season and non-breeding season muskrats differed significantly in the roles that they play in their hosts. ## 4. Discussion As a typical seasonal breeding animal, changes in the gut microbiota of the muskrat with the seasons are of great curiosity. There are many studies describing the phenomenon of seasonal changes in gut microbiota, such as rats (Rattus norregicus) [53], Tibetan macaques (Macaca thibetana) [54], and Siberian hamsters (Phodopus sungorus) [55]. Seasonal differences were also found in the gut microbiota of forest musk deer (Moschus spp.), an animal capable of musk secretion similar to the muskrat [56]. The present study analyzed the changes in the gut microbiota of muskrats during different breeding seasons. The results showed that there was significant variation in the structure and function of the gut microbiota between the breeding and non-breeding muskrats. These results suggested the possible involvement of gut microbes in meeting the various demands of seasonal breeding in muskrats. Seasonal changes may influence the structure of the muskrats’ gut microbiota. Results of alpha diversity analysis showed that the breeding and non-breeding seasons differed significantly and that the gut microbial richness was higher in the non-breeding season. Previous studies have shown that a higher abundance of gut microbiota helps hosts adapt to the external environment and enhances their resistance to adverse external conditions [57,58]. Therefore, we hypothesized that the increased abundance of gut microbiota in the non-breeding muskrats during winter may enhance resistance to factors such as the cold. The dominant phyla of the gut microbiota of the muskrat were Bacteroidetes and Firmicutes, which was consistent with our previous findings on another rodent seasonal breeder, the wild ground squirrel (Spermophilus dauricus) [59]. The dominant phylum of other rodents such as arctic ground squirrels (Urocitellus parryii) was also Bacteroidetes and Firmicutes [60]. Bacteroidetes and Firmicutes are widespread in animals and are a major component of healthy gut microbiota [61]. Their main function is to break down carbohydrates in the gut and synthesize short-chain fatty acids (SCFAs) to provide energy to the host [62]. Many studies have shown that changes in the ratio of Bacteroidetes and Firmicutes may be associated with obesity and that an increased Firmicutes/Bacteroidetes ratio may lead to obesity and other diseases [63,64]. The abundance of Bacteroidetes increased in the breeding season, whilst Firmicutes significantly increased in the non-breeding season. This change may lead to more stable energy metabolism during the breeding season and safeguard the breeding activity of muskrats. Muskrats in the breeding season have a higher protein requirement, about 7.2 g of protein per 36 g of dry matter, which is greater than in the non-breeding season (winter). The winter diet should not be high in protein and fat to avoid over-fattening the animal [65]. Genus Prevotella and genus Alistipes were enriched in the breeding season, genus Oscillibacter, order Lactobacillales, family Lachnospiraceae, and family Ruminococcaceae were enriched in the non-breeding season. In a previous study, Prevotella and Alistipes were enriched in Brandt’s voles (Lasiopodomys brandtii) with long photoperiods [22], which were reflected in the present study for the breeding season. Alistipes were positively correlated with cholesterol metabolism [66], which may be related to the high-fat, high-protein diet structure of breeding muskrats. Prevotella had a higher diversity, and it has been suggested that it may promote increased glycogen stores [67], while the study by Chen et al. also demonstrated that Prevotella was positively associated with fat accumulation in pigs [68]. Genus Oscillibacter has been reported to be positively associated with a reduction in body weight, which is an important reference for weight control in non-breeding muskrats [69]. The enrichment of some bacteria of order Lactobacillales has been found to cause a decrease in sperm viability, which may be the reason why they were enriched in the non-breeding season rather than the breeding season [70]. Additionally, the family Lachnospiraceae, order Lactobacillales, and family Ruminococcaceae can hydrolyze starch and other sugars to produce butyrate and other SCFAs [71,72]. The non-breeding muskrats with increased carbohydrate content in their diet and reduced total food intake are likely to encounter energy-deficient conditions [65], and these gut microorganisms provide effective energy for muskrats over the winter. The present study also predicted and analyzed the function of the muskrat gut microbiota based on the KEGG and eggNOG databases. The results of the analysis showed that most of the functions of muskrat gut microbiota were enriched in metabolism-related pathways, especially carbohydrate metabolism. Sixty percent of the muskrat’s diet is composed of carbohydrates, which are the main source of energy for the muskrat. Other major pathways involved are amino acid metabolism, metabolism of cofactors and vitamins, energy metabolism, lipid metabolism, etc. There are significant seasonal differences in all these functions. Muskrats, similar to other monogastric animals, can automatically regulate the amount of food that they eat to meet their energy requirements. During the breeding season, the muskrat’s diet contains more animal feed and higher levels of proteins, lipids, and other nutrients, so the metabolism pathways associated with amino acids, lipids, vitamins, and cofactors increase significantly during the breeding season to meet the complex food structure of the muskrat during that season. In contrast, during the non-breeding season, muskrats eat less and have a high proportion of plant-based feeds, while the lower ambient temperature was not conducive to their energy supply, and the enriched carbohydrate metabolic pathways at this time can help muskrats better survive in this period. This study investigated seasonal differences in the gut microbiota of muskrats using a metagenomic approach. The present study initially explored the correlation between gut microbiota and seasonal breeding and speculated on the possible regulatory mechanisms of gut microbiota on seasonal breeding in muskrats. This research can provide a new theoretical basis for muskrat captive breeding and help breeders understand the seasonal changes in muskrat metabolism. They can use this as a basis to better regulate the diet structure of muskrats in different seasons. Furthermore, more work needs to be conducted in the future to clarify the composition and function of the gut microbes in muskrats and to further explore the important role of the gut microbiota in seasonal breeding animals. ## 5. Conclusions There were notable seasonal differences in the species composition and structure of the gut microbiota of muskrats. Gut microbiota richness increased in muskrats during the non-breeding season, which may help muskrats to increase their resistance to the external environment. At the phylum level, the relative abundance of Bacteroidetes was increased and the relative abundance of Firmicutes was decreased in the gut microbiota of breeding muskrats. The genus Prevotella and genus Alistipes enriched in the breeding season, and the genus Oscillibacter, order Lactobacillales, family Lachnospiraceae, and family Ruminococcaceae enriched in the non-breeding season. The muskrat’s gut microbiota was highly enriched in nutrient metabolic pathways. Differential microbiotas during the breeding season were mainly enriched in amino acid, lipid, vitamin, and cofactor metabolism pathways to help muskrats better reproductive activities. This study may provide a new theoretical basis for managing muskrat captive breeding and lay the foundation for further ensuring the efficient breeding of muskrats. ## References 1. Hsiao W.W., Metz C., Singh D.P., Roth J.. The microbes of the intestine: An introduction to their metabolic and signaling capabilities. Endocrinol. Metab. Clin. North Am. (2008) 37 857-871. DOI: 10.1016/j.ecl.2008.08.006 2. Ley R.E., Peterson D.A., Gordon J.I.. Ecological and evolutionary forces shaping microbial diversity in the human intestine. Cell (2006) 124 837-848. DOI: 10.1016/j.cell.2006.02.017 3. Morrison D.J., Preston T.. Formation of short chain fatty acids by the gut microbiota and their impact on human metabolism. Gut Microbes (2016) 7 189-200. DOI: 10.1080/19490976.2015.1134082 4. O’Hara A.M., Shanahan F.. The gut flora as a forgotten organ. EMBO Rep. (2006) 7 688-693. DOI: 10.1038/sj.embor.7400731 5. Chen K., Pachter L.. Bioinformatics for whole-genome shotgun sequencing of microbial communities. PLoS Comput. Biol. (2005) 1 106-112. DOI: 10.1371/journal.pcbi.0010024 6. Handelsman J., Rondon M.R., Brady S.F., Clardy J., Goodman R.M.. Molecular biological access to the chemistry of unknown soil microbes: A new frontier for natural products. Chem Biol. (1998) 5 R245-R249. DOI: 10.1016/S1074-5521(98)90108-9 7. Gill S.R., Pop M., Deboy R.T., Eckburg P.B., Turnbaugh P.J., Samuel B.S., Gordon J.I., Relman D.A., Fraser-Liggett C.M., Nelson K.E.. Metagenomic analysis of the human distal gut microbiome. Science (2006) 312 1355-1359. DOI: 10.1126/science.1124234 8. Fan C., Zhang L., Jia S., Tang X., Fu H., Li W., Liu C., Zhang H., Cheng Q., Zhang Y.. Seasonal variations in the composition and functional profiles of gut microbiota reflect dietary changes in plateau pikas. Integr. Zool (2022) 17 379-395. DOI: 10.1111/1749-4877.12630 9. Guo N., Wu Q., Shi F., Niu J., Zhang T., Degen A.A., Fang Q., Ding L., Shang Z., Zhang Z.. Seasonal dynamics of diet-gut microbiota interaction in adaptation of yaks to life at high altitude. NPJ Biofilms Microbiomes (2021) 7 38. DOI: 10.1038/s41522-021-00207-6 10. O’Toole P.W.. Changes in the intestinal microbiota from adulthood through to old age. Clin. Microbiol. Infect. (2012) 18 44-46. DOI: 10.1111/j.1469-0691.2012.03867.x 11. Gomaa E.Z.. Human gut microbiota/microbiome in health and diseases: A review. Antonie Van Leeuwenhoek (2020) 113 2019-2040. DOI: 10.1007/s10482-020-01474-7 12. Sun Y., Sun Y., Shi Z., Liu Z., Zhao C., Lu T., Gao H., Zhu F., Chen R., Zhang J.. Gut Microbiota of Wild and Captive Alpine Musk Deer (Moschus chrysogaster). Front. MicroBiol. (2019) 10 3156. DOI: 10.3389/fmicb.2019.03156 13. Wong S., Rawls J.F.. Intestinal microbiota composition in fishes is influenced by host ecology and environment. Mol. Ecol. (2012) 21 3100-3102. DOI: 10.1111/j.1365-294X.2012.05646.x 14. Ridlon J.M., Ikegawa S., Alves J.M.P., Zhou B., Kobayashi A., Iida T., Mitamura K., Tanabe G., Serrano M., De Guzman A.. J. Lipid Res. (2013) 54 2437-2449. DOI: 10.1194/jlr.M038869 15. Colldén H., Landin A., Wallenius V., Elebring E., Fändriks L., Nilsson M.E., Ryberg H., Poutanen M., Sjögren K., Vandenput L.. The gut microbiota is a major regulator of androgen metabolism in intestinal contents. Am. J. Physiol Endocrinol. Metab. (2019) 317 E1182-E1192. DOI: 10.1152/ajpendo.00338.2019 16. Matthews C.D., Guerin M.V., Deed J.R.. Melatonin and photoperiodic time measurement: Seasonal breeding in the sheep. J. Pineal Res. (1993) 14 105-116. DOI: 10.1111/j.1600-079X.1993.tb00493.x 17. Malpaux B., Tricoire H., Mailliet F., Daveau A., Migaud M., Skinner D.C., Pelletier J., Chemineau P.. Melatonin and seasonal reproduction: Understanding the neuroendocrine mechanisms using the sheep as a model. Reprod Suppl. (2002) 59 167-179. PMID: 12698980 18. Dardente H., Lomet D., Robert V., Decourt C., Beltramo M., Pellicer-Rubio M.T.. Seasonal breeding in mammals: From basic science to applications and back. Theriogenology (2016) 86 324-332. DOI: 10.1016/j.theriogenology.2016.04.045 19. Nakao N., Ono H., Yoshimura T.. Thyroid hormones and seasonal reproductive neuroendocrine interactions. Reproduction (2008) 136 1-8. DOI: 10.1530/REP-08-0041 20. Smith J.T.. The role of kisspeptin and gonadotropin inhibitory hormone in the seasonal regulation of reproduction in sheep. Domest. Anim. Endocrinol. (2012) 43 75-84. DOI: 10.1016/j.domaniend.2011.11.003 21. Smith J.T., Clarke I.J.. Seasonal breeding as a neuroendocrine model for puberty in sheep. Mol. Cell Endocrinol. (2010) 324 102-109. DOI: 10.1016/j.mce.2010.03.007 22. Zhu H., Li G., Liu J., Xu X., Zhang Z.. Gut microbiota is associated with the effect of photoperiod on seasonal breeding in male Brandt’s voles (Lasiopodomys brandtii). Microbiome (2022) 10 194. DOI: 10.1186/s40168-022-01381-1 23. Zhang F., Liu Q., Wang Z., Xie W., Sheng X., Zhang H., Yuan Z., Han Y., Weng Q.. Seasonal Expression of Oxytocin and Oxytocin Receptor in the Scented Gland of Male Muskrat (Ondatra zibethicus). Sci. Rep. (2017) 7 16627. DOI: 10.1038/s41598-017-16973-3 24. Chen Y.S., Jin S.D., Dong W.C., Tong Y.R., Hua S.F., Zhao H.Y.. Effect of muskrat on the cardiovascular system in anesthetized dogs. Zhongguo Zhong Yao Za Zhi (1989) 14 241-256. PMID: 2505807 25. Li Y., Zhang T., Fan M., Zhou J., Yang S., Zhang M., Qi L., Lin S., Hu D., Liu S.. Comparison of amino acid profiles and metabolic gene expression in muskrat scented glands in secretion and non-secretion season. Sci. Rep. (2017) 7 41158. DOI: 10.1038/srep41158 26. Li Y., Zhang T., Zhou J., Yang S., Fan M., Sun X., Zhang M., Xu S., Cha M., Hu X.. Transcriptome analysis of muskrat scented glands degeneration mechanism. PLoS ONE (2017) 12. DOI: 10.1371/journal.pone.0176935 27. Cao H., Wang L., Zhang S., Lu L., Sheng X., Han Y., Yuan Z., Weng Q.. Seasonal Expression of Prolactin Receptor in the Scented Gland of Male Muskrat (Ondatra zibethicus). Sci. Rep. (2015) 5 15036. DOI: 10.1038/srep15036 28. Hua S.F.. Muskrat Breeding Technology (1989) 52-53 29. Wang Z.Y.. Practical Breeding Techniques for Muskrat (2010) 21-30 30. Shomer N.H., Allen-Worthington K.H., Hickman D.L., Jonnalagadda M., Newsome J.T., Slate A.R., Valentine H., Williams A.M., Wilkinson M.. Review of Rodent Euthanasia Methods. J. Am. Assoc. Lab Anim. Sci. (2020) 59 242-253. DOI: 10.30802/AALAS-JAALAS-19-000084 31. Li D., Liu C.-M., Luo R., Sadakane K., Lam T.-W.. MEGAHIT: An ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics (2015) 31 1674-1676. DOI: 10.1093/bioinformatics/btv033 32. Zeller G., Tap J., Voigt A.Y., Sunagawa S., Kultima J.R., Costea P.I., Amiot A., Böhm J., Brunetti F., Habermann N.. Potential of fecal microbiota for early-stage detection of colorectal cancer. Molecular Systems Biology (2014) 10 766. DOI: 10.15252/msb.20145645 33. Qin J., Li R., Raes J., Arumugam M., Burgdorf K.S., Manichanh C., Nielsen T., Pons N., Levenez F., Yamada T.. A human gut microbial gene catalogue established by metagenomic sequencing. Nature (2010) 464 59-65. DOI: 10.1038/nature08821 34. Nielsen H.B., Almeida M., Juncker A.S., Rasmussen S., Li J., Sunagawa S., Plichta D.R., Gautier L., Pedersen A.G., Le Chatelier E.. Identification and assembly of genomes and genetic elements in complex metagenomic samples without using reference genomes. Nat. Biotechnol. (2014) 32 822-828. DOI: 10.1038/nbt.2939 35. Daniel R.M., Alison S.W., Shinichi S., Aino I.J., Aino I.J., Michelle M.Y.C., Manimozhiyan A., Jeroen R., Peer B.. Assessment of metagenomic assembly using simulated next generation sequencing data. PLoS ONE (2012). DOI: 10.1371/journal.pone.0031386 36. Julia O., Allyson L.B., Clay D., Sean C., Nisc Comparative Sequencing P., Heidi H.K., Julia A.S.. Biogeography and individuality shape function in the human skin metagenome. Nature (2014) 514 59-64. DOI: 10.1038/nature13786 37. Jun L., Junhua L., Huijue J., Xianghang C., Huanzi Z., Qiang F., Shinichi S., Manimozhiyan A., Jens Roat K., Edi P.. An integrated catalog of reference genes in the human gut microbiome. Nat. Biotechnol. (2014) 32 834-841. DOI: 10.1038/nbt.2942 38. Georg Z., Julien T., Anita Y.V., Shinichi S., Jens Roat K., Paul I.C., Aurelien A., Jürgen B., Francesco B., Nina H.. Potential of fecal microbiota for early-stage detection of colorectal cancer. Mol. Syst. Biol. (2014) 10 766. DOI: 10.15252/msb.20145645 39. Nan Q., Fengling Y., Ang L., Edi P., Yanfei C., Li S., Jing G., Jian Y., Lingjiao W., Jiawei Z.. Alterations of the human gut microbiome in liver cirrhosis. Nature (2014) 513 59-64. DOI: 10.1038/nature13568 40. Rob P., Geet D., Michael I.L., Rafael A.I., Carl K.. Salmon provides fast and bias-aware quantification of transcript expression. Nat. Methods (2017) 14 417-419. DOI: 10.1038/nmeth.4197 41. Powell S., Forslund K., Szklarczyk D., Trachana K., Roth A., Huerta-Cepas J., Gabaldón T., Rattei T., Creevey C., Kuhn M.. eggNOG v4.0: Nested orthology inference across 3686 organisms. Nucleic Acids Res. (2014) 42 D231-D239. DOI: 10.1093/nar/gkt1253 42. Jensen L.J., Julien P., Kuhn M., von Mering C., Muller J., Doerks T., Bork P.. eggNOG: Automated construction and annotation of orthologous groups of genes. Nucleic Acids Res. (2008) 36 D250-D254. DOI: 10.1093/nar/gkm796 43. Kanehisa M., Goto S.. KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Res. (2000) 28 27-30. DOI: 10.1093/nar/28.1.27 44. Kanehisa M., Goto S., Sato Y., Kawashima M., Furumichi M., Tanabe M.. Data, information, knowledge and principle: Back to metabolism in KEGG. Nucleic Acids Res. (2014) 42 D199-D205. DOI: 10.1093/nar/gkt1076 45. Buchfink B., Xie C., Huson D.H.. Fast and sensitive protein alignment using DIAMOND. Nat. Methods (2015) 12 59-60. DOI: 10.1038/nmeth.3176 46. Buchfink B., Reuter K., Drost H.-G.. Sensitive protein alignments at tree-of-life scale using DIAMOND. Nat. Methods (2021) 18 366-368. DOI: 10.1038/s41592-021-01101-x 47. Karlsson F.H., Tremaroli V., Nookaew I., Bergström G., Behre C.J., Fagerberg B., Nielsen J., Bäckhed F.. Gut metagenome in European women with normal, impaired and diabetic glucose control. Nature (2013) 498 99-103. DOI: 10.1038/nature12198 48. Bağcı C., Patz S., Huson D.H.. DIAMOND+MEGAN: Fast and Easy Taxonomic and Functional Analysis of Short and Long Microbiome Sequences. Curr. Protoc. (2021) 1 e59. DOI: 10.1002/cpz1.59 49. Segata N., Izard J., Waldron L., Gevers D., Miropolsky L., Garrett W.S., Huttenhower C.. Metagenomic biomarker discovery and explanation. Genome Biol. (2011) 12 R60. DOI: 10.1186/gb-2011-12-6-r60 50. Chen T., Liu Y.-X., Huang L.. ImageGP: An easy-to-use data visualization web server for scientific researchers. iMeta (2022) 1 e5. DOI: 10.1002/imt2.5 51. Parks D.H., Tyson G.W., Hugenholtz P., Beiko R.G.. STAMP: Statistical analysis of taxonomic and functional profiles. Bioinformatics (2014) 30 3123-3124. DOI: 10.1093/bioinformatics/btu494 52. White J.R., Nagarajan N., Pop M.. Statistical methods for detecting differentially abundant featuRes. in clinical metagenomic samples. PLoS Comput. Biol. (2009) 5. DOI: 10.1371/journal.pcbi.1000352 53. Liu J., Song Y., Lu X., Chen T., Guo W., Fan Z., Kang X., Wang Y., Wang Y.. Seasonal Variation Influences on Intestinal Microbiota in Rats. Curr. MicroBiol. (2018) 75 1006-1010. DOI: 10.1007/s00284-018-1480-6 54. Sun B., Wang X., Bernstein S., Huffman M.A., Xia D.P., Gu Z., Chen R., Sheeran L.K., Wagner R.S., Li J.. Marked variation between winter and spring gut microbiota in free-ranging Tibetan Macaques (Macaca thibetana). Sci. Rep. (2016) 6 26035. DOI: 10.1038/srep26035 55. Shor E.K., Brown S.P., Freeman D.A.. A novel role for the pineal gland: Regulating seasonal shifts in the gut microbiota of Siberian hamsters. J. Pineal Res. (2020) 69 e12696. DOI: 10.1111/jpi.12696 56. Jiang F., Gao H., Qin W., Song P., Wang H., Zhang J., Liu D., Wang D., Zhang T.. Marked Seasonal Variation in Structure and Function of Gut Microbiota in Forest and Alpine Musk Deer. Front MicroBiol. (2021) 12 699797. DOI: 10.3389/fmicb.2021.699797 57. Stoffel M.A., Acevedo-Whitehouse K., Morales-Durán N., Grosser S., Chakarov N., Krüger O., Nichols H.J., Elorriaga-Verplancken F.R., Hoffman J.I.. Early sexual dimorphism in the developing gut microbiome of northern elephant seals. Mol. Ecol. (2020) 29 2109-2122. DOI: 10.1111/mec.15385 58. Videvall E.. Microbiome maturation during a unique developmental window. Mol. Ecol. (2020) 29 1941-1943. DOI: 10.1111/mec.15436 59. Yang X., Yao Y., Zhang X., Zhong J., Gao F., Zhang H., Han Y., Weng Q., Yuan Z.. Seasonal Changes in the Distinct Taxonomy and Function of the Gut Microbiota in the Wild Ground Squirrel (Spermophilus dauricus). Animals (2021) 11. DOI: 10.3390/ani11092685 60. Stevenson T.J., Duddleston K.N., Buck C.L.. Effects of season and host physiological state on the diversity, density, and activity of the arctic ground squirrel cecal microbiota. Appl. Environ. MicroBiol. (2014) 80 5611-5622. DOI: 10.1128/AEM.01537-14 61. Jandhyala S.M., Talukdar R., Subramanyam C., Vuyyuru H., Sasikala M., Nageshwar Reddy D.. Role of the normal gut microbiota. World J. Gastroenterol (2015) 21 8787-8803. DOI: 10.3748/wjg.v21.i29.8787 62. Macfarlane S., Macfarlane G.T.. Regulation of short-chain fatty acid production. Proc. Nutr. Soc. (2003) 62 67-72. DOI: 10.1079/PNS2002207 63. Indiani C., Rizzardi K.F., Castelo P.M., Ferraz L.F.C., Darrieux M., Parisotto T.M.. Childhood Obesity and Firmicutes/Bacteroidetes Ratio in the Gut Microbiota: A Systematic Review. Child Obes. (2018) 14 501-509. DOI: 10.1089/chi.2018.0040 64. Magne F., Gotteland M., Gauthier L., Zazueta A., Pesoa S., Navarrete P., Balamurugan R.. The Firmicutes/Bacteroidetes Ratio: A Relevant Marker of Gut Dysbiosis in Obese Patients?. Nutrients (2020) 12. DOI: 10.3390/nu12051474 65. Feng X.L.. Characteristics of nutritional metabolism and energy balance of muskrats. Mod. Livest. Poult. Breed. Ind. (2012) 10 48-49. DOI: 10.14070/j.cnki.15-1150.2012.10.050 66. Le Roy T., Lécuyer E., Chassaing B., Rhimi M., Lhomme M., Boudebbouze S., Ichou F., Haro Barceló J., Huby T., Guerin M.. The intestinal microbiota regulates host cholesterol homeostasis. BMC Biol. (2019) 17. DOI: 10.1186/s12915-019-0715-8 67. Kovatcheva-Datchary P., Nilsson A., Akrami R., Lee Y.S., De Vadder F., Arora T., Hallen A., Martens E., Björck I., Bäckhed F.. Dietary Fiber-Induced Improvement in Glucose Metabolism Is Associated with Increased Abundance of Prevotella. Cell Metab. (2015) 22 971-982. DOI: 10.1016/j.cmet.2015.10.001 68. Chen C., Fang S., Wei H., He M., Fu H., Xiong X., Zhou Y., Wu J., Gao J., Yang H.. Prevotella copri increases fat accumulation in pigs fed with formula diets. Microbiome (2021) 9 175. DOI: 10.1186/s40168-021-01110-0 69. Konikoff T., Gophna U.. Oscillospira: A Central, Enigmatic Component of the Human Gut Microbiota. Trends MicroBiol. (2016) 24 523-524. DOI: 10.1016/j.tim.2016.02.015 70. Wang H., Chen T., Chen Y., Luo T., Tan B., Chen H., Xin H.. Evaluation of the inhibitory effects of vaginal microorganisms on sperm motility in vitro. Exp. Ther. Med. (2020) 19 535-544. DOI: 10.3892/etm.2019.8237 71. Vacca M., Celano G., Calabrese F.M., Portincasa P., Gobbetti M., De Angelis M.. The Controversial Role of Human Gut Lachnospiraceae. Microorganisms (2020) 8. DOI: 10.3390/microorganisms8040573 72. Zhuang M., Shang W., Ma Q., Strappe P., Zhou Z.. Abundance of Probiotics and Butyrate-Production Microbiome Manages Constipation via Short-Chain Fatty Acids Production and Hormones Secretion. Mol. Nutr. Food Res. (2019) 63 e1801187. DOI: 10.1002/mnfr.201801187
--- title: The Associations between Meeting 24-Hour Movement Guidelines (24-HMG) and Mental Health in Adolescents—Cross Sectional Evidence from China authors: - Lin Luo - Xiaojin Zeng - Yunxia Cao - Yulong Hu - Shaojing Wen - Kaiqi Tang - Lina Ding - Xiangfei Wang - Naiqing Song journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9966615 doi: 10.3390/ijerph20043167 license: CC BY 4.0 --- # The Associations between Meeting 24-Hour Movement Guidelines (24-HMG) and Mental Health in Adolescents—Cross Sectional Evidence from China ## Abstract [1] Background: This study determined the prevalence of adolescents that meet 24-HMGs alone and in combination, and their association with the risk of developing adolescent anxiety and depression. [ 2] Methods: Participants were drawn from 9420 K8 grade adolescents (age 14.53 ± 0.69 years; $54.78\%$ boys) from the China Education Tracking Survey (CEPS) 2014–2015 tracking data. Data on depression and anxiety were collected from the results of the questionnaire in the CEPS for the adolescent mental health test. Compliance with the 24-HMG was defined as: physical activity time (PA) ≥ 60 min/day was defined as meeting the PA. Screen time (ST) ≤ 120 min/day was defined as meeting the ST. Adolescents aged 13 years achieved 9–11 h of sleep per night and adolescents aged 14–17 years achieved 8–10 h of sleep per night, defined as meeting sleep. Logistic regression models were used to examine the association between meeting and not meeting the recommendations and the risk of depression and anxiety in adolescents. [ 3] Results: Of the sample studied, $0.71\%$ of adolescents met all three recommendations, $13.54\%$ met two recommendations and $57.05\%$ met one recommendation. Meeting sleep, meeting PA+ sleep, meeting ST + sleep, and meeting PA + ST + sleep were associated with a significantly lower risk of anxiety and depression in adolescents. Logistic regression results showed that differences in the effects of gender on the odds ratio (ORs) for depression and anxiety in adolescents were not significant. [ 4] Conclusions: This study determined the risk of developing depression and anxiety in adolescents who met the recommendations for 24-HMG alone and in combination. Overall, meeting more of the recommendations in the 24-HMGs was associated with lower anxiety and depression risk outcomes in adolescents. For boys, reducing the risk of depression and anxiety can be prioritised by meeting PA + ST + sleep, meeting ST + sleep and meeting sleep in the 24-HMGs. For girls, reducing the risk of depression and anxiety may be preferred by meeting PA + ST + sleep or meeting PA+ sleep and meeting sleep in 24-HMGs. However, only a small proportion of adolescents met all recommendations, highlighting the need to promote and support adherence to these behaviours. ## 1. Introduction Previous studies have extensively investigated the relationship between adequate physical activity, limited screen time and adequate sleep time, as well as the physical and mental health of adolescents [1,2,3,4,5]. More recently, physical activity time, screen time and sleep time have been proposed jointly in relevant studies as three key components of the 24-hour movement guidelines because of their interactions and interdependence [6,7]. Using the subjects’ use of time of day as an entry point for the study [8], the researchers proposed a framework for the study. This framework integrates physical activity time, screen time and sleep time, as this combination provides more reliable information on health and behaviour changes [9]. In Australia and Canada, 24-hour movement guidelines (24-HMG) for children and adolescents containing three recommendations for physical activity time, screen time and sleep time were published in 2015 and 2016, respectively [6,7]. Both versions of the 24-HMG recommend that children and adolescents engage in at least 60 min of moderate-to-vigorous physical activity per day. Screen time should not exceed 2 h per day. Children and adolescents aged 5–13 years should have at least 9–11 h of sleep per night, and adolescents aged 14–17 years should have at least 8–10 h of sleep per night [10]. A growing amount of research evidence showed that adolescents who consistently meets 24-HMG can achieve additional health benefits, such as preventing obesity, improving emotional health, improving metabolic indicators and promoting social adjustment [11,12,13]. Depression and anxiety are common mental health problems among adolescents, and depression is often accompanied by anxiety. From 2001 to 2020, the global prevalence of depression among adolescents increased by $34\%$, with the highest prevalence of depression in the Middle East, Africa and Asia [14]. In a survey of 38 countries/regions, Erskine et al. reported a $6.2\%$ prevalence of depression and a $3.2\%$ prevalence of anxiety in children and adolescents aged 5–17 years [15]. Previous studies have shown that the three recommendations of adequate physical activity (PA ≥ 60 min/day) [16,17], limited screen time (ST ≤ 120 min/day) [18,19] and good sleep duration (meeting age-specific needs) [20,21] in the 24-HMG are associated with lower depression and anxiety symptoms in adolescents, respectively. For example, a study by McDowell et al. [ 2017] reported that adolescents aged 14–17 years ($$n = 481$$) who achieved PA ≥ 60 min/day had $30\%$ lower odds of depression and $46\%$ lower odds of anxiety compared to adolescents with PA < 60 min/day (insufficient time or insufficient days) [22]. A meta-analysis study conducted by Mougharbel et al. [ 2020] showed that higher screen time was associated with more severe depressive symptoms in adolescents who engaged in ST > 120 min/day. However, its relationship with anxiety symptoms was not stable and may be related to gender [23]. A study by Baum et al. [ 2014] found that insufficient sleep duration affected anxiety and fatigue in healthy adolescents aged 14–17 years [24]. Ojio et al. [ 2020] showed that weekday sleep duration was independently associated with levels of depressive symptoms in Japanese adolescents aged 12–15 years ($$n = 942$$) [25]. Thus, adolescents who have adequate physical activity time, less screen time and adequate sleep time may have better mental health and be less likely to experience anxiety and depressive symptoms [12]. However, these behavioural studies mainly examined the recommendations in isolation and not the combined effects on adolescent depression and anxiety. With the release of the 24-HMG, more researchers have begun to examine adolescent mental health from the perspective of integrating physical activity time, screen time and sleep time, rather than examining multiple determinants in isolation. For example, a systematic evaluation by Sampasa-Kanyinga et al. [ 2020] of 115,540 adolescents aged 5–17 years in 12 countries, including the USA and the UK, found that meeting the three recommendations in the 24-HMG held a good association with better mental health indicators (depression and other mental health indicators) [26]. The study by Lu et al. [ 2021] of 5357 adolescents in grades K4 and K5 in China found that meeting the 24-HMG recommendations was associated with better mental health outcomes [27]. It has been suggested that holistic surveys that integrate physical activity, screen time and sleep time may improve their validity in predicting individual health outcomes [8]. From a practical application perspective, it may be possible to compare the differences in meeting the impact of different recommendations on adolescent mental health [28]. Despite the benefits of meeting the 24-HMG on adolescent mental health, research evidence is still very limited, which may reduce the generalizability of the findings. Therefore, there is a need to replicate and extend this study in other culturally diverse populations and across age groups to further confirm the association between 24-HMG and adolescent mental health. While previous studies have conducted preliminary explorations of these factors as key to adolescent mental health in small samples [26], the use of a nationally representative sample in this study will help to improve the generalisability of the findings, which can inform policy makers when designing effective interventions. Several previous studies have found that there may be gender differences in the association of these factors with mental health. Therefore, the aims of this study included: (a) to examine the relationship between the 24-HMG recommendations met by Chinese adolescents and the risk of depression and anxiety disorders; and (b) to determine whether this association was influenced by gender. ## 2.1. Study Design and Participants This study analysed primary data from the China Education Tracking Survey (CEPS), a national cohort study designed to document the attributes of developmental and educational experiences of Chinese adolescents. The CEPS study adopted a multi-stage stratified design with county (or equivalent administrative area), school and class as primary, secondary and tertiary sampling units, respectively. Primary sampling units were stratified by region and size of the migrant population, with counties in Shanghai or other regions with high migrant populations being oversampled. Within each sampled county, four schools were sampled using the probability proportion method. the CEPS team developed survey weights to address the probability of inequality of choice [29]. This study used the CEPS data from 2014–2015. The study involved data on youth and parent demographic characteristics, physical exercise time, screen time, sleep time, and depression and anxiety scores. The 2014–2015 CEPS youth (grade K8) sample size was 10,750. A total of 301 classes were surveyed in the 2014–2015 survey. This included a sample of 9653 for physical activity time, 9865 for screen time, 9649 for sleep time and a sample size of 9888 for the mental health data. Excluding data with missing physical exercise time, screen time, sleep time, depression and anxiety scale scores, the final study sample size for this study was 9420. This study used the free public database of the CEPS. The ethical review of the CEPS was approved by the ethics committee of the People’s University of China. Prior to the CEPS, all participants signed a consent form and their parents also signed a consent form to participate in the study. Students were not rewarded for their participation. Details of the CEPS data, and information on its ethical review, can be found at http://ceps.ruc.edu.cn/, accessed on 19 January 2023. ## 2.2. Measures This study used the Canadian 24-HMG as a reference standard. The contents of the Canadian 24-HMG are shown in Table 1. This study used time for moderate to vigorous physical exercise from the physical activity recommendations, sleep time from the sleep recommendations and screen time from the sedentary recommendations [30]. Then, we analysed their relationship with depression and anxiety in adolescents. ## 2.2.1. Physical Activity Time The CEPS investigated the frequency of physical activity per week and the duration of a single session of physical activity in adolescents, and we referenced the Canadian 24-HMG, where PA ≥ 60 min/day was defined as meeting physical activity guidelines (meeting PA). ## 2.2.2. Screen Time The CEPS investigated the daily screen time of adolescents from Monday to the weekend. The amount of time adolescents spent watching television, surfing the internet and playing computer games each day was aggregated to obtain the amount of screen time adolescents spent each day. Referring to the Canadian 24-HMG, screen time ≤ 120 min/day was defined as meeting sedentary guidelines (meeting ST). ## 2.2.3. Sleep Time The CEPS surveyed adolescents’ self-reported daily sleep duration. Referring to the Canadian 24-HMG, satisfying sleep was defined as sleeping at least 9–11 h per night for 13 year olds and 8–10 h per night for 14–17 year olds. ## 2.2.4. Depression and Anxiety Scores The CEPS measures anxiety and depression in adolescents with a total of nine items. The scale was designed with reference to the generalized anxiety scale (GAD-7) [31,32] and the health questionnaire depressive symptom scale (PHQ-9) [33]. Three of the questions measured anxiety, and Cronbach’s alpha coefficient for this questionnaire was 0.815. Six questions measured depression, and Cronbach’s alpha coefficient for this questionnaire was 0.918. Raw scores for the anxiety-related questions were summed to obtain a total score for anxiety. The raw scores for the depression-related questions were summed to obtain a total score for depression. Referring to the previous literature, we defined the score at position P75 as the critical score. A total anxiety score exceeding the sample P75 score was defined as having anxiety symptoms. A total depression score exceeding the sampleP75 score was defined as having depressive symptoms [34]. ## 2.2.5. Covariates Study participants were asked to self-report demographic data, including sex, age, ethnicity, single child, residence, father’s highest education, mother’s highest education, body mass index (BMI) and perceived household economic status. BMI was calculated by dividing weight (kg) by the square of height (m) [35]. In this study, WS/T 586-2018 screening criteria for overweight and obesity in school-aged children and adolescents were used. This is a set of Chinese national standards (WS/T 586-2018) recommended by the National Health Council of China for adolescents aged 6 to 18 years [36]. BMI is classified as not overweight (including low and normal weight) and overweight (including overweight and obesity). ## 2.3. Statistical Analysis All data analyses were carried out using Stata 17.0 software. Descriptive statistics were used to characterise the sample. Frequency percentages were used to describe the categorical variables. Continuous variables were tested with the Shapiro–Wilk test, satisfying the normality distribution, and the variables were described by the mean ± SD. Gender differences between variables were performed using chi-square tests or t-tests. Logistic regression analysis was used to test the relationship between physical activity time (PA ≥ 60 min/day), screen time (ST ≤ 120 min/day) and sleep time (9–11 h at age 13, 14–17 years—8–10 h) in meeting the 24-HMG and the occurrence of anxiety and depression. Adolescents were divided into eight groups based on their meeting of the 24-HMG recommendations: meeting none, only meeting PA, only meeting ST, only meeting sleep, meeting PA + ST, meeting PA + sleep, meeting ST + sleep, meeting PA + ST + sleep. All logistic regression models were adjusted for covariates and analysed extraordinarily to see if this association was affected by gender. Statistical significance was defined as $p \leq 0.05.$ ## 3.1. Sample Characteristics The final study sample size for this study was 9420 Chinese adolescents in grade K8. The mean age was 14.53 ± 0.69 years. The overall sample consisted of $54.81\%$ boys and $44.86\%$ single children (Table 2). Participants had a $29.44\%$ prevalence of anxiety and a $26.93\%$ prevalence of depression. The proportion meeting the recommendations for physical activity time, screen time and sleep time were $5.31\%$, $20.32\%$ and $60.54\%$, respectively. Only $0.71\%$ of participants met all three recommendations in the 24-HMG, $13.54\%$ met any two recommendations, $57.05\%$ met one guideline and $28.70\%$ did not meet any of the recommendations. A higher proportion of boys than girls met all three recommendations in the 24-HMG.The differences between boys and girls were significant in terms of age, ethnicity, single child, residence, father’s highest education, mother’s highest education, perceived household economic status, BMI, meeting PA, meeting ST, meeting sleep and meeting 24-HMG categories. There were no significant differences in the prevalence of anxiety and depression. A description of participant characteristics by gender is shown in Table 2. ## 3.2. Prevalence of Depression and Anxiety among Adolescents in Different 24-HMG Categories The prevalence of depression and anxiety in different groups of adolescents is shown in Figure 1 and Figure 2. Overall, the three groups with the lowest incidence of depression were, in order, the meeting PA + ST + sleep, meeting PA + sleep group, and meeting ST + sleep group. The three groups with the lowest incidence of anxiety were, in order, the meeting PA + ST + sleep, meeting PA + ST group, and meeting ST + sleep group. Among boys, the three groups with the lowest prevalence of depression were, in order, the meeting PA + ST + sleep group ($9.43\%$), the meeting ST + sleep group ($21.15\%$) and the meeting PA + sleep group ($22.68\%$). Among the girls, the three groups with the lowest prevalence of depression, in order, were the meeting PA + ST group ($11.11\%$), the meeting PA + sleep group ($212.50\%$) and the meeting ST + sleep group ($19.56\%$). Among boys, the three groups with the lowest prevalence of anxiety were, in order, the meeting PA + ST + sleep group ($7.55\%$), the meeting ST + sleep group ($22.44\%$) and the meeting PA + sleep group ($23.71\%$). Among the girls, the three groups with the lowest prevalence of anxiety were, in order, the meeting PA + sleep group ($9.8\%$), the meeting ST + sleep group ($21.22\%$) and the meeting PA + ST + sleep group ($21.43\%$). ## 3.3. Meeting One of the 24-HMG Recommendations in Relation to Anxiety and Depression Table 3 presents the relationship between meeting one of 24-HMG recommendations and adolescent anxiety and depression obtained through a binary logistic regression. The results of the study showed that the ORs for depression and anxiety in the only meeting sleep group were 0.65 ($p \leq 0.01$) and 0.68 ($p \leq 0.01$), respectively, compared to the not-meeting group. This indicated that the effect of only meeting sleep on depression was more significant than the effect on anxiety. Logistic regression results showed that differences in the effects of gender on the ORs for depression and anxiety in adolescents were not significant. In the boys sample, the ORs for depression and anxiety in the only meeting sleep group were 0.60 ($p \leq 0.01$) and 0.64 ($p \leq 0.01$), respectively. In the girls sample, the ORs for depression and anxiety in the only meeting sleep group were 0.70 ($p \leq 0.01$) and 0.73 ($p \leq 0.01$), respectively. ## 3.4. Meeting Two of the 24-HMG Recommendations in Relation to Anxiety and Depression Table 4 presents the relationship between meeting two 24-HMG recommendations and adolescents’ anxiety and depression obtained through a binary logistic regression. The results of the study showed that the ORs for depression and anxiety in the meeting PA + sleep group were 0.52 and 0.59, respectively, compared to the meeting none group ($p \leq 0.01$). This indicates that the effect of meeting PA and sleep guidelines on depression was more significant than the effect on anxiety. Compared to the meeting none group, the ORs for depression and anxiety in the meeting ST + sleep group were 0.46 and 0.57, respectively ($p \leq 0.01$). Logistic regression results showed that differences in the effects of gender on the ORs for depression and anxiety in adolescents were not significant. In the boys sample, the ORs for depression and anxiety in the meeting PA + sleep group were 0.55 and 0.64, respectively, compared to the meeting none group ($p \leq 0.01$). This indicates that meeting PA + sleep had a more significant effect on depression than on anxiety. The ORs for depression and anxiety in the meeting ST + sleep group were 0.42 and 0.54, respectively, compared to the meeting none group ($p \leq 0.01$). For boys, this indicates that the effect of meeting ST and sleep guidelines on depression was more significant than the effect on anxiety. The effect of meeting ST and sleep guidelines on depression and anxiety was more significant than the effect of meeting PA and sleep guidelines on depression and anxiety. In the girls sample, the ORs for depression and anxiety in the meeting PA + sleep group were 0.29 and 0.24, respectively, compared to the meeting none group ($p \leq 0.01$). This indicates that the effect of meeting PA and sleep guidelines on anxiety was more significant than that on depression. Compared to the meeting none group, the ORs for depression and anxiety in the meeting ST + sleep group were 0.42 and 0.59, respectively ($p \leq 0.01$). For girls, This indicates that the effect of meeting ST and sleep guidelines on depression was more significant than the effect on anxiety. The effect of meeting PA and sleep guidelines on depression was more significant than the effect of meeting ST and sleep guidelines on depression. The effect of meeting ST and sleep guidelines on anxiety was more significant than the effect of meeting PA and sleep guidelines on anxiety. The effect of meeting PA + and sleep guidelines on depression was more significant than the effect of meeting PA + and sleep guidelines on anxiety. The effect of meeting ST and sleep guidelines on anxiety was more significant than the effect of meeting PA and sleep guidelines on anxiety. ## 3.5. Meeting Three of the 24-HMG Recommendations in Relation to Anxiety and Depression Table 5 presents the relationship between meeting the three 24-HMG recommendations and adolescents’ anxiety and depression obtained through a binary logistic regression. The findings show that the ORs for depression and anxiety in the meeting PA + ST + sleep group were 0.27 and 0.18, respectively, compared to the meeting none group ($p \leq 0.01$). This indicates that meeting PA, ST and sleep guidelines had a more significant effect on anxiety than on depression. In the boys sample, the ORs for depression and anxiety in the meeting PA + ST + sleep group were 0.20 and 0.08, respectively, compared to the meeting none group ($p \leq 0.01$). This indicates that meeting PA, ST and sleep guidelines had a more significant effect on anxiety than on depression. In the girls sample, the ORs for depression and anxiety in the meeting PA + ST + sleep group were 0.52 and 0.60, respectively, compared to the meeting none group ($p \leq 0.01$). This indicates that the effects of meeting PA, ST and sleep guidelines were more significant for anxiety than for depression. Logistic regression results showed that the differences in the effects of gender on the ORs for depression and anxiety in adolescents were not significant. ## 4. Discussion This study used a K8 grade sample based on the CEPS (2014–2015) to determine the relationship between meeting the 24-HMG and students’ self-rated anxiety and depression. In this study, only $0.71\%$ of the adolescents in the study sample met the three recommendations of the 24-HMG. This indicated a low adherence to the guideline among Chinese adolescents in grade K8 compared to previous studies. A previous study by Liu et al. [ 2022] of 14–17 year old adolescents in grades K9-K12 in the USA found that the proportion of adolescents meeting all three of the 24-HMG was $3\%$ [37]. The study of Janssen et al. [ 2017] of 17,000 Canadian adolescents aged 10–17 years found that less than $3\%$ of adolescents met all three of the 24-HMG [38]. In a study of 3772 Spanish minors aged 4–14 years, López-Gil et al. [ 2022] found that the proportion of children and adolescents meeting all three of the 24-HMG was $13.5\%$ [39]. This result was also lower than that reported in a previous study by Chen et al. [ 2021]. Their study of 114,072 Chinese adolescents aged 6–13 years in grades K4-K12 found that the proportion of adolescents meeting all three of the 24-HMG was $5\%$ [40]. The reason for the inconsistency between the present study and the results of these previous studies, however, may be related to the grade range of the population from which the sample was collected and the different testing instruments. For example, some studies used accelerometers to estimate the PA status [8], while others used self-reported PA data (as in this study). In this study, there was a gender difference in the proportion meeting the recommendations in 24-HMG. Previous studies have reported that among younger children, boys were more likely to meet all three recommendations than girls. However, among adolescents, there were no gender differences in meeting all three recommendations. However, regarding the components of the 24-HMG, adolescent boys were more likely than girls to meet the physical activity recommendations, and females were more likely than males to meet the screen time recommendations (Sampasa-Kanyinga et al., 2020) [41]. In this study, boys were more likely than girls to meet the PA and sleep recommendations, and girls were more likely than boys to meet the ST recommendations, but a higher proportion of boys met all three recommendations. Previously, Ying et al. [ 2020] did not find gender differences in meeting sleep and screen time, in a survey of Chinese high school students meeting the 24-HMG recommendations [42]. Therefore, the relationship between meeting the 24-HMG recommendations and gender needs to be further developed in future studies. However, it is worth noting that low adherence to 24-HMG among Chinese adolescents may threaten clinical and public health outcomes. Previous studies have found that compliance with 24-HMG has beneficial effects on adolescent health development. Therefore, this issue should be addressed through an effective public health approach. However, there is limited evidence on how to optimise adolescent health behaviours and therefore future research on intervention strategies to promote adolescent compliance with 24-HMG should be encouraged. In a study by de Castro et al. [ 2023] on depression and anxiety among adolescents in 26 low- and middle-income countries, $5.5\%$ of adolescents ($$n = 123$$,975, 10–17 years) had symptoms of anxiety and $3.1\%$ had symptoms of depression [43].In a study by Merikangas et al. [ 2010], the prevalence of anxiety among adolescents in the United States ($$n = 10$$, 123, 13–18 years) was $31.9\%$ [44]. Ma et al. [ 2021] reported a prevalence of depression and anxiety of $29\%$ and $26\%$, respectively, in adolescents ($$n = 57$$,927) [45]. The differences in the prevalence of depression and anxiety among adolescents in these studies were related to the depression and anxiety testing instruments used in the different studies. In the present study, $26.93\%$ and $29.44\%$ of adolescents self-reported symptoms of depression and anxiety, respectively, and no significant gender differences in the prevalence of depression and anxiety were observed. This data, is close to the previous study by Tang et al. [ 2019]. Their study of Chinese adolescents ($$n = 144$$,060) reported a $24.3\%$ ($95\%$ CI, $21.3\%$–$27.6\%$) prevalence of depression in adolescents [46]. Previous studies had systematically evaluated the independent effects of physical activity, screen time and sleep on adolescent mental health outcomes [16,17,18,19,20,21,22]. In contrast, not many studies evaluated their combined effects of the three jointly on adolescent mental health. Consistent with past studies [9,47,48], our study showed that meeting all three of the 24-HMG was associated with a lower risk of anxiety and depression in adolescents. These studies highlighted the importance of helping adolescents meet more of the 24-HMG recommendations, which could be considered as a way to prevent or treat mental health problems in adolescents in this age group. In line with the findings of Zhu et al. [ 2019], the present study confirmed that the higher the number meeting the 24-HMG recommendations, the lower the odds of anxiety and depression in adolescents [48]. This finding is indirectly supported by the study of Sampasa-Kanyinga et al. [ 2017] [9], showing that adolescents who met the optimal combination of PA, screen time and sleep time had the lowest odds of depression or anxiety. This is consistent with the findings of Lu et al. [ 2021] in a study of Chinese children aged 10–13 years [27]. The results of this study suggested that meeting the 24-HMG guidelines would be beneficial in maintaining better mental health in Chinese adolescents at the junior high school level. These findings may further help refine and update the cross-cultural use of the 24-HMG. This study found that meeting sleep recommendations significantly reduced the risk of depression and anxiety in adolescents. The relationship between insufficient sleep duration and anxiety and depression has been confirmed in many studies [49,50]. Insomnia is the most common sleep disorders among adolescents (Johnson et al., 2006) [51]. Inadequate sleep duration in adolescents had serious implications for future health and functioning (Brand et al., 2009) [52] and was thought to trigger and maintain many emotional and behavioural problems, particularly anxiety and depression (Dahl and Harvey, 2007) [53]. Sleep, arousal and emotion represent overlapping regulatory systems, with dysregulation of one system affecting the others, so that sleep disruption during critical periods of maturational development may provide a pathway for later emotional dysregulation, and vice versa (Dahl, 1996) [54]. Sleep deprivation has been shown to increase negative emotions, decrease positive emotions and alter the way adolescents understand, express and regulate their emotions (Palmer et al., 2016) [55]. Several studies have shown an increase in depression (Fredriksen et al., 2004) [56] and anxiety disorder (Sagaspe et al., 2006) [56] symptoms in a sample of healthy adolescents following sleep deprivation. Excessive sleep duration, such as narcolepsy, has also been found to be associated with an increased occurrence of depression and anxiety in adolescents. For example, Kaplan and Harvey [2009] reported that narcolepsy may be an important mechanism contributing to the maintenance of symptoms of mood disorders [56]. These findings suggest that the prevention and optimisation of adolescent anxiety and depressive symptoms requires attention on the impact of adolescent sleep duration. However, this study did not use an objective instrument to measure sleep and therefore could not determine the duration of continuous sleep in adolescents, which also needs to be further explored in depth. This study found that the greater the number of 24-HMG recommendations met, the lower the risk of depression and anxiety in adolescents. Meeting PA + sleep or meeting ST + sleep of the 24-HMG recommendations reduced the risk of anxiety and depression in adolescents compared to not meeting any of the recommendations. This is consistent with previous findings from Zhu et al. [ 2019] [48]. The results of this study suggest that the combined form of meeting PA + sleep or meeting ST + sleep may have a significant role in the prevention of anxiety and depression. Meeting PA+ST + sleep was more effective in reducing the risk of depression and anxiety in adolescents. Meeting the 24-HMG recommendation categories differed in their effect on reducing the risk of depression or anxiety in adolescents. For example, meeting one 24-HMG recommendation for sleep duration alone significantly reduced the risk of depression and anxiety in adolescents. In two 24-HMG recommendations, meeting only physical activity time and sleep time together, and meeting both screen time and sleep time together significantly reduced the risk of adolescents developing anxiety. This further analysis, in which specific behavioural combinations are most beneficial to adolescents’ mental health, enables the implementation of more effective interventions. This suggests that the promotion and guidance of these specific behavioural combinations should be taken into account when implementing integrated or holistic interventions with adolescents, and that priority should be given to guiding which behavioural recommendations to meet if 24-HMG cannot be met in its entirety. To our knowledge, this is the first time that the effects of meeting specific behavioural combination categories of the 24-HMG recommendations on adolescent mental health have been observed in a sample of Chinese adolescents, and is a further development and extension of the findings of Lu et al. ’s [2021] study of children [27]. Our study could add new research evidence to the field of research on promoting adolescent mental health. This study measured the association between meeting physical activity, screen time and sleep time recommendations in the 24-HMG and the risk of developing anxiety and depression in Chinese adolescents in grade K8, using data from the national adolescent sample of the CEPS and using the Canadian 24-HMG as a reference standard. From the results of this study, it was suggested that encouraging an optimal combination of healthy behaviours may be an effective intervention to reduce anxiety and depression in adolescents. However, there were limitations to this study. Firstly, $55\%$ of the schools surveyed in CEPS were in the eastern region of China, $19\%$ in the central region and $24\%$ in the western region. Due to the limitations of this sampling method of the CEPS, the results of this study need to be further validated with data from other large samples of Chinese adolescents. Secondly, the data in this study were derived from self-reported data of adolescents, and therefore the findings may be subject to bias from meetings and social expectations. Thirdly, the instruments used in this study to measure anxiety and depression were not derived from commonly used psychometric instruments of anxiety or depression. Although the validity of the CEPS’s approach to measuring adolescent mental health has been validated in a number of studies, we were unable to obtain additional data evidence from its development process. Fourthly, the measurements of physical activity, sedentary behaviour and sleep in this study were not based on objective measurement work. This may have implications for the stability of the estimates from this study. Fifthly, this study selected few control covariates and only examined adolescents’ sex, age, ethnicity, single child, residence, father’s highest level of education, mother’s highest level of education, perceived household economic status, and BMI were controlled for covariates. Previous studies have found that academic stress and social support may all have an impact on adolescent depression or anxiety, and the study did not control for these covariates. Finally, due to the cross-sectional design used in this study, no conclusions could be drawn regarding causality. In future studies, we recommend the use of objective measures of physical activity, quantity and quality of sleep, and validate depression and anxiety questionnaires, such as the depression anxiety spectrum disorder questionnaire (DASS). ## 5. Conclusions This study determined the risk of developing depression and anxiety in adolescents who met the recommendations for 24-HMG alone and in combination. Overall, meeting more of the recommendations in the 24-HMG was associated with lower anxiety and depression risk outcomes in adolescents. For boys, reducing the risk of depression and anxiety can be prioritised by meeting PA + ST + sleep, meeting ST+ sleep and meeting sleep in the 24-HMG. For girls, reducing the risk of depression and anxiety may be preferred by meeting PA + ST + sleep, meeting PA + sleep and meeting sleep in 24-HMG.However, only a small proportion of adolescents met all recommendations, highlighting the need to promote and support adherence to these behaviours. ## References 1. Biddle S.J.H., Asare M.. Physical activity and mental health in children and adolescents: A review of reviews. Br. J. Sports Med. (2019) 45 886-895. DOI: 10.1136/bjsports-2011-090185 2. Babic M.J., Smith J.J., Morgan P.J., Eather N., Plotnikoff R.C., Lubans D.R.. Longitudinal associations between changes in screen-time and mental health outcomes in adolescents. Ment. Health Phys. Act. (2017) 12 124-131. DOI: 10.1016/j.mhpa.2017.04.001 3. Zhang J., Paksarian D., Lamers F., Hickie I.B., He J., Merikangas K.R.. Sleep Patterns and Mental Health Correlates in US Adolescents. J. Pediatr. (2017) 182 137-143. DOI: 10.1016/j.jpeds.2016.11.007 4. Tremblay M.S., Carson V., Chaput J.P., Connor Gorber S., Dinh T., Duggan M., Faulkner G., Gray C.E., Gruber R., Janson K.. Canadian 24-hour movement guidelines for children and youth: An integration of physical activity, sedentary behaviour, and sleep. Appl. Physiol. Nutr. Metab. (2016) 41 S311-S327. DOI: 10.1139/apnm-2016-0151 5. Khan A., Lee E.-Y., Tremblay M.S.. Meeting 24-h movement guidelines and associations with health related quality of life of Australian adolescents. J. Sci. Med. Sport (2020) 24 468-473. DOI: 10.1016/j.jsams.2020.10.017 6. Rollo S., Antsygina O., Tremblay M.S.. The whole day matters: Understanding 24-hour movement guideline adherence and relationships with health indicators across the lifespan. J. Sport Health Sci. (2020) 9 493-510. DOI: 10.1016/j.jshs.2020.07.004 7. Sampasa-Kanyinga H., Standage M., Tremblay M., Katzmarzyk P., Hu G., Kuriyan R., Maher C., Maia J., Olds T., Sarmiento O.L.. Associations between meeting combinations of 24-h movement guidelines and health-related quality of life in children from 12 countries. Public Health (2017) 153 16-24. DOI: 10.1016/j.puhe.2017.07.010 8. Parrish A.-M., Tremblay M.S., Carson S., Veldman S.L.C., Cliff D., Vella S., Chong K.H., Nacher M., Cruz B.D.P., Ellis Y.. Comparing and assessing physical activity guidelines for children and adolescents: A systematic literature review and analysis. Int. J. Behav. Nutr. Phys. Act. (2020) 17 1-22. DOI: 10.1186/s12966-020-0914-2 9. López-Gil J., Tremblay M., Brazo-Sayavera J.. Changes in Healthy Behaviors and Meeting 24-h Movement Guidelines in Spanish and Brazilian Preschoolers, Children and Adolescents during the COVID-19 Lockdown. Children (2021) 8. DOI: 10.3390/children8020083 10. Kim H., Ma J., Harada K., Lee S., Gu Y.. Associations between Adherence to Combinations of 24-h Movement Guidelines and Overweight and Obesity in Japanese Preschool Children. Int. J. Environ. Res. Public Health (2020) 17. DOI: 10.3390/ijerph17249320 11. Guimarães R.D.F., Gilbert J.-A., Lemoyne J., Mathieu M.-E.. Better health indicators of FitSpirit participants meeting 24-h movement guidelines for Canadian children and youth. Health Promot. Int. (2020) 36 836-845. DOI: 10.1093/heapro/daaa102 12. Shorey S., Ng E.D., Wong C.H.J.. Global prevalence of depression and elevated depressive symptoms among adolescents: A systematic review and meta-analysis. Br. J. Clin. Psychol. (2021) 61 287-305. DOI: 10.1111/bjc.12333 13. Erskine H.E., Baxter A.J., Patton G., Moffitt T.E., Patel V., Whiteford H.A., Scott J.G.. The global coverage of prevalence data for mental disorders in children and adolescents. Epidemiology Psychiatr. Sci. (2016) 26 395-402. DOI: 10.1017/S2045796015001158 14. Wiles N.J., Jones G.T., Haase A.M., Lawlor D.A., Macfarlane G.J., Lewis G.. Physical activity and emotional problems amongst adolescents. Soc. Psychiatry Psychiatr. Epidemiol. (2008) 43 765-772. DOI: 10.1007/s00127-008-0362-9 15. Rothon C., Edwards P., Bhui K., Viner R.M., Taylor S., A Stansfeld S.. Physical activity and depressive symptoms in adolescents: A prospective study. BMC Med. (2010) 8. DOI: 10.1186/1741-7015-8-32 16. Oberle E., Ji X.R., Kerai S., Guhn M., Schonert-Reichl K.A., Gadermann A.M.. Screen time and extracurricular activities as risk and protective factors for mental health in adolescence: A population-level study. Prev. Med. (2020) 141 106291. DOI: 10.1016/j.ypmed.2020.106291 17. Twenge J.M., Campbell W.K.. Associations between screen time and lower psychological well-being among children and adolescents: Evidence from a population-based study. Prev. Med. Rep. (2018) 12 271-283. DOI: 10.1016/j.pmedr.2018.10.003 18. Fuligni A.J., Arruda E.H., Krull J.L., Gonzales N.A.. Adolescent Sleep Duration, Variability, and Peak Levels of Achievement and Mental Health. Child Dev. (2017) 89 e18-e28. DOI: 10.1111/cdev.12729 19. Yeo S.C., Jos A.M., Erwin C., Lee S.M., Lee X.K., Lo J.C., Chee M., Gooley J.J.. Associations of sleep duration on school nights with self-rated health, overweight, and depression symptoms in adolescents: Problems and possible solutions. Sleep Med. (2018) 60 96-108. DOI: 10.1016/j.sleep.2018.10.041 20. McDowell C.P., MacDonncha C., Herring M.P.. Brief report: Associations of physical activity with anxiety and depression symptoms and status among adolescents. J. Adolesc. (2017) 55 1-4. DOI: 10.1016/j.adolescence.2016.12.004 21. Mougharbel F., Goldfield G.S.. Psychological Correlates of Sedentary Screen Time Behaviour Among Children and Adolescents: A Narrative Review. Curr. Obes. Rep. (2020) 9 493-511. DOI: 10.1007/s13679-020-00401-1 22. Baum K.T., Desai A., Field J., Miller L.E., Rausch J., Beebe D.W.. Sleep restriction worsens mood and emotion regulation in adolescents. J. Child Psychol. Psychiatry (2014) 55 180-190. DOI: 10.1111/jcpp.12125 23. Ojio Y., Kishi A., Sasaki T., Togo F.. Association of depressive symptoms with habitual sleep duration and sleep timing in junior high school students. Chrono- Int. (2020) 37 877-886. DOI: 10.1080/07420528.2020.1746796 24. Sampasa-Kanyinga H., Colman I., Goldfield G.S., Janssen I., Wang J., Podinic I., Tremblay M.S., Saunders T.J., Sampson M., Chaput J.-P.. Combinations of physical activity, sedentary time, and sleep duration and their associations with depressive symptoms and other mental health problems in children and adolescents: A systematic review. Int. J. Behav. Nutr. Phys. Act. (2020) 17 1-16. DOI: 10.1186/s12966-020-00976-x 25. Lu S., Cheval B., Yu Q., Hossain M., Chen S.-T., Taylor A., Bao R., Doig S., Li J., Wang T.. Associations of 24-Hour Movement Behavior with Depressive Symptoms and Anxiety in Children: Cross-Sectional Findings from a Chinese Sample. Healthcare (2021) 9. DOI: 10.3390/healthcare9111532 26. Chong K., Parrish A.-M., Cliff D., Dumuid D., Okely A.. Cross-Sectional and Longitudinal Associations between 24-Hour Movement Behaviours, Recreational Screen Use and Psychosocial Health Outcomes in Children: A Compositional Data Analysis Approach. Int. J. Environ. Res. Public Health (2021) 18. DOI: 10.3390/ijerph18115995 27. Li M., Xue H., Wang W., Wang Y.. Parental Expectations and Child Screen and Academic Sedentary Behaviors in China. Am. J. Prev. Med. (2017) 52 680-689. DOI: 10.1016/j.amepre.2016.12.006 28. Ross R., Chaput J.P., Giangregorio L.M., Janssen I., Saunders T.J., Kho M.E., Poitras V.J., Tomasone J.R., El-Kotob R., McLaughlin E.C.. Canadian 24-Hour Movement Guidelines for Adults aged 18–64 years and Adults aged 65 years or older: An integration of physical activity, sedentary behaviour, and sleep. Appl. Physiol. Nutr. Metab. (2020) 45 S57-S102. DOI: 10.1139/apnm-2020-0467 29. Gong J., Lu Y., Song H.. The Effect of Teacher Gender on Students’ Academic and Noncognitive Outcomes. J. Labor Econ. (2018) 36 743-778. DOI: 10.1086/696203 30. Williams N.. The GAD-7 questionnaire. Occup. Med. (2014) 64 224. DOI: 10.1093/occmed/kqt161 31. Kroenke K., Spitzer R.L., Williams J.B.. The PHQ-9: Validity of a brief depression severity measure. J. Gen. Intern. Med. (2001) 16 606-613. DOI: 10.1046/j.1525-1497.2001.016009606.x 32. Guo Y.H., Li S.P., Yang W.N., Liu C.F.. The impact of classroom girl ratio on students’ psychological health-Based on data from the China Education Tracking Survey (CEPS). Peking Univ. Educ. Rev. (2021) 19 19-188 33. Ma L., Gao L., Chiu D.T., Ding Y., Wang Y., Wang W.. Overweight and Obesity impair academic performance in adolescence: A national cohort study of 10,279 adolescents in China. Obesity (2020) 28 1301-1309. DOI: 10.1002/oby.22801 34. 34. National Health Commission Screening for Overweight and Obesity among School-Age Children and AdolescentsNational Health CommissionBeijing, China2018. Screening for Overweight and Obesity among School-Age Children and Adolescents (2018) 35. Liu S., Yu Q., Hossain M., Doig S., Bao R., Zhao Y., Yan J., Luo X., Yang J., Kramer A.F.. Meeting 24-h Movement Guidelines is Related to Better Academic Achievement: Findings from the YRBS 2019 Cycle. Int. J. Ment. Health Promot. (2022) 24 13-24. DOI: 10.32604/IJMHP.2021.017660 36. Janssen I., Roberts K.C., Thompson W.. Is adherence to the Canadian 24-Hour Movement Behaviour Guidelines for Children and Youth associated with improved indicators of physical, mental, and social health?. Appl. Physiol. Nutr. Metab. (2017) 42 725-731. DOI: 10.1139/apnm-2016-0681 37. López-Gil J.F., Roman-Viñas B., Aznar S., Tremblay M.S.. Meeting 24-h movement guidelines: Prevalence, correlates, and associations with socioemotional behavior in Spanish minors. Scand. J. Med. Sci. Sports (2022) 32 881-891. DOI: 10.1111/sms.14132 38. Chen S.-T., Liu Y., Tremblay M.S., Hong J.-T., Tang Y., Cao Z.-B., Zhuang J., Zhu Z., Wu X., Wang L.. Meeting 24-h movement guidelines: Prevalence, correlates, and the relationships with overweight and obesity among Chinese children and adolescents. J. Sport Health Sci. (2021) 10 349-359. DOI: 10.1016/j.jshs.2020.07.002 39. Sampasa-Kanyinga H., Chaput J.-P., Goldfield G.S., Janssen I., Wang J., Hamilton H.A., Colman I.. 24-hour movement guidelines and suicidality among adolescents. J. Affect. Disord. (2020) 274 372-380. DOI: 10.1016/j.jad.2020.05.096 40. Ying L., Zhu X., Haegele J., Wen Y.. Movement in High School: Proportion of Chinese Adolescents Meeting 24-Hour Movement Guidelines. Int. J. Environ. Res. Public Health (2020) 17. DOI: 10.3390/ijerph17072395 41. de Castro F., Cappa C., Madans J.. Anxiety and Depression Signs Among Adolescents in 26 Low- and Middle-Income Countries: Prevalence and Association with Functional Difficulties. J. Adolesc. Health (2023) 72 S79-S87. DOI: 10.1016/j.jadohealth.2022.03.022 42. Merikangas K.R., He J.-P., Burstein M., Swanson S.A., Avenevoli S., Cui L., Benjet C., Georgiades K., Swendsen J.. Lifetime Prevalence of Mental Disorders in U.S. Adolescents: Results from the National Comorbidity Survey Replication–Adolescent Supplement (NCS-A). J. Am. Acad. Child Adolesc. Psychiatry (2010) 49 980-989. DOI: 10.1016/j.jaac.2010.05.017 43. Ma L., Mazidi M., Li K., Li Y., Chen S., Kirwan R., Zhou H., Yan N., Rahman A., Wang W.. Prevalence of mental health problems among children and adolescents during the COVID-19 pandemic: A systematic review and meta-analysis. J. Affect. Disord. (2021) 293 78-89. DOI: 10.1016/j.jad.2021.06.021 44. Tang X., Tang S., Ren Z., Wong D.F.K.. Prevalence of depressive symptoms among adolescents in secondary school in mainland China: A systematic review and meta-analysis. J. Affect. Disord. (2019) 245 498-507. DOI: 10.1016/j.jad.2018.11.043 45. Patte K.A., Faulkner G., Qian W., Duncan M., Leatherdale S.T.. Are one-year changes in adherence to the 24-hour movement guidelines associated with depressive symptoms among youth?. BMC Public Health (2020) 20 1-12. DOI: 10.1186/s12889-020-08887-z 46. Zhu X., Haegele J.A., Healy S.. Movement and mental health: Behavioral correlates of anxiety and depression among children of 6–17 years old in the U.S. Ment. Health Phys. Act. (2019) 16 60-65. DOI: 10.1016/j.mhpa.2019.04.002 47. Alvaro P.K., Roberts R., Harris J.K.. A Systematic Review Assessing Bidirectionality between Sleep Disturbances, Anxiety, and Depression. Sleep (2013) 36 1059-1068. DOI: 10.5665/sleep.2810 48. Gregory A.M., Sadeh A.. Sleep, emotional and behavioral difficulties in children and adolescents. Sleep Med. Rev. (2012) 16 129-136. DOI: 10.1016/j.smrv.2011.03.007 49. Johnson E.O., Roth T., Schultz L., Breslau N.. Epidemiology of DSM-IV Insomnia in Adolescence: Lifetime Prevalence, Chronicity, and an Emergent Gender Difference. Pediatrics (2006) 117 e247-e256. DOI: 10.1542/peds.2004-2629 50. Brand S., Hatzinger M., Beck J., Holsboer-Trachsler E.. Perceived parenting styles, personality traits and sleep patterns in adolescents. J. Adolesc. (2009) 32 1189-1207. DOI: 10.1016/j.adolescence.2009.01.010 51. Dahl R.E., Harvey A.G.. Sleep in Children and Adolescents with Behavioral and Emotional Disorders. Sleep Med. Clin. (2007) 2 501-511. DOI: 10.1016/j.jsmc.2007.05.002 52. Dahl R.E.. The regulation of sleep and arousal: Development and psychopathology. Dev. Psychopathol. (1996) 8 3-27. DOI: 10.1017/S0954579400006945 53. Reddy R., Palmer C.A., Jackson C., Farris S.G., Alfano C.A.. Impact of sleep restriction versus idealized sleep on emotional experience, reactivity and regulation in healthy adolescents. J. Sleep Res. (2017) 26 516-525. DOI: 10.1111/jsr.12484 54. Fredriksen K., Rhodes J., Reddy R., Way N.. Sleepless in Chicago: Tracking the Effects of Adolescent Sleep Loss During the Middle School Years. Child Dev. (2004) 75 84-95. DOI: 10.1111/j.1467-8624.2004.00655.x 55. Sagaspe P., Sanchez-Ortuno M., Charles A., Taillard J., Valtat C., Bioulac B., Philip P.. Effects of sleep deprivation on Color-Word, Emotional, and Specific Stroop interference and on self-reported anxiety. Brain Cogn. (2006) 60 76-87. DOI: 10.1016/j.bandc.2005.10.001 56. Kaplan K.A., Harvey A.G.. Hypersomnia across mood disorders: A review and synthesis. Sleep Med. Rev. (2009) 13 275-285. DOI: 10.1016/j.smrv.2008.09.001
--- title: 'Antibacterial Effects of X-ray and MRI Contrast Media: An In Vitro Pilot Study' authors: - Michael P. Brönnimann - Lea Hirzberger - Peter M. Keller - Monika Gsell-Albert journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966632 doi: 10.3390/ijms24043470 license: CC BY 4.0 --- # Antibacterial Effects of X-ray and MRI Contrast Media: An In Vitro Pilot Study ## Abstract Some radiological contrast agents have been shown to have effects on bacterial growth. In this study, the antibacterial effect and mechanism of action of iodinated X-ray contrast agents (Ultravist 370, Iopamiro 300, Telebrix Gastro 300 and Visipaque) and complexed lanthanide MRI contrast solutions (MultiHance and Dotarem) were tested against six different microorganisms. Bacteria with high and low concentrations were exposed to media containing different contrast media for various lengths of time and at pH 7.0 and 5.5. The antibacterial effect of the media was examined in further tests using agar disk diffusion analysis and the microdilution inhibition method. Bactericidal effects were found for microorganisms at low concentrations and low pH. Reductions were confirmed for Staphylococcus aureus and Escherichia coli. ## 1. Introduction CT- or MR-guided interventions often require luminal administration of contrast media for diagnostic purposes such as characteristics of the abscess cavity, positional control of drainage, or detection of a fistula tract to adjacent structures. The main components are the chemically bound iodine in X-ray contrast media and gadolinium in MRI contrast media. First, in 1839, John Davis, a surgeon, used a tincture of iodine on the battlefield as an antiseptic and disinfectant [1]. The high toxicity of free lanthanides in the body, such as gadolinium, has been known since the middle of the last century [2]. As the development of contrast agents has progressed, every effort has been made to create stable compounds with these cytotoxic elements bound therein. The gadolinium ion (Gd3+) is highly toxic in its free form. Therefore, it is chelated with an organic ligand molecule to produce highly soluble nontoxic complexes [3]. Nevertheless, small amounts of these components (iodine, gadolinium) are known to dissociate freely [1,2]. The extent to which presumed free components have a bacteriostatic or bactericidal effect has been investigated little and stretches selectively over the last century. Different results with older, ionic X-ray contrast agents and common bacteriostatic effects of newer, nonionic X-ray contrast agents, primarily against Gram-negative bacteria, have been reported [4,5,6,7]. The data situation regarding MRI contrast agents is even thinner, with Green, Moustache [8] demonstrating a lack of proliferation of organisms and Beussink et al. [ 9] showing antimicrobial activity in tested MRI contrast agents. This in vitro pilot study aims to evaluate the antibacterial effect of commonly used X-ray and MRI contrast media from different chemical classification groups, considering pH, incubation time and carbon substrate utilization. Such diversification in contrast agents and other laboratory parameters has never been studied. We hypothesize that bacteriostatic effects can be shown. ## 2.1. Antibacterial Efficacy Tests in Blood Culture Bottles Blood culture bottles were inoculated with bacterial suspensions of the microorganisms Staphylococcus aureus, Pseudomonas aeruginosa and Bacillus subtilis in two different densities that were previously incubated with contrast agents Ultravist, Iopamiro, Telebrix Gastro or Visipaque. The time until growth of pathogens was recorded (Table 1). Negative controls with sodium chloride solution were integrated. For the pathogen S. aureus, all blood culture bottles showed growth after a few hours of incubation (3.7 and 9.9 h for the high bacterial count and between 11.1 and 16.5 for the low bacterial count) with all X-ray contrast media. Incubation with X-ray contrast media for 48 h resulted in growth always being delayed compared to incubation for 24 h. Similarly, the blood culture bottles with the lower bacterial counts (5 mL of a suspension with 102 CFU/mL was placed in 10 mL of blood culture liquid) always became positive later. The negative controls showed no growth anymore after 48 h incubation. A similar picture was seen for the pathogen B. subtilis. However, with the X-ray contrast medium Ultravist 370, no growth was seen after 24 h at the low pH of 5.5 with the low bacterial count, but after 48 h, growth was seen again (Figure 1). The negative controls showed no growth after 48 h. The results were similar for P. aeruginosa. Only the contrast medium Teblebrix Gastro 300 showed no growth at both pH values with the low bacterial count after 24 h; likewise, after 48 h at pH 7.1 (Figure 2). The negative controls with the low pH showed no more growth after 48 h. ## 2.2. Agar Disk Diffusion Filter paper disks soaked in X-ray contrast medium were placed on plates inoculated with B. subtilis. The X-ray contrast agent’s inhibitory or bactericidal effect could be shown due to the inhibition zone that formed around the filter paper disks. Erythromycin-soaked filter paper disks were used as controls. The inhibition zone that formed around them had a diameter between 27 and 38 mm. At pH 7.4, only Ultravist 370 showed a minimal inhibition zone, whereas, at pH 5.5, small inhibition zones were detected for the contrast agents Ultravist 370, Telebrix Gastro 300, Multihance and Dotarem. In a second test series, fresh suspensions of S. aureus and P. aeruginosa were prepared and plated onto agar plates. Filter paper disks soaked with X-ray contrast medium were placed onto these plates. The X-ray contrast agent’s inhibitory or bactericidal effect could be shown due to the inhibition zone that formed around the filter paper disks. Erythromycin-soaked filter paper plates were used as controls. They had a diameter between 17 and 25 mm for the inhibition of S. aureus, whereas P. aeruginosa was not sensitive to erythromycin and showed no inhibition. All the contrast agents showed no inhibition either. They all had a diameter of 6 mm (corresponding to the diameter of the filter paper disk). ## 2.3. Broth Microdilution in Minimal Medium Iodine and gadolinium ions are complexed with organic ligands and are, therefore, not available for bacteria. Thus, we tried to show that if the carbon source in the medium has been used up, the contrast agent itself can be used as a carbon source. By breaking up the contrast media, iodine or gadolinium ions are released and can have a toxic effect. For the broth microdilution experiment, microorganisms that have long doubling times were added. Dense, homogenous suspensions of E.coli, S. aureus, M. smegmatis and F. necrophorum were prepared. Afterwards, these cultures were transferred into minimal medium M9 supplemented with $2\%$ glucose for culturing. A quantity of 1.5 mL was taken out and pelleted by centrifugation. The pellet was washed with PBS and transferred into minimal medium M9. A quantity of 100 µL of all bacterial strains in the two media were placed onto a microplate and incubated with 100 µL of each contrast agent. Growth controls in M9 supplemented with $2\%$ glucose and M9 of all bacterial strains were included. After up to 48 h incubation, the growth was examined. After 48 h, the pathogen M. smegmatis showed growth with the following five X-ray contrast media; Ultravist, Iopamiro, Telebrix Gastro, Multihance and Dotarem (Table 2), Visipaque being the only contrast agent that has not shown growth. S. aureus has not grown in any of the X-ray contrast media. E.coli showed growth with the following X-ray contrast media: Iopamiro, Visipaque, Multihance ($\frac{1}{3}$) and Dotarem($\frac{1}{3}$). ## 3. Discussion The effect of representatives of all classification groups of iodinated X-ray contrast agents (monomeric, dimeric, high osmolar, low osmolar, ionic and nonionic) as well as gadolinium-based contrast media (macrocyclic vs. linear) on bacteria was investigated in our in vitro study by using three different tests. The most important findings are the lack of growth detection in broth microdilution in a minimal medium of M. smegmatis in Visipaque, of E. coli in Ultravist 370 and Telebrix Gastro, and of S. aureus in all tested contrast media. The additional lack of effect on S. aureus in the other tests (antibacterial efficacy test in the blood culture bottle and agar disk diffusion) supports our hypothesis that the components (free iodine and gadolinium) are consumed and a cytotoxic effect results. The lack of growth of negative controls with NaCl after 48 h in low suspension density also supports this. The bacteria probably died because of the lack of nutritious substances. The lack of a carbon source was not compensated. Still, we consider the controls to be adequate, since there was growth after 24 h. That the duration of contact between contrast media and microorganisms is an important factor was already assumed by Blake and Halasz [4]. This could also explain why in our results Multihance and Dotarem showed a lack of growth in S. aureus and E. coli in only one-third of our controls. Since our incubation period was only 86 h and not up to 28 days as in Beussink, Godat and Seaton [9]. Furthermore, this aspect is supported by our following vital results, such as a bacteriostatic effect on B. subtilis and the only bactericidal effect of Telebrix Gastro on P. aeruginosa (Figure 1 and Figure 2). On the one hand, we were able to reproduce the previously described bactericidal effect of X-ray contrast media, especially on Gram-negative bacteria [4,10,11]. Thus, we support the hypothesis that the different cell wall of gram-negative bacteria presumably is more strongly affected by the presence of X-ray contrast media [4]. On the other hand, this is presumably only true for the first 48 h of contact, as our other tests (agar disk diffusion and broth microdilution in minimal medium) could not substantiate this, and later, the released components (iodine or gadolinium) from the complex may cause the subsequent main effect. This is also supported by the fact that the free iodine content of ionic X-ray contrast media is up to 12 times higher than that of the more stable, nonionic X-ray contrast media [12]. Based on this, our inconsistent findings concerning ionic and nonionic X-ray contrast agents in the different tests can be explained and are in line with the previous literature [4]. This argument is also supported by the fact that even our bactericidal and bacteriostatic results were initially demonstrated only at low organism concentrations (low inocula of 1.5 × 103–5.0 × 103 CFU/mL) (e.g., [4]). Instead, the bound amount of potentially cytotoxic significant components that play a critical role could also be explained by the lack of growth of M. smegmatis under Visipaque in broth microdilution in a minimal medium. This contains one of the highest iodine concentrations of all tested iodine-containing agents (320 mg iodine per ml compared to 300 mg iodine per ml). Furthermore, the results of the antibacterial efficacy tests in blood culture bottles, as well as those of the agar disk diffusion, showed that a higher concentration of iodine in X-ray contrast media to achieve the described effects is necessary, which has been assumed in previous studies [4,10,13]. With regard to MRI contrast agents (Multihance with 334 mg/mL gadobenic acid and Dotarem with 279 mg/mL gadoteric acid), we found no equal evidence for this. In contrast, the lack of effect of MRI contrast agents on M. smegmatis may be related to the fact that this pathogen is also often found in soil samples and, as an environmental microorganism, is better able to deal with heavy metals. Our other key results focus on the unique findings concerning the effect of pH on contrast media. Only when the pH was reduced from 7.3 to 5.5 could a bacteriostatic effect be shown in B. subtilis with Ultravist 370. Likewise, when the pH was reduced from 7.4 to 5.5, inhibition zones were no longer detected not only with Ultravist 370 but also with Telebrix Gastro 300, Multihance and Dotarem. There is no sporadic literature available regarding pH changes and the impact on microorganisms in the presence of contrast agents. So far, only Narins and Chase [6] have reported otherwise, that pH changes do not affect the effect of Hypaque (diatrizoate sodium, drug class Ionic-iodinated contrast media) on microorganisms. Whether the free components or the possibly provoked instability of the complexes with the release of the bound components is responsible for this effect cannot be clearly assigned, and further studies are needed. So far, however, it is assumed that the stability of MRI contrast agents is pH-dependent, but it is unclear how much free gadolinium is or will be present [14]. Our opposite result of Telebrix Gastro 300 and bacteriostatic effect on P. aeruginosa at pH 5.5 and bactericidal effect at pH 7 needs an explanation in this context. A possible explanation could be that the higher pH, which leads to a shift of the equilibrium from elemental iodine to hypoiodous acid, consequently acts more effectively against specific bacterial classes such as P. aeruginosa [15,16,17,18]. The disinfection effectiveness of different iodine species could already be shown [18,19,20]. Furthermore, the lack of interclass difference (agar disk diffusion at pH 5.5 and broth microdilution in a minimal medium) suggests that the specific properties of the contrast media compositions, such as thermodynamic stability and kinetic inertness (macrocyclic Gd(III) complexes vs. linear and ionic vs. nonionic iodinated contrast media) do not achieve the main effect in abscess collections [21,22]. Another critical aspect is that only about three relevant papers on this topic have been published in the last 20 years. On the one hand, this is because the contrast media most commonly used to date (e.g., Telebrix 1972), including the MRI contrast media (e.g., Dotarem 1989) were developed over several decades towards the end of the last century and the corresponding studies on patient safety had to be published between invention and approval [23,24,25,26]. On the other hand, the importance of image-guided interventions increased with an evident time lag [27]. Consequently, relevant interdisciplinary studies on this specific topic are lacking. Considering that the significance of percutaneous, CT- and MRI-guided interventions with luminal use of contrast medium to visualize abscess collections and possible associated fistula tracts is rapidly gaining over surgical repair, our results are a first step to filling this important gap. ## Limitations Our limitations are the lack of direct transferability from in vitro to in vivo systems and the lack of testing of anaerobic pathogens, which are also major representatives of abscess collections. For example, it is possible that the altered phagocyte activity is due to the presence of contrast agents; the results in vivo again differ greatly [28]. Furthermore, we did not measure the effective free concentration of the major components (iodine and gadolinium), which will be of great interest to future studies ## 4. Materials and Methods Six radiographic contrast media were tested in total. Four of them were iodinated contrast media: Ultravist 370 (agent: iopromid, Bayer Schweiz AG, Zurich, Switzerland, monomeric, nonionic, low osmolar, containing 300 mg iodine/mL), Iopamiro 300 (agent: iopamidol, Bracco Suisse SA, Ticino, Switzerland, monomeric, nonionic, low osmolar, containing 300 mg iodine/mL), Telebrix Gastro 300 (agent: meglumin, Guerbet AG, Zurich, Switzerland, monomeric, ionic, high osmolar, containing 300 mg iodine/mL), Visipaque (agent: iodixanol, dimeric, nonionic, iso osmolar, containing 320 mg iodine/mL). Furthermore, the two products containing gadolinium ions were MultiHance 0.5 mmol/mL (agent: gadobenic acid, Bracco Suisse SA, Ticino, Switzerland, containing 334 mg gadobenic acid/mL) and Dotarem (agent: gadoteric acid, Guerbet AG, Zurich, Switzerland containing 279.32 mg gadoteric acid/mL). All iodinated contrast agents consist of a central element, the tri-iodinated benzene ring. Three iodine atoms covalently bonded to the benzene ring, on the one hand, create a local concentration of iodine, and on the other hand, this organic, functional group reduces the risk of free iodine [21]. The potentially highly reactive and consequently toxic benzene ring is protected from oxidation by side chains [29]. Gd(III) ions are also toxic and consequently bound by chelates or ligands, which are arranged linearly or cyclically [30]. Descriptive statistical analysis was performed after the following tests. ## 4.1. Bacterial Strains The selected bacterial strains are clinically relevant entities with high loads in abscesses. They are responsible for many abscesses in the clinic. M. smegmatis was chosen, as it has a long doubling time, which was helpful in certain experiments. Six different bacterial strains were used in the different experiments of this study: S. aureus (ATCC 25923), P. aeruginosa (ATCC 27853), B. subtilis (DSM 618) and B. subtilis (spore suspension, Merck 110649, number of germinable spores 8 × 106 to 5 × 107 CFU/mL), M. smegmatis (ATCC 35798) and E. coli (ATCC 25922). The following methods were chosen similarly to disinfectant and antimicrobial activity tests (e.g., ISO 22196:2011 and DIN EN 13727:2012). The methods were technically available in the laboratory. ## 4.2. Antibacterial Efficacy Tests in Blood Culture Bottles Suspensions of fresh overnight cultures of S. aureus, P. aeruginosa and B. subtilis were prepared (2 different densities of 1.5 × 108 CFU/mL to 5.0 × 108 CFU/mL and 1.5 × 103 CFU/mL to 5.0 × 103 CFU/mL). One milliliter of this suspension was incubated with 9 mL of the nondiluted contrast agent Ultravist, Iopamiro, Telebrix Gastro or Visipaque. Two different pH values of 5.5 und 7.0 were chosen to show possible differences in growth. A pH value of 7.0 is optimal for all used microorganisms, and contrast agents are stable at this pH. On the other hand, a pH value of 5.5 simulates the pH value in abscess cavities, and instabilities of the contrast agents may be possible [31,32,33]. The bacterial suspensions were adjusted drop by drop with 0.5 molar HCL. The resulting pH value therefore varied slightly. These solutions were incubated at 37 °C. After 24 h and 48 h, 5 mL was taken out and incubated in an aerobic blood culture bottle (bact/alert R PF Plus, BioMérieux, Marcy-L’Etoile, France) for 10 to 15 h in an automated system for blood cultures. As a negative control, 1 mL of the bacterial test solution was incubated in sterile $0.9\%$ sodium chloride. ## 4.3. Agar Disk Diffusion A 2 mL ampule of B. subtilis spore suspension was added to agar (Merck 110663) before pouring plates (spore concentration of 6400 CFU to 40,000 CFU/mL). The 2 pH conditions of 5.5 (with 0.5 molar HCL) and 7.4 (with 0.5 molar NaOH) were adjusted in the agar. The bacteria were metabolically active, i.e., there was a variation from well to well depending on growth. Four nonimpregnated paper disks (BioRad, Art. No. 66101) were distributed uniformly onto the plates. Ten microliters of contrast medium (Ultravist, Iopamiro, Telebrix Gastro, Visipaque, Multihance and Dotarem) were pipetted, undiluted, onto paper disks. As positive control, erythromycin (5 mg/L, BioRad, Hercules, CA, USA) was used. All experiments were conducted as duplicates and repeated on two consecutive days. In a second test series, the bacterial strains S. aureus and P. aeruginosa were used. Of each microorganism dense, homogenous suspensions were prepared (McFarland standard of 0.5 for S. aureus and 1.0 for P. aeruginosa). A quantity of 200 µL of this suspension was plated onto Müller–Hinton agar plates. Two pH conditions of 5.5 and 7.4 were adjusted in the agar. Four nonimpregnated paper disks (BioRad, Art. No. 66101) were distributed uniformly onto the plates. 10 µL of contrast medium were pipetted undiluted onto the paper disks. As a positive control, erythromycin (5 mg/L, BioRad) was used. Blank controls without antibiotics were not included as we know from other experiments that the filter paper disks themselves do not inhibit bacterial growth. This was checked in the daily internal QC of our diagnostic laboratory. ## 4.4. Broth Microdilution in Minimal Medium Dense, homogenous suspensions of the microorganisms E.coli, S. aureus, M. smegmatis and F. necrophorum were prepared in Müller–Hinton Broth. The growth time in Müller–Hinton broth growth was 8–12 h for E. coli and S. aureus and 24 h for M. smegmatis. According to our earlier experiments, these species are in the log growth phase after this time span. Then the cultures were transferred into a minimal medium M9 supplemented with $2\%$ glucose (produced in-house in its own media production unit). The bacteria were cultured in this medium until dense suspensions were reached. Next, 1.5 mL was taken out and pelleted by centrifugation. The pellet was washed with PBS and transferred into minimal medium M9. A quantity of 100 µL of all bacterial strains in the two media were placed onto a microplate and incubated with 100 µL of each contrast agent (Ultravist, Iopamiro, Telebrix Gastro, Visipaque, Multihance and Dotarem). Negative and positive controls were included. The plates were covered by a film and incubated at 37 °C aerobic conditions for E. coli, S. aureus and M. smegmatis and anaerobic conditions for F. necrophorum. ## 5. Conclusions The possible correlations obtained from our new results concerning the antibacterial effect of X-ray as well as MRI contrast agents may have far-reaching consequences. Diagnostics for abscess pathogens in materials with contrast agents should involve molecular analytics, i.e., PCR-based methods in case of inhibited culture growth. Furthermore, so far, the use of X-ray contrast media in body cavities has to be monitored only fluoroscopically as well as the MRI contrast media used in this study are not explicitly prohibited for use in body cavities (e.g., [34,35]). Since the number of minimally invasive, percutaneous interventions is steadily increasing compared to surgery, further research must be carried in this direction. Consequently, there is a possibility that, on the one hand, additional therapeutic options may be gained by luminal administration of contrast agents into abscess cavities and, on the other hand, that the mode of administration of these contrast agents may need to be reevaluated. ## References 1. Kelly F.C.. Iodine in Medicine and Pharmacy Since its Discovery—1811–1961. Proc. R. Soc. Med. (1961.0) 54 831-836. DOI: 10.1177/003591576105401001 2. Haley T.J., Raymond K., Komesu N., Upham H.C.. Toxicological and Pharmacological Effects of Gadolinium and Samarium Chlorides. Br. J. Pharmacol. Chemother. (1961.0) 17 526-532. DOI: 10.1111/j.1476-5381.1961.tb01139.x 3. Baykara M., Ozcan M., Bilgen M., Kelestimur H.. Effects of gadolinium and gadolinium chelates on intracellular calcium signaling in sensory neurons. Neurosci. Lett. (2019.0) 707 134295. DOI: 10.1016/j.neulet.2019.134295 4. Blake M., Halasz S.. The effects of X-ray contrast media on bacterial growth. Australas. Radiol. (1995.0) 39 10-13. DOI: 10.1111/j.1440-1673.1995.tb00223.x 5. Kuhns L.R., Baublis J.V., Gregory J., Poznanski A.K.. In Vitro Effect of Cystographic Contrast Media on Urinary Tract Pathogens. Investig. Radiol. (1972.0) 7 112-117. DOI: 10.1097/00004424-197203000-00008 6. Narins D.J., Chase R.M.. The Effect of Hypaque upon Urine Cultures. J. Urol. (1971.0) 105 433-435. DOI: 10.1016/S0022-5347(17)61543-X 7. Speck U.. Kontrastmittel: Übersicht, Anwendung und Pharmazeutische Aspekte (2013.0) 8. Green K.A., Mustachi B., Schoer K., Moro D., Blend R., McGeer A.. Gadolinium-based MR contrast media: Potential for growth of microbial contaminants when single vials are used for multiple patients. Am. J. Roentgenol. (1995.0) 165 669-671. DOI: 10.2214/ajr.165.3.7645493 9. Beussink D.R., Godat J.F., Seaton T.. Antimicrobial properties of magnetic resonance imaging contrast media. Am. J. Health Pharm. (2000.0) 57 48-50. DOI: 10.1093/ajhp/57.1.48 10. Bruins M.J., Zwiers J.H., Verheyen C.C.P.M., Wolfhagen M.J.H.M.. In vitro study of the antimicrobial effects of radiological contrast agents used in arthrography. J. Bone Jt. Surg. (2011.0) 93-B 126-130. DOI: 10.1302/0301-620X.93B1.24968 11. Dawson P., Becker A., Holton J.M.. The effect of contrast media on the growth of bacteria. Br. J. Radiol. (1983.0) 56 809-815. DOI: 10.1259/0007-1285-56-671-809 12. Rendl J., Saller B.. Schilddruse und Rontgenkontrastmittel-Pathophysiologie, Haufigkeit und Prophylaxe der jodinduzierten Hyperthyreose. Dtsch. Arztebl-Arztl. Mitt. Ausg. A (2001.0) 98 402-406 13. Langer R., Usmani A., van Gorkom K., Lorke D., Petroianu G., Azimullah S., Nurulain S.. In vitro assessment of the antibiotic efficacy of contrast media and antibiotics and their combinations at various dilutions. Br. J. Radiol. (2010.0) 83 394-400. DOI: 10.1259/bjr/62389328 14. Caravan P., Ellison J.J., McMurry T.J., Lauffer R.B.. Gadolinium (III) chelates as MRI contrast agents: Structure, dynamics, and applications. Chem. Rev. (1999.0) 99 2293-2352. DOI: 10.1021/cr980440x 15. Black A.P., Kinman R.N., Keirn M.A., Smith J.J., Harlan W.E.. The disinfection of swimming pool waters. I. Comparison of iodine and chlorine as swimming pool disinfectants. Am. J. Public Health Nations Heath (1970.0) 60 535-545. DOI: 10.2105/AJPH.60.3.535 16. Chang S.L.. The use of active iodine as a water disinfectant. J. Am. Pharm. Assoc. (Sci. Ed.) (1958.0) 47 417-423. DOI: 10.1002/jps.3030470610 17. Ellis K., Van Vree H.. Iodine used as a water-disinfectant in turbid waters. Water Res. (1989.0) 23 671-676. DOI: 10.1016/0043-1354(89)90198-X 18. 18. World Health Organization Iodine in Drinking-WaterWorld Health OrganizationGeneva, Switzerland2020. Iodine in Drinking-Water (2020.0) 19. Clarke N., Berg G., Kabler P., Chang S., TAFT R.A.. Human enteric viruses in water: Source, survival and removability. Advances in Water Pollution Research (1964.0) 523-541 20. Taylor G.R., Butler M.. A comparison of the virucidal properties of chlorine, chlorine dioxide, bromine chloride and iodine. Epidemiol. Infect. (1982.0) 89 321-328. DOI: 10.1017/S0022172400070856 21. Pasternak J.J., Williamson E.E.. Clinical Pharmacology, Uses, and Adverse Reactions of Iodinated Contrast Agents: A Primer for the Non-radiologist. Mayo Clin. Proc. (2012.0) 87 390-402. DOI: 10.1016/j.mayocp.2012.01.012 22. Wahsner J., Gale E.M., Rodríguez-Rodríguez A., Caravan P.. Chemistry of MRI Contrast Agents: Current Challenges and New Frontiers. Chem. Rev. (2018.0) 119 957-1057. DOI: 10.1021/acs.chemrev.8b00363 23. Sheet P.. Berlex Ultravist Imaging Agent Will Be Launched in July; Approved May 10, Two Years after Advisory Committee 24. Dawson P., Clauss W.. Contrast Media in Practice: Questions and Answers (2012.0) 25. Dotarem 26. News I.T.. Bracco’s MultiHance to Show Improved MR Visualization 27. Cleary K., Peters T.M.. Image-Guided Interventions: Technology Review and Clinical Applications. Annu. Rev. Biomed. Eng. (2010.0) 12 119-142. DOI: 10.1146/annurev-bioeng-070909-105249 28. Böhm I., Speck U., Schild H.. Cell-dependent influence on the phagocytosis induced by non-ionic contrast medium injection. Br. J. Radiol. (2008.0) 81 199-203. DOI: 10.1259/bjr/13906058 29. Baert A.L.. Encyclopedia of Imaging (2008.0) Volume 2 30. Li H., Meade T.J.. Molecular Magnetic Resonance Imaging with Gd(III)-Based Contrast Agents: Challenges and Key Advances. J. Am. Chem. Soc. (2019.0) 141 17025-17041. DOI: 10.1021/jacs.9b09149 31. Bessman A.N., Page J., Thomas L.J.. In vivo pH of induced soft-tissue abscesses in diabetic and nondiabetic mice. Diabetes (1989.0) 38 659-662. DOI: 10.2337/diab.38.5.659 32. Huemer M., Mairpady Shambat S., Bergada-Pijuan J., Söderholm S., Boumasmoud M., Vulin C., Gómez-Mejia A., Antelo Varela M., Tripathi V., Götschi S.. Molecular reprogramming and phenotype switching in. Proc. Natl. Acad. Sci. USA (2021.0) 118 e2014920118. DOI: 10.1073/pnas.2014920118 33. Nekoofar M.H., Namazikhah M.S., Sheykhrezae M.S., Mohammadi M.M., Kazemi A., Aseeley Z., Dummer P.. pH of pus collected from periapical abscesses. Int. Endod. J. (2009.0) 42 534-538. DOI: 10.1111/j.1365-2591.2009.01550.x 34. Dotarem Inj Solution 35. Ultravist Inj Solution 370 mg
--- title: Long-Term Ingestion of Sicilian Black Bee Chestnut Honey and/or D-Limonene Counteracts Brain Damage Induced by High Fat-Diet in Obese Mice authors: - Simona Terzo - Pasquale Calvi - Domenico Nuzzo - Pasquale Picone - Mario Allegra - Flavia Mulè - Antonella Amato journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966634 doi: 10.3390/ijms24043467 license: CC BY 4.0 --- # Long-Term Ingestion of Sicilian Black Bee Chestnut Honey and/or D-Limonene Counteracts Brain Damage Induced by High Fat-Diet in Obese Mice ## Abstract Obesity is linked to neurodegeneration, which is mainly caused by inflammation and oxidative stress. We analyzed whether the long-term intake of honey and/or D-limonene, which are known for their antioxidant and anti-inflammatory actions, when ingested separately or in combination, can counteract the neurodegeneration occurring in high fat diet (HFD)-induced obesity. After 10 weeks of HFD, mice were divided into: HFD-, HFD + honey (HFD-H)-, HFD + D-limonene (HFD-L)-, HFD + honey + D-limonene (HFD-H + L)-fed groups, for another 10 weeks. Another group was fed a standard diet (STD). We analyzed the brain neurodegeneration, inflammation, oxidative stress, and gene expression of Alzheimer’s disease (AD) markers. The HFD animals showed higher neuronal apoptosis, upregulation of pro-apoptotic genes Fas-L, Bim P27 and downregulation of anti-apoptotic factors BDNF and BCL2; increased gene expression of the pro-inflammatory IL-1β, IL-6 and TNF-α and elevated oxidative stress markers COX-2, iNOS, ROS and nitrite. The honey and D-limonene intake counteracted these alterations; however, they did so in a stronger manner when in combination. Genes involved in amyloid plaque processing (APP and TAU), synaptic function (Ache) and AD-related hyperphosphorylation were higher in HFD brains, and significantly downregulated in HFD-H, HFD-L and HFD-H + L. These results suggest that honey and limonene ingestion counteract obesity-related neurodegeneration and that joint consumption is more efficacious than a single administration. ## 1. Introduction Neurodegenerative diseases (NDs), including Alzheimer’s disease (AD), are characterized by the progressive loss of neurons in areas of the brain, leading to cognitive and functional deterioration. NDs represent a serious problem because they affect about 50 million patients worldwide and this number is estimated to reach 115 million in 2050 [1]. Different factors contribute to the onset and progression of neurodegeneration such as aging, genetics, environment [2] and oxidative stress and inflammation [3,4]. Moreover, obesity and diabetes increase the risk of developing dementia and AD. In fact, the neurodegenerative process is exacerbated by obesity or diabetes, leading to the concept of metabolism-dependent neurodegeneration [5]. Indeed, insulin receptor down-regulation has been observed in the brains of patients with AD [6] confirming the theory that AD may be considered as “type 3 diabetes” [7]. Furthermore, studies on animal models pointed out that obesity affects learning and memory [8,9] and long-term ingestion of high-fat diet (HFD) in rodents is responsible for neuronal loss and synaptic plasticity damage [10,11,12,13]. NDs are as yet incurable and strongly debilitating for the patients. Nevertheless, current research is pushing towards effective therapies [14,15]. Numerous nutraceuticals and/or functional foods are considered protective and/or therapeutic against the metabolic dysfunctions and the related neurodegeneration [16,17,18,19,20]. For example, foods rich in antioxidants, micronutrients, phytochemicals, essential oils, and probiotics have been found to be helpful in maintaining body weight and reducing the incidence of neurodegenerative diseases [1,20]. In particular, honey might prove useful in the treatment of chronic diseases linked to oxidative stress and inflammation due to its high content in polyphenols [21]. Although the composition of honey is variable depending on various factors such as botanical origin, geographical region and climatic conditions, most of the polyphenols present in honey are flavonoids and phenolic acid derivatives that possess anti-inflammatory and neuroprotective properties [22]. More specifically, our recent investigation demonstrated the ability of honey consumption to prevent HFD-dependent neuronal injury. In particular, a 16-week-intake of Sicilian black bee chestnut honey, which is particularly rich in kaempferol and quercetin [22], prevented peripheral and central insulin resistance and neuroinflammation in mice fed with a hyperlipidic diet [23]. This neuroprotective effect proved to be mainly due to the positive modulation of brain genes involved in insulin signaling, neuroinflammation and apoptosis [23]. However, it remains to be investigated whether the long-term ingestion of honey is able to revert obesity-related metabolic dysfunctions and related neurodegeneration. Recently, D-limonene (1-methyl-4-(1-methylethenyl) cyclohexane), a monocyclic monoterpene that is the major constituent of citrus essential oils, has also received notable scientific interest due to its ability to mitigate inflammation and oxidative stress and reduce apoptotic cell death [24]. In fact, it possesses antidiabetic, antioxidant, anti-inflammatory, antinociceptive and anticancer properties [25]. In animal models, D-limonene has been reported to alleviate obesity-related metabolic disorders [26,27]. However, although D-limonene has recently been shown in vitro to inhibit acetylcholinesterase [28] and to exert beneficial effects in the Drosophila AD model by reducing oxidative stress and neuroinflammation [29], data on neuroprotective actions against neuronal damage caused by HFD are lacking. Moreover, a recent study suggested that the usage of D-limonene together with other drugs, such as aminoguanidine, is more efficient in the prevention of secondary complications in diabetes in comparison to single treatment [30]. Therefore, the present research was undertaken with the purpose of exploring whether honey, administered alone or in combination with D-limonene, can represent a potential dietary supplement that can aid in ameliorating or reverting HFD-caused brain damage. In this view, we investigated the effects of the long-term ingestion of honey and D-limonene, separately or in combination, on brain damage in HFD mice when the pathological conditions were overt. ## 2.1. Body Weight, Glycaemia and Serum Lipids As shown in Figure 1A, at the end of the experimental protocol, HFD mice were significantly heavier than STD mice. The weight gain of HFD-L and HFD-H + L mice was significantly lower than that of HFD animals. The fasting blood glucose concentration of HFD mice was significantly higher than that of the STD group. HFD-H and HFD-H + L mice had similar fasting blood glucose concentrations to the HFD group. However, the D-limonene supplementation markedly reduced the fasting blood glucose levels induced by HFD (Figure 1B). The lipid profile of mice that were fed with the different diets is represented in Figure 1C. Total cholesterol and triglyceride levels, that were high in the plasma of the HFD mice compared to STD group, did not significantly differ in HFD-H, HFD-L and HFD-H + L mice, suggesting that honey and D-limonene, alone or in combination, did not impact on the lipid metabolism of obese mice. ## 2.2. Neurodegeneration: TUNEL Assay Neurodegeneration has been suggested to be associated with cell apoptosis. To identify whether apoptotic cells were present in the brain tissues of the different groups of mice, we used the TUNEL assay. A higher number of TUNEL-positive cells was observed in the cortex of HFD mice in comparison with STD mice. As shown in Figure 2, neuronal apoptosis resulting from a high-fat diet was significantly decreased in the cortex of HFD-H, HFD-L, and HFD-H + L mice, suggesting that both honey and D-limonene contributed to neuroprotective effects. Interestingly, the diet containing honey and D-limonene together was more efficacious than the single supplement. ## 2.3. Pro-Apoptosis and Anti-Apoptosis Genes Expression In this work, the gene expression of the most important regulators of apoptosis. The pro-apoptotic factors FAS-L, P27, and BIM were significantly upregulated in mouse brain tissues from the HFD group compared to the STD group. A high-fat diet supplemented with honey, D-limonene or honey plus D-limonene significantly decreased the gene expression levels of all investigated factors, suggesting a reduced presence of neurons that undergo programmed cell death (Figure 3A,B). On the contrary, the brain gene expression of factors that help neuronal survival, such as BDNF and BCL2, was decreased in the HFD mice compared to the STD group. This down-regulation induced by HFD was counteracted by the simultaneous ingestion of honey or D-limonene. HFD-H + L proved to be the most efficacious diet to increase BCL2 and BDNF expression (Figure 3C,D). ## 2.4. Brain Pro-Inflammatory Gene and Protein Expression To determine whether honey and D-limonene, together or separately, reduced neuroinflammation, we examined the brain expression of some pro-inflammatory cytokines and other proteins, which are markers of inflammation. The IL-1β, IL-6 e TNF-α increased expression, found in HFD brains, was reduced by honey or D-limonene ingested separately, and it returned to control levels in the brain of HFD-H + L mice, suggesting that the combined administration of honey and D-limonene was more efficacious than single administration (Figure 4A,B). Moreover, the elevated expression of COX-2 and iNOS induced by HFD, was mitigated by honey, D-limonene, and honey plus D-limonene (Figure 5A,B). ## 2.5. Brain Oxidative Stress Increasingly, studies have demonstrated that oxidative stress is critical for neuronal injury. Therefore, we determined the effect of the different supplemented diets on ROS generation and nitrite content in the brain of the different groups of animals. After a 20-week HFD administration, ROS generation assessed with H2DCF-DA was significantly increased in HFD brain compared not only to STD, but also to HFD-H, HFD-L, and HFD-H + L (Figure 6A). Moreover, we found a significant increase in nitrite levels in the brains of HFD obese animals in comparison with STD animals. HFD-H, HFD-L, and HFD-H + L mice showed nitrite values that were significantly lower than those of HFD mice (Figure 6B). ## 2.6. Expression of Genes Involved in AD Using a Mouse Alzheimer’s Disease RT2 Profiler PCR Array we analyzed expression changes of genes involved in the onset, development and progression of Alzheimer’s disease in the different groups of animals. Among them, there are genes that contribute to amyloid beta-peptide (Aβ) generation, clearance and degradation but also genes related to neuronal toxicity. The list of genes is shown in Table S1. We focused on the gene expression levels that were affected more than two-fold among the analyzed groups. The results showed that in the HFD brains, various genes involved in the processing of Amiloid β Precusor (APP) and TAU (Aplp1, Aplp2, App, Apba3, Apbb2, Apoe, Ckk5, Clu, Ctsl, Mapt, Prkca, Prkce and Hsd17b10), in synaptic function (Ache), in AD-related iperphosphorylation (Gsk3α, GCdk5 and Prkca), and in inflammation (MPO and Il-1α) (Table 1) were upregulated in comparison with lean brains. These abnormal expressions were significantly ameliorated in the brain of obese animals fed with honey, D-limonene and honey plus D-limonene with a major improvement in the HFD-H group (Table 1). ## 3. Discussion The results of the present study suggest that long-term intake of Sicilian black bee chestnut honey and/or D-limonene, ingested separately or in combination, can protect central neurons against HFD-induced cerebral damage by reducing oxidative stress and neuroinflammation. To our knowledge, our study is the first report on the neuroprotective effects of D-limonene against the damage induced by HFD. Epidemiological human studies pointed out that a high-calorie diet is associated with worse performance on cognitive tasks [31]. It increases the risk of dementia because high lipid content causes oxidative stress and neuronal dysfunctions [32]. Indeed, high stress oxidative triggers the up-regulation of pro-inflammatory factors leading to neuroinflammation [33]. However, different biological mechanisms including insulin resistance, developmental disturbances, altered membrane functioning, and altered vascularization have been involved in HFD-induced neuronal damage and cognitive decline [5,32]. In our experiments we used mice which, following chronic consumption of HFD, developed obesity accompanied by hyperglycemia, dyslipidaemia, insulin resistance [34,35,36], activation of amyloidogenic pathways, neuroinflammation and neurodegeneration [10,18,37,38,39]; consequently, they are suitable for verifying the potential effects of functional food/phytochemicals on neuronal survival. First of all, we analyzed the presence of apoptosis in the cerebral cortex and the gene expression of pro- and anti-apoptotic factors in the brains of the different animal groups. It is well known that apoptosis plays a key role in the pathogenesis of neurodegenerative diseases [40], involving mainly the BCL-2 protein family. This family includes proteins that control the mithocondrion membrane permeability such as Bax, Bim (pro-apoptotic proteins) and BCL-2, Bcl-xL, Bcl-w (anti-apoptotic proteins). Additionally, FAS ligand (FAS-L) has been involved in neuronal death [41] and P-27, an inhibitor of cyclin-dependent kinase, has been reported to promote neuronal apoptosis induced by the neurotoxic αβ42 peptide [42]. According to our previous reports [10,17,18], our results confirmed the presence of neurodegeneration caused by HFD as suggested by the increase of apoptotic neurons in the brain cortex of obese mice in comparison with STD mice. In HFD-H or HFD-L cerebral cortexes, the level of apoptotic neurons was significantly reduced suggesting that the daily ingestion of honey or D-limonene inhibits programmed cellular death. Moreover, honey and D-limonene ingested in combination further decreased the apoptotic neuron number, suggesting a synergistic neuroprotective action. The results from molecular analysis also supported our hypothesis on the neuroprotective effect of honey and D-limonene. In fact, the pro-apoptotic gene up-regulation and the anti-apoptotic gene down-regulation that was found in the HFD brain was attenuated in HFD-H, HFD-L and HFD-H + L animal groups. We also found a down-regulation of BDNF in the HFD brain, which was in accordance with previous studies that reported reductions in levels of BDNF in the hippocampus of obese rodents [11,43] as a consequence of increased oxidative stress [44]. However, honey and D-limonene when separately ingested increased the BDNF gene expression; even more so when ingested in combination, suggesting that an increase of survival factors can also be responsible for the observed beneficial effects. Neurodegeneration can be triggered by various pro-inflammatory and neurotoxic mediators, such as IL-1β, IL-6, and TNF-α, and neuroinflammation is strictly associated with oxidative stress [45]. Indeed, several studies demonstrated that neuroinflammation is linked to high levels of ROS and high expression of AD biomarkers in the brains of HFD mice [10,46,47]. Because both honey and D-limonene have been reported to possess anti-inflammatory and antioxidant properties, leading to the assumption that they could be used as a supplement in anti-inflammatory therapies [21,48,49] we examined and compared the expression of pro-inflammatory factors, the levels of oxidative stress and nitrite in the brains of the different animal groups. The results suggested that HFD increases the gene expression of inflammatory cytokines (IL-1β, IL-6, TNF-α) and other proteins, markers of inflammation (i-NOS and COX-2) and ROS and nitrite levels in the brain as previously shown [10,18,32,50,51]. Interestingly, long-term ingestion of honey or D-limonene, and even more so, the combined ingestion of honey and D-limonene reduced the inflammatory and oxidative stress markers suggesting once more a beneficial action against damage induced by HFD in the brain. We can only speculate about the honey compounds responsible for the observed beneficial effects, which generally have been attributed to polyphenols [21]. However, it is noteworthy that we used Sicilian black bee chestnut honey, whose kaempferol and quercetin levels corresponded to $69\%$ of the total content [22]. Quercetin as well as kaempferol can cross the blood–brain barrier [52]. Quercetin has been reported to protect neurons from oxidative stress and inflammation and to have beneficial properties against mechanisms involved in AD in different in vitro and in vivo models [53], and kaempferol can act positively in various models of neurodegenerative diseases [54,55]. Although recent research using a Drosophila AD model suggested that D-limonene has a neuroprotective action against Aβ42-induced toxicity associated with its antioxidant and anti-inflammatory properties [29], the effects of D-limonene on AD have not been well-studied yet. Therefore, by using a mouse Alzheimer’s disease microarray, we have analyzed and compared the expression of genes involved in amyloid beta-peptide (Aβ) generation and processing and/or genes related to neuronal toxicity in the brains of different mouse groups. The results clearly suggest that long-term HFD feeding promotes the expression of genes associated with AD, including Ache, App, Apba3, Apbb2, Aplp1, Aplp2, Apoe, CdK5, Clu, Ctls, GSK3α, Hsd17b10, Mapt, Psen1, Prkca,Prkcb and genes linked to inflammation such as Mpo and Il1α [56,57]. However, these deleterious changes in gene expression were counteracted in the brains of HFD-H, HFD-L and HFD-H + L, suggesting that the increased neurotoxicity induced by HFD may be mitigated by long-term ingestion of honey and D-limonene, both separately and in combination. In particular, the down regulation of App, Apba3, Apbb2, Aplp1, Aplp2, Apoe and Psen1 could suggest that the eventual endogenous APP generation and processing were reduced after the long-term ingestion of honey and D-limonene [58]. Moreover, Cdk5, a promoter of neuronal death [59] and Clu, encoding clusterin, a protein involved in several processes such as suppression of the complement system, lipid transport, and neuronal cell death and cell-survival mechanisms, whose levels are increased in AD [60], were mitigated by the intake of honey and D-limonene either alone or in combination. ## 4.1. Animals and Diets Male C57BL/6 mice, purchased from Envigo (S.Pietro al Natisone, Udine, Italy) were maintained in the ATeN center animal house according to the European guide lines. The animals (4-weeks old) were housed (2 mice/cage) in a temperature- (23 ± 1 °C) and relative humidity ($55\%$ ± $5\%$)-controlled facility, under a 12-h light–dark cycle, according to the Italian legislative decree n. $\frac{26}{2014}$ and were approved by the Ministry of Health (Rome, Italy; Authorization n. $\frac{891}{2018}$-PR). After two weeks of acclimatization, 8 mice were fed a standard diet (STD) (negative control) containing protein $20.0\%$, fat $10.0\%$, carbohydrate $70.0\%$, w/w, and water (code 4RF25, Mucedola, Milan, Italy), and 32 mice were fed a HFD, containing protein $20.0\%$, fat $60.0\%$, carbohydrate $20.0\%$, w/w (PF4215, Mucedola, Milan, Italy) for 10 weeks to induce obesity. Subsequently, HFD mice were divided randomly into four groups Then, four groups ($$n = 8$$/group) were created from the HFD mice: one group received HFD, the second group received HFD supplemented with honey (45 mg per day/mouse) (HFD-H), the third received HFD supplemented with D-limonene ($0.5\%$ w/w) (HFD-L) and the last one received HFD supplemented with honey and D-limonene in combination at the same doses (HFD-H + L), for another 10 weeks. The doses of D-limonene (Sigma—St. Louis, MO, USA) and honey (Prezzemolo and Vitale Supermarket, Palermo, Italy) were taken from the literature [23,27,61] and added to the HFD cow in a percentage amount that was useful so as not to change the HFD caloric value. Body weight and food intake were monitored every week. At the end of the experimental protocol (20th week), biochemical analyses were performed on blood collected from the tail vein and then the animals were sacrificed. The aorta was perfused with a buffer solution of Dulbecco and the right atrium was incised to allow outflow. Brains were rapidly explanted, weighed and coronally cut into two halves. One part was fixed in $4\%$ formalin was utilized for histological investigation; the other half was ice-covered and used for molecular analysis. ## 4.2. Biochemical Analyses Glucose concentration was measured using a glucometer (GlucoMen LX meter, Menarini, Florence, Italy) in overnight fasting mice. Plasma total cholesterol and triglyceride concentrations were determined using the ILAB 600 Analyzer (Instrumentation Laboratory, Bedford, MA, USA). ## 4.3. Apoptosis Investigation The Tunel assay was used to determine the level of apoptosis (Promega, Madison, WI, USA) in the cerebral cortex sections, following the manufacturer’s instructions. The values of the damaged nuclei were counted by two blind investigators and the ratio of apoptotic nuclei in respect of normal nuclei was calculated. ## 4.4. Reactive Oxygen Species Analysis To determine the reactive oxygen species (ROS), 5 mg of brain tissue was homogenized with 1 mL of cold PBS1X and 10 μL of protease inhibitors (Amersham Life Science, Munich, Germany). The preparate brain homogenates were incubated with 1 mM dichlorofluorescein diacetate (DCFH-DA) at room temperature in the dark for 15 min, then the fluorescence was measured by fluorimeter (GloMax® Plate Reader, Promega, Milano, Italy) with an excitation filter set at 485 nm and an emission filter set at 530 nm. ROS levels were expressed as a percentage of the fluorescence emitted by STD cerebral samples. ## 4.5. Determination of Nitric Oxide (NO) Levels The level of nitric oxide (NO) in the brains was evaluated by using Griess reagent (Thermo Fisher Scientific Inc., Waltham, MA, USA). Briefly, 5 mg of brain tissue was homogenized with 1 mL of PBS1X and centrifuged at 14,000 rpm, for 30 min at 4 °C. 100 μL of supernatant was incubated with equal volumes of Griess reagent ($1\%$ sulphanilamide in $5\%$ phosphoric acid and $0.1\%$ N-(1-naphthyl)-ethylenediamine), the absorbance was immediately read at 520 nm in a microplate reader (GloMax® Plate Reader, Promega). ## 4.6. Molecular Analyses Whole brain was used to extract RNA by using a RNeasy plus Mini Kit (Qiagen, Valencia, CA, USA). Subsequently, by using High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Waltham, MA, USA). cDNA was prepared by 2 ng of total RNA. Then the expression of target genes was performed by using Reverse Transcription Polymerase Chain Reaction (RT-PCR) with the subsequent primers: β-actin For 5′-CGGGATCCCCGCCCTAGGCACCAGGGT-3′; Rev 5′-GGAATTCGGCTGGGGTGTTGAAGGTCTCAAA-3′; for pro-inflammatory factors: IL-1β For 5′-CATGGGATGATGATAACCTGCT-3′; Rev 5′-CCCATACTTTAGGAAGACACGATT-3′; IL-6 For 5′-CTGGTGACAACCACGGCCTTCCCT-3′; Rev 5′-ATGCTTAGGCATAACGCACTAGGT-3′; TNF-α For 5′-AGCCCACGTCGTAGCAAACCA-3′; Rev 5′-GCAGGGGCTCTTGACGGCAG-3′; for pro-apoptotic factors: FAS-L For 5′-CAAGTCCAACTCAAGGTCCATGCC-3′; Rev 5′-AGAGAGAGCTCAGATACGTTTGAC-3′; BIM For 5′-AACCTTCTGATGTAAGTTCT-3′; Rev 5′-GTGATTGCCTTCAGGATTAC-3′; p27 For 5′-TGCGAGTGTCTAACGGGAG-3′; Rev 5′-GTTTGACGTCTTCTGAGGCC-3′; for anti-apoptosis factors: BCL-2 For 5′-ATGTGTGTGGAGAGCGTCAA-3′; Rev 5′-AGAGACAGCCAGGAGAAATCA-3′; BDNF For 5′-GGCTGACACTTTTGAGCACGTC-3′; Rev 5′-CTCCAAAGGCACTTGACTGCTG-3′. The amplification cycles comprised denaturation (45 s at 95 °C), annealing (45 s at 52 °C) and elongation (45 s at 72 °C), for 40 cycles. The amplification products were visualized by ultraviolet light using E-Gel GelCapture (Thermo Fisher Scientific, Monza, Italy) after separation on agarose gel. The quantification of gene expression was obtained by using E-Gel GelQuant Express Analysis Software (version 1.14.6.0 (Dongle)) (Thermo Fisher Scientific, Monza, Italy). The signal intensity of the products was normalized to its respective β-actin signal intensity. Protein expression. Brains dissected from the experimental animals were homogenized in ice-cold solubilization buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM DDT, $1\%$ Triton X-100, 24 mM sodium deoxycholate, $0.01\%$ SDS, 10 mM sodium pyrophosphate, 100 mM sodium fluoride, 10 mM sodium orthovanadate, 1.5 µM aprotinin, 1 mM phenylmethanesulfonylfluoride and 2.1 µM leupeptin) and centrifuged at 12,000× g at 4 °C for 30 min. Then, the supernatants were used for protein determination, as previously described [62]. Samples containing 50 µg protein were resolved by SDS-PAGE electrophoresis on $12\%$ acrylamide gels and transferred to nitrocellulose membranes. After blocking for 2 h in $5\%$ (w/v) skimmed dry milk, the membranes were incubated in the presence of primary antibodies overnight at 4 °C (Santa Cruz, Milan, Italy, 1:1000 dilution): anti-COX-2 (sc-376861), anti-iNOS (sc-7271). Subsequently, the samples were incubated with the secondary for 90 min. HRP-conjugated antibodies (Dako, Milan, Italy, 1:10,000 dilution) and chemiluminescent bands were detected by a C-Digit Blot Scanner (LI-COR, Lincoln, NE, USA) and densitometric analysis was used to analyze band intensities, by using LI-COR Image Studio 4.0. ## 4.7. RT2 Profiler PCR Array Mouse Alzheimer’s disease array (Alzheimer’s Disease RT2 Profiler PCR Array, QIAGEN, Monza, Italy) was used in order to analyze factors in HFD brains that are involved in the onset, development and progression of AD. The 96 genes reported in the plate are listed in the Supplementary data (Table S1). RNA from whole brains was utilized. A High-Capacity cDNA Reverse Transcription kit (Applied Biosystems, Bedford, MA, USA) was used to synthetize cDNA from 2 ng of RNA. The array was executed by using a StepOne Real-Time instrument (Applied Biosystem) and the results were obtained through the relative quantification method (2−ΔΔCT). We chose to highlight only the genes showing changes in the expression levels that were more than two-fold among the different groups analyzed (HFD vs. Lean; HFD-H vs. HFD; HFD-L vs. HFD; HFD-H + L vs. HFD). ## 4.8. Statistical Analysis The results are presented as mean values ± the standard error of the mean SEM. The number of animals is indicated with the letter ‘n’. The comparison between the groups was performed by ANOVA, and then a Bonferroni post hoc test was used. All the analyses were obtained using Prism 6.0, GraphPad software (San Diego, CA, USA). Results with a p-value ≤ 0.05 were considered statistically significant. ## 5. Conclusions In conclusion, our results confirm that HFD causes detrimental effects on AD-related neuropathological and neuroinflammatory pathways leading to neurodegeneration. However, the long-term ingestion of honey and D-limonene, either separately or in combination, is able to counteract and to ameliorate the cerebral stressing conditions related to HFD-induced metabolic disorders. ## References 1. Livingston G., Huntley J., Sommerlad A., Ames D., Ballard C., Banerjee S., Brayne C., Burns A., Cohen-Mansfield J., Cooper C.. Dementia prevention, intervention, and care: 2020 report of the Lancet Commission. Lancet (2020) 396 413-446. DOI: 10.1016/S0140-6736(20)30367-6 2. Reitz C., Brayne C., Mayeux R.. Epidemiology of Alzheimer disease. Nat. Rev. Neurol. (2011) 7 137-152. DOI: 10.1038/nrneurol.2011.2 3. Singh A., Kukreti R., Saso L., Kukreti S.. Oxidative stress: A key modulator in neurodegenerative diseases. Molecules (2019) 24. DOI: 10.3390/molecules24081583 4. Stephenson J., Nutma E., van der Valk P., Amor S.. Inflammation in CNS neurodegenerative diseases. Immunology (2018) 154 204-219. DOI: 10.1111/imm.12922 5. Terzo S., Amato A., Mulè F.. From obesity to Alzheimer’s disease through insulin resistance. J. Diabetes Complicat. (2021) 35 108026. DOI: 10.1016/j.jdiacomp.2021.108026 6. Frazier H., Ghoweri A., Anderson K., Lin R., Porter N.M., Thibault O.. Broadening the definition of brain insulin resistance in aging and Alzheimer’s disease. Exp. Neurol. (2019) 313 79-87. DOI: 10.1016/j.expneurol.2018.12.007 7. De la Monte S.. Type 3 diabetes is sporadic Alzheimer’s disease: Mini-review. Eur. Neuropsychopharmacol. (2014) 24 1954-1960. DOI: 10.1016/j.euroneuro.2014.06.008 8. Farr S., Yamada K., Butterfield D., Abdul H.M., Xu L., Miller N.E., Banks W.A., Morley J.E.. Obesity and hypertriglyceridemia produce cognitive impairment. Endocrinology (2008) 149 2628-2636. DOI: 10.1210/en.2007-1722 9. Valladolid-Acebes I., Stucchi P., Cano V., Fernández-Alfonso M., Merino B., Gil-Ortega M., Fole A., Morales L., Ruiz-Gayo M., Del Olmo N.. High-fat diets impair spatial learning in the radial-arm maze in mice. Neurobiol. Learn. Mem. (2011) 95 80-85. DOI: 10.1016/j.nlm.2010.11.007 10. Nuzzo D., Picone P., Baldassano S., Caruana L., Messina E., Marino Gammazza A., Cappello F., Mulè F., Di Carlo M.. Insulin resistance as common molecular denominator linking obesity to Alzheimer’s disease. Curr. Alzheimer Res. (2015) 12 723-735. DOI: 10.2174/1567205012666150710115506 11. Stranahan A., Norman E., Lee K., Cutler R.G., Telljohann R.S., Egan J.M., Mattson M.P.. Diet-induced insulin resistance impairs hippocampal synaptic plasticity and cognition in middle-aged rats. Hippocampus (2008) 18 1085-1088. DOI: 10.1002/hipo.20470 12. Siino V., Amato A., Di Salvo F., Caldara G.F., Filogamo M., James P., Vasto S.. Impact of diet-induced obesity on the mouse brain phosphoproteome. J. Nutr. Biochem. (2018) 58 102-109. DOI: 10.1016/j.jnutbio.2018.04.015 13. Siino V., Jensen P., James P., Vasto S., Amato A., Mulè F., Accardi G., Larsen M.R.. Obesogenic Diets Cause Alterations on Proteins and Theirs Post-Translational Modifications in Mouse Brains. Nutr. Metab. Insights (2021) 14 11786388211012405. DOI: 10.1177/11786388211012405 14. Salman M., Al-Obaidi Z., Kitchen P., Loreto A., Bill R.M., Wade-Martins R.. Advances in Applying Computer-Aided Drug Design for Neurodegenerative Diseases. Int. J. Mol. Sci. (2021) 22. DOI: 10.3390/ijms22094688 15. Aldewachi H., Al-Zidan R., Conner M., Salman M.. High-Throughput Screening Platforms in the Discovery of Novel Drugs for Neurodegenerative Diseases. Bioengineering (2021) 8. DOI: 10.3390/bioengineering8020030 16. Carvalho J., Fernandes C., Daleprane J., Alves M.S., Stien D., Dhammika Nanayakkara N.P.. Role of Natural Antioxidants from Functional Foods in Neurodegenerative and Metabolic Disorders. Oxid. Med. Cell Longev. (2018) 18 459753. DOI: 10.1155/2018/1459753 17. Nuzzo D., Amato A., Picone P., Terzo S., Galizzi G., Bonina F.P., Mulè F., Di Carlo M.. A natural dietary supplement with a combination of nutrients prevents neurodegeneration induced by a high fat diet in mice. Nutrients (2018) 10. DOI: 10.3390/nu10091130 18. Nuzzo D., Galizzi G., Amato A., Terzo S., Picone P., Cristaldi L., Mulè F., Di Carlo M.. Regular intake of pistachio mitigates the deleterious effects of a high fat-diet in the brain of obese mice. Antioxidants (2020) 9. DOI: 10.3390/antiox9040317 19. Storz M.. Is there a lack of support for whole-food, plant-based diets in the medical community?. Perm. J. (2018) 23 18-068. DOI: 10.7812/TPP/18-068 20. González-Castejón M., Rodriguez-Casado A.. Dietary phytochemicals and their potential effects on obesity: A review. Pharm. Res. (2011) 64 438-455. DOI: 10.1016/j.phrs.2011.07.004 21. Terzo S., Mulè F., Amato A.. Honey and obesity-related dysfunctions: A summary on health benefits. J. Nutr. Biochem. (2020) 82 108401. DOI: 10.1016/j.jnutbio.2020.108401 22. Lo Dico G.A., Ulrici A., Pulvirenti A., Camilleri G., Macaluso A.. Multivariate statistical analysis of the polyphenols content for the discrimination of honey produced in Sicily (Southern Italy). J. Food Compos. Anal. (2019) 82 103225. DOI: 10.1016/j.jfca.2019.05.008 23. Terzo S., Calvi P., Nuzzo D., Picone P., Galizzi G., Caruana L., Di Carlo M., Lentini L., Puleio R., Mulè F.. Preventive impact of long-term ingestion of chestnut honey on glucose disorders and neurodegeneration in obese mice. Nutrients (2022) 14. DOI: 10.3390/nu14040756 24. Eddin L., Jha N., Meeran M., Kesari K.K., Beiram R., Ojha S.. Neuroprotective Potential of Limonene and Limonene Containing Natural Products. Molecules (2021) 26. DOI: 10.3390/molecules26154535 25. Vieira A., Beserra F., Souza M., Totti B.M., Rozza A.L.. Limonene: Aroma of innovation in health and disease. Chem. Biol. Interact. (2018) 283 97-106. DOI: 10.1016/j.cbi.2018.02.007 26. Jesudoss V.A.S., Jayachitra J., Shenbagam M., Nalini N.. Dietary d-limonene alleviates insulin resistance and oxidative stress-induced liver injury in high-fat diet and L-NAME-treated rats. Eur. J. Nutr. (2012) 51 57-68. DOI: 10.1007/s00394-011-0182-7 27. Jing L., Zhang Y., Fan S., Gu M., Guan Y., Lu X., Huang C., Zhou Z.. Preventive and ameliorating effects of citrus D-limonene on dyslipidemia and hyperglycemia in mice with high-fat diet-induced obesity. Eur. J. Pharm. (2013) 715 46-55. DOI: 10.1016/j.ejphar.2013.06.022 28. Szwajgier D., Baranowska-Wójcik E.. Terpenes and phenylpropanoids as acetyl- and butyrylcholinesterase inhibitors: A comparative study. Curr. Alzheimer Res. (2019) 16 963-973. DOI: 10.2174/1567205016666191010105115 29. Shin M., Liu Q., Choi B., Shin C., Lee B., Yuan C., Song Y.J., Yun H.S., Lee I.S., Koo B.S.. Neuroprotective Effects of Limonene (+) against Aβ42-Induced Neurotoxicity in a Drosophila Model of Alzheimer’s Disease. Biol. Pharm. Bull. (2020) 43 409-417. DOI: 10.1248/bpb.b19-00495 30. Joglekar M., Bavkar L., Sistla S., Arvindekar A.. Effective inhibition of protein glycation by combinatorial usage of limonene and aminoguanidine through differential and synergistic mechanisms. Int. J. Biol. Macromol. (2017) 99 563-569. DOI: 10.1016/j.ijbiomac.2017.02.104 31. Ortega R., Requejo A., Andres P., López-Sobaler A.M., Quintas M.E., Redondo M.R., Navia B., Rivas T.. Dietary intake and cognitive function in a group of elderly people. Am. J. Clin. Nutr. (1997) 66 803-809. DOI: 10.1093/ajcn/66.4.803 32. Freeman L.R., Haley-Zitlin V., Rosenberger D.S., Granholm A.C.. Damaging effects of a high- fat diet to the brain and cognition: A review of proposed mechanisms. Nutr. Neurosci. (2014) 17 241-251. DOI: 10.1179/1476830513Y.0000000092 33. Yeo E., Wong K., See M., See M.L., Wong K.Y., Yap J.K.Y., Gan S.Y.. Piper sarmentosum Roxb. confers neuroprotection on beta-amyloid (Aβ)-induced microglia-mediated neuroinflammation and attenuates tau hyperphosphorylation in SH-SY5Y cells. J. Ethnopharmacol. (2018) 217 187-194. DOI: 10.1016/j.jep.2018.02.025 34. Collins S., Martin T.L., Surwit R.S., Robidoux J.. Genetic vulnerability to diet-induced obesity in the C57BL/6J mouse: Physiological and molecular characteristics. Physiol. Behav. (2004) 81 243-248. DOI: 10.1016/j.physbeh.2004.02.006 35. Amato A., Caldara G., Nuzzo D., Baldassano S., Picone P., Rizzo M., Mulè F., Di Carlo M.. NAFLD and Atherosclerosis Are Prevented by a Natural Dietary Supplement Containing Curcumin, Silymarin, Guggul, Chlorogenic Acid and Inulin in Mice Fed a High-Fat Diet. Nutrients (2017) 9. DOI: 10.3390/nu9050492 36. Baldassano S., Rappa F., Amato A., Cappello F., Mulè F.. GLP-2 as Beneficial Factor in the Glucose Homeostasis in Mice Fed a High Fat Diet. J. Cell. Physiol. (2015) 230 3029-3036. DOI: 10.1002/jcp.25039 37. Glass C., Saijo K., Winner B., Marchetto M.C., Gage F.H.. Mechanisms underlying inflammation in neurodegeneration. Cell (2010) 140 918-934. DOI: 10.1016/j.cell.2010.02.016 38. Petrov D., Pedrós I., Artiach G., Sureda F.X., Barroso E., Pallàs M., Casadesús G., Beas-Zarate C., Carro E., Ferrer I.. High-fat diet-induced deregulation of hippocampal insulin signaling and mitochondrial homeostasis deficiences contribute to Alzheimer disease pathology in rodents. Biochim. Biophys. Acta (2015) 1852 1687-1699. DOI: 10.1016/j.bbadis.2015.05.004 39. Busquets O., Ettcheto M., Pallàs M., Beas-Zarate C., Verdaguer E., Auladell C., Folch J., Camins A.. Long-term exposition to a high fat diet favors the appearance of β-amyloid depositions in the brain of C57BL/6J mice. A potential model of sporadic Alzheimer’s disease. Mech. Ageing Dev. (2017) 162 38-45. DOI: 10.1016/j.mad.2016.11.002 40. Ghavami S., Shojaei S., Yeganeh B., Ande S.R., Jangamreddy J.R., Mehrpour M., Christoffersson J., Chaabane W., Moghadam A.R., Kashani H.H.. Autophagy and apoptosis dysfunction in neurodegenerative disorders. Prog. Neurobiol. (2014) 112 24-49. DOI: 10.1016/j.pneurobio.2013.10.004 41. Planells-Ferrer L., Urresti J., Coccia E., Galenkamp K.M., Calleja-Yagüe I., López-Soriano J., Carriba P., Barneda-Zahonero B., Segura M.F., Comella J.X.. Fas apoptosis inhibitory molecules: More than death-receptor antagonists in the nervous system. J. Neurochem. (2016) 139 11-21. DOI: 10.1111/jnc.13729 42. Jaiswal S., Sharma P.. Role and regulation of p27 in neuronal apoptosis. J. Neurochem. (2017) 140 576-588. DOI: 10.1111/jnc.13918 43. Park H.R., Park M., Choi J., Park K.Y., Chung H.Y., Lee J.. A high-fat diet impairs neurogenesis: Involvement of lipid peroxidation and brain-derived neurotrophic factor. Neurosci. Lett. (2010) 482 235-239. DOI: 10.1016/j.neulet.2010.07.046 44. Wu A., Ying Z., Gomez-Pinilla F.. The interplay between oxidative stress and brain-derived neurotrophic factor modulates the outcome of a saturated fat diet on synaptic plasticity and cognition. Eur. J. Neurosci. (2004) 19 1699-1707. DOI: 10.1111/j.1460-9568.2004.03246.x 45. Kempuraj D., Thangavel R., Natteru P., Selvakumar G.P., Saeed D., Zahoor H., Zaheer S., Iyer S.S., Zaheer A.. Neuroinflammation induces neurodegeneration. J. Neurol. Neurosurg. Spine (2016) 1 1003. PMID: 28127589 46. Duffy C., Hofmeister J., Nixon J., Butterick T.. High fat diet increases cognitive decline and neuroinflammation in a model of orexin loss. Neurobiol. Learn. Mem. (2019) 157 41-47. DOI: 10.1016/j.nlm.2018.11.008 47. Spencer S., D’Angelo H., Soch A., Watkins L.R., Maier S.F., Barrientos R.M.. High-fat diet and aging interact to produce neuroinflammation and impair hippocampal and amygdala-dependent memory. Neurobiol. Aging (2017) 58 88-101. DOI: 10.1016/j.neurobiolaging.2017.06.014 48. D’Alessio P., Ostan R., Bisson J., Schulzke J.D., Ursini M.V., Béné M.C.. Oral administration of d-limonene controls inflammation in rat colitis and displays anti-inflammatory properties as diet supplementation in humans. Life Sci. (2013) 92 1151-1156. DOI: 10.1016/j.lfs.2013.04.013 49. Roberto D., Micucci P., Sebastian T., Graciela F., Anesini C.. Antioxidant activity of limonene on normal murine lymphocytes: Relation to H2O2 modulation and cell proliferation. Basic Clin. Pharm. Toxicol. (2010) 106 38-44. DOI: 10.1111/j.1742-7843.2009.00467.x 50. Nuzzo D., Baldassano S., Amato A., Picone P., Galizzi G., Caldara G.F., Di Carlo M., Mulè F.. Glucagon-like peptide-2 reduces the obesity-associated inflammation in the brain. Neurobiol. Dis. (2019) 121 296-304. DOI: 10.1016/j.nbd.2018.10.012 51. Nuzzo D., Presti G., Picone P., Galizzi G., Gulotta E., Giuliano S., Mannino C., Gambino V., Scoglio S., Di Carlo M.. Effects of the aphanizomenon flos-aquae extract (Klamin®) on a neurodegeneration cellular model. Oxid. Med. Cell Longev. (2018) 17 9089016. DOI: 10.1155/2018/9089016 52. Shimazu R., Anada M., Miyaguchi A., Nomi Y., Matsumoto H.. Evaluation of Blood-Brain Barrier Permeability of Polyphenols, Anthocyanins, and Their Metabolites. Agric. Food Chem. (2021) 69 11676-11686. DOI: 10.1021/acs.jafc.1c02898 53. Khan H., Ullah H., Aschner M., Cheang W., Akkol E.. Neuroprotective Effects of Quercetin in Alzheimer’s Disease. Biomolecules (2019) 10. DOI: 10.3390/biom10010059 54. Siddique Y.R.. Neurodegenerative Diseases and Flavonoids: Special Reference to Kaempferol. CNS Neurol. Disord. Drug Targets (2021) 20 327-342. DOI: 10.2174/1871527320666210129122033 55. Beg T., Jyoti S., Naz F., Rahul Ali F., Ali S.K., Reyad A.M., Siddique Y.H.. Protective Effect of Kaempferol on the Transgenic Drosophila Model of Alzheimer’s Disease. CNS Neurol. Disord. Drug Targets (2018) 17 421-429. DOI: 10.2174/1871527317666180508123050 56. Vinklarova L., Schmidt M., Benek O., Kuca K., Gunn-Moore F., Musilek K.. Friend or enemy? Review of 17β-HSD10 and its role in human health or disease. J. Neurochem. (2020) 155 231-249. DOI: 10.1111/jnc.15027 57. Nikolac Perkovic M., Pivac N.. Genetic Markers of Alzheimer’s Disease. Adv. Exp. Med. Biol. (2019) 1192 27-52. DOI: 10.1007/978-981-32-9721-0_3 58. Serrano-Pozo A., Das S., Hyman B.T.. APOE and Alzheimer’s disease: Advances in genetics, pathophysiology, and therapeutic approaches. Lancet Neurol. (2021) 20 68-80. DOI: 10.1016/S1474-4422(20)30412-9 59. Liu S.L., Wang C., Jiang T., Tan L., Xing A., Yu J.T.. The Role of Cdk5 in Alzheimer’s Disease. Mol. Neurobiol. (2016) 53 4328-4342. DOI: 10.1007/s12035-015-9369-x 60. Uddin M.S., Kabir M.T., Begum M.M., Islam M.S., Behl T., Ashraf G.M.. Exploring the Role of CLU in the Pathogenesis of Alzheimer’s Disease. Neurotox. Res. (2021) 39 2108-2119. DOI: 10.1007/s12640-020-00271-4 61. Nasuti C., Gabbianelli R., Falcioni G., Cantalamessa F.. Antioxidative and gastroprotective activities of anti-inflammatory formulations derived from chestnut honey in rats. Nutr. Res. (2006) 26 130-137. DOI: 10.1016/j.nutres.2006.02.007 62. Terzo S., Attanzio A., Calvi P., Mulè F., Tesoriere L., Allegra M., Amato A.. Indicaxanthin from Opuntia ficus-indica Fruit Ameliorates Glucose Dysmetabolism and Counteracts Insulin Resistance in High-Fat, Diet-Fed Mice. Antioxidants (2021) 11. DOI: 10.3390/antiox11010080
--- title: Focal Photocoagulation as an Adjunctive Therapy to Reduce the Burden of Intravitreal Injections in Macula Edema Patients, the LyoMAC2 Study authors: - Lucas Séjournet - Laurent Kodjikian - Sandra Elbany - Benoit Allignet - Emilie Agard - Mayeul Chaperon - Jérémy Billant - Philippe Denis - Thibaud Mathis - Carole Burillon - Corinne Dot journal: Pharmaceutics year: 2023 pmcid: PMC9966640 doi: 10.3390/pharmaceutics15020308 license: CC BY 4.0 --- # Focal Photocoagulation as an Adjunctive Therapy to Reduce the Burden of Intravitreal Injections in Macula Edema Patients, the LyoMAC2 Study ## Abstract Aim: To assess the efficacy of focal photocoagulation of capillary macroaneurysms (CMA) to reduce the burden of intravitreal injections (IVI) in patients with macular edema (ME). Materials and Methods: Retrospective multicenter study in patients with diabetic ME or ME secondary to retinal vein occlusion (ME-RVO). CMA associated with ME were selectively photocoagulated. Patients were followed for one year after photocoagulation. Results: 93 eyes of 76 patients were included in this study. At 6 months after the laser ($$n = 93$$), there was a significant decrease in mean macular thickness (from 354 µm to 314 µm, $p \leq 0.001$) and in mean IVI number (from 2.52 to 1.52 at 6 months, $p \leq 0.001$). The mean BCVA remained stable (0.32 and 0.31 logMAR at baseline and 6 months, $$p \leq 0.95$$). At 12 months ($$n = 81$$/93), there was a significant decrease in mean macular thickness (from 354 µm to 314 µm, $p \leq 0.001$) and in mean IVI number (from 4.44 to 2.95 at 12 months, $p \leq 0.001$), while the mean BCVA remained stable (0.32 and 0.30 logMAR at baseline and 12 months, $$p \leq 0.16$$). Conclusion: Focal laser photocoagulation of CMA seems to be effective and safe for reducing the burden of IVI in patients with ME. Their screening during the follow-up should be considered closely. ## 1. Introduction Diabetic retinopathy and retinal vein occlusion (RVO) are the leading causes of long-lasting macular edema (ME), which can lead to an irreversible loss of vision in the absence of treatment [1,2]. Focal vascular abnormalities such as capillary microaneurysms may develop in these retinal vascular diseases [3,4]. The diameter of these microaneurysms is usually less than 90 µm; however, the diameter of some aneurysms can reach up to several hundred microns [5,6]. These large abnormalities, referred to as capillary macroaneurysms (CMA) [6,7] or telangiectatic capillaries [8] in the medical literature, may be associated with chronic refractory ME and hard exudates [5]. The prevalence of CMA in diabetic ME (DME) and RVO is about $60\%$, and indocyanine green angiography (ICGA) and optical coherence tomography (OCT) have been previously shown to effectively detect CMA [9,10]. Treatment of ME is mainly based on intravitreal injections (IVIs) of anti-vascular endothelial growth factor (VEGF) or corticosteroids. Because of its chronicity, IVIs and physical examinations are needed in the long term, leading to a significant therapeutic burden. In recent years, the main objective has been to increase the interval between IVIs through the use of new injection protocols such as the treat-and-extend regimen [11,12]. A potential alternative could be to photocoagulate CMA. Several publications have shown that focal laser photocoagulation of CMA effectively improves the visual acuity and reduces the macular thickness [6,13,14,15]. These studies are mostly retrospective, non-controlled, non-blinded, and monocentric. However, in these previous studies, the sample size was small, and no patients previously treated with IVIs were included. None of these studies has assessed the long-term efficacy of photocoagulation or the need for IVIs after focal laser photocoagulation. The aim of this retrospective study was to assess the relevance of focal laser photocoagulation of CMA in both DME and ME-RVO to reduce the burden of intravitreal injections in patients with long-lasting macular edema (ME) by determining the number of IVIs received after laser treatment. The secondary outcomes were the best-corrected visual acuity (BVCA), the focal retinal thickness, and the macular thickness at 6 and 12 months after focal laser photocoagulation. ## 2.1. Patient Selection A multicenter, retrospective, non-randomized, and non-comparative study was conducted in adult patients with DME or ME-RVO treated with focal laser photocoagulation for CMA between November 2019 and June 2022 in the ophthalmology departments of Desgenettes Military Hospital, Croix Rousse, and Edouard Herriot University Hospital, Lyon (France). This study was conducted in accordance with the principles outlined in the Declaration of Helsinki. All patients received oral and written information and gave their consent before being treated for any laser treatment. The Ethics Committee of the French Society of Ophthalmology approved the study conduct (IRB 00008855 Société Française d’Ophtalmologie IRB 1). ## 2.2. Data Collection All patients treated with focal laser photocoagulation were included in this study, only if they had been treated with IVIs at least 12 months prior to the laser treatment. The medical history, the number and type of IVIs received, and prior ophthalmological findings were recorded retrospectively. The dataset was collected using anonymized Excel files before statistical analysis. The systematic screening of CMA has been standardized in our departments since 2019. In all patients, an ophthalmological examination and retinal imaging, including fluorescein angiography (FA), ICGA, and OCT (spectral domain [SD]-OCT, Heidelberg Engineering, Heidelberg, Germany), were performed. CMA were identified on the late-phase ICGA and on the OCT scans (Figure 1). All lesions related to ME located more than 750 µm from the fovea were treated. Patients with CMA in the central 1500-µm macular area (<750 µm) were also included if other lesions were accessible for focal photocoagulation. The CMA diameter was measured on the OCT B-scan (Spectralis HRA, “pole post” pattern, scan length of 8.3 mm). The mean central macular thickness, the mean focal edema thickness (i.e., the thickness of the retina around the CMA), and the mean distance between the CMA and the fovea were also recorded. ## 2.3. Outcome Measures Laser photocoagulation was performed with a Centralis® contact lens (Volk) using a 532-nm laser with the following parameters: spot size of 50 µm; duration of 0.2 s. The power was increased from 100 mW until whitening of the macroaneurysms was observed. Three consecutive impacts were made. In our daily practice, OCT is not performed after laser therapy. IVIs were administered according to a treat-and-extend protocol with a 2-week adjustment period. Injection intervals were decreased when the retinal fluid was stable or increased on the OCT B-scan or in cases of worsening of the BCVA. Injection intervals were increased in the absence of retinal fluid on the OCT B-scan. Treatment intervals were clinically assessed by an ophthalmologist at each visit. Aflibercept, ranibizumab, and dexamethasone implant (DEXi) were used for both DME and ME-RVO. A treatment switch was possible in the absence of response after three consecutive anti-VEGF IVIs. IVIs could be discontinued when a dry macula persisted 4 months after the last injection of aflibercept and ranibizumab and 6 months after the last DEXi injection. Follow-up examinations were performed 3, 6, and 12 months after photocoagulation. Opacification of the CMA lumen was assessed by comparing the pre- and post-laser OCT scans using the eye tracking tool. Patients could be retreated in the absence of signs of photothrombosis and when intraretinal fluid persisted 3 months after laser treatment. ## 2.4. Statistics Data are presented as a mean ± standard deviation (SD) or a mean (range) and as a count (percentage) for quantitative and qualitative variables, respectively. A paired-sample Wilcoxon rank-sum test or a t test was used for comparisons of continuous variables. All tests were two-sided, and a p-value < 0.05 was considered significant. Statistical analyses were performed using R software version 4.1.2 (R Foundation for Statistical Computing, Vienna, Austria). ## 3.1. Patients’ Characteristics Ninety-three eyes of 76 patients were included in this study. Patients’ characteristics are shown in Table 1. The mean age was 70.8 (range: 31–90), and 44 patients ($56.6\%$) were men. Regarding the distribution, 62 cases were DME and 31 cases were ME-RVO. The mean ME duration before laser treatment was 23.9 months (range: 12–48). All patients received IVIs for at least 12 months before inclusion. Patients received a mean number of 4.43 ± 2.44 IVIs and 2.52 ± 1.44 IVIs in the 12 and 6 months prior to photocoagulation, respectively. Patients were treated with anti-VEGF ($\frac{39}{93}$, $39.8\%$), DEXi ($\frac{39}{93}$, $41.9\%$), or both ($\frac{17}{93}$, $18.3\%$) in the 12 months prior to photocoagulation. At baseline, the mean BCVA was 0.32 ± 0.15 logMAR, the mean central macular thickness was 354 ± 114 µm, and the mean focal retinal thickness was 462 ± 92.7 µm. The mean number of CMA per eye was 2.6 (range: 1–7). The mean diameter of CMA was 171 ± 52 µm. The mean closest distance between the lesions and the fovea was 1690 µm (range: 750–3100 µm). At baseline, hard exudates and circinate exudates were detected in $\frac{71}{93}$ and $\frac{14}{93}$ eyes, respectively. ## 3.2. Primary and Secondary Outcomes Six months after photocoagulation, there was a significant decrease in the mean number of IVIs received (from 2.52 IVIs in the 6 months prior to photocoagulation to 1.52 IVIs 6 months after photocoagulation, $p \leq 0.001$), in the mean foveal thickness (from 354 µm at baseline to 314 µm at 6 months, $p \leq 0.001$), and in the mean focal thickness (from 462 µm at baseline to 399 µm at 6 months, $p \leq 0.001$). The mean BVCA was not significantly improved six months after photocoagulation compared to the baseline (0.32 and 0.31 logMAR, respectively; $$p \leq 0.95$$). The same results were observed 12 months after photocoagulation for the mean number of IVIs (from 4.44 IVIs in the 12 months before photocoagulation to 2.95 IVIs 12 months after photocoagulation, $p \leq 0.001$), the mean foveal thickness (from 354 µm at baseline to 299 µm at 12 months, $p \leq 0.001$), and the mean focal thickness (from 462 µm at baseline to 399 µm at 12 months, $p \leq 0.001$). The mean BVCA was not significantly improved 12 months after photocoagulation compared to baseline (0.32 and 0.30 logMAR, respectively; $$p \leq 0.16$$) (Figure 2 and Figure 3). Differences in central retinal thickness and focal thickness are statistically significant ($p \leq 0.001$). The BCVA was not significantly improved ($p \leq 0.05$). CMA were still detected in $\frac{15}{93}$ eyes at 6 months because they were too close to fovea ($\frac{4}{15}$), or they could not be targeted due to poor patient fixation, cataract, or vitreous hemorrhage ($\frac{11}{15}$). At 6 months, this subgroup of patients with persistent CMA (CMA+) received a mean number of 2.11 IVIs compared to 1.37 IVI in patients with occluded CMA (CMA-) ($$p \leq 0.02$$), had a poorer BVCA (0.37 versus 0.29 logMAR in the CMA+ and CMA- subgroups, $p \leq 0.01$), a higher foveal thickness (354 versus 303 µm in the CMA+ and CMA- subgroups, $$p \leq 0.01$$), and a higher focal thickness (469 versus 382 µm in the CMA+ and CMA- subgroups, $p \leq 0.01$). The same results were observed at 12 months with a mean number of 4.1 IVIs in the CMA+ subgroup versus 2.6 IVIs in the CMA- subgroup ($$p \leq 0.04$$), a higher foveal thickness (323 versus 291 µm in the CMA+ and CMA- subgroups, $p \leq 0.01$), a poorer BVCA (0.39 versus 0.27 logMAR in the CMA+ and CMA- subgroups, $p \leq 0.01$), and a higher focal thickness (449 versus 385 µm in the CMA+ and CMA- subgroups, $p \leq 0.01$). Hard exudates had disappeared in $\frac{26}{71}$ ($37\%$) eyes at 6 months and in $\frac{34}{66}$ ($50\%$) eyes at 12 months. Overall, IVIs were fully discontinued in $\frac{23}{93}$ eyes in $\frac{21}{76}$ ($27.6\%$) patients. The results of the primary and secondary outcomes are shown in Table 2. At 12 months, treatment was switched in $\frac{5}{93}$ eyes from anti-VEGF to DEXi ($$n = 4$$) and from DEXi to aflibercept ($$n = 1$$), and three patients received an acetonide fluocinolone implant in addition to DEXi. In $\frac{30}{93}$ ($32\%$) eyes of 28 patients, laser photocoagulation was repeated because OCT showed CMA with persistent retinal fluid. In $\frac{5}{30}$ eyes of 5 patients, laser photocoagulation was repeated a third time. No complications from photocoagulation were reported. ## 3.3. Functional and Anatomical Outcomes According to the Type of IVI Agent After 6 months, patients treated with anti-VEGF showed a significant decrease in the mean number of IVIs (from 3.62 IVIs 6 months before photocoagulation to 1.9 IVIs at 6 months, $p \leq 0.001$), in the mean central foveal thickness (from 301 µm at baseline to 283 µm at 6 months, $$p \leq 0.04$$) and in the mean focal thickness (from 416 µm at baseline to 376 µm at 6 months, $p \leq 0.001$) compared to baseline. There was no significant difference in BVCA six months after photocoagulation (0.22 and 0.25 logMAR at baseline and at 6 months; $$p \leq 0.35$$). After 12 months, patients treated with anti-VEGF showed a significant decrease in the mean number of IVIs (from 6.19 IVIs 12 months before photocoagulation to 3.70 IVIs at 12 months, $p \leq 0.001$), in the mean central foveal thickness (from 301 µm at baseline to 237 µm at 12 months, $p \leq 0.01$), and in the mean focal thickness (from 416 µm at baseline to 358 µm at 12 months, $p \leq 0.01$). The BVCA was not significantly improved 12 months after photocoagulation compared to baseline (0.22 and 0.24 logMAR at baseline and 6 months; $$p \leq 0.80$$). After 6 months, patients treated with DEXi showed a significant decrease in the mean number of IVIs (from 1.69 IVIs 6 months before photocoagulation to 1.26 IVIs at 6 months, $$p \leq 0.02$$), in the mean central thickness (from 403 µm at baseline to 336 µm at 6 months, $p \leq 0.01$), and in the mean focal thickness (from 492 µm at baseline to 427 µm at 12 months, $p \leq 0.01$). There was no significant difference in BVCA six months after photocoagulation (0.47 and 0.43 logMAR at baseline and 6 months, $$p \leq 0.08$$). After 12 months, there was a significant decrease in the mean number of IVIs (from 3.05 IVIs 12 months before photocoagulation to 2.42 IVIs at 12 months, $$p \leq 0.02$$), in the mean central thickness (from 403 µm at baseline to 317 µm at 12 months, $p \leq 0.01$), and in the mean focal thickness (from 492 µm at baseline to 425 µm at 12 months, $p \leq 0.01$). The BVCA was not significantly improved compared to baseline (0.47 and 0.39 logMAR at baseline and 12 months, $$p \leq 0.11$$). These results are shown in Table 3. ## 3.4. Functional and Anatomical Outcomes According to the Size of CMA CMA larger than 150 µm were found in $\frac{60}{93}$ eyes. A total of 6 months after photocoagulation, there was a significant decrease in the mean number of IVIs (from 2.30 IVIs 6 months before photocoagulation to 1.40 at 6 months, $p \leq 0.001$), in the mean central foveal thickness (from 373 µm at baseline to 317 µm at 6 months, $p \leq 0.001$), and in the mean focal thickness (from 474 µm at baseline to 401 µm at 6 months, $p \leq 0.001$). The BVCA was not significantly improved 6 months after photocoagulation compared to baseline (0.36 and 0.34 logMAR at baseline and at 6 months; $$p \leq 0.97$$). Similar results were obtained 12 months after photocoagulation for the mean number of IVIs (from 4.05 IVIs 12 months before photocoagulation to 2.84 IVIs at 12 months, $p \leq 0.001$), the mean foveal thickness (from 373 µm at baseline to 309 µm at 12 months, $p \leq 0.001$), and the mean focal thickness (from 474 µm at baseline to 402 µm at 12 months, $p \leq 0.001$). The BVCA was not significantly improved 12 months after photocoagulation compared to baseline (0.36 and 0.33 logMAR at baseline and at 12 months; $$p \leq 0.16$$). CMA smaller than 150 µm were found in $\frac{33}{93}$ eyes. A total of 6 months after photocoagulation, there was a significant decrease in the mean number of IVIs (from 2.91 IVIs 6 months before photocoagulation to 1.73 IVIs at 6 months, $p \leq 0.001$) and in the mean focal thickness (from 441 µm at baseline to 397 µm at 6 months, $p \leq 0.001$). The BVCA and the mean central foveal thickness were not significantly improved six months after photocoagulation compared to baseline (0.26 and 0.26 logMAR at baseline and at 6 months, $$p \leq 0.95$$, and 319 and 306 µm at baseline and at 6 months, $$p \leq 0.34$$, respectively). Similar results were obtained 12 months after photocoagulation for the mean number of IVIs (from 5.15 IVIs 12 months before photocoagulation to 3.13 IVIs at 12 months, $p \leq 0.01$), the mean foveal thickness (from 319 µm at baseline to 280 µm at 12 months, $p \leq 0.01$), and the mean focal thickness (from 441 µm at baseline to 395 µm at 12 months, $p \leq 0.01$). The BVCA was not significantly improved 12 months after photocoagulation compared to baseline (0.26 and 0.25 logMAR at baseline and at 12 months; $$p \leq 0.60$$). These results are shown in Table 4. ## 3.5. Functional and Anatomical Outcomes According to the Cause of ME In DME patients, after 6 months, there was a significant decrease in the mean number of IVIs (from 2.52 IVIs 6 months before photocoagulation to 1.53 IVIs at 6 months, $p \leq 0.01$), in the mean central foveal thickness (from 360 µm at baseline to 319 µm at 6 months; $p \leq 0.01$), and in the mean focal thickness (from 476 µm at baseline to 415 µm at 6 months, $p \leq 0.01$). There was no significant difference in BVCA six months after photocoagulation (0.32 and 0.30 logMAR at baseline and at 6 months; $$p \leq 0.73$$). After 12 months, DME patients showed a significant decrease in the mean number of IVIs (from 4.37 IVIs 12 months before photocoagulation to 3.15 IVIs at 6 months, $p \leq 0.01$), in the mean central foveal thickness (from 360 µm at baseline to 299 µm at 12 months, $p \leq 0.01$), and in the mean focal thickness (from 476 µm at baseline to 413 µm at 12 months, $p \leq 0.01$). The BVCA was not significantly improved 12 months after photocoagulation compared to baseline (0.32 and 0.28 logMAR at baseline and at 12 months, $$p \leq 0.06$$). In ME-RVO patients, after 6 months, there was a significant decrease in the mean number of IVIs (from 2.51 IVIs 6 months before photocoagulation to 1.48 IVIs at 6 months, $p \leq 0.01$), in mean central thickness (from 342 µm at baseline to 302 µm at 6 months, $$p \leq 0.02$$), and in mean focal thickness (from 436 µm at baseline to 371 µm at 6 months, $p \leq 0.01$). There was no significant difference in BVCA six months after photocoagulation (0.33 and 0.33 logMAR at baseline and at 6 months, $$p \leq 0.70$$). After 12 months, there was a significant decrease in the mean number of IVIs (from 4.58 IVIs 12 months before photocoagulation to 2.59 IVIs at 12 months, $p \leq 0.01$), in mean central thickness (from 342 µm at baseline to 298 µm at 12 months, $p \leq 0.01$), and in mean focal thickness (from 436 µm at baseline to 373 µm at 12 months, $p \leq 0.01$). The BVCA was not significantly improved compared to baseline (0.33 and 0.33 logMAR at baseline and at 12 months, $$p \leq 0.68$$). These results are shown in Table 5. ## 3.6. Functional and Anatomical Outcomes According to the Distance between CMA and the Fovea The whole cohort of patients was divided into tertiles based on the closest distance between the CMA and the fovea. Patients whose closest CMA was located less than 750 µm from the fovea were excluded. The first tertile (T1) included 30 eyes with the closest distance between CMA and the fovea ranging between 750 and 1400 µm; the second tertile (T2) included 31 eyes with the closest distance between CMA and the fovea ranging between 1400 and 1900 µm; and the third tertile (T3) included 28 eyes with the closest distance between CMA and the fovea ranging between 1900 and 3100 µm. Patients in all tertiles showed a significant decrease in the mean number of IVIs from baseline, regardless of the timepoint analyzed (at 6 and 12 months). At 6 months, the mean number of IVIs decreased from 2.17 to 1.67 in T1 ($$p \leq 0.03$$), from 2.55 to 1.51 in T2 ($p \leq 0.01$), and from 2.93 to 1.35 in T3 ($p \leq 0.01$) compared to 6 months before photocoagulation. At 12 months, the mean number of IVIs decreased from 4.17 to 3.20 in T1 ($p \leq 0.01$), from 4.61 to 2.80 in T2 ($p \leq 0.01$), and from 4.64 to 2.80 in T3 ($p \leq 0.01$) compared to 12 months before photocoagulation. Moreover, the number of IVIs decreased when the distance to the fovea increased, with a mild but significant correlation between the distance to the fovea and the number of IVIs at 6 months ($31.3\%$; $$p \leq 0.003$$). However, there was no significant correlation between the distance to the fovea and the number of IVIs at 12 months ($$p \leq 0.42$$). ## 4. Discussion In our non-comparative retrospective study, we reached our primary endpoint, showing that focal laser photocoagulation of CMA effectively reduced the number of IVIs in patients with chronic DME or ME-RVO. To our knowledge, this was the largest real-life study conducted on this topic. In chronic ME, the grid laser has been historically performed with an overall modest effect on vision [16,17,18]. In our study, targeted, selective laser photocoagulation appeared to be a more appropriate approach. However, in current guidelines, CMA treatment is not yet explicitly mentioned [19,20]. Moreover, our subgroup analysis according to the cause of ME showed that targeting CMA was effective in both DME and ME-RVO patients. In the literature, there is no consensus on the definition of CMA. Some authors have defined CMA as any microvascular lesion with late focal hyperfluorescence on ICGA [5] or use a diameter threshold of 150 µm [13,21]. A retrospective study conducted in 2013 found a lower efficacy of laser photocoagulation on lesions <150 µm [22]. Nevertheless, we chose the smaller size of 100 µm to treat all the lesions eligible for laser therapy, as explained in a previous study recently published by our team [9]. Indeed, it may be challenging to see and effectively target CMA <100 µm when a traditional laser is used, compared to a navigated retinal laser. We showed consistent results between patients with CMA >150 µm and <150 µm in a subgroup analysis. It may be challenging to locate CMA for laser treatment. Only lesions located at more than 750 µm from the fovea were treated in our study to reduce the risk of lesion scar evolution and foveal damage. No complication was reported in our study after focal coagulation. Navigated retinal lasers, such as the Navilas® system, could allow treating lesions closer to the fovea if needed [21]. The fact that patients with lesions closer than 750 µm were also included in this study when another CMA was eligible for treatment could explain the poorer occlusion rate compared to other studies [13,21]. As expected, patients with persistently active CMA at six months received a higher number of IVIs than patients with fully treated CMA and had poorer BVCA and foveal and focal thicknesses. We also showed that focal photocoagulation decreased the mean number of IVIs in patients treated with anti-VEGF and DEXi. Patients treated with DEXi had a more severe ME, as shown by their baseline BVCA, mean foveal thickness, and mean focal thickness. Additionally, it should be noted that the number of IVIs was reduced by almost $50\%$ 12 months after laser therapy in patients treated with anti-VEGFs without any complications, whereas this reduction was less marked in patients treated with DEX-i ($25\%$). This could be explained by the longer release of this implant (4 months) and the severity of the ME. Nevertheless, the gain in DEX injections was about 1 IVI at 12 months. In addition, IVIs could be discontinued in $25\%$ of eyes, suggesting that, in some cases, ME was mainly due to CMA and that its appropriate diagnosis and treatment could be very relevant to decreasing or even discontinuing IVIs. On the other hand, some patients had persistently active CMA despite several focal photocoagulation sessions. This could be explained by the location of some CMA that were too close to the fovea to be targeted or by the difficulty of performing laser treatment due to a poor fixation of patients, cataract, or vitreous hemorrhage. This could have contributed to increasing the number of IVIs in these patients and decreasing the overall efficacy of treatment. We used a 532-nm laser with standard parameters as previously described [6]. Other types of lasers have shown their efficacy in microaneurysm and DME treatment, such as yellow lasers and subthreshold micropulse lasers [23,24]. To date, no study has assessed these lasers for CMA treatment in ME-RVO and DME, despite their potential interest. Indeed, these types of lasers are not yet available in all departments. Infrared reflectance image-guided focal laser photocoagulation of CMA has recently been shown to be effective in persistent DME [14,15]. We have highlighted the good detection of CMA on en face OCT in the LyoMAC1 study [12]. This could be, in some cases, a new multimodal way, including OCT-angiography, to detect CMA in DME and ME-RVO before planning laser treatment. As expected, the BVCA was not improved after focal photocoagulation, despite a trend in DME patients ($$p \leq 0.06$$). However, we showed a decrease in mean macular thickness after CMA treatment. A focal laser treatment could be considered an adjuvant treatment. We could assume that combining focal laser photocoagulation with IVIs could improve the long-term functional outcomes since it is known that fluctuations are deleterious in DME [25]. IVIs are an effective but expensive treatment for DME and ME-RVO [26,27]. The cost-effectiveness of the systematic detection and laser focal photocoagulation of CMA was not assessed in this study but should raise interest because the number of IVIs was significantly reduced in our patients after one year. In addition, its effect on quality of life and decrease in therapeutic burden will also need to be assessed. This study has some limitations. Our patients were likely to have more severe retinal disorders at baseline because they were followed at three tertiary centers. Three different university centers were involved, all in the same city, reflecting a local approach and not reflecting the diversity of medical practice worldwide. Furthermore, this was a non-blinded and non-comparative study. This can induce some measurement bias, as investigators may have been influenced by the fact that the laser was conducted and could have been less stringent about edema recurrence. Patients were under their own control during the year after laser treatment. The reduction of injections with time is well known; nevertheless, our patients had chronic and old macular edema with repeated injections for at least 2 years. Our study was also limited by the small size of our cohort, even though this is the largest series described to date. As focal photocoagulation was not performed by a single investigator, this could explain some differences between centers, but this reflects our real-life experience. The retrospective nature of this study could also have led to missing data or missing confounding factors. As several investigators were involved in this study, the treatment decision could have been partially subjective, despite being confirmed by a senior retinal specialist. Treatments could be switched, stopped, or intensified at the discretion of physicians, which could reflect to some degree the doctor’s preference. However, common guidelines for recurrences were followed in all centers to limit this bias. Moreover, we did not find a strong correlation between distance to the fovea and outcome measurements. Although it seems intuitive to think that the closer the CMA is to the fovea, the better the laser treatment would work, our results did not support this. This could be explained by the small size of sub-groups. All patients achieved our primary outcome at 6 months, but some patients were lost to follow-up before the 12-month visit. The number of patients lost to follow-up was low, and its impact on our results should be low. Switching to DEXi or a fluocinolone acetonide implant could have artificially reduced the number of IVIs over 6 and 12 months, but only a few patients underwent the switch, and it should therefore have no significant impact on our results. The strengths of our study are the systematic use of ICGA for the diagnosis of CMA and the use of multimodal imaging for laser decisions. The same laser wavelength and protocol were used to treat. Patients were under their own control, excluding the influence of other comorbidities common in diabetes. The one-year follow-up after photocoagulation allowed us to identify a reduction in treatment burden. Additionally, our study was a real-life study reflecting what we are supposed to face in our daily practice. Our results could help better define focal photocoagulation procedures in order to improve their efficacy. Systematic screening and treatment of CMA in chronic or resistant ME could be of interest to reduce the therapeutic burden for patients and the medical cost for society. ## 5. Conclusions Focal laser photocoagulation of CMA in patients with DME and ME-RVO appears to be relevant to extending the interval between IVIs and reducing the burden of intravitreal injections in patients with long-lasting ME. It should be noted that the number of IVIs was reduced by almost $50\%$. A total of 12 months after laser therapy, in patients treated with anti-VEGFs and IVIs, could be totally discontinued in $25\%$ of eyes. This approach could be beneficial for patients, especially in terms of quality of life and medico-economic aspects. As our study is non-randomized, non-blinded, and retrospective, further larger prospective controlled studies are needed to confirm our results. ## References 1. Bhagat N., Grigorian R.A., Tutela A., Zarbin M.A.. Diabetic Macular Edema: Pathogenesis and Treatment. Surv. Ophthalmol. (2009) 54 1-32. DOI: 10.1016/j.survophthal.2008.10.001 2. Spaide R.F.. Retinal vascular cystoid macular edema: Review and New Theory. Retina (2016) 36 1823-1842. DOI: 10.1097/IAE.0000000000001158 3. Stitt A.W., A Gardiner T., Archer D.B.. Histological and ultrastructural investigation of retinal microaneurysm development in diabetic patients. Br. J. Ophthalmol. (1995) 79 362-367. DOI: 10.1136/bjo.79.4.362 4. Murakami T., Nishijima K., Sakamoto A., Ota M., Horii T., Yoshimura N.. Foveal Cystoid Spaces Are Associated with Enlarged Foveal Avascular Zone and Microaneurysms in Diabetic Macular Edema. Ophthalmology (2011) 118 359-367. DOI: 10.1016/j.ophtha.2010.03.035 5. Bourhis A., Girmens J.-F., Boni S., Pecha F., Favard C., Sahel J.-A., Paques M.. Imaging of macroaneurysms occurring during retinal vein occlusion and diabetic retinopathy by indocyanine green angiography and high resolution optical coherence tomography. Graefe’s Arch. Clin. Exp. Ophthalmol. (2009) 248 161-166. DOI: 10.1007/s00417-009-1175-6 6. Paques M., Philippakis E., Bonnet C., Falah S., Ayello-Scheer S., Zwillinger S., Girmens J.-F., Dupas B.. Indocyanine-green-guided targeted laser photocoagulation of capillary macroaneurysms in macular oedema: A pilot study. Br. J. Ophthalmol. (2017) 101 170-174. DOI: 10.1136/bjophthalmol-2015-308142 7. Parodi M.B., Bondel E., Ravalico G.. Capillary Macroaneurysms in Central Retinal Vein Occlusion. Ophthalmologica (1995) 209 248-250. DOI: 10.1159/000310624 8. Farías D.C., Serrano R.M., Gancharov J.B., Cuadras U.D.D., Sahel J., Wiechers F.G., Dupas B., Paques M.. Indocyanine green angiography for identifying telangiectatic capillaries in diabetic macular oedema. Br. J. Ophthalmol. (2019) 104 509-513. DOI: 10.1136/bjophthalmol-2019-314355 9. Chaperon M., Kodjikian L., Agard E., Mathis T., Billant J., El-Chehab H., Pradat P., Dot C.. Screening of telangiectatic capillaries in chronic macular edema based on multimodal imaging: A study of 101 eyes. LyoMAC1 study. Graefe’s Arch. Clin. Exp. Ophthalmol. (2022) 260 2501-2508. DOI: 10.1007/s00417-022-05592-y 10. Hamada M., Ohkoshi K., Inagaki K., Ebihara N., Murakami A.. Visualization of microaneurysms using optical coherence tomography angiography: Comparison of OCTA en face, OCT B-scan, OCT en face, FA, and IA images. Jpn. J. Ophthalmol. (2018) 62 168-175. DOI: 10.1007/s10384-018-0570-0 11. Pennington B., Alshreef A., Flight L., Metry A., Poku E., Hykin P., Sivaprasad S., Prevost A.T., Vasconcelos J.C., Murphy C.. Cost Effectiveness of Ranibizumab vs Aflibercept vs Bevacizumab for the Treatment of Macular Oedema Due to Central Retinal Vein Occlusion: The LEAVO Study. Pharmacoeconomics (2021) 39 913-927. DOI: 10.1007/s40273-021-01026-5 12. Ross E.L., Hutton D.W., Stein J., Bressler N.M., Jampol L.M., Glassman A.R.. Diabetic Retinopathy Clinical Research Network Cost-effectiveness of Aflibercept, Bevacizumab, and Ranibizumab for Diabetic Macular Edema Treatment. JAMA Ophthalmol (2016) 134 888-896. DOI: 10.1001/jamaophthalmol.2016.1669 13. Ogura S., Yasukawa T., Kato A., Kuwayama S., Hamada S., Hirano Y., Uemura A., Yoshida M., Ogura Y.. Indocyanine Green Angiography-Guided Focal Laser Photocoagulation for Diabetic Macular Edema. Ophthalmologica (2015) 234 139-150. DOI: 10.1159/000437360 14. Datlinger F., Datlinger A., Pollreisz A., Sacu S., Schmidt-Erfurth U., Datlinger P.. Intraprocedural OCT monitoring of the immediate treatment response during indocyanine green angiography-guided laser therapy of teleangiectatic capillaries in diabetic macular edema. Sci. Rep. (2022) 12 2315. DOI: 10.1038/s41598-022-05950-0 15. Roh H.C., Lee C., Kang S.W., Choi K.J., Eun J.S., Hwang S.. Infrared reflectance image-guided laser photocoagulation of telangiectatic capillaries in persistent diabetic macular edema. Sci. Rep. (2021) 11 1-8. DOI: 10.1038/s41598-021-01183-9 16. Evaluation of grid pattern photocoagulation for macular edema in central vein occlusion. The Central Vein Occlusion Study Group M report. Ophthalmology (1995) 102 1425-1433. DOI: 10.1016/S0161-6420(95)30849-4 17. Elman M.J., Ayala A., Bressler N.M., Browning D., Flaxel C.J., Glassman A.R., Jampol L.M., Stone T.W.. Intravitreal Ranibizumab for Diabetic Macular Edema with Prompt versus Deferred Laser Treatment: 5-Year Randomized Trial Results. Ophthalmology (2014) 122 375-381. DOI: 10.1016/j.ophtha.2014.08.047 18. Brown D.M., Schmidt-Erfurth U., Do D.V., Holz F.G., Boyer D.S., Midena E., Heier J.S., Terasaki H., Kaiser P.K., Marcus D.M.. Intravitreal Aflibercept for Diabetic Macular Edema. Ophthalmology (2015) 122 2044-2052. DOI: 10.1016/j.ophtha.2015.06.017 19. Mitchell P., Wong T.Y.. Diabetic Macular Edema Treatment Guideline Working Group Management Paradigms for Diabetic Macular Edema. Am. J. Ophthalmol. (2013) 157 505-513.e8. DOI: 10.1016/j.ajo.2013.11.012 20. Glanville J., Patterson J., McCool R., Ferreira A., Gairy K., Pearce I.. Efficacy and safety of widely used treatments for macular oedema secondary to retinal vein occlusion: A systematic review. BMC Ophthalmol. (2014) 14 7. DOI: 10.1186/1471-2415-14-7 21. Amoroso F., Pedinielli A., Colantuono D., Jung C., Capuano V., Souied E.H., Miere A.. Selective Photocoagulation of Capillary Macroaneurysms by Navigated Focal Laser. Ophthalmic Surg. Lasers Imaging Retin. (2021) 52 366-373. DOI: 10.3928/23258160-20210628-02 22. Na Lee S., Chhablani J., Chan C.K., Wang H., Barteselli G., El-Emam S., Gomez M.L., Kozak I., Cheng L., Freeman W.R.. Characterization of Microaneurysm Closure After Focal Laser Photocoagulation in Diabetic Macular Edema. Am. J. Ophthalmol. (2013) 155 905-912.e2. DOI: 10.1016/j.ajo.2012.12.005 23. Inagaki K., Ohkoshi K., Ohde S., Deshpande G.A., Ebihara N., Murakami A.. Comparative efficacy of pure yellow (577-nm) and 810-nm subthreshold micropulse laser photocoagulation combined with yellow (561–577-nm) direct photocoagulation for diabetic macular edema. Jpn. J. Ophthalmol. (2014) 59 21-28. DOI: 10.1007/s10384-014-0361-1 24. Vujosevic S., Gatti V., Muraca A., Brambilla M., Villani E., Nucci P., Rossetti L., De Cilla S.. Optical coherence tomography angiography changes after subthreshold micropulse yellow laser in diabetic macular edema. Retina (2020) 40 312-321. DOI: 10.1097/IAE.0000000000002383 25. Starr M.R., Salabati M., Mahmoudzadeh R., Patel L.G., Ammar M.J., Hsu J., Garg S., Ho A.C., Kuriyan A.E.. Fluctuations in Central Subfield Thickness Associated With Worse Visual Outcomes in Patients With Diabetic Macular Edema in Clinical Trial Setting. Am. J. Ophthalmol. (2021) 232 90-97. DOI: 10.1016/j.ajo.2021.06.030 26. Lin J., Gibbons A., Smiddy W.E.. Cost-Utility of Anti–Vascular Endothelial Growth Factor Treatment for Macular Edema Secondary to Central Retinal Vein Occlusion. Ophthalmol. Retin. (2020) 5 656-663. DOI: 10.1016/j.oret.2020.09.017 27. Curry B.A., Sanfilippo P.G., Chan S., Hewitt A.W., Verma N.. Clinical Outcomes of a Treat and Extend Regimen with Intravitreal Aflibercept Injections in Patients with Diabetic Macular Edema: Experience in Clinical Practice. Ophthalmol. Ther. (2020) 9 87-101. DOI: 10.1007/s40123-019-00224-x
--- title: Relationship between Coffee, Tea, and Carbonated Beverages and Cardiovascular Risk Factors authors: - Hye-Ji An - Yejin Kim - Young-Gyun Seo journal: Nutrients year: 2023 pmcid: PMC9966641 doi: 10.3390/nu15040934 license: CC BY 4.0 --- # Relationship between Coffee, Tea, and Carbonated Beverages and Cardiovascular Risk Factors ## Abstract We aimed to analyze the relationship between coffee, tea, and carbonated beverages and cardiovascular risk factors. We used data from the fourth to eighth Korea National Health and Nutrition Examination Surveys (2007–2016, 2019–2020). We categorized the frequency of intake into three groups (<1 time/week, 1 time/week to <1 time/day, and ≥1 time/day). Subsequently, logistic regression analyses by sex were performed to assess cardiovascular risk factors (hypertension (HTN), diabetes mellitus (DM), dyslipidemia (DL), or metabolic syndrome (MetS)) according to the frequency of coffee, tea, and carbonated beverage intake. For HTN, coffee intake showed an inverse relationship and tea intake showed a direct relationship. For DM, coffee intake showed an inverse relationship, and tea and carbonated beverage intake showed a direct relationship. For DL, coffee intake showed an inverse relationship, whereas tea intake demonstrated a direct relationship. In addition, carbonated beverage intake showed a direct relationship with MetS. Coffee intake showed an inverse relationship with HTN, DM, and DL. However, tea intake showed a direct relationship with HTN, DM, and DL, whereas carbonated beverage intake showed a direct relationship with DM and MetS. ## 1. Introduction Coffee is the most popular beverage worldwide, with a $37\%$ increase in per capita coffee consumption worldwide over the past 20 years [1], and $66\%$ of Americans drink coffee daily [2]. In the case of Koreans, according to a survey on beverage and water intake of adults in 2019, coffee ranked first among beverages excluding water [3]. In addition, black and green teas are mainly consumed in West and East Asia, respectively [4], and tea consumption is increasing globally [5]. Regarding carbonated beverages, production in Korea increased by $13.9\%$ in 2018 compared with 2014, accounting for the largest scale ($45.5\%$) of the global beverage market [6]. Coffee contains micronutrients (magnesium, potassium, niacin, and vitamin E), caffeine, and chlorogenic acid [7]. Caffeine is involved in metabolic processes, such as increasing thermogenesis and metabolic rate or stimulating fat oxidation in peripheral tissues [8], and chlorogenic acid is expected to have the effect of preventing inflammation and oxidation. In addition, tea contains less caffeine, catechins, and other bioactive polyphenols (typically kaempferol) than coffee [9], and catechins and other polyphenols exhibit anti-tumor effects through interfering with cell division, activating programmed cell death, and inducing autophagocytosis [10]. Therefore, it was expected that tea and coffee would have health benefits. Many meta-analyses have demonstrated an association between tea or coffee and chronic metabolic diseases or cancer. A meta-analysis of patients with nonalcoholic fatty liver reported that coffee intake was inversely related to the liver fibrosis [11]. Consistent consumption of three to four cups of coffee per day was related to a reduced risk of all-cause and cardiovascular (CV) death, CV disease and cancer risk, and metabolic and liver disease compared to none at all [12]. In addition, two meta-analysis studies found high coffee intake to be inversely related to liver [13] and prostate [14] cancer, respectively. Another meta-analysis reported that tea intake was inversely related to all-cause, CV, and cancer deaths [15]. Thus, these studies support a trend toward increased consumption of coffee and tea worldwide. However, controversies still exist, as studies have shown that coffee intake is not significantly related to cancer death [16] or that tea consumption negatively affects disease [17]. In addition, there is a clinical study that cafestol, a diterpene contained in coffee, increases plasma triglycerides and low-density lipoprotein cholesterol [18]. However, various positive anti-inflammatory, anti-carcinogenic, and anti-diabetic effects of cafestol have also been reported [19]. Carbonated beverages contain various additives for flavor or food color, including artificial sweeteners for a sweet taste. Previous reviews have demonstrated that artificial sweeteners are associated with cardiometabolic risk [20,21]. Furthermore, beverages made with artificial sweeteners induce weight gain by reducing satiety [22], rapidly increasing blood sugar [23], and increasing the morbidity and mortality of CV disease [24]. In addition, although carbonated beverages also contain caffeine, the amount is not as high as that in coffee [25], and most caffeine intake in the United States population is driven by coffee and tea, while carbonated beverages make a small contribution to total caffeine intake [26]. CV disease is the leading cause of death, and obesity, metabolic syndrome (MetS), dyslipidemia (DL), diabetes mellitus (DM), and hypertension (HTN) are risk factors [27]. It is also associated with lifestyle factors such as lack of physical activity, excessive consumption of saturated fats, and tobacco and alcohol use [28]. Therefore, the aim of this study was to analyze the relationship between coffee, tea, and carbonated beverages and CV risk factors, including MetS, DL, DM, and HTN, using the data from the Korea National Health and Nutrition Examination Surveys (KNHANES). ## 2.1. Study Design and Participants The KNHANES has been performed by the Korea Disease Control and Prevention Agency (KDCA) since 1998 to produce representative statistics on the health and nutritional status of Koreans. 75,160 participants aged 19 years or older were selected from the fourth to eighth surveys (2007–2016, 2019–2020; the food intake frequency survey was stopped from 2017 to 2018 to review the survey method). Among them, 32,547 were excluded according to the following exclusion criteria: missing test results or survey records; inadequate water intake (≥90 g/kg of body weight); inadequate nutritional intake (>5000 or <500 kcal/day); inadequate fasting time before test sampling (<8 or >24 h); renal dysfunction (estimated glomerular filtration rate <30); and a history of diagnosed cancer. Consequently, data from 42,613 participants (17,311 men and 25,302 women) comprised the final dataset. All procedures were approved by the Ethics Committee of the KDCA (approval numbers 2018-01-03-2C-A, 2018-01-03-C-A, 2013-12EXP-03-5C, 2013-07CON-03-4C, 2012-01EXP-01-2C, 2011-02CON-06-C, 2010-02CON-21-C, 2009-01CON-03-2C, 2008-04EXP-01-C, and 2007-02CON-04-P) and were conducted in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. All KNHANES participants signed informed consent forms. The KNHANES data were made publicly available. ## 2.2. Coffee, Tea, and Carbonated Beverages Intake Information on the frequency of coffee, tea, and carbonated beverage intake over the past year was obtained using the questionnaire. It consisted of the following 10 categories in 2007–2011: almost no intake, 6–11 times per year (($\frac{8.5}{52.1}$)/week), 1 time per month (($\frac{1}{4.3}$)/week), 2–3 times per month (($\frac{2.5}{4.3}$)/week), 1 time per week (1/week), 2–3 times per week (2.5/week), 4–6 times per week (5/week), 1 time per day (1/day), 2 times per day (2/day), and 3 times per day (3/day). It consisted of the following nine categories in 2012–2016: almost no drinking, 1 time per month (($\frac{1}{4.3}$)/week), 2–3 times per month (($\frac{2.5}{4.3}$)/week), 1 time per week (1/week), 2–4 times per week (3/week), 5–6 times per week (5.5/week), 1 time per day (1/day), 2 times per day (2/day), and 3 times per day (3/day). It consisted of the following seven categories in 2019–2020: <1 time per week (0.5/week), 1 time per week (1/week), 2–4 times per week (3/week), 5–6 times per week (5.5/week), 1 time per day (1/day), 2 times per day (2/day), and 3 times per day (3/day). We categorized the frequency of intake as follows: group 1 (<1/week), group 2 (1/week to <1/day), and group 3 (≥1/day). ## 2.3. Other Variables We used the following variables: age, daily nutritional intake (total energy, carbohydrate, protein, and fat intake), average monthly household income, education (≤elementary school, middle school, high school, or ≥college), smoking (no, past, or current smoker), alcohol drinking frequency (<1 or ≥1/month), walking (<30 or ≥30 min/day, 5 days/week), body mass index (BMI; <25 or ≥25 kg/m2), menopause status, and comorbidities (doctor-diagnosed HTN, DM, or DL). Based on the 2001 National Cholesterol Education Program/Adult Treatment Panel III [29] and the 2005 American Heart Association/National Heart, Lung, and Blood Institute [30] criteria, we defined MetS when three or more of the following five factors were satisfied: [1] waist circumference ≥85 cm for women or ≥90 cm for men (Korean cutoff for abdominal obesity [31]); [2] serum triglyceride level ≥150 mg/dL or under treatment with drugs for DL; [3] serum high-density lipoprotein cholesterol level <50 mg/dL for women or <40 mg/dL for men; [4] blood pressure (BP) ≥$\frac{130}{85}$ mm Hg or under treatment with drugs for high BP; and [5] fasting plasma glucose level ≥100 mg/dL or under treatment with drugs for high glucose levels. ## 2.4. Statistical Analysis We used STATA version 14.0 (StataCorp., College Station, TX, USA) for statistical analysis, and $p \leq 0.05$ was set as the statistical significance level. The KNHANES was conducted using a two-stage stratified cluster sampling method. Therefore, we assigned weights to the stratified data in our analysis. Linear regression analyses and χ2 tests were used to analyze and compare the participants’ various characteristics according to sex and group. Logistic regression analyses by sex were used to assess CV risk factors (HTN, DM, DL, or MetS) according to coffee, tea, and carbonated beverage intake frequency. Adjusted odds ratios (ORs) were derived after controlling for potential confounding variables, such as age, daily nutritional intake (total and fat), average monthly household income, education, smoking, alcohol drinking, walking, BMI status, the frequency of intake of coffee, tea, and carbonated beverages, and menopausal status (only in women). In addition, the proportion of age groups according to the frequency of intake of coffee, tea, and carbonated beverages was derived. As a sensitivity analysis, logistic regression analyses were also conducted for participants 20 to less than 60 years. ## 3.1. General Characteristics Table S1 shows the participants’ general characteristics according to their sex. The average age of the 42,613 participants was 41.87 years, and $40.62\%$ were men. The proportion of men who drink coffee, tea, and carbonated beverages once a day or more was higher compared to women. Total energy intake, protein and fat intake, and average monthly household income were higher, and carbohydrate intake was lower in men compared to women. Furthermore, the proportion of highly educated participants (≥college), current smokers, alcohol drinkers (≥1/month), participants walking more than 30 min per day, 5 days a week, participants with BMI ≥ 25 kg/m2, and participants with HTN, DM, and MetS were higher in men compared to women. Tables S2–S7 show the participants’ general characteristics according to their coffee, tea, and carbonated beverage intake frequencies. In coffee, the proportion of men with HTN and DM was higher in group 1 (<1/week) than in groups 2 (1/week to <1/day) and 3 (≥1/day). The proportion of men with DL and MetS was higher in group 3 (≥1/day) than in groups 1 (<1/week) and 2 (1/week to <1/day) (Table S2). The proportion of women with HTN, DM, DL, and MetS was higher in group 1 (<1/week) than in groups 2 (1/week to <1/day) and 3 (≥1/day) (Table S3). In tea, the proportion of men with HTN, DM, DL, and MetS was higher in group 3 (≥1/day) than in groups 1 (<1/week) and 2 (1/week to <1/day) (Table S4). The proportion of women with HTN, DM, and DL was higher in group 3 (≥1/day) than in groups 1 (<1/week) and 2 (1/week to <1/day). The proportion of women with MetS was higher in group 1 (<1/week) than in groups 2 (1/week to <1/day) and 3 (≥1/day) (Table S5). In carbonated beverages, the proportion of men (Table S6) and women (Table S7) with HTN, DM, DL, and MetS was higher in group 1 (<1/week) than in groups 2 (1/week to <1/day) and 3 (≥1/day). ## 3.2. Cardiovascular Risk Factors by the Frequency of Intake of Coffee, Tea, and Carbonated Beverages The multivariable logistic regression analysis results for HTN according to the frequency of coffee, tea, and carbonated beverage intake after adjusting for potential confounding variables are shown in Table 1. For coffee consumption in men, groups 2 (OR, 0.66; $95\%$ CI, 0.49–0.90) and 3 (OR, 0.66; $95\%$ CI, 0.52–0.84) had lower adjusted odds of doctor-diagnosed HTN than group 1. For tea consumption in men, groups 2 (OR, 1.46; $95\%$ CI, 1.20–1.77) and 3 (OR, 2.60; $95\%$ CI, 2.02–3.34) had higher adjusted odds of doctor-diagnosed HTN than group 1. For tea consumption in women, groups 2 (OR, 1.58; $95\%$ CI, 1.30–1.92) and 3 (OR, 3.08; $95\%$ CI, 2.21–4.30) had higher adjusted odds of doctor-diagnosed HTN than group 1. The multivariable logistic regression analysis results for DM according to the frequency of coffee, tea, and carbonated beverage intake after adjusting for potential confounding variables are shown in Table 2. For coffee intake in men, group 3 (OR, 0.64; $95\%$ CI, 0.47–0.87) had lower adjusted odds of doctor-diagnosed DM than group 1. For tea consumption in men, groups 2 (OR, 1.84; $95\%$ CI, 1.40–2.42) and 3 (OR, 3.62; $95\%$ CI, 2.64–4.96) had higher adjusted odds of doctor-diagnosed DM than group 1. For coffee consumption in women, groups 2 (OR 0.66; $95\%$ CI, 0.50–0.88) and 3 (OR, 0.56; $95\%$ CI, 0.44–0.71) had lower adjusted odds of doctor-diagnosed DM than group 1. For tea consumption in women, groups 2 (OR, 1.54; $95\%$ CI, 1.19–1.99) and 3 (OR, 3.16; $95\%$ CI, 2.16–4.62) had higher adjusted odds of doctor-diagnosed DM than group 1. For carbonated beverages in women, group 3 (OR, 4.29; $95\%$ CI, 1.27–14.46) had higher adjusted odds of doctor-diagnosed DM than group 1. The multivariable logistic regression analysis results for DL according to the frequency of coffee, tea, and carbonated beverage intake after adjusting for potential confounding variables are shown in Table 3. For coffee consumption in men, groups 2 (OR, 0.63; $95\%$ CI, 0.46–0.85) and 3 (OR, 0.65; $95\%$ CI, 0.50–0.84) had lower adjusted odds of doctor-diagnosed DL than group 1. For tea consumption in men, groups 2 (OR, 1.55; $95\%$ CI, 1.26–1.92) and 3 (OR, 2.56; $95\%$ CI, 1.93–3.41) had higher adjusted odds of doctor-diagnosed DL than group 1. For tea consumption in women, groups 2 (OR, 1.43; $95\%$ CI, 1.18–1.73) and 3 (OR, 2.20; $95\%$ CI, 1.60–3.01) had higher adjusted odds of doctor-diagnosed DL than group 1. The multivariable logistic regression analysis results for MetS according to the frequency of coffee, tea, and carbonated beverage intake after adjusting for potential confounding variables are shown in Table 4. For carbonated beverages in women, groups 2 (OR, 1.19; $95\%$ CI, 1.03–1.37) and 3 (OR, 2.10; $95\%$ CI, 1.35–3.28) had higher adjusted odds of MetS than group 1. ## 3.3. Cardiovascular Risk Factors for Participants Aged 20 to Less than 60 Years Although there were differences by age, most of them drank more coffee than other beverages, and the age group with a high percentage by category of intake pattern was limited to those in their 20s to 50s (Table S8); therefore, only this age group was reanalyzed. For HTN, results that were similar to those of all age groups were obtained (Table S9). For DM, results generally similar to those of all age groups were obtained, except for carbonated beverages in women, which provided non-significant results (Table S10). For DL, results that were similar to those of all age groups were obtained (Table S11). For MetS, results generally similar to those of all age groups were obtained, except for coffee in women, which provided statistically significant results (group 3: OR, 0.82; $95\%$ CI, 0.72–0.95) (Table S12). ## 3.4. Cardiovascular Risk Factors When the Frequency of Intake Is Categorized into Four Groups The median frequency of intake was 1 time per day (1/day) (inter-quartile range [IQR], 1 time per week (1/week)–2 times per day (2/day)) for coffee, 6–11 times per year ($\frac{8.5}{52.1}$)/week (IQR, 0–1 time per week (1/week)) for tea, and 1 time per month ($\frac{1}{4.3}$)/week (IQR, 0–1 time per week (1/week)) for carbonated beverages. Therefore, we further categorized the frequency of coffee, tea, and carbonated beverages intake as follows: coffee group 1 (<1/week), coffee group 2 (1/week to <1/day), coffee group 3 (1/day to <2/day), and coffee group 4 (≥2/day); tea group 1 (<($\frac{8.5}{52.1}$)/week), tea group 2 (($\frac{8.5}{52.1}$)/week) to <1/week), tea group 3 (1/week to <1/day), and tea group 4 (≥1/day); carbonated beverages group 1 (<($\frac{1}{4.3}$)/week), carbonated beverages group 2 (($\frac{1}{4.3}$)/week) to <1/week), carbonated beverages group 3 (1/week to <1/day), and carbonated beverages group 4 (≥1/day). For HTN, results generally similar to those of the three groups were obtained, except for coffee group 3 in men, which provided non-significant results, and coffee group 4 and tea group 2 in women, which provided statistically significant results (coffee group 4 in women: OR, 0.77; $95\%$ CI, 0.63–0.94; and tea group 2 in women: OR, 1.27; $95\%$ CI, 1.04–1.57) (Table S13). For DM, results generally similar to those of the three groups were obtained, except for coffee group 3 in both men and women, which provided non-significant results (Table S14). For DL, results generally similar to those of three groups were obtained, except for coffee group 3 in men, which provided non-significant results, and tea group 2 in men and tea group 2 and carbonated beverages group 4 in women, which provided statistically significant results (tea group 2 in men: OR, 1.37; $95\%$ CI, 1.07–1.75, and tea group 2 in women: OR, 1.28; $95\%$ CI, 1.04–1.57, and carbonated beverages group 4 in women: OR, 2.72; $95\%$ CI, 1.01–7.35) (Table S15). For MetS, results generally similar to those of the three groups were obtained, except for coffee group 3 in women, which provided non-significant results, and carbonated beverages group 2 in women, which provided statistically significant results (carbonated beverages group 2 in women: OR, 1.19; $95\%$ CI, 1.05–1.34) (Table S16). ## 4. Discussion In this study, HTN and DL were related to low coffee and high tea consumption, respectively. DM was related to low coffee, high tea, and high carbonated beverage intake in women. MetS has been linked to a high intake of carbonated beverages in women. When the frequency of coffee intake was subdivided, coffee consumption between once and less than twice a day was not associated with cardiovascular risk factors. The following points are in the same direction as the previous studies: [1] coffee intake is inversely related to the risk of HTN [32] and DM [33], and [2] beverages made with artificial sweeteners are related to the risk of DM [34]. [ 3] In addition, both sweetened and artificially sweetened beverages, which are also a type of carbonated beverage, are associated with MetS [35]. However, the following points are inconsistent with the previous studies: [1] tea consumption was able to reduce blood levels of fasting glucose in participants younger than 55 years old [36]. [ 2] both tea and coffee intake are less likely to develop MetS [11]. [ 3] Coffee may be related to the risk of DL. [ 37] The following pathophysiology partially supports the findings of our study: [1] A prospective cohort study found that the pesticide residue (1,1,1-trichloro-2,2-bis(p-chlorophenyl) ethane) of oolong tea may have been involved in the association between long-term oolong tea intake and the DM risk [38]. [ 2] Chlorogenic acid, which is a major component of coffee, is involved in lipid metabolism, increases fat oxidation by upregulating peroxisome proliferator-activated receptor alpha (fenofibrate-like action), and reduces fatty acid synthesis through inhibiting 3-hydroxy-3-methylglutaryl-coenzyme A reductase and hepatic fatty acid synthase (statin-like action) [39]. However, further studies are needed to clarify these differences from previous studies. Although meta-analyses of prospective studies have found that tea intake is related to a reduced risk of DM and MetS [11,36], a few cross-sectional studies have found that tea intake is related to a higher BMI and metabolic abnormalities [17,40]. This is consistent with our study, in that tea intake was related to higher odds of HTN, DM, and DL in both women and men. This might be explained by the reverse causality that participants with chronic diseases are more likely to care for their diets and receive nutritional counseling than those without chronic diseases. For example, in a Canadian population of 98,733 adolescents to the elderly, participants with DM, heart disease, or cancer were more inclined to select or avoid foods based on health considerations or food content than subjects without these conditions [41]. Similarly, in a study of 1399 Italians, participants with DM had a healthier diet than participants without DM in both food and beverage choices; patients with DM had more calories from vegetables, fruits, and eggs than those without DM and consumed less juice but more water [42]. Considering that tea is one of Asia’s most consumed beverages and its health benefits are well known, tea intake may increase when the participant’s health deteriorates. Another possible explanation is the presence of plastic substances in tea bags. Tea bags are mainly made of plastic, such as nylon [43], which enables the release of significant amounts of plastic-related particles [44]. Additionally, the migration of cyclic oligomers was observed in a polyamide-6 (nylon 6) tea bag in hot water [45]. A recent study detected that one plastic tea bag at brewing temperature discharged 11.6 billion micro- and 3.1 billion nano-plastics into a single cup, respectively [46]. Although the association between microplastics and nanoplastics and health risk is unclear, microplastics and nanoplastics can induce oxidative stress, inflammation [47], and carcinogenesis [48]. Our study found that individuals who consumed more carbonated beverages had a higher risk of DM and MetS than those who consumed fewer carbonated beverages. This can be explained by the main sources of carbonated beverages, fructose, and liquid carbohydrates, which increase the overall calorie intake due to decreased satiety and induce central obesity, insulin resistance, and CV disease [22,23,34]. In contrast, maintaining or reducing the consumption of carbonated beverages or replacing them with water, tea, or coffee was related to a reduced risk of DM [49]. Moreover, frequent consumption of carbonated beverages can be a proxy indicator of an unhealthy lifestyle, such as lower physical activity levels, a poor diet, and smoking [50,51]. Individuals who consumed carbonated beverages once a day or more had higher calorie intake, current smoking, and alcohol consumption than those who consumed less than once a week in our study. However, the association between the consumption of carbonated beverages, DM, and MetS was only statistically significant in women. This is consistent with a 10-year prospective study in Korean adults, where soft beverage consumption was associated with an increased incidence of MetS and elevated its components only in women [52]. Additionally, a recent cross-sectional study found that sugar-sweetened beverage intake is related to metabolic risk, particularly in women [53]. This sex difference may be explained by sex hormone levels. Estrogen is crucial for fundamental sex differences when lipids and carbohydrates are used as fuel sources [54]. Estrogen affects the renin-angiotensin system, which activates the transport of fat and increases triglyceride and total cholesterol levels, whereas androgens have the opposite effect in men [55]. This can cause the response to carbohydrate or fat intake to become more sensitive in women. Therefore, sex hormones may have contributed to the relationship between carbonated beverage intake and DM and MetS, particularly in women. In our study, a significant difference was found in the association between the frequency of coffee and tea consumption and the adjusted odds of doctor-diagnosed HTN, DM, and DL. However, no difference was observed in the relationship with MetS. Due to the nature of MetS, which is a cluster of three or more CV risk factors, coffee and tea consumption are insufficient to influence more than three components at once. Since many factors cause MetS, it is impossible to control them when conducting a study; therefore, it might be challenging to demonstrate our thoughts. There were some limitations in this study. First, as this was a cross-sectional design, we could not evaluate the cause-and-effect relationship between beverages (coffee, tea, and carbonated beverages) and each disease (HTN, DM, DL, and MetS). Therefore, it is necessary to conduct a randomized controlled trial to confirm whether the same results can be obtained in clinical practice. Second, it did not reflect the diversity of the detailed types and ingredients of beverages. In the case of coffee, there are various types, and the content of components such as caffeine and chlorogenic acid differs depending on the type of coffee. There are also various types of tea, as well as the three main types of green tea, black tea, and oolong tea, according to the degree of fermentation, and each tea has a different type and amount of polyphenols. Therefore, studies should be performed to confirm the association between more subcategorized beverage types and diseases. Third, when collecting data, the format of the survey on dietary habits was changed in the middle (24-h recall survey in 2007–2017, open type; dietary survey in 2018–2019, categorical type). Therefore, there is a possibility of data inconsistency. Future studies need to consider these limitations. ## 5. Conclusions In this study, coffee intake showed an inverse relationship with HTN, DM, and DL. However, tea intake showed a direct relationship with HTN, DM, and DL, and carbonated beverage intake showed a direct relationship with DM and MetS. Therefore, we expect to prevent metabolic diseases, such as HTN, DM, DL, and MetS, in adults and further reduce CV disease morbidity by improving the amount and type of beverage consumption. ## References 1. Quadra G.R., Brovini E.M., dos Santos J.A., Paranaíba J.R.. Caffeine consumption over time. Handbook of Substance Misuse and Addictions: From Biology to Public Health (2022.0) 1-18. DOI: 10.1007/978-3-030-67928-6_78-1 2. National Coffee DATA trends Fall 2022 3. Han G.. Status of beverage and water intake among adults in Korea-Data from Korea National health and nutrition examination survey 2019. Korean J. Food Nutr. (2021.0) 34 430-440. DOI: 10.9799/KSFAN.2021.34.5.430 4. Yang C.S., Landau J.M.. Effects of tea consumption on nutrition and health. J. Nutr. (2000.0) 130 2409-2412. DOI: 10.1093/jn/130.10.2409 5. Tea Fact Sheet 2022 6. 2019 Processed Foods Segmented Market Status: Beverage Market’ Report 7. Higdon J.V., Frei B.B.. Coffee and health: A review of recent human research. Crit. Rev. Food Sci. Nutr. (2006.0) 46 101-123. DOI: 10.1080/10408390500400009 8. Shang F., Li X., Jiang X.. Coffee consumption and risk of the metabolic syndrome: A meta-analysis. Diabetes Metab. (2016.0) 42 80-87. DOI: 10.1016/j.diabet.2015.09.001 9. Sirotkin A.V., Kolesárová A.. The anti-obesity and health-promoting effects of tea and coffee. Physiol. Res. (2021.0) 70 161-168. DOI: 10.33549/physiolres.934674 10. Singh A.K., Bishayee A., Pandey A.K.. Targeting histone deacetylases with natural and synthetic agents: An emerging anticancer strategy. Nutrients (2018.0) 10. DOI: 10.3390/nu10060731 11. Marventano S., Salomone F., Godos J., Pluchinotta F., Del Rio D., Mistretta A., Grosso G.. Coffee and tea consumption in relation with nonalcoholic fatty liver and metabolic syndrome: A systematic review and meta-analysis of observational studies. Clin. Nutr. (2016.0) 35 1269-1281. DOI: 10.1016/j.clnu.2016.03.012 12. Poole R., Kennedy O.J., Roderick P., Fallowfield J.A., Hayes P.C., Parkes J.. Coffee consumption and health: Umbrella review of meta-analyses of multiple health outcomes. BMJ (2017.0) 359 j5024. DOI: 10.1136/bmj.j5024 13. Sang L.-X., Chang B., Li X.-H., Jiang M.. Consumption of coffee associated with reduced risk of liver cancer: A meta-analysis. BMC Gastroenterol. (2013.0) 13. DOI: 10.1186/1471-230X-13-34 14. Chen X., Zhao Y., Tao Z., Wang K.. Coffee consumption and risk of prostate cancer: A systematic review and meta-analysis. BMJ Open (2021.0) 11 e038902. DOI: 10.1136/bmjopen-2020-038902 15. Tang J., Zheng J.S., Fang L., Jin Y., Cai W., Li D.. Tea consumption and mortality of all cancers, CVD and all causes: A meta-analysis of eighteen prospective cohort studies. Br. J. Nutr. (2015.0) 114 673-683. DOI: 10.1017/S0007114515002329 16. Crippa A., Discacciati A., Larsson S.C., Wolk A., Orsini N.. Coffee consumption and mortality from all causes, cardiovascular disease, and cancer: A dose-response meta-analysis. Am. J. Epidemiol. (2014.0) 180 763-775. DOI: 10.1093/aje/kwu194 17. Wan C.J., Lin L.Y., Yu T.H., Sheu W.H.H.. Metabolic syndrome associated with habitual indulgence and dietary behavior in middle-aged health-care professionals. J. Diabetes Investig. (2010.0) 1 259-265. DOI: 10.1111/j.2040-1124.2010.00055.x 18. De Roos B., Van Tol A., Urgert R., Scheek L.M., Van Gent T., Buytenhek R., Princen H.M., Katan M.B.. Consumption of French-press coffee raises cholesteryl ester transfer protein activity levels before LDL cholesterol in normolipidaemic subjects. J. Intern. Med. (2000.0) 248 211-216. DOI: 10.1046/j.1365-2796.2000.00728.x 19. Ren Y., Wang C., Xu J., Wang S.. Cafestol and kahweol: A review on their bioactivities and pharmacological properties. Int. J. Mol. Sci. (2019.0) 20. DOI: 10.3390/ijms20174238 20. Malik V.S., Popkin B.M., Bray G.A., Després J.P., Willett W.C., Hu F.B.. Sugar-sweetened beverages and risk of metabolic syndrome and type 2 diabetes: A meta-analysis. Diabetes Care (2010.0) 33 2477-2483. DOI: 10.2337/dc10-1079 21. Malik V.S., Hu F.B.. Sugar-sweetened beverages and cardiometabolic health: An update of the evidence. Nutrients (2019.0) 11. DOI: 10.3390/nu11081840 22. Malik V.S., Popkin B.M., Bray G.A., Després J.P., Hu F.B.. Sugar-sweetened beverages, obesity, type 2 diabetes mellitus, and cardiovascular disease risk. Circulation (2010.0) 121 1356-1364. DOI: 10.1161/CIRCULATIONAHA.109.876185 23. Solomi L., Rees G.A., Redfern K.M.. The acute effects of the non-nutritive sweeteners aspartame and acesulfame-K in UK diet cola on glycaemic response. Int. J. Food Sci. Nutr. (2019.0) 70 894-900. DOI: 10.1080/09637486.2019.1585418 24. Yin J., Zhu Y., Malik V., Li X., Peng X., Zhang F.F., Shan Z., Liu L.. Intake of sugar-sweetened and low-calorie sweetened beverages and risk of cardiovascular disease: A meta-analysis and systematic review. Adv. Nutr. (2021.0) 12 89-101. DOI: 10.1093/advances/nmaa084 25. McCusker R.R., Goldberger B.A., Cone E.J.. Caffeine content of energy drinks, carbonated sodas, and other beverages. J. Anal. Toxicol. (2006.0) 30 112-114. DOI: 10.1093/jat/30.2.112 26. Drewnowski A., Rehm C.D.. Sources of caffeine in diets of US children and adults: Trends by beverage type and purchase location. Nutrients (2016.0) 8. DOI: 10.3390/nu8030154 27. Fryar C.D., Chen T.C., Li X.. Prevalence of uncontrolled risk factors for cardiovascular disease: United States, 1999–2010 (2012.0) 1-8 28. Heart Disease 29. Executive summary of the third report of the National Cholesterol Education Program (NCEP) expert panel on detection, evaluation, and treatment of high blood cholesterol in adults (Adult Treatment Panel III). JAMA (2001.0) 285 2486-2497. DOI: 10.1001/jama.285.19.2486 30. Grundy S.M., Cleeman J.I., Daniels S.R., Donato K.A., Eckel R.H., Franklin B.A., Gordon D.J., Krauss R.M., Savage P.J., Smith S.C.. Diagnosis and management of the metabolic syndrome: An American Heart Association/National Heart Lung and Blood Institute scientific statement. Circulation (2005.0) 112 2735-2752. DOI: 10.1161/CIRCULATIONAHA.105.169404 31. Lee S., Park H.S., Kim S.M., Kwon H.S., Kim D.Y., Kim D.J., Cho G.J., Han J.H., Kim S.R., Park C.Y.. Cutoff points of waist circumference for defining abdominal obesity in the Korean population. Korean J. Obes. (2006.0) 15 1-9 32. Navarro A.M., Martinez-Gonzalez M.A., Gea A., Ramallal R., Ruiz-Canela M., Toledo E.. Coffee consumption and risk of hypertension in the SUN Project. Clin. Nutr. (2019.0) 38 389-397. DOI: 10.1016/j.clnu.2017.12.009 33. Kolb H., Martin S., Kempf K.. Coffee and lower risk of Type 2 diabetes: Arguments for a causal relationship. Nutrients (2021.0) 13. DOI: 10.3390/nu13041144 34. Hirahatake K.M., Jacobs D.R., Shikany J.M., Jiang L., Wong N.D., Steffen L.M., Odegaard A.O.. Cumulative intake of artificially sweetened and sugar-sweetened beverages and risk of incident type 2 diabetes in young adults: The Coronary artery Risk Development in Young Adults (CARDIA) Study. Am. J. Clin. Nutr. (2019.0) 110 733-741. DOI: 10.1093/ajcn/nqz154 35. Narain A., Kwok C.S., Mamas M.A.. Soft drink intake and the risk of metabolic syndrome: A systematic review and meta-analysis. Int. J. Clin. Pract. (2017.0) 71 e12927. DOI: 10.1111/ijcp.12927 36. Kondo Y., Goto A., Noma H., Iso H., Hayashi K., Noda M.. Effects of coffee and tea consumption on glucose metabolism: A systematic review and network meta-analysis. Nutrients (2018.0) 11. DOI: 10.3390/nu11010048 37. Du Y., Lv Y., Zha W., Hong X., Luo Q.. Effect of coffee consumption on dyslipidemia: A meta-analysis of randomized controlled trials. Nutr. Metab. Cardiovasc. Dis. (2020.0) 30 2159-2170. DOI: 10.1016/j.numecd.2020.08.017 38. Hayashino Y., Fukuhara S., Okamura T., Tanaka T., Ueshima H.. High oolong tea consumption predicts future risk of diabetes among Japanese male workers: A prospective cohort study. Diabet. Med. (2011.0) 28 805-810. DOI: 10.1111/j.1464-5491.2011.03239.x 39. Santos R.M.M., Lima D.R.A.. Coffee consumption, obesity and type 2 diabetes: A mini-review. Eur. J. Nutr. (2016.0) 55 1345-1358. DOI: 10.1007/s00394-016-1206-0 40. Yu S., Guo X., Yang H., Zheng L., Sun Y.. An update on the prevalence of metabolic syndrome and its associated factors in rural northeast China. BMC Public Health (2014.0) 14. DOI: 10.1186/1471-2458-14-877 41. Ree M., Riediger N., Moghadasian M.H.. Factors affecting food selection in Canadian population. Eur. J. Clin. Nutr. (2008.0) 62 1255-1262. DOI: 10.1038/sj.ejcn.1602863 42. Tatoli R., Lampignano L., Bortone I., Donghia R., Castellana F., Zupo R., Tirelli S., De Nucci S., Sila A., Natuzzi A.. Dietary patterns associated with diabetes in an older population from Southern Italy using an unsupervised learning approach. Sensors (2022.0) 22. DOI: 10.3390/s22062193 43. Jha D.K., Dhekne P.P., Patwardhan A.W.. Characterization and evaluation of tea bag papers. J. Food Sci. Technol. (2020.0) 57 3060-3070. DOI: 10.1007/s13197-020-04339-z 44. Xu J.L., Lin X., Hugelier S., Herrero-Langreo A., Gowen A.A.. Spectral Imaging for characterization and detection of plastic substances in branded teabags. J. Hazard. Mater. (2021.0) 418 126328. DOI: 10.1016/j.jhazmat.2021.126328 45. Heimrich M., Nickl H., Bönsch M., Simat T.J.. Migration of cyclic monomer and oligomers from polyamide 6 and 66 food contact materials into food and food simulants: Direct food contact. Packag. Technol. Sci. (2015.0) 28 123-139. DOI: 10.1002/pts.2094 46. Hernandez L.M., Xu E.G., Larsson H.C.E., Tahara R., Maisuria V.B., Tufenkji N.. Plastic teabags release billions of microparticles and nanoparticles into tea. Environ. Sci. Technol. (2019.0) 53 12300-12310. DOI: 10.1021/acs.est.9b02540 47. Hu M., Palić D.. Micro- and nano-plastics activation of oxidative and inflammatory adverse outcome pathways. Redox Biol. (2020.0) 37 101620. DOI: 10.1016/j.redox.2020.101620 48. Kumar R., Manna C., Padha S., Verma A., Sharma P., Dhar A., Ghosh A., Bhattacharya P.. Micro(nano)plastics pollution and human health: How plastics can induce carcinogenesis to humans?. Chemosphere (2022.0) 298 134267. DOI: 10.1016/j.chemosphere.2022.134267 49. Drouin-Chartier J.P., Zheng Y., Li Y., Malik V., Pan A., Bhupathiraju S.N., Tobias D.K., Manson J.E., Willett W.C., Hu F.B.. Changes in Consumption of Sugary Beverages and Artificially Sweetened Beverages and Subsequent Risk of Type 2 Diabetes: Results From Three Large Prospective U.S. Cohorts of Women and Men. Diabetes Care (2019.0) 42 2181-2189. DOI: 10.2337/dc19-0734 50. Barrett P., Imamura F., Brage S., Griffin S.J., Wareham N.J., Forouhi N.G.. Sociodemographic, lifestyle and behavioural factors associated with consumption of sweetened beverages among adults in Cambridgeshire, UK: The Fenland Study. Public Health Nutr. (2017.0) 20 2766-2777. DOI: 10.1017/S136898001700177X 51. Ejtahed H.S., Bahadoran Z., Mirmiran P., Azizi F.. Sugar-sweetened beverage consumption is associated with metabolic syndrome in Iranian adults: Tehran lipid and glucose study. Endocrinol. Metab. (2015.0) 30 334-342. DOI: 10.3803/EnM.2015.30.3.334 52. Kang Y., Kim J.. Soft drink consumption is associated with increased incidence of the metabolic syndrome only in women. Br. J. Nutr. (2017.0) 117 315-324. DOI: 10.1017/S0007114517000046 53. Kuo C.-T., Chen D.-R., Chan C.-C., Yeh Y.-P., Chen H.-H.. Sex differences in the association between sugar-sweetened beverages consumption and metabolic risks among the working-age population in Taiwan. Public Health Nutr. (2022.0) 1-19. DOI: 10.1017/S1368980022001549 54. Mauvais-Jarvis F.. Sex differences in metabolic homeostasis, diabetes, and obesity. Biol. Sex Differ. (2015.0) 6 14. DOI: 10.1186/s13293-015-0033-y 55. White M.C., Fleeman R., Arnold A.C.. Sex differences in the metabolic effects of the renin-angiotensin system. Biol. Sex Differ. (2019.0) 10 31. DOI: 10.1186/s13293-019-0247-5
--- title: Oral Wound Healing Potential of Polygoni Cuspidati Rhizoma et Radix Decoction—In Vitro Study authors: - Jakub Hadzik - Anna Choromańska - Bożena Karolewicz - Adam Matkowski - Marzena Dominiak - Adrianna Złocińska - Izabela Nawrot-Hadzik journal: Pharmaceuticals year: 2023 pmcid: PMC9966654 doi: 10.3390/ph16020267 license: CC BY 4.0 --- # Oral Wound Healing Potential of Polygoni Cuspidati Rhizoma et Radix Decoction—In Vitro Study ## Abstract Polygoni Cuspidati Rhizoma et Radix (syn. rhizomes of Reynoutria japonica Houtt.) is a pharmacopoeial raw material in Europe and China. In traditional medicine, one of the applications for Reynoutria japonica rhizomes is wound healing. In a recent in vitro study, we demonstrated that ethanol and acetone extracts from this herbal drug have the potential to heal oral gum wounds. However, considering that a majority of herbal medicines have been traditionally administered as water decoctions, in the present study, a decoction of Reynoutria japonica rhizomes was prepared and detailed tests to determine its in vitro gingival wound healing activity were conducted. We used the primary human gingival fibroblasts (HGF) incubated with a decoction to determine cell viability (MTT assay), cell proliferation (the confocal laser scanning microscope—CLSM), and cell migration (wound healing assay). Moreover, the collagen type III expression was examined using immunocytochemical staining. The studied decoction was qualitatively and quantitatively characterized using the validated HPLC/DAD/ESI-HR-QTOF-MS method. The Folin–Ciocalteu assay was used to determine the total phenols and tannins content. Additionally, HPLC-RI analysis of decoction and the previously obtained ethanol and acetone extracts was used to determine the composition of saccharides. Low concentration (from 50 to 1000 µg/mL) of decoction after 24 h incubation caused a significant increase in HGF cell viability. No cytotoxic effect was observed at any tested concentration (up to 2000 µg/mL). The lowest active concentration of decoction (50 µg/mL) was selected for further experiments. It significantly stimulated human gingival fibroblasts to proliferate, migrate, and increase the synthesis of collagen III. Phytochemical analysis showed significantly fewer polyphenols in the decoction than in the ethanol and acetone extracts tested earlier. In contrast, high levels of polysaccharides were observed. In our opinion, they may have a significant effect on the oral wound healing parameters analyzed in vitro. The results obtained encourage the use of this raw material in its traditional, safe form—decoction. ## 1. Introduction The rhizome with roots of Reynoutria japonica Houtt. is a traditional Chinese medicinal herb (hu zhang in pinyin Chinese) listed in the European Pharmacopoeia under the name Polygoni cuspidati rhizoma et radix. This herbal medicine was included in the European Pharmacopoeia in 2017. According to Pharmacopoeia, a minimum of $1.0\%$ emodin content and $1.5\%$ of polydatin content in the dried herb are required as a quality check [1]. Traditionally, hu zhang has been used in East Asia against many diseases, such as hyperlipemia, inflammation, infection, and for wound healing [2]. According to current knowledge, the most important bioactive compounds in Polygoni cuspidati rhizoma et radix are stilbenes (with a high concentration of resveratrol), anthraquinones, procyanidins, phenylpropanoid disaccharide esters, and other polyphenols [2,3,4,5]. To obtain a high content of these constituents, different percentages of ethanol/methanol or acetone are used in the extraction process [3,6]. In the previous in vitro study, our research team reported the gingival wound healing activity of $25\%$ and $40\%$ ethanolic as well as $60\%$ acetone extracts from the Reynoutria japonica rhizomes [7]. The extracts stimulated human gingival fibroblasts (HGF) to proliferate and migrate, as well as increasing the synthesis of collagen III, but with different potency. The highest stimulation of proliferative and migratory activity was observed after incubation with $25\%$ EtOH extract. This effect may have been related to the highest resveratrol content and the favorable composition of procyanidins. Currently, our attention has been drawn to the fact that although the overwhelming majority of Chinese herbal medicines (CHMs) are traditionally administered as water decoctions (tangs), relatively little research is done on them [8]. This experiment was designed to test whether a traditionally used Polygoni cuspidati rhizoma et radix decoction could have a stimulating effect on the oral wound healing process. The intraoral healing process is different from skin wound healing, so a different therapeutic approach is required. Intraoral wound healing aids are desirable for patients whose natural wound healing process is impaired, i.e., immunocompromised, post-transplant, post-radiotherapy or chemotherapy patients, and persons with systemic diseases [9,10]. Also, periodontal diseases, including periodontal and peri-implant disease [11], predispose to impaired wound healing due to high levels of oxidative stress, which impairs the proliferation and migration of HGFs [12,13]. Also, older age affects oral wound healing by impairing fibroblast proliferation, migration, and differentiation, and reducing collagen synthesis [14]. Moreover, prolonged wound healing in the oral cavity predisposes to bacterial invasion and wound infection, which can further inhibit the proper wound healing process. Due to the unsatisfactory results of currently available treatments, there is a need to search for new, effective therapies that promote intraoral healing [15,16]. In this study, we have prepared a decoction of Reynoutria japonica rhizomes according to the traditional recipe. We conducted a detailed study to determine its activity in healing gingival wounds in vitro using the same raw material and the same research methods as in our previous study [7]. Additionally, an HPLC-RI analysis of the decoction and the previously obtained extracts was used to determine the composition of saccharides. ## 2.1. Cell Viability—MTT Assay The Reynoutria japonica rhizome decoction was not cytotoxic to HGF cells at any tested concentrations and incubation time (Figure 1). It was different from the previously tested ethanol and acetone extracts, which at high concentrations—from 1000 µg/mL to 2000 µg/mL—showed a significant reduction in the viability of fibroblasts after 24 h of incubation [7]. The decoction at concentrations from 50 µg/mL to 1000 µg/mL, after 24 h of incubation, significantly increased the viability of HGF cells, up to $124\%$ compared with that of the control. This stimulating effect on cellular metabolism has also been observed in previous studies with ethanol and acetone extracts from R. japonica rhizomes [5,7]. To check if fibroblasts proliferate after treatment with the decoction, we used a confocal laser scanning microscope to visualize changes in histone 3 expression. The lowest concentration (50 µg/mL) demonstrating a significant stimulation of the fibroblasts’ viability in the MTT test was selected for further studies (the same as in our previous research with solvent extracts [7]). ## 2.2. Confocal Laser Microscopy Study Microphotographs of cells with immunofluorescent staining of histone H3 after 24 h incubation with tested decoction (at a concentration of 50 µg/mL) and without decoction (untreated cells) are presented in Figure 2. The primary rabbit polyclonal antibody anti-phospho-histone H3 was used to present an abundance of phosphorylation of serine 10 on histone (H3S10ph), the increase of which is observed during cell division [17,18]. Figure 3 shows the mean fluorescence intensity in arbitrary units for treated cells [22,488] and for untreated cells (4.515). The results indicate that the studied decoction significantly stimulated human gingival fibroblasts to divide. Importantly, the fluorescence intensity after incubation with the decoction was even higher than with $25\%$ EtOH extract, which was the most active in our previous study [7]. ## 2.3. Wound Healing Assay Proper proliferation and migration of gingival fibroblast cells play a significant role in the healing of oral cavity wounds. Using a cell migration-based wound healing assay, we analyzed the motility of the fibroblasts after incubation with 50 µg/mL of R. japonica decoction. Figure 4A shows the changes in the cell-covered area (gap closure) over time. Figure 4B illustrates the percentage of a healed wound as a function of time. Table 1 presents the percentage of wound (gap) closure over time. ## 2.4. Immunocytochemical Staining Gingival fibroblasts synthesize and secrete collagen type III, a high expression of which is observed during wound healing. Over time, type III collagen is reabsorbed and replaced by type I collagen, a key component of the extracellular matrix. Microphotographs (Figure 5) of HGF with immunocytochemically stained collagen type III show the influence of R. japonica decoction on expression of collagen type III. According to Table 2, which shows the semi-quantitative results, the 50 µg/mL of R. japonica decoction had a similar stimulating effect on collagen III production as 2 µM betulinic acid, used as the positive control. In comparison to our previous study, the effect of R. japonica decoction on stimulation of fibroblasts to produce collagen III was similar to $60\%$ acetone extract, slightly weaker than $40\%$ EtOH, but stronger than $25\%$ EtOH extract [7]. ## 2.5. HPLC/DAD/ESI-HR-TOF-MS Analysis To determine the composition of the compounds that contributed to the observed effects, qualitative and quantitative HPLC/DAD/ESI-HR-QTOF-MS analyses of decoction was performed. The UHPLC-QTOF-MS analysis revealed a total of 37 different compounds (Table 3, Figure 6) that belong to carbohydrates, stilbenes, flavan-3-ols, procyanidins, anthraquinones, organic acids, and naphthalenes. Among all of the detected chromatographic peaks, seven remained unassigned and without a clear indication of their chemical nature. Two peaks were tentatively defined as carbohydrates. Most of the identified compounds were previously detected in extracts and fractions from rhizomes of R. japonica, as described in our previous reports [3,5,7,19]. The decoction was less diverse in terms of the content of compounds than the ethanol and acetone extracts analyzed in previous studies. It did not contain phenylpropanoids. Moreover, the trace amount of resveratrol was only noticed after extraction from the raw chromatogram. Stilbene glycosides and anthraquinone glycosides were present in greater amounts than their aglycones. In contrast, compound 31 with a high peak in the BPC (base peak chromatogram), labeled as unknown, with a deprotonated ion at m/z 253.0513 [M∓H]− and predicted formula of C15H9O4 was not observed in the previously studied extracts. Product ions at m/z 225.0545 (C14H9O3, 0.3 ppm), m/z 224.0481 (C14H8O3, 1.26 ppm), m/z 209.0611 (C14H9O2, −1.2 ppm), m/z 197.0606 (C13H9O2, 0.9 ppm), m/z 169.0664 (C12H9O, −3.0 ppm), and m/z 135.009 (C7H3O3, −1.5 ppm) suggest that it may be some kind of anthraquinone, i.e., rubiadin (HMDB0257354). In order to determine the structure of this compound, it must be isolated and subjected to detailed studies. A previously developed, validated analytical method was used to quantify the selected compounds [19]. Piceid was detected in much lower amounts in the decoction than in previously tested ethanol or acetone extracts (Table 4). There was almost six times less piceid in the decoction than in the previously tested $25\%$ EtOH extract and over ten times less than in the $60\%$ acetone extract. Resveratrol was not detectable without extraction from the chromatogram and was not quantifiable. Vanicoside A and B were not detected even in trace amounts Also, emodin and physcion were detected in much lower amounts than in ethanol and acetone extracts. Their content was sufficient for detection but not for correct quantification. ## 2.6. Total Polyphenols and Tannins Content According to the modified Folin–Ciocalteu assay (Figure 7), total polyphenols and tannins make up a small part of the decoction. ## 2.7. HPLC-RI Analysis The low content of polyphenols, including tannins, stilbenes, anthraquinones, phenylpropanoids, as well as the viscous consistency of the decoction encouraged us to examine its saccharide content. To compare the decoction with ethanol and acetone extracts from the previous study, the latter were also analyzed. The composition of saccharides in the studied samples was determined by high-pressure liquid chromatography with a refractometric detector (Figure 8A). All the extracts tested previously—$25\%$ EtOH (B), $40\%$ EtOH (C), and $60\%$ acetone (D) showed similar HPLC-RI chromatograms, which revealed the presence of three main compounds—glucose, xylose, and an unknown compound 1. The HPLC-RI chromatogram of the decoction differs from the extracts by the presence of a much larger unknown peak number 1 which also appears at a shorter retention time. To find out what kind of compound the unknown peak 1 is, multiple sugars were tested, among others: two monosaccharides (xylose (A) and glucose (B)), a disaccharide (maltose (C)), trisaccharide (maltotriose (D)), and polysaccharide (pectins (E)) (Figure 8B). As shown above, under the chosen analysis conditions, the analytes flow from the Rezex ROA—Organic Acid H + column (Phenomenex) in the following order: first the most complex saccharides, such as the polysaccharide pectins, followed by trisaccharides, then disaccharides, and finally monosaccharides. We also confirmed this dependence on other compounds belonging to these sugar groups (data not presented). This is in accordance with the product specification of the Rezex ROA—Organic Acid H + column (Phenomenex). Our research shows that the decoction contained the greatest amount of polysaccharides. The exact identification of the polysaccharides and other saccharides present in the decoction requires further extensive research. At this point, however, it is important to consider whether they may have affected the activities under study. The term polysaccharides covers a large group of compounds that are composed of monosaccharides (glucose, mannose, galactose, fructose, etc.). Depending on the type of monosaccharide, the polysaccharides can be divided into a homo-polysaccharide containing one type of monosaccharide and a hetero-polysaccharide containing two or more different types of monosaccharides. Monosaccharides are linked by glycosidic bonds to form long linear or branched structures. The literature can confirm the influence of polysaccharides on wound healing. There are reviews available [20,21] that show the skin wound healing activity of polysaccharides, most often obtained by hot water extraction, among others from such plants as Trigonella foenum-graecum, Hammada scoparia, Linus usitatissimum, Avena sativa, Caesalpinia pulcherrima, Sanguisorba officinalis, Glycyrrhiza uralensis, Pimpinella anisum as well as Bletilla striata, Konjac, Eucommia ulmoides, and Mesona procumbens [22]. Most of these studies have not established the exact chemical structure of the polysaccharides. Fewer data are available on the activity of polysaccharides in gingival wound healing. However, one of the polysaccharides with known gingival wound healing properties is acemannan from Aloe vera. Acemannan, a β-[1,4]-acetylated soluble polymannose has gingival wound healing properties proven in vitro and in vivo as well as in clinical studies [23]. A study by Jettanacheawchankit et al. [ 24] showed that acemannan induces proliferation and upregulation of growth factors (KGF-1, VEGF), and type I collagen expression in gingival fibroblasts as good reduction of wound area of experimental animals at day 7 after treatment. Another polysaccharide with a proven wound healing effect is β -glucan (glucose polymers linked by 1,3; 1,4 or 1,6 β-glycosidic bonds), which is a constituent of grains, yeast, and other fungi. In vitro studies have shown that β-glucans induced the proliferation and migration of keratinocytes and fibroblasts through specific receptors such as Dectin-1, CR3 or TLR. Preclinical animal studies have confirmed that it is effective as a wound healing agent [25]. The biological activity of the studied Polygoni cuspidati rhizoma et radix decoction was higher than that demonstrated by the ethanol and acetone extracts in our previous study, even though the decoction contained significantly fewer polyphenols [7]. On the other hand, the content of carbohydrates was significantly higher in the decoction than in either organic solvent extract. According to our best contemporary knowledge, there are no reports in the literature on polysaccharide contents in rhizomes of R. japonica and their biological activity. Notwithstanding, the methods used in this study were insufficient to determine the identity of the carbohydrates. Further research is therefore required to confirm our assumptions that polysaccharides in the decoction are largely responsible for stimulating gingival fibroblast cells to proliferate, migrate, and increase collagen III synthesis. ## 3. Materials and Methods The present study was designed as an in vitro study on the human fibroblast cell line. The experimental protocol of a study “influence of extracts from medicinal plants Reynoutria on the functions of oral fibroblasts” was submitted to a Bioethics Commission of Wrocław Medical University (Wrocław, Poland) by a study principal investigator JH, and approved by the commission with the number KB-$\frac{134}{2020.}$ The gingival biopsies for the cell study were provided by the Dental Clinical and Teaching facility of Dental Surgery Department at Wrocław Medical University. The study was conducted in full compliance with the GCP ICH: E6 (R2) and Declaration of Helsinki. ## 3.1. Plant Material and Decoction Preparation Reynoutria japonica rhizomes harvested in Wroclaw (Poland) were used in the study. Details on the identification of the plant species, location, time of collection of raw material, and storage were presented in our previous study [7]. The same raw material was used in our previous study to prepare ethanol and acetone extracts [7]. Fifty grams of air-dried and powdered rhizomes of R. japonica were flooded with 500 mL of cold water and boiled under reflux for 20 min. After cooling down, the decoction was filtered through a paper filter and the solvent was evaporated under reduced pressure. An amount of 2.41 g of dry decoction was obtained. Stock solutions were prepared from the dried decoction by dissolving 100 mg of decoction in 1 mL of dimethyl sulfoxide (DMSO, Sigma-Aldrich, St. Louis, MO, USA). Different concentrations of the decoction (5–2000 µg/mL) were tested by taking the appropriate amount of the stock solution. As a solvent to prepare different concentrations of decoction, the cell culture medium was used. ## 3.2. Cell Culture The study was carried out on the primary human gingival fibroblasts (HGF) obtained from the connective tissue of a dental patient’s hard palate. The details of the cell culture study were presented in our previous paper [7]. ## 3.3. Cell Viability Assay Fibroblast cell viability was assessed using the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) colorimetric assay (Sigma-Aldrich, Merck Group, Darmstadt, Germany). The assay was carried out as described in our previous work [7], except that a decoction was used instead of ethanol or acetone extracts. The decoction was used in the concentration range from 5 µg/mL to 2000 µg/mL. The concentration of DMSO in the samples was equal to or less than $2\%$, therefore, also $2\%$ DMSO was used as a control in the study. For the experiment, the cells were seeded in 96-well microculture plates at 1 × 104 cells/well. After the incubation with selected concentrations of decoction, the experiments were conducted according to the manufacturer’s protocol. The absorbance was determined at 570 nm. ## 3.4. Confocal Laser Microscopy Study A confocal laser scanning microscope (CLSM, Olympus FluoView FV1000, Tokyo, Japan) was used to visualize histone 3 expression change after incubation with 50 µg/mL decoction. The study was performed exactly as described in our previous article [7], except that a decoction was used instead of acetone and ethanol extracts. ## 3.5. In Vitro Wound Healing Assay Decoction at 50 µg/mL concentration was used for the wound healing assay as a fibroblast migration capacity test. The study was performed exactly as described in our previous article [7], except that a decoction was used instead of acetone and ethanol extracts. ## 3.6. Immunocytochemical Staining Immunocytochemical staining was used to evaluate the expression of collagen type III after 24 h incubation with decoction at 50 µg/mL. Betulinic acid (2 µM) was used as a positive control. The study was performed exactly as described in our previous paper [7], except that a decoction was used instead of acetone and ethanol extracts. The intensity of immunohistochemical staining was evaluated as negative (−), weak (+), moderate (++), or strong (+++), as in the previous article [7,26]. ## 3.7. HPLC/DAD/ESI-HR-QTOF-MS Analysis The obtained decoction was prepared for the qualitative and quantitative analysis. An amount of 50 mg of dried decoction was dissolved in $80\%$ methanol (MeOH, Merck / MilliporeSigma, Darmstadt, Germany) in a volumetric flask to obtain a 5 mg/mL concentration. After filtering the prepared solutions through a 0.22 μm syringe membrane (Chromafil, Macherey-Nagel, Düren, Germany) into vials, 4 μL of the sample was injected by autosampler into the high-pressure liquid chromatography (HPLC) system. The same HPLC-DAD-MS system was used as well as the same qualitative analysis conditions as in our previous study [7]. A previously developed, validated analytical method was used to quantify the selected compounds [19]. Linearity, the LOD (limit of detection), and LOQ (limit of quantification) for all quantified compounds were presented in our previous study [19]. ## 3.8. Total Polyphenols and Tannins Content The content of polyphenols and tannins was determined using the modified Folin–Ciocalteu assay based on Singleton and Rossi method [27]. The study was conducted in the same way as described in our previous articles [28,29]. ## 3.9. HPLC-RI Analysis The composition of saccharides in the studied samples was determined by high-pressure liquid chromatography (HPLC) according to the procedure described in the previous paper [30]. Briefly, the chromatographic method was performed using the Dionex Ultimate 3000 system (Thermo Fisher Scientific, Waltham, MA, USA). The content of saccharides was determined using a Rezex ROA—Organic Acid H+ column with the following dimensions: 300 mm × 7.8 mm i.d. and $8\%$ cross-linked H + (Phenomenex, Torrance, CA, USA). The column was kept at 60 °C. The mobile phase was 5 mM sulfuric acid, previously filtered through a 0.45 mm membrane filter (Millipore). The flow rate was 0.6 mL/min. A refractometric detector ERC RefractoMax 520 (DataApex, Prague, Czech Republic) was used for analyte detection. All analytical determinations were performed in triplicate and mean values are given. An amount of 50 µL of decoction or extract (at 2.5 mg/mL concentration) dissolved in the mobile phase and filtered through a 0.22 µm Chromafil PTFE hydrophilic syringe membrane was injected into the HPLC system by an autosampler. Among others, the following compounds were analyzed as standards: glucose, xylose, maltose, maltotriose, and pectins (all from Merck/MilliporeSigma, Darmstadt, Germany). The standards were dissolved in the mobile phase in a volumetric flask to obtain a concentration of 1 mg/mL. ## 3.10. Statistical Analysis All assays were performed in at least triplicate and results are presented as the mean of the replicates ± SD. The Shapiro–Wilk test was used to evaluate the distribution of results. Two-way ANOVA and Tukey’s multiple comparisons tests (GraphPad Prism v. 9, San Diego, CA, USA) were used to evaluate significant differences between the obtained values. In studies where only the means between the control and treatments were compared, the t-test was used. ## 4. Conclusions The decoction of Reynoutria japonica rhizomes prepared according to a traditional recipe stimulated gingival fibroblasts to proliferate, migrate, and increase the synthesis of collagen III. Based on the phytochemical research (HPLC/DAD/ESI-HR-QTOF-MS, HPLC-RI analysis, and Folin–Ciocalteu assay), we can conclude that the high content of polysaccharides observed in the decoction may have an impact on their high wound-healing activity in vitro. Because this is the first published report on the presence of polysaccharides in this plant material, further research is needed to confirm our assumptions. Importantly, the decoction was found to be noncytotoxic to HGF cells at any tested concentrations and incubation time. The safety and high activity of the studied Polygoni cuspidati rhizoma et radix decoction encourage its future use in oral wound healing. ## References 1. 1.European Pharmacopoeia9th ed.European Director- ate for the Quality of MedicinesStrasburg, France201714811483. European Pharmacopoeia (2017) 1481-1483 2. Peng W., Qin R., Li X., Zhou H.. Botany, phytochemistry, pharmacology, and potential application of. J. Ethnopharmacol. (2013) 148 729-745. DOI: 10.1016/j.jep.2013.05.007 3. Nawrot-Hadzik I., Ślusarczyk S., Granica S., Hadzik J., Matkowski A.. Phytochemical Diversity in Rhizomes of Three Reynoutria Species and their Antioxidant Activity Correlations Elucidated by LC-ESI-MS/MS Analysis. Molecules (2019) 24. DOI: 10.3390/molecules24061136 4. Nawrot-Hadzik I., Zmudzinski M., Matkowski A., Preissner R., Kęsik-Brodacka M., Hadzik J., Drag M., Abel R.. Reynoutria Rhizomes as a Natural Source of SARS-CoV-2 Mpro Inhibitors–Molecular Docking and In Vitro Study. Pharmaceuticals (2021) 14. DOI: 10.3390/ph14080742 5. Nawrot-Hadzik I., Hadzik J., Fleischer M., Choromańska A., Sterczała B., Kubasiewicz-Ross P., Saczko J., Gałczyńska-Rusin M., Gedrange T., Matkowski A.. Chemical composition of east Asian invasive knotweeds, their cytotoxicity and antimicrobial efficacy against cariogenic pathogens: An in-vitro study. Med. Sci. Monit. (2019) 25 3279-3287. DOI: 10.12659/MSM.913855 6. Alperth F., Melinz L., Fladerer J.P., Bucar F.. Uhplc analysis of reynoutria japonica houtt. Rhizome preparations regarding stilbene and anthranoid composition and their antimycobacterial activity evaluation. Plants (2021) 10. DOI: 10.3390/plants10091809 7. Nawrot-Hadzik I., Matkowski A., Pitułaj A., Sterczała B., Olchowy C., Szewczyk A., Choromańska A.. In Vitro Gingival Wound Healing Activity of Extracts from Reynoutria japonica Houtt Rhizomes. Pharmaceutics (2021) 13. DOI: 10.3390/pharmaceutics13111764 8. Sheridan H., Krenn L., Jiang R., Sutherland I., Ignatova S., Marmann A., Liang X., Sendker J.. The potential of metabolic fingerprinting as a tool for the modernisation of TCM preparations. J. Ethnopharmacol. (2012) 140 482-491. DOI: 10.1016/j.jep.2012.01.050 9. Buranasukhon W., Athikomkulchai S., Tadtong S., Chittasupho C.. Wound healing activity of Pluchea indica leaf extract in oral mucosal cell line and oral spray formulation containing nanoparticles of the extract. Pharm. Biol. (2017) 55 1767-1774. DOI: 10.1080/13880209.2017.1326511 10. Field E.A., Allan R.B.. Review article: Oral ulceration--aetiopathogenesis, clinical diagnosis and management in the gastrointestinal clinic. Aliment. Pharmacol. Ther. (2003) 18 949-962. DOI: 10.1046/j.1365-2036.2003.01782.x 11. Caton J.G., Armitage G., Berglundh T., Chapple I.L.C., Jepsen S., Kornman K.S., Mealey B.L., Papapanou P.N., Sanz M., Tonetti M.S.. A new classification scheme for periodontal and peri-implant diseases and conditions—Introduction and key changes from the 1999 classification. J. Clin. Periodontol. (2018) 45 S1-S8. DOI: 10.1111/jcpe.12935 12. Buranasin P., Mizutani K., Iwasaki K., Mahasarakham C.P.N., Kido D., Takeda K., Izumi Y.. High glucose-induced oxidative stress impairs proliferation and migration of human gingival fibroblasts. PLoS ONE (2018) 13. DOI: 10.1371/journal.pone.0201855 13. Nawrot-Hadzik I., Matkowski A., Kubasiewicz-Ross P., Hadzik J.. Proanthocyanidins and Flavan-3-ols in the Prevention and Treatment of Periodontitis—Immunomodulatory Effects, Animal and Clinical Studies. Nutrients (2021) 13. DOI: 10.3390/nu13010239 14. Cáceres M., Oyarzun A., Smith P.C.. Defective wound-healing in aging gingival tissue. J. Dent. Res. (2014) 93 691-697. DOI: 10.1177/0022034514533126 15. Toma A.I., Fuller J.M., Willett N.J., Goudy S.L.. Oral wound healing models and emerging regenerative therapies. Transl. Res. (2021) 236 17-34. DOI: 10.1016/j.trsl.2021.06.003 16. Kubasiewicz-Ross P., Fleischer M., Pitułaj A., Hadzik J., Nawrot-Hadzik I., Bortkiewicz O., Dominiak M., Jurczyszyn K.. Evaluation of the three methods of bacterial decontamination on implants with three different surfaces. Adv. Clin. Exp. Med. (2020) 29 177-182. DOI: 10.17219/acem/112606 17. Clément C., Orsi G.A., Gatto A., Boyarchuk E., Forest A., Hajj B., Miné-Hattab J., Garnier M., Gurard-Levin Z.A., Quivy J.P.. High-resolution visualization of H3 variants during replication reveals their controlled recycling. Nat. Commun. (2018) 9 3181. DOI: 10.1038/s41467-018-05697-1 18. Komar D., Juszczynski P.. Rebelled epigenome: Histone H3S10 phosphorylation and H3S10 kinases in cancer biology and therapy. Clin. Epigenetics (2020) 12 147. DOI: 10.1186/s13148-020-00941-2 19. Nawrot-Hadzik I., Granica S., Domaradzki K., Pecio Ł., Matkowski A.. Isolation and Determination of Phenolic Glycosides and Anthraquinones from Rhizomes of Various Reynoutria Species. Planta Med. (2018) 84 1118-1126. DOI: 10.1055/a-0605-3857 20. Albuquerque P.B.S., de Oliveira W.F., dos Santos Silva P.M., dos Santos Correia M.T., Kennedy J.F., Coelho L.C.B.B.. Skincare application of medicinal plant polysaccharides—A review. Carbohydr. Polym. (2022) 277 118824. DOI: 10.1016/j.carbpol.2021.118824 21. Li Q., Niu Y., Xing P., Wang C.. Bioactive polysaccharides from natural resources including Chinese medicinal herbs on tissue repair. Chin. Med. (2018) 13 7. DOI: 10.1186/s13020-018-0166-0 22. Fan S.L., Lin J.A., Chen S.Y., Lin J.H., Lin H.T., Chen Y.Y., Yen G.C.. Effects of Hsian-tsao (: Mesona procumbens Hemsl.) extracts and its polysaccharides on the promotion of wound healing under diabetes-like conditions. Food Funct. (2021) 12 119-132. DOI: 10.1039/D0FO02180F 23. Liu C., Cui Y., Pi F., Cheng Y., Guo Y., Qian H.. Extraction, purification, structural characteristics, biological activities and pharmacological applications of acemannan, a polysaccharide from aloe vera: A review. Molecules (2019) 24. DOI: 10.3390/molecules24081554 24. Jettanacheawchankit S., Sasithanasate S., Sangvanich P., Banlunara W., Thunyakitpisal P.. Acemannan stimulates gingival fibroblast proliferation; expressions of keratinocyte growth factor-1, vascular endothelial growth factor, and type I collagen; and wound healing. J. Pharmacol. Sci. (2009) 109 525-531. DOI: 10.1254/jphs.08204FP 25. Majtan J., Jesenak M.. β-Glucans: Multi-functional modulator of wound healing. Molecules (2018) 23. DOI: 10.3390/molecules23040806 26. Sterczala B., Kulbacka J., Saczko J., Dominiak M.. The effect of dental gel formulation on human primary fibroblasts—An in vitro study. Folia Histochem. Cytobiol. (2020) 58 156-161. DOI: 10.5603/FHC.a2020.0013 27. Singleton V.L., Rossi J.A.. Colorimetry of Total Phenolics with Phosphomolybdic-Phosphotungstic Acid Reagents. Am. J. Enol. Vitic. (1965) 16 144-158 28. Cieslak A., Zmora P., Matkowski A., Nawrot-Hadzik I., Pers-Kamczyc E., El-Sherbiny M., Bryszak M., Szumacher-Strabel M.. Tannins from sanguisorba officinalis affect in vitro rumen methane production and fermentation. J. Anim. Plant Sci. (2016) 26 54-62 29. Nawrot-Hadzik I., Granica S., Abel R., Czapor-Irzabek H., Matkowski A.. Analysis of antioxidant polyphenols in loquat leaves using HPLC-based activity profiling. Nat. Prod. Commun. (2017) 12 163-166. DOI: 10.1177/1934578X1701200205 30. Pińkowska H., Krzywonos M., Wolak P., Złocińska A.. Pectin and Neutral Monosaccharides Production during the Simultaneous Hydrothermal Extraction of Waste Biomass from Refining of Sugar—Optimization with the Use of Doehlert Design. Molecules (2019) 24. DOI: 10.3390/molecules24030472
--- title: Inflammation as Prognostic Hallmark of Clinical Outcome in Patients with SARS-CoV-2 Infection authors: - Diana Fuzio - Angelo Michele Inchingolo - Vitalba Ruggieri - Massimo Fasano - Maria Federico - Manuela Mandorino - Lavinia Dirienzo - Salvatore Scacco - Alessandro Rizzello - Maurizio Delvecchio - Massimiliano Parise - Roberto Rana - Nicola Faccilongo - Biagio Rapone - Francesco Inchingolo - Antonio Mancini - Maria Celeste Fatone - Antonio Gnoni - Gianna Dipalma - Giovanni Dirienzo journal: Life year: 2023 pmcid: PMC9966655 doi: 10.3390/life13020322 license: CC BY 4.0 --- # Inflammation as Prognostic Hallmark of Clinical Outcome in Patients with SARS-CoV-2 Infection ## Abstract Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is often characterized by a life-threatening interstitial pneumonia requiring hospitalization. The aim of this retrospective cohort study is to identify hallmarks of in-hospital mortality in patients affected by Coronavirus Disease 19 (COVID-19). A total of 150 patients admitted for COVID-19 from March to June 2021 to “F. Perinei” Murgia Hospital in Altamura, Italy, were divided into survivors ($$n = 100$$) and non-survivors groups ($$n = 50$$). Blood counts, inflammation-related biomarkers and lymphocyte subsets were analyzed into two groups in the first 24 h after admission and compared by Student’s t-test. A multivariable logistic analysis was performed to identify independent risk factors associated with in-hospital mortality. Total lymphocyte count and CD3+ and CD4+ CD8+ T lymphocyte subsets were significantly lower in non-survivors. Serum levels of interleukin-6 (IL-6), lactate dehydrogenase (LDH), C-reactive protein (CRP) and procalcitonin (PCT) were significantly higher in non-survivors. Age > 65 years and presence of comorbidities were identified as independent risk factors associated with in-hospital mortality, while IL-6 and LDH showed a borderline significance. According to our results, markers of inflammation and lymphocytopenia predict in-hospital mortality in COVID-19. ## 1. Introduction From March 2019 to the present, the pandemic caused by a novel betacoronavirus, named severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has quickly spread from the city of Wuhan (China) to the entire world, with a variable trend characterized by flattening periods of the pandemic curve and periods of re-emergency [1]. Up-to-date information from the World Health Organization (WHO) reports 663,248,631 confirmed cases of Coronavirus Disease 19 (COVID-19) and 6,709,387 deaths (https://covid19.who.int/, accessed on 19 January 2023). The clinical course of COVID-19 is characterized by markedly divergent clinical manifestations. Aside from asymptomatic or pauci-symptomatic cases, nearly $20\%$ of patients had severe bilateral interstitial pneumonia, which was associated with a rapid deterioration of clinical condition and necessitated hospitalization [2]. In COVID-19, multiple organ dysfunction syndrome (MODS) is triggered by an immune-mediated systemic inflammation, which often leads to death [3]. This pro-inflammatory state is reflected in specific laboratory parameters, such as blood count and markers of systemic inflammation. In severe COVID-19, some altered parameters are common to septic states, such as C-reactive protein (CRP), procalcitonin (PCT), ferritin, D-dimers and fibrinogen, while others indicate, more specifically, a hyper-activation of the immune system and are often present in systemic autoimmune diseases, such as pro-inflammatory cytokine levels, e.g., interleukin-6 (IL-6) and interleukin-10 (IL-10), and reduction of both total lymphocytes and specific lymphocyte subsets [4,5,6]. The understanding of the clinical and biochemical effects of COVID-19 is constantly expanding. Several strategies for dealing with this global emergency have been developed, ranging from preventive measures, from individual protection devices, isolation of positive cases and vaccines, to adjuvant therapies and specific antiviral treatments [7,8,9]. Due to the extreme variety of the spectrum of clinical manifestations of COVID-19, it is advisable to tailor preventive and therapeutic choices according to the characteristics of the patients. For instance, it is worth noting that elder patients and patients with pre-existing pathological conditions and comorbidities are at a higher risk of developing a severe outcome of COVID-19 [10]. Thus, it is crucial for physicians to distinguish, as soon as possible, the patients who will suffer from severe illness and are at risk of death, especially in the hospital setting during pandemic waves. The aim of this retrospective study is to detail the clinical and laboratoristic features of hospitalized COVID-19 patients in a center of Southern Italy, identifying significant differences between the groups of survivors and non-survivors, and independent factors of death, in an attempt to define threshold values indicative of patient outcome. ## 2.1. Study Design and Participants A total of 150 patients hospitalized from March to June 2021 in the public hospital “F. Perinei” Murgia Hospital in Altamura, Italy, were enrolled; patients were divided into survivors ($$n = 100$$, group 1) and non-survivors ($$n = 50$$, group 2) according to the clinical outcome. This study was approved by the Ethics Committee of “Azienda Ospedaliero Universitaria Consorziale Policlinico”, Bari, Italy (0015987, 17 February 2022), and it was conducted in accordance with the Declaration of Helsinki for human studies. ## 2.2. COVID-19 RT-PCR Assay for Nasal and Pharyngeal Swab Specimens Real-time reverse-transcriptase polymerase-chain reaction (RT-PCR) assay for nasal and pharyngeal swab specimens was employed to confirm COVID-19. Briefly, RNA extraction was performed using a NeoPlex COVID-19 Detection kit (GeneMatrix Inc., Temecula, CA, USA) on a KingFisher Extraction System (ThermoFisher Scientific, Waltham, MA, USA), according to the manufacturer’s instructions. Amplification conditions included reverse transcription at 50 °C for 30 min, denaturation at 95 °C for 15 min and 40 cycles of 95 °C for 15 s and 60 °C for 60 s for fluoresce detection. A cycle threshold value (Ct-value) ≤ 38 was defined as a positive test, following Centers of Disease Control and Prevention (CDC) recommendations. ## 2.3. Laboratory Medicine Analyses and Clinical Data Collection Laboratory medicine analyses (e.g., blood routine, lymphocyte subsets and inflammation-related biomarkers) were performed on patients’ blood samples collected and analyzed in the first 24 h after admission to the Infective Diseases Department of “F. Perinei” Murgia Hospital. The total number of lymphocytes in peripheral blood was counted with an automated hematology analyzer (Pentra ABX, HORIBA, Kyoto, Japan). An Olympus AU680 (Beckman Coulter Brea, CA, USA) was used to collect LDH and CRP data. PCT was determined by Liaison (DiaSorin S.p. A., Saluggia, Italy). IL-6 and ferritin were measured by ADVIA Centaur XP Immunoassay System (SIEMENS Health, Erlangen, Germany, GmbH). D-Dimers and Fibrinogen values were measured by ACL TOP 500 (Werfen, Bedford, MA USA). Peripheral blood lymphocyte typing was performed in patients’ whole blood samples by cytofluorimetric analysis using AQUIOS CL Flow Cytometer (AQUIOS—Beckman Coulter CA, USA). Antibodies used for cell staining were TETRA-1 Panel (CD45, CD4, CD8 and CD3), and obtained data were analyzed using flow cytometry analysis software (Aquios system software, V2.2.0). ## 2.4. Statistical Analysis Statistical analysis was performed by setting categorical variables as frequency rates and percentages, and continuous variables as means and $95\%$ confidence intervals ($95\%$ CI) or median and interquartile range (IQR) values. The comparison of means for continuous variables that were normally distributed was performed with Student’s t-test. The Mann–Whitney U test was used for continuous variables there were not normally distributed. Proportions for categorical variables were compared using the χ2 test, while receiver operating characteristic (ROC) curve analysis was performed using the Wilson/Brown method ($95\%$ confidence interval and standard error). Multinomial binary logistic regression analysis results were reported as Odds Ratios (OR) with $95\%$ CI. Statistical analyses were performed by MedCalc® (Mariakerke, Belgium) and GraphPad Prism version 8.2 (GraphPad Software Inc., San Diego, CA, USA). Two-sided p-values lower than 0.05 were considered statistically significant. ## 3. Results The median age of the patients was 70 years, showing a statistically significant difference between non-survivors and survivors (79 vs. 65 years, $p \leq 0.001$) (Table 1). Male patients were significantly older than females (79 vs. 71 years), while the median period from the onset of the symptoms to hospital admission was the same in females and males (7 days). One hundred and twenty-six patients ($84\%$) had comorbidities such as hypertension ($55.3\%$), diabetes mellitus ($24\%$), chronic cardiac disease ($22.6\%$), malignancies ($4.6\%$), obesity ($35.3\%$), chronic pulmonary disease ($8\%$), chronic kidney disease ($7.3\%$) and chronic neurological disorders ($16\%$) (Table 2). All the patients with severe and moderate disease were given empirical antimicrobial treatment (cephalosporin, azithromycin and levofloxacin). Nineteen patients ($12\%$) received antiviral therapy with remdesivir. In addition, all severe and moderate cases were administered corticosteroids (CTS) during hospitalization. Nine patients ($6\%$) received hyperimmune plasma, and thirty-four patients ($22\%$) required admission to the intensive care unit (ICU) (Table 3). At baseline, 102 patients ($68\%$) needed respiratory support by continuous positive airway pressure, 25 patients ($16.7\%$) by a Venturi-type mask, 18 patients ($12\%$) by a simple face mask. Compared with the reference range, significant differences in blood count, lymphocyte subsets and inflammatory-related biomarkers were observed between survivor and non-survivor groups (Table 4). The median lymphocyte count was lower in the non-survivors group ($p \leq 0.0001$). When we tested different subsets of T cells, we found that even though both helper T cells (CD3+CD4+) and suppressor T cells (CD3+CD8+) in patients with COVID-19 were below reference range (CD3+CD4+: 500–1700 cells/µL, CD3+CD8+: 244–1100 cells/µL), the lowering of helper T cells was considerably pronounced in fatal cases (184 vs. 353 cells/uL; $p \leq 0.0001$). Suppressor T cells also showed a decreasing trend (83 vs. 172 cells/uL; $p \leq 0.0001$). Conversely, T-helper and T-suppressor ratio (CD4+/CD8+ ratio) remained in the normal range and showed no difference between the two subgroups. As shown in Figure 1, the area under the curve (AUC) derived from CD8+ T cells was as large as that derived from CD3+ cells or CD4+ cells (AUC CD8+ = 0.741 [0.655–0.827] vs. AUC CD3+ = 0.769 [0.690–0.848] or AUC CD4+ = 0.752 [0.670–0.833], $p \leq 0.001$). Non-survivors had significantly higher serum levels of IL-6, LDH, CRP and PCT than survivors (Table 4). Conversely, except for fibrinogen, no significant differences were found in the levels of D-dimers and ferritin between the two groups. Figure 2 shows that the AUC of IL-6 was 0.735 [0.651–0.818] and LDH was 0.784 [0.703–0.864] ($p \leq 0.001$), and age and comorbidities had AUCs of 0.805 [0.736–0.873] and 0.709 [0.622–0.800], respectively. By multivariable logistic regression analysis, two indicators were identified to be independent risk factors associated with in-hospital mortality: age > 65 years (OR = 1.14; $95\%$ CI, 1.07–1.22, $$p \leq 0.0001$$) and number of comorbidities (OR = 1.84; $95\%$ CI, 1.11–3.05; $$p \leq 0.0178$$). IL-6 levels > 20 pg/mL (OR = 1.03; $95\%$ CI, 1.00–1.06) and LDH levels > 489 U/L (OR = 1.01; $95\%$ CI, 1.00–1.01) showed a borderline $95\%$ CI (Table 5). ## 4. Discussion Infection by SARS-CoV-2 can cause sustained responses of pro-inflammatory cytokines and chemokines (namely, a “cytokine storm”), leading to a life-threatening immune-mediated MODS [3]. The identification of specific immunological and inflammatory profiles of patients, and their association with COVID-19 severity, is a challenge in order to promptly block systemic inflammation with targeted therapeutic interventions and, on the other hand, minimize unnecessary treatments, especially during pandemic waves. In this study, we investigated the predictive values of markers of inflammation and lymphocytopenia in hospitalized severe COVID-19 patients, already assessed in other previous studies, with, in part, conflicting results. Our research showed that non-surviving hospitalized COVID-19 patients had significantly higher PCT, CRP, LDH and IL-6 levels, which are indicators of both in-hospital mortality and inflammation. Age and comorbidities—particularly hypertension, obesity, diabetes and chronic heart disease—were found to be independent risk factors linked to in-hospital mortality, while LDH and IL-6 showed a borderline significance. PCT, the precursor of the hormone calcitonin, is an acute-phase glycoprotein produced by C-cells of thyroids and monocytes. PCT dramatically increases during bacterial and fungal infections while slightly increasing during viral infections, making it an important biomarker of sepsis. In our series of hospitalized COVID-19 patients, PCT levels correlated with disease severity. Data on the value of PCT as prognostic markers for COVID-19 are contradictory in the literature. When compared to moderate illness, PCT is four times higher in severe patients and eight times higher in critical patients, according to Hu R. et al. [ 11]. Likewise, several authors found that PCT levels are increased in patients with a fatal outcome of severe COVID-19 both at admission and during the course of hospitalization [10,12,13]. In addition, Sayah W. et al. assessed that PCT and the neutrophil-to-lymphocyte ratio are not influenced by the administration of CTS. For this reasons, they may constitute valid alternative markers to assess severe forms in patients already undergoing CTS [14]. The markedly increased levels of PCT in COVID-19 patients can be explained by several factors: a coexistent bacterial infection, prolonged invasive mechanical ventilation and the up-regulation of the signal transducer and activator of the transcription 3 (STAT3)-dependent pathway, which stimulates angiotensin-converting enzyme 2 (ACE2) and PCT production in monocytes [15,16,17]. Accordingly, other clinical trials refuted the negative prognostic value of PCT in severe COVID-19 [18,19]. CRP is a non-specific acute-phase glycoprotein produced by the liver in response to trauma, myocardial ischemia and infections. Bacterial infections usually determine a marked increase of CRP, while viral infections are associated with a mild increase in CRP levels. In our study, CRP was significantly higher in non-survivor COVID-19 patients compared with survivor groups. CRP proved to be one of the earliest negative prognostic markers in COVID-19 because its levels increased before the appearance of radiologic findings at chest computer tomography (CT) [20]. Furthermore, CRP levels increased both at the beginning and during the progression of COVID-19 disease, and correlated with severity and mortality [20,21]. Several authors tried to improve the predictive value of CRP by relating it to other parameters, such as the CRP/albumin ratio and the CRP/lymphocytes ratio [22,23]. In particular, CRP is associated with in-hospital mortality due to venous thromboembolism and acute kidney injury [24], and a value of CRP equal to or higher than 40 mg/L is considered life-threatening in COVID-19-hospitalized patients [25]. LDH is an enzyme of the oxidoreductase class produced by different cells that is released into the blood as a result of cell damage or high turnover, such as in cancer, trauma, inflammation or infection. Other authors have assessed that LDH is correlated with poor prognosis in hospitalized COVID-19 patients, also indicating lung and other tissue injuries [26,27]. Thus, COVID-19 may lead to inadequate tissue perfusion and MODS, causing LDH elevation [28]. Thus, high values of LDH could represent a valid biomarker of mortality due to widespread infection in COVID-19. In our study, we also showed that lymphocyte counts in COVID-19 patients and the CD3+ and CD4+CD8+ subsets were significantly lower, especially in the non-survivor group. Lymphocytopenia indicates a dysregulation of the immune system and has been observed in COVID-19 patients showing different spectrums of clinical disease [5,24]. Consisting with the literature data, we found a significant decrease of total peripheral lymphocyte counts and T cell main subsets (CD3+ and CD4+CD8+) in both the survivor and non-survivor groups, even if these parameters were significantly lower in the latter group. The etiopathogenesis of lymphocytopenia in COVID-19 patients has been related to different causes. ACE2, identified as the main cell entry receptor for SARS-CoV-2, is low-expressed by lymphocytes, and the viral genome is rarely detectable in peripheral blood of infected patients [29]. Thus, it is reasonable to speculate that the decrease of peripheral lymphocytes is not ascribable to the direct damage of SARS-CoV-2 on lymphocytes, but rather to an exhaustion of them in terms both of number and function due to the persistent exposure to viral antigens [30,31]. An alternative explanation is that the decrease of peripheral lymphocytes is a result of activation-induced apoptosis or aggressive migration from peripheral blood to the lungs, where robust viral replication occurs [32]. Lymphocytopenia in COVID-19 may be related to hyper-activation of STING (stimulator of interferon genes) due to DNA damage following acute distress respiratory syndrome (ARDS). STING is able to activate the NF–κB pathway and also to determine a progressive CD4+CD8+ T lymphocytopenia, similar to what occurs in STING-associated vasculopathy with onset in infancy (SAVI) syndromes [33]. Several studies detected a significant reduction in total lymphocytes count and in CD3+ and CD4+CD8+ T-cell subsets, both at the early stages and in severe forms of COVID-19-associated disease in deceased hospitalized patients compared to survivors [34,35,36,37]. Interestingly, a Brazilian study found that reduction of T-cell subtypes is a prognostic factor not only of death but also of need for intubation [38]. In addition to lymphocyte subsets, a high neutrophil-to-lymphocyte ratio has been considered a sensible parameter of negative outcome in COVID-19 [38,39]. The reduction of lymphocytes in COVID-19 also affects B cells and NK (natural killer) cells, as suggested by other studies, highlighting a significant reduction of CD19+ and NK cell count in hospitalized patients, which correlated with progression of disease and death [34,40]. Cytokines have been thought to play an important role in immunity and immunopathology during virus infections. “ Cytokines storm” is a phenomenon of excessive inflammatory reaction in which cytokines are rapidly produced in large amounts in response to microbial infection, as well as to therapies, pathogens, cancers, autoimmune conditions and monogenic disorders [41]. This phenomenon has been considered an important contributor to ARDS and MODS in COVID-19 patients [3]. It has been reported that the levels of IL-6, interleukin-2 (IL-2), interleukin-7 (IL-7), IL-10, tumor necrosis factor-alpha (TNF-α), granulocyte colony-stimulating factor (G-CSF), interferon gamma induced protein (IP-10), monocyte chemoattractant protein-1 (MCP-1) and macrophage inflammatory protein-1 alpha (MIP-1α) were significantly higher in COVID-19 patients [42,43,44,45,46]. IL-6 is a protein produced by various types of cells, particularly T lymphocytes, macrophages, mature adipocytes and myocytes, in response to tissue damage from trauma, infection or inflammation, exerting both pro-inflammatory and anti-inflammatory effects [47]. In several studies, IL-6 has been proven to be an independent factor predictive of in-hospital mortality, and its levels are not influenced by the administration of CTS [14,42,43]. In particular, high levels of IL-6 are indicative of lung involvement, acute kidney injury, brain damage, cardiovascular events, and, finally, intestinal permeability, which allows viruses to become widespread in the general circulation [48,49,50,51,52]. The value of IL-6 has been included in a score system that predicts the need for non-invasive ventilation by combining systemic inflammatory biomarkers and a chest CT severity score [53]. Besides IL-6 and LDH, other laboratory parameters are cited in the literature as independent risk factors of death in COVID-19, such as neutrophil count, platelet count, CRP, D-dimers, troponin-I and low total testosterone [54,55,56,57]. We found that age and comorbidities—primarily hypertension and obesity, followed by diabetes, chronic cardiac disease, chronic neurological disorders, chronic pulmonary disease, chronic kidney disease and malignancies—are independent risk factors of COVID-19 mortality [25,50]. Consistently, other studies reported that patients older than 65 suffering from comorbidities experienced more severe symptoms, MODS and death [27,58]. Hypertension and obesity are the most frequent comorbidities in patients with severe or fatal COVID-19, and the main cause of death after ARDS is a cardiovascular acute event, such as myocardial dysfunction, arrhythmia or shock [10,12,13,26,59,60]. Instead, according to other authors, COVID-19-deceased patients presented, at admission, more frequently with chronic kidney disease and neurological diseases [13]. Obesity, diabetes and chronic kidney disease are also independent risk factors for intubation, as well as death [61]. Other pre-existing pathological conditions, such as anemia, hypotension, dyslipidemia, hyperglycemia and use of CTS, have been reported as independent risk factors of severe COVID-19 [26,53,62,63]. ## 5. Conclusions SARS-CoV-2 induces serious infectious diseases and becomes a continuous threat to human health. A rapid and well-coordinated immune response is the first line of defense against viral infections. However, when the immune response is dysregulated, it will result in excessive inflammation, even causing death. The higher expression of proinflammatory cytokines in COVID-19 patients, especially in severe cases, with the consumption of CD4+CD8+ T cells might result in aggravated inflammatory responses, the production of cytokine storms and clinical conditions worsening. Our findings demonstrated that lymphocyte counts and CD3+ and CD4+CD8+ subsets were significantly lower in COVID-19 patients, particularly in the non-survivor group. IL-6, LDH, CRP and PCT levels were significantly higher in non-survivor hospitalized patients affected by COVID-19, representing not only markers of inflammation but also in-hospital mortality. Age and comorbidities, especially hypertension, obesity, diabetes and chronic heart disease, were identified as independent risk factors associated with in-hospital mortality. Elevated LDH and IL-6 levels may be substantial risk factors for in-hospital mortality even though they did not achieve significance as independent risk variables. Even though our findings resulted from a monocentric study and the time frame for patient enrollment was limited, they highlighted the pivotal role of IL-6, LDH, CRP and PCT in predicting mortality for hospitalized COVID-19 patients. Nevertheless, the results of this study should be confirmed in further larger controlled trials, while also developing scoring systems to assess the severity of COVID-19 disease, correlating every stage to a specific risk of mortality. On the other hand, it appears critical to establish prognostic survival factors for COVID-19 patients in order to screen the most vulnerable patients, organize targeted therapeutic interventions and reduce healthcare costs and waste. ## References 1. Khan M., Adil S.F., Alkhathlan H.Z., Tahir M.N., Saif S., Khan M., Khan S.T.. COVID-19: A Global Challenge with Old History, Epidemiology and Progress So Far. Molecules (2020) 26. DOI: 10.3390/molecules26010039 2. Jiang S., Du L., Shi Z.. An emerging coronavirus causing pneumonia outbreak in Wuhan, China: Calling for developing therapeutic and prophylactic strategies. Emerg. Microbes Infect. (2020) 9 275-277. DOI: 10.1080/22221751.2020.1723441 3. Montazersaheb S., Hosseiniyan Khatibi S.M., Hejazi M.S., Tarhriz V., Farjami A., Ghasemian Sorbeni F., Farahzadi R., Ghasemnejad T.. COVID-19 infection: An overview on cytokine storm and related interventions. Virol. J. (2022) 19 92. DOI: 10.1186/s12985-022-01814-1 4. Ji P., Zhu J., Zhong Z., Li H., Pang J., Li B., Zhang J.. Association of elevated inflammatory markers and severe COVID-19: A meta-analysis. Medicine (2020) 99 e23315. DOI: 10.1097/MD.0000000000023315 5. Tavakolpour S., Rakhshandehroo T., Wei E.X., Rashidian M.. Lymphopenia during the COVID-19 infection: What it shows and what can be learned. Immunol. Lett. (2020) 225 31-32. DOI: 10.1016/j.imlet.2020.06.013 6. Fatone M.C.. COVID-19: A Great Mime or a Trigger Event of Autoimmune Manifestations?. Curr. Rheumatol. Rev. (2021) 17 7-16. PMID: 33019935 7. Meng J., Li R., Zhang Z., Wang J., Huang Q., Nie D., Fan K., Guo W., Zhao Z., Han Z.. A Review of Potential Therapeutic Strategies for COVID-19. Viruses (2022) 14. DOI: 10.3390/v14112346 8. Inchingolo A.D., Inchingolo A.M., Bordea I.R., Malcangi G., Xhajanka E., Scarano A., Lorusso F., Farronato M., Tartaglia G.M., Isacco C.G.. SARS-CoV-2 Disease Adjuvant Therapies and Supplements Breakthrough for the Infection Prevention. Microorganisms (2021) 9. DOI: 10.3390/microorganisms9030525 9. Inchingolo A.D., Dipalma G., Inchingolo A.M., Malcangi G., Santacroce L., D’Oria M.T., Isacco C.G., Bordea I.R., Candrea S., Scarano A.. The 15-Months Clinical Experience of SARS-CoV-2: A Literature Review of Therapies and Adjuvants. Antioxidants (2021) 10. DOI: 10.3390/antiox10060881 10. Pan F., Yang L., Li Y., Liang B., Li L., Ye T., Li L., Liu D., Gui S., Hu Y.. Factors associated with death outcome in patients with severe coronavirus disease-19 (COVID-19): A case-control study. Int. J. Med. Sci. (2020) 17 1281-1292. DOI: 10.7150/ijms.46614 11. Hu R., Han C., Pei S., Yin M., Chen X.. Procalcitonin levels in COVID-19 patients. Int. J. Antimicrob. Agents (2020) 56 106051. DOI: 10.1016/j.ijantimicag.2020.106051 12. Todor S.B., Bîrluțiu V., Topîrcean D., Mihăilă R.G.. Role of biological markers and CT severity score in predicting mortality in patients with COVID-19: An observational retrospective study. Exp. Ther. Med. (2022) 24 698. DOI: 10.3892/etm.2022.11634 13. Poskurica M., Stevanović Đ., Zdravković V., Čekerevac I., Ćupurdija V., Zdravković N., Nikolić T., Marković M., Jovanović M., Popović M.. Admission Predictors of Mortality in Hospitalized COVID-19 Patients-A Serbian Cohort Study. J. Clin. Med. (2022) 11. DOI: 10.3390/jcm11206109 14. Sayah W., Berkane I., Guermache I., Sabri M., Lakhal F.Z., Yasmine Rahali S., Djidjeli A., Lamara Mahammed L., Merah F., Belaid B.. Interleukin-6, procalcitonin and neutrophil-to-lymphocyte ratio: Potential immune-inflammatory parameters to identify severe and fatal forms of COVID-19. Cytokine (2021) 141 155428. DOI: 10.1016/j.cyto.2021.155428 15. Lugito N.P.H.. Is procalcitonin a part of human immunological response to SARS-CoV-2 infection or “just” a marker of bacterial coinfection?. Curr. Res. Transl. Med. (2021) 69 103289. DOI: 10.1016/j.retram.2021.103289 16. Pink I., Raupach D., Fuge J., Vonberg R.P., Hoeper M.M., Welte T., Rademacher J.. C-reactive protein and procalcitonin for antimicrobial stewardship in COVID-19. Infection (2021) 49 935-943. DOI: 10.1007/s15010-021-01615-8 17. Heer R.S., Mandal A.K., Kho J., Szawarski P., Csabi P., Grenshaw D., Walker I.A., Missouris C.G.. Elevated procalcitonin concentrations in severe Covid-19 may not reflect bacterial co-infection. Ann. Clin. Biochem. (2021) 58 520-527. DOI: 10.1177/00045632211022380 18. Dolci A., Robbiano C., Aloisio E., Chibireva M., Serafini L., Falvella F.S., Pasqualetti S., Panteghini M.. Searching for a role of procalcitonin determination in COVID-19: A study on a selected cohort of hospitalized patients. Clin. Chem. Lab. Med. (2020) 59 433-440. DOI: 10.1515/cclm-2020-1361 19. Liu F., Li L., Xu M., Wu J., Luo D., Zhu Y., Li B., Song X., Zhou X.. Prognostic value of interleukin-6, C-reactive protein, and procalcitonin in patients with COVID-19. J. Clin. Virol. (2020) 127 104370. DOI: 10.1016/j.jcv.2020.104370 20. Tan C., Huang Y., Shi F., Tan K., Ma Q., Chen Y., Jiang X., Li X.. C-reactive protein correlates with computed tomographic findings and predicts severe COVID-19 early. J. Med. Virol. (2020) 92 856-862. DOI: 10.1002/jmv.25871 21. Mousavi-Nasab S.D., Mardani R., Nasr Azadani H., Zali F., Ahmadi Vasmehjani A., Sabeti S., Alavi Darazam I., Ahmadi N.. Neutrophil to lymphocyte ratio and C-reactive protein level as prognostic markers in mild versus severe COVID-19 patients. Gastroenterol. Hepatol. Bed. Bench. (2020) 13 361-366. PMID: 33244379 22. Yang M., Chen X., Xu Y.. A Retrospective Study of the C-Reactive Protein to Lymphocyte Ratio and Disease Severity in 108 Patients with Early COVID-19 Pneumonia from January to March 2020 in Wuhan, China. Med. Sci. Monit. (2020) 26 e926393. DOI: 10.12659/MSM.926393 23. Li Y., Li H., Song C., Lu R., Zhao Y., Lin F., Han D., Chen L., Pan P., Dai M.. Early Prediction of Disease Progression in Patients with Severe COVID-19 Using C-Reactive Protein to Albumin Ratio. Dis. Markers (2021) 2021 6304189. DOI: 10.1155/2021/6304189 24. Smilowitz N.R., Kunichoff D., Garshick M., Shah B., Pillinger M., Hochman J.S., Berger J.S.. C-reactive protein and clinical outcomes in patients with COVID-19. Eur. Heart J. (2021) 42 2270-2279. DOI: 10.1093/eurheartj/ehaa1103 25. Stringer D., Braude P., Myint P.K., Evans L., Collins J.T., Verduri A., Quinn T.J., Vilches-Moraga A., Stechman M.J., Pearce L.. The role of C-reactive protein as a prognostic marker in COVID-19. Int. J. Epidemiol. (2021) 50 420-429. DOI: 10.1093/ije/dyab012 26. Turan O., Mirici A., Duru Akçalı S., Turan P.A., Batum Ö., Şengül A., Ekici Ünsal Z., Işık Kabakoğlu N., Ogan N., Torun Ş.. Characteristics of hospitalised COVID-19 patients and parameters associated with severe pneumonia. Int. J. Clin. Pract. (2021) 75 e14786. DOI: 10.1111/ijcp.14786 27. Guan X., Zhang B., Fu M., Li M., Yuan X., Zhu Y., Peng J., Guo H., Lu Y.. Clinical and inflammatory features based machine learning model for fatal risk prediction of hospitalized COVID-19 patients: Results from a retrospective cohort study. Ann. Med. (2021) 53 257-266. DOI: 10.1080/07853890.2020.1868564 28. Vincent J.L., Taccone F.S.. Understanding pathways to death in patients with COVID-19. Lancet Respir. Med. (2020) 8 430-432. DOI: 10.1016/S2213-2600(20)30165-X 29. Inchingolo A.D., Inchingolo A.M., Bordea I.R., Malcangi G., Xhajanka E., Scarano A., Lorusso F., Farronato M., Tartaglia G.M., Isacco C.G.. SARS-CoV-2 Disease through Viral Genomic and Receptor Implications: An Overview of Diagnostic and Immunology Breakthroughs. Microorganisms (2021) 9. DOI: 10.3390/microorganisms9040793 30. Scialo F., Daniele A., Amato F., Pastore L., Matera M.G., Cazzola M., Castaldo G., Bianco A.. ACE2: The Major Cell Entry Receptor for SARS-CoV-2. Lung (2020) 198 867-877. DOI: 10.1007/s00408-020-00408-4 31. Signore A., Lauri C., Colandrea M., Di Girolamo M., Chiodo E., Grana C.M., Campagna G., Aceti A.. Lymphopenia in patients affected by SARS-CoV-2 infection is caused by margination of lymphocytes in large bowel: An [18F] FDG PET/CT study. Eur. J. Nucl. Med. Mol. Imaging (2022) 49 3419-3429. DOI: 10.1007/s00259-022-05801-0 32. Ashrafi F., Nematollahi P., Salmasi M., Hedayat A., Amra B.. Association of lymphocyte subsets with mortality in severe COVID-19 pneumonia patients. J. Clin. Lab. Anal. (2021) 35 e24046. DOI: 10.1002/jcla.24046 33. Berthelot J.M., Lioté F., Maugars Y., Sibilia J.. Lymphocyte Changes in Severe COVID-19: Delayed Over-Activation of STING?. Front. Immunol. (2020) 11 607069. DOI: 10.3389/fimmu.2020.607069 34. Cantenys-Molina S., Fernández-Cruz E., Francos P., de Quirós J.C.L.B., Muñoz P., Gil-Herrera J.. Lymphocyte subsets early predict mortality in a large series of hospitalized COVID-19 patients in Spain. Clin. Exp. Immunol. (2021) 203 424-432. DOI: 10.1111/cei.13547 35. Zhao C., Hu X., Xue Q., Chen L.. Reduced Counts of Various Subsets of Peripheral Blood T Lymphocytes in Patients with Severe Course of COVID-19. Bull. Exp. Biol. Med. (2022) 172 721-724. DOI: 10.1007/s10517-022-05464-9 36. Iannetta M., Buccisano F., Fraboni D., Malagnino V., Campogiani L., Teti E., Spalliera I., Rossi B., Di Lorenzo A., Palmieri R.. Baseline T-lymphocyte subset absolute counts can predict both outcome and severity in SARS-CoV-2 infected patients: A single center study. Sci. Rep. (2021) 11 12762. DOI: 10.1038/s41598-021-90983-0 37. Balzanelli M.G., Distratis P., Dipalma G., Vimercati L., Catucci O., Amatulli F., Cefalo A., Lazzaro R., Palazzo D., Aityan S.K.. Immunity Profiling of COVID-19 Infection, Dynamic Variations of Lymphocyte Subsets, a Comparative Analysis on Four Different Groups. Microorganisms (2021) 9. DOI: 10.3390/microorganisms9102036 38. de Oliveira D.C., Spiri B.S., Schluga Y.C., Justus J.L.P., Neto F.D.N.L., de Azambuja A.P.. Evaluation of lymphocyte count, T-cell subsets and neutrophil-to-lymphocyte ratio as early predictors for severity and outcome of COVID-19 disease-a report from a highly complex hospital in Brazil. Hematol. Transfus. Cell Ther. (2022) 39. Montiel-Cervantes L.A., Medina G., Pilar Cruz-Domínguez M., Pérez-Tapia S.M., Jiménez-Martínez M.C., Arrieta-Oliva H.I., Carballo-Uicab G., López-Pelcastre L., Camacho-Sandoval R.. Poor Survival in COVID-19 Associated with Lymphopenia and Higher Neutrophile-Lymphocyte Ratio. Isr. Med. Assoc. J. (2021) 23 153-159. PMID: 33734627 40. Li M., Guo W., Dong Y., Wang X., Dai D., Liu X., Wu Y., Li M., Zhang W., Zhou H.. Elevated Exhaustion Levels of NK and CD8+ T Cells as Indicators for Progression and Prognosis of COVID-19 Disease. Front. Immunol. (2020) 11 580237. DOI: 10.3389/fimmu.2020.580237 41. Yuan S., Jiang S.C., Zhang Z.W., Fu Y.F., Hu J., Li Z.L.. Quantification of Cytokine Storms During Virus Infections. Front. Immunol. (2021) 12 659419. DOI: 10.3389/fimmu.2021.659419 42. Balzanelli M.G., Distratis P., Lazzaro R., Cefalo A., Catucci O., Aityan S.K., Dipalma G., Vimercati L., Inchingolo A.D., Maggiore M.E.. The Vitamin D, IL-6 and the eGFR Markers a Possible Way to Elucidate the Lung–Heart–Kidney Cross-Talk in COVID-19 Disease: A Foregone Conclusion. Microorganisms (2021) 9. DOI: 10.3390/microorganisms9091903 43. Dhar S.K., Damodar S., Gujar S., Das M.. IL-6 and IL-10 as predictors of disease severity in COVID-19 patients: Results from meta-analysis and regression. Heliyon (2021) 7 e06155. DOI: 10.1016/j.heliyon.2021.e06155 44. Mulla S., Molla M.M.A., Ahmed S.M.A., Akhtaruzzaman A.K.M., Saleh A.A., Anwar S.. Association of interferon gamma inducible protein-10, monocyte chemoattractant protein-1, macrophage inflammatory protein-1 alpha, interleukin-6, and rs12252 single nucleotide polymorphism of interferon-induced transmembrane protein-3 gene with the severity of COVID-19 infection. Egypt J. Intern. Med. (2022) 34 53. PMID: 35822078 45. Ergoren M.C., Paolacci S., Manara E., Dautaj A., Dhuli K., Anpilogov K., Camilleri G., Suer H.K., Sayan M., Tuncel G.. A pilot study on the preventative potential of alpha-cyclodextrin and hydroxytyrosol against SARS-CoV-2 transmission. Acta Biomed. (2020) 91 e2020022 46. Balzanelli M.G., Distratis P., Aityan S.K., Amatulli F., Catucci O., Cefalo A., De Michele A., Dipalma G., Inchingolo F., Lazzaro R.. An Alternative “Trojan Horse” Hypothesis for COVID-19: Immune Deficiency of IL-10 and SARS-CoV-2 Biology. Endocr. Metab. Immune Disord. Drug Targets (2022) 22 1-5. DOI: 10.2174/1871530321666210127141945 47. Hirano T.. IL-6 in inflammation, autoimmunity and cancer. Int. Immunol. (2021) 33 127-148. DOI: 10.1093/intimm/dxaa078 48. Basheer M., Saad E., Kananeh M., Asad L., Khayat O., Badarne A., Abdo Z., Arraf N., Milhem F., Bassal T.. Cytokine Patterns in COVID-19 Patients: Which Cytokines Predict Mortality and Which Protect Against?. Curr. Issues Mol. Biol. (2022) 44 4735-4747. DOI: 10.3390/cimb44100323 49. Medeiros T., Guimarães G.M.C., Carvalho F.R., Alves L.S., Faustino R., Campi-Azevedo A.C., Peruhype-Magalhães V., Teixeira-Carvalho A., de Souza Gomes M., do Amaral L.R.. Acute kidney injury associated to COVID-19 leads to a strong unbalance of circulant immune mediators. Cytokine (2022) 157 155974. DOI: 10.1016/j.cyto.2022.155974 50. Hernandez-Diaz S., Huybrechts K.F., Desai R.J., Cohen J.M., Mogun H., Pennell P.B., Bateman B.T., Patorno E.. Topiramate use early in pregnancy and the risk of oral clefts: A pregnancy cohort study. Neurology (2018) 90 e342-e351. DOI: 10.1212/WNL.0000000000004857 51. Kokkoris S., Stamataki E., Emmanouil G., Psachoulia C., Ntaidou T., Maragouti A., Kanavou A., Malachias S., Christodouli F., Papachatzakis I.. Serum inflammatory and brain injury biomarkers in COVID-19 patients admitted to intensive care unit: A pilot study. eNeurologicalSci (2022) 29 100434. DOI: 10.1016/j.ensci.2022.100434 52. Nguyen N., Nguyen H., Ukoha C., Hoang L., Patel C., Ikram F.G., Acharya P., Dhillon A., Sidhu M.. Relation of interleukin-6 levels in COVID-19 patients with major adverse cardiac events. Proc. Bayl. Univ. Med. Cent. (2022) 35 6-9. DOI: 10.1080/08998280.2021.1961571 53. Halmaciu I., Arbănași E.M., Kaller R., Mureșan A.V., Arbănași E.M., Bacalbasa N., Suciu B.A., Cojocaru I.I., Runcan A.I., Grosu F.. Chest CT Severity Score and Systemic Inflammatory Biomarkers as Predictors of the Need for Invasive Mechanical Ventilation and of COVID-19 Patients’ Mortality. Diagnostics (2022) 12. DOI: 10.3390/diagnostics12092089 54. Luo X., Zhou W., Yan X., Guo T., Wang B., Xia H., Ye L., Xiong J., Jiang Z., Liu Y.. Prognostic Value of C-Reactive Protein in Patients With Coronavirus 2019. Clin. Infect. Dis. (2020) 71 2174-2179. DOI: 10.1093/cid/ciaa641 55. Chapanduka Z.C., Abdullah I., Allwood B., Koegelenberg C.F., Irusen E., Lalla U., Zemlin A.E., Masha T.E., Erasmus R.T., Jalavu T.P.. Haematological predictors of poor outcome among COVID-19 patients admitted to an intensive care unit of a tertiary hospital in South Africa. PLoS ONE (2022) 17. DOI: 10.1371/journal.pone.0275832 56. Yousfi N., Fathallah I., Attoini A., Jones M., Henchir M., Ben Hassine Z., Kouraichi N., Ben Salah N.. Prognostic Value of Routine Blood Parameters in Intensive Care Unit COVID-19 Patients. EJIFCC (2022) 33 121-130. PMID: 36313910 57. Peralta-Amaro A.L., Pecero-García E.I., Valadez-Calderón J.G., Ramírez-Ventura J.C., Coria-Moctezuma L.A., Hernández-Utrera J.E.. Low testosterone levels as an independent risk factor for mortality in male patients with COVID-19: Report of a Single-Center Cohort Study in Mexico. Rev. Int. Androl. (2022) 58. Chen T., Dai Z., Mo P., Li X., Ma Z., Song S., Chen X., Luo M., Liang K., Gao S.. Clinical Characteristics and Outcomes of Older Patients with Coronavirus Disease 2019 (COVID-19) in Wuhan, China: A Single-Centered, Retrospective Study. Newman A, editor. J. Gerontol. Ser. A (2020) 75 1788-1795. DOI: 10.1093/gerona/glaa089 59. Heubner L., Petrick P.L., Güldner A., Bartels L., Ragaller M., Mirus M., Rand A., Tiebel O., Beyer-Westendorf J., Rößler M.. Extreme obesity is a strong predictor for in-hospital mortality and the prevalence of long-COVID in severe COVID-19 patients with acute respiratory distress syndrome. Sci. Rep. (2022) 12 18418. DOI: 10.1038/s41598-022-22107-1 60. Vimercati L., De Maria L., Quarato M., Caputi A., Gesualdo L., Migliore G., Cavone D., Sponselli S., Pipoli A., Inchingolo F.. Association between Long COVID and Overweight/Obesity. J. Clin. Med. (2021) 10. DOI: 10.3390/jcm10184143 61. Nader M., Zmerli O., Platt D.E., Hamdan H., Hamdash S., Tayeh R.A., Azar J., Kadi D., Sultan Y., Bazarbachi T.. Observational study of factors associated with morbidity and mortality from COVID-19 in Lebanon, 2020-2021. PLoS ONE (2022) 17. DOI: 10.1371/journal.pone.0275101 62. Kurian N.M., BYadav S.C., Gurjar M., Mishra P., Sanjeev O.P., Azim A., Gupta D., Hashim Z., Verma A., Bhatia E.. Association of hyperglycaemia at-admission & diabetes mellitus with 28 day mortality in patients admitted with moderate-severe SARS-CoV-2 infection: A retrospective study. Indian J. Med. Res. (2022) 155 554-564. PMID: 36348602 63. Caira-Chuquineyra B., Fernandez-Guzman D., Alvarez-Arias P.M., Zarate-Curi Á.A., Herrera-Añazco P., Benites-Zapata V.A.. Association between prehospital medication and fatal outcomes in a cohort of hospitalized patients due to coronavirus disease-2019 in a referral hospital in Peru. Travel. Med. Infect. Dis. (2022) 50 102472. DOI: 10.1016/j.tmaid.2022.102472
--- title: Modulation of Skin Inflammatory Responses by Aluminum Adjuvant authors: - Yanhang Liao - Lixiang Sun - Meifeng Nie - Jiacheng Li - Xiaofen Huang - Shujun Heng - Wenlu Zhang - Tian Xia - Zhuolin Guo - Qinjian Zhao - Ling-juan Zhang journal: Pharmaceutics year: 2023 pmcid: PMC9966661 doi: 10.3390/pharmaceutics15020576 license: CC BY 4.0 --- # Modulation of Skin Inflammatory Responses by Aluminum Adjuvant ## Abstract Aluminum salt (AS), one of the most commonly used vaccine adjuvants, has immuno-modulatory activity, but how the administration of AS alone may impact the activation of the skin immune system under inflammatory conditions has not been investigated. Here, we studied the therapeutic effect of AS injection on two distinct skin inflammatory mouse models: an imiquimod (IMQ)-induced psoriasis-like model and an MC903 (calcipotriol)—induced atopic dermatitis-like model. We found that injection of a high dose of AS not only suppressed the IMQ-mediated development of T-helper 1 (Th1) and T-helper 17 (Th17) immune responses but also inhibited the IMQ-mediated recruitment and/or activation of neutrophils and macrophages. In contrast, AS injection enhanced MC903-mediated development of the T-helper 2 (Th2) immune response and neutrophil recruitment. Using an in vitro approach, we found that AS treatment inhibited Th1 but promoted Th2 polarization of primary lymphocytes, and inhibited activation of peritoneal macrophages but not bone marrow derived neutrophils. Together, our results suggest that the injection of a high dose of AS may inhibit Th1 and Th17 immune response-driven skin inflammation but promote type 2 immune response-driven skin inflammation. These results may provide a better understanding of how vaccination with an aluminum adjuvant alters the skin immune response to external insults. ## 1. Introduction The skin, the primary interface with the environment, functions as an important barrier, protecting the body against pathogens, chemicals, allergens, and mechanical insults. Prolonged exposure of epidermal keratinocytes to environmental insults may lead to activation of the immune system, and ultimately the development of inflammatory skin disorders such as psoriasis and atopic dermatitis (AD), the most common dermatologic conditions [1,2]. T cells, the central component of adaptive immunity, play a critical role in host defense against pathogens. During infections, the cytokine milieu modulates the differentiation and polarization of CD4+ T cells into distinct effector phenotypes, including interferon γ (IFNγ) producing-T-helper 1 (Th1) cells that mediate the clearance of infected cells, interleukin 4 (IL4)- and IL13-producing Th2 cells that play a role in parasite expulsion and driving allergic response, and IL17-producing Th17 cells that mediate anti-fungal response and promote autoimmunity [3,4]. Psoriasis and atopic dermatitis are considered to be T cell-mediated chronic relapsing inflammatory skin diseases, mediated by different effector mechanisms. Psoriatic inflammation is primarily driven by the Th17 immune response, although Th1 cells are also present [5]. In contrast, the overactivation of Th2 cells plays a dominant role in driving acute skin inflammation in atopic dermatitis [1,5,6,7,8]. In the early 1900s, investigators found that injecting aluminum-precipitated antigen induced a stronger antibody and protective immune response compared to the response generated by free antigen injection [9,10]. Aluminum salts have now become one of the most commonly used adjuvants in human vaccines. Despite their longstanding use, the mechanisms by which aluminum adjuvants enhance immune responses are still not fully understood. Studies have shown that aluminum adjuvants promote the differentiation of CD4 T cells into Th2 effector cells but do not support Th1 cell differentiation in vivo or in vitro [11,12,13,14]. Dysregulation of T cell differentiation is the central disease mechanism for both psoriasis and atopic dermatitis, while aluminum salts exhibit their potential immunomodulatory effect through effector T cell differentiation/activation. Therefore, we aimed to determine whether the administration of aluminum salt influences the activation of the immune system in mouse models of psoriasis or atopic dermatitis. We also tested the in vitro effect of aluminum salt on effector T cell differentiation and the activation of key myeloid cells (neutrophils and macrophages). The results of our study provide insights into how vaccination may lead to the development of skin side effects by altering the skin’s immune response to external insults. ## 2.1. Chemicals, Antibodies and Dyes Aluminum salt, a suspension of aluminum hydroxyphosphate containing 1680 μg aluminum/mL and 9.33 mM/L inorganic phosphorus, is prepared by mixing AlCl3, Na2HPO4, and NaOH (to adjust the PH to 6~7) as described previously [15,16,17]. The HiScript II Q RT SuperMix kit for RNA reverse-transcription was purchased from Vazyme (Nanjing, China); 2× SYBR Green qPCR Master Mix for quantitative reverse transcription PCR was purchased from Bimake (Houston, TX, USA). Imiquimod Cream (IMQ) was purchased from the Med-Shine Corporation (Hongkong). Methotrexate disodium (MTX) and MC903 (calcipotriol) were purchased from Selleckchem (Houston, TX, USA). Rat anti-Ki67 antibody and ProLong™ Gold Antifade Mountant with DAPI (4′,6-diamidino-2-phenylindole, dihydrochloride) were purchased from ThermoFisher Scientific (Waltham, MA, USA). Alexa Fluor 647 AffiniPure Donkey Anti-Rat IgG (H+L) secondary antibody was purchased from Jackson ImmunoResearch (West Grove, PA, USA). Zombie violet viability dye, PECy7 anti-CD45, APC anti-CD11C, PerCP-Cy5.5 anti-LY6C, PECy7 anti-CD11B, APC/Cyanine7 anti-CD3, AF488 anti-IL4 and APC anti-TCRγ/δ antibodies were purchased from BioLegend (San Diego, CA, USA); CD16/CD32 Monoclonal Antibody, AF488 anti-LY6G, PE anti-F$\frac{4}{80}$, AF700 anti-MHCII, AF488 anti-IL13, PerCP-Cy5.5 anti-IFNγ, AF700 anti-CD4 antibodies and the fixation and permeabilization buffer set for fluorescence activated cell sorting (FACS) analysis were purchased from ThermoFisher Scientific. A hematoxylin and eosin staining kit was obtained from ZSGB-BIO Corporation (Beijing, China). Recombinant mouse IL-4 was obtained from BioLegend. Recombinant mouse IL-2, mouse IL-12, mouse IFNγ, mouse IL-6, mouse IL1β, mouse TGFβ1 were purchased from R&D Systems (Minneapolis, MN, USA). LPS and FSL were purchased from ThermoFisher Scientific. ## 2.2. Animal Cares and Animal Models C57BL/6 mice used in this study were purchased from GemPharmatech (Nanjing, China), then bred and maintained in the standard pathogen-free (SPF) environment of the Laboratory Animal Center in Xiamen University. All animal experiments were approved by the Institutional Animal Care and Use Committee of Xiamen University. For the IMQ-induced psoriasis-like model, 7~8-week-old female C57BL/6 mice were anesthetized, shaved, and depilated. Then, the backs of the mice were treated daily with a topical dose of 50 mg IMQ cream for 7 days. A single dose of aluminum salt (500 μL volume containing 840 μg AL3+) was injected peritoneally at day 2.5 post IMQ application, and daily i.p. injection of methotrexate (MTX), a commonly used systemic anti-inflammatory drug to treat psoriasis, was used as a positive control [18]. For the MC903-induced dermatitis-like model, the backs of the mice were shaved, depilated, and treated daily with a topical dose of 45 μL MC903 (100 μM, dissolved in ethanol) for 10 days. A single dose of aluminum salt was injected peritoneally at day 4 post-MC903 application. Daily administration of methotrexate (MTX) at 1 mg/kg was used as a positive anti-psoriatic agent [18]. The appearances of the lesions were recorded using a digital camera. The severity of skin inflammation was evaluated daily, including the measurements for skin redness, scaling, and epidermal thickness, and scored separately using a five-point scale from 0 to 4 (0, no incidence; 1, slight; 2, moderate; 3, marked; and 4, most severe) [19,20]. Additionally, the cumulative score (redness plus scaling plus epidermal thickness) was depicted [19,20]. ## 2.3. Histology and Immunohistochemistry (IHC) The biopsies of lesional skin were embedded in OCT (#4583, SAKURA, Torrance, CA, USA) followed by frozen sectioning. Frozen mouse skin sections were subjected to hematoxylin and eosin staining according to the manufacturer’s protocol. For IHC staining, OCT-embedded sections were permeabilized with $0.1\%$ saponin (#47036, Sigma, Tokyo, Japan) for 10 min, then blocked in $5\%$ BSA (#4240GR100, Biofroxx, Einhausen, Germany) solution for 1 h. Blocked sections were incubated with the indicated primary antibodies at 4 °C overnight, followed by appropriate fluorophore-coupled secondary antibodies in the dark for 4–6 h at 4 °C. Finally, the sections were mounted and observed using a Zeiss LSM 880 Laser Confocal Microscope (Zeiss, Jena, Germany). ## 2.4. Flow Cytometry and Analysis (FACS) FACS analysis was performed to analyze immune cell populations of innate immunity and adaptive immunity in inflammatory skin. Briefly, skin tissues were digested with collagenase D (V900893, Sigma) and Dnase 1 (D8071, Solarbio, Beijing, China) to prepare a single cell suspension. The cell suspension was then stained with zombie violet viability dye, and blocked non-specific Fc-mediated interactions with CD16/CD32 Monoclonal Antibody. Then, cells were incubated with the indicated antibody cocktail mix (Panel A or B, listed in Table 1) for 1 h at 4 °C with shaken every 10 min gently. Finally, the cells were resuspended in stabilizing fixative buffer (#338036, BD biosciences, San Jose, CA, USA). FACS analysis was performed using the Thermo Attune NxT machine (Waltham, MA, USA) and further analyzed using FlowJo V10 software. ## 2.5. Quantitative Reverse Transcription-Quantitative PCR (qRT-PCR) Analyses Total cellular RNA was extracted from skin tissues or cultured cells using Trizol (#T9424, Sigma), chloroform, and the RNAExpress Total RNA Kit (#M050, NCM Biotech, Newport, UK), and purified RNA was reversed transcribed to cDNA using the HiScript II Q RT SuperMix kit. Quantitative PCRs were performed by SYBR Green qPCR Master Mix (#B21202, Bimake) on the Qtower real-time machine (Analytikjena, Swavesey, Cambridge, UK). All primers used in our study were designed to span exon—exon junctions and are shown in Table 2. The expression of the Tbp (TATA-Box Binding Protein) gene was used as a housekeeping gene to normalize the target gene expression. The ratio of target mRNA to Tbp was calculated using the −2ΔCT (ΔCT = CT of target gene—CT of Tbp) method based on the published relative quantification method [21]. ## 2.6. In Vitro T Cell Differentiation Firstly, naïve T cells were isolated and purified from the inguinal lymph nodes of 8-week wild-type mice, and the cells were seeded in a 24-well cell culture plate that was pre-coated overnight with anti-CD3&CD28 (3 μg/mL). Next, T cells were incubated with the indicated cytokine cocktail mix including rm-IL-12 (15 ng/mL), rm-Il-2 (10 ng/mL) and anti–IL-4 (10 μg/mL) for Th1 differentiation, or rm-IL-4 (10 ng/mL), rm-IL-2 (10 ng/mL), anti-IFN-γ (15 μg/mL) and anti-IL-12 (15 μg/mL) for Th2 development, or rm-IL-6 (20 ng/mL), rm-TGFβ1 (2 ng/mL) and rm-IL-1β (10 ng/mL) to promote Th17 differentiation, with AS (1:250 dilution, ~6.74 μg/mL Al3+). The cells were cultured at 37 °C for 144 h. Then the cells were stimulated with PMA (50 ng/mL), Ionomycin (500 ng/mL), and GolgiPlug for 3 h. Finally, the cell suspensions were centrifuged, and the supernatant and precipitated cells were collected and used for further experiments. ## 2.7. Neutrophil and Macrophage Cultures To isolate neutrophils, bone marrow cells were first flushed from mouse femurs and tibias using RPMI-1640 medium containing $10\%$ FBS. Bone marrow cells were treated with red blood cell lysis buffer and washed once with PBS. Cells were then overlaid on top of the Histopaque gradient (1077–1119) and centrifuged for 30 min at 872× g at room temperature, without a break. Neutrophils were collected at the interface of the Histopaque 1119 and Histopaque 1077 layers, and isolated neutrophils were cultured in RPMI-1640 medium containing $10\%$ FBS and $1\%$ penicillin/streptomycin, and then treated as indicated. The purity of neutrophils was >$90\%$ as determined by flow cytometry. For treatment, cells were pretreated with AS (1:250 dilution from the 1680 μg/mL stock = 6.74 μg/mL Al3+) for 2 h and then stimulated with FSL (50 ng/mL) for 6 h. Peritoneal macrophages were isolated by injecting the mice with 5 mL of PBS containing $1\%$ FBS, gently massaging the belly, and then aspirating the fluid. This process was repeated three times in total. The cells were treated with red blood cell lysis buffer and resuspended in DMEM containing $10\%$ FBS and $1\%$ penicillin/streptomycin. Isolated macrophages were seeded at a cell density of 5 × 105 cells per well of a 24-well plate. Twenty-four hours after being seeded, nonadherent cells were removed, and the adherent macrophages were incubated with DMEM containing $10\%$ FBS and $1\%$ penicillin/streptomycin for all experiments. For treatment, cells were pretreated with AS (1:250 or 1:1000 dilution from the 1680 ug AL3+/mL stock) for 2 h and then stimulated with LPS (0.5 μg/mL) for 12 h. ## 2.8. Statistics Experiments were repeated at least 3 times independently. All statistical analyses were performed using GraphPad Prism version 9 software. For comparisons between more than two groups, statistical analysis was performed by one-way analysis of variance (ANOVA) followed by a Dunnett test [22], or two way ANOVA followed by a Bonferroni test [23], to correct for multiple comparisons as listed in the legend. For Figure 1b, a one-way ANOVA analysis was performed, followed by a two-stage linear step-up procedure of Benjamini, Krieger, and Yekutieli to correct for multiple comparisons by controlling the False Discovery Rate [24]. Quantitative results are presented as mean ± standard error of mean (SEM). A p value less than 0.05 was considered statistically significant and indicated with asterisks, * $p \leq 0.05$, $p \leq 0.01$, * $p \leq 0.001$, ** $p \leq 0.0001.$ ## 3.1. Administration of Aluminum Salt Alleviates the Development of Psoriasis-like Skin Inflammation in Mice To determine the effect of an aluminum adjuvant (see characterization results in Figure S1a–d) on the development of skin inflammation, we first employed an imiquimod (IMQ)-induced psoriasis-like skin inflammation mouse model (Figure 1a), in which both Th1 and Th17 cell-mediated adaptive immunity play a role in driving epidermal hyperplasia and scaling [25]. A dose of 840 μg aluminum/mouse was used in our animal study, and this dose is within the recommended range in vaccines for clinical use [26]. It is also a common practice for determining vaccine immunogenicity in a mouse model as an in vivo potency assay, in which a full human dose or even doubled the human dose was used for intraperitoneal (i.p.) injection into each mouse [27,28]. The injection of this high dose of AS had a minimal systemic cytotoxic effect in mice (Figure S1e,f). We found a single i.p. injection of aluminum salt (AS) during the application of IMQ was effective in alleviating the development of psoriasis phenotypes characterized by erythema, thickness, and scales (Figure 1b). However, it was less effective than the daily i.p. injection of methotrexate (MTX) (Figure 1b), a commonly used systemic anti-inflammatory drug for psoriasis. In addition, AS injections blocked the development of systemic inflammatory responses, as shown by the reduced enlargement of the lymph nodes and spleen tissues (Figure 1d,e). AS was more effective than MTX in reducing lymph node enlargement (Figure 1d). Histological analysis of the skin sections revealed that epidermal thickness and dermal cell infiltration in IMQ-treated skin were significantly lower in AS- and MTX-treated mice (Figure 1f–i). Immunostaining for Ki67, a proliferation-associated nuclear antigen, showed that IMQ treatment increased the number of Ki67+ basal and/or supra-basal keratinocytes, and this increase in epidermal cell proliferation was partially inhibited by either AS or MTX injections (Figure 1j,k). Antimicrobial peptides, including defensins (DEFBs), are strongly induced in activated keratinocytes in psoriasis, and defensins participate in cutaneous inflammation by promoting keratinocyte migration, proliferation, and the production of inflammatory cytokines [29]. Analysis of the expression of key defensins, including Defb3, Defb4, and Defb14, showed that AS or MTX partially inhibited the IMQ-mediated induction of defensin genes (Figure 1l,m and Figure S1g). Interestingly, we found that the expression of type 1 collagen (Col1a1) was significantly reduced in IMQ-treated skin, and this effect was reversed by AS or MTX injections (Figure S1h), suggesting that IMQ-triggered dysregulation of dermal homeostasis can be restored by AS. Together, these results demonstrate that the administration of aluminum salt alleviates the development of psoriasis-like skin inflammation in mice. ## 3.2. Administration of Aluminum Salt Inhibited the Development of Th1 and Th17 Immune Responses in the IMQ-Induced Psoriasis Model Next, we aimed to investigate the effect of AS on lymphocyte activation in an IMQ-induced psoriasis model. In mouse skin, γδ T cells are the major IL17-producing cells after infection, wounding or IMQ application. On the other hand, αβ T cells, either CD4+ or CD8+, are capable of producing large amounts of the Th1 cytokine IFNγ or Th2 cytokines (IL4 and IL13) under different inflammatory conditions [8,30]. FACS analysis of skin CD4 T cells revealed that IMQ treatment inhibited Th2 but promoted Th1 polarization of CD4+ T cells in the skin, and increased the percentage of IL17A-producing γδ T cells (Figure 2a–c). These effects were largely reversed by either AS or MTX injections (Figure 2a–c). In line with these FACS results, qRT-PCR analysis showed that the IMQ-dependent induction of Il17a and Il17f was also significantly inhibited by either AS or MTX injections (Figure 2d,e). However, the IMQ-mediated induction of Il23 and Il22 did not appear to be influenced by AS or MTX injections (Figure S2a,b). ## 3.3. Administration of Aluminum Salt Reduced the Recruitment and Activation of Myeloid Cells Abnormal dermal infiltration and activation of myeloid cells, including neutrophils and macrophages, is one of the histologic hallmarks of psoriasis [31]. Flow cytometry (FACS) analysis showed that AS or MTX injections reduced the percentage of infiltrated CD11B+Ly6G+ neutrophils in IMQ-treated skin samples (Figure S3a and Figure 3a,b). In line with this FACS result, qRT-PCR analysis showed that the expression levels of neutrophil marker genes, including Ly6g and S100A8, were also reduced in skin samples from AS or MTX-treated mice (Figure 3c,d). FACS analysis of CD11B+F$\frac{4}{80}$+ macrophages revealed that in IMQ-treated skin, macrophages expressed Ly6C (Figure S3a and Figure 3a,e), a phenotypic marker for pro-inflammatory macrophages [32], indicating that macrophages shift from a resting to an inflammatory state during the development of psoriasis. Furthermore, we found that the IMQ-mediated increase in Ly6Chi macrophages was inhibited by either AS or MTX injection (Figure 3a,e). Additionally, IMQ-mediated induction of Il1b, Cxcl1, and Saa3, key genes related to myeloid cell activation and/or chemotaxis [33,34], was also significantly inhibited by AS or MTX injections (Figure 3a,e and Figure S3b). Together these results show that the systemic administration of aluminum salt potently suppresses the activation of myeloid cells, including neutrophils and macrophages, in the IMQ-induced psoriasis model. ## 3.4. Administration of Aluminum Salt Promoted the Development of MC903-Induced Atopic Dermatitis-like Skin Inflammation To determine whether aluminum salt promotes the development of the Th2 immune response in a mouse model of dermatitis, we adapted the MC903-induced dermatitis model, one of the most well-characterized murine models of atopic dermatitis in which T cells are preferentially polarized toward the Th2 phenotype [35]. Daily topical application of MC903 to the back skin led to reddening and scaling of the skin, which could be partially inhibited by systemic administration of MTX; in contrast, MC903-induced redness and scaling were markedly increased by the injection of aluminum salt (Figure 4a–d). Histological analysis of skin sections revealed that epidermal thickness and dermal cell infiltration in MC903-treated skin were significantly higher in the AS injected mice (Figure 4e–g). In addition, qRT-PCR analysis showed that AS injection enhanced the MC903-mediated induction of genes associated with epidermal keratinocyte activation (Defb4 and Defb3) (Figure 4h,i). Similar to the IMQ-induced psoriasis model, MC903 application also led to suppression of Col1a1 expression in the skin, but AS injection failed to restore Col1a1 expression in MC903-treated skin (Figure S4a). ## 3.5.1. Administration of Aluminum Salt Promoted the Development of Type 2 Immune Response in the MC903-Induced Dermatitis Model FACS analysis of skin cells revealed that AS injection significantly enhanced the expression of Th2 cytokines (IL4 and IL13) from CD45+SSClo T cells in MC903-treated skin (Figure 5a,b). In addition, qRT-PCR analysis showed that MC903-mediated induction of Il4 was further enhanced by AS injection but inhibited by MTX injection (Figure 5c). These results showed that the administration of aluminum salt promoted the development of type 2 inflammation in the MC903-induced dermatitis mouse model. ## 3.5.2. Administration of Aluminum Salt Altered Myeloid Cell Activation in the MC903-Induced Dermatitis Model FACS analysis of myeloid cells showed that, in contrast to daily IMQ application, daily application of MC903 led to only a small increase in CD11B+Ly6G+ neutrophils and no increase in CD11B+F$\frac{4}{80}$+Ly6C+ pro-inflammatory macrophages (Figure S5a and Figure 5d). AS injection increased MC903-mediated infiltration of neutrophils, but not the inflammatory macrophages (Figure S5a and Figure 5d). In line with the FACS results, qRT-PCR analysis showed that AS injection increased the MC903-mediated induction of Ly6g (neutrophil marker gene) (Figure 5e). These results showed that the administration of aluminum salt promoted the infiltration of neutrophils, but not inflammatory macrophages, into MC903-treated skin. ## 3.6. The In Vitro Effect of Aluminum Salt in Modulating T Cell Differentiation and Myeloid Cell Activation We showed that injection of AS promoted the development of type 2 skin inflammation, and differentially altered myeloid cell activation in two distinct dermatitis models in vivo. Next, we aimed to determine whether AS directly alters the activation of lymphocytes and myeloid cells in vitro. ## 3.6.1. Aluminum Salt Inhibited Th1 but Promoted Th2 Polarization during In Vitro T Cell Differentiation First, to determine whether aluminum salt can directly alter Th1 or Th2 polarization in differentiating lymphocytes, we subjected lymph node-derived primary lymphocytes to in vitro differentiation assays in the presence of CD3/CD28 antibody under Th1, Th2, or Th17 polarizing conditions. We found that the addition of AS significantly inhibited IFNγ production under Th1 polarizing conditions from CD4+ T cells, but promoted IL4 and IL13 production under Th2 polarizing conditions from CD4+ T cells (Figure 6a,b and Figure S6a,b). In addition, we found that under Th2 polarizing conditions, AS treatment promoted a Th17 to Th2 shift in γδ T cells, the major IL17-producing cell type in the skin (Figure S6c,d). However, under Th17 skewing conditions, in which γδ T cells were robustly shifted into IL17A producing cells, AS treatment only mildly reduced the expression of IL17A in γδ T cells without altering the expression of IFNγ or IL4/IL13 (Figure S6e,f). ## 3.6.2. Aluminum Salt Inhibited Macrophage but Not Neutrophil Activation In Vitro Next, primary neutrophils derived from bone marrow were stimulated with FSL, and primary peritoneal macrophages were stimulated with LPS with or without AS (Figure 6c,d). We found that while AS had no effect on neutrophil activation (Figure 6c), it significantly suppressed macrophage activation even at low concentration (1:1000 dilution from the original stock), as shown by qRT-PCR analysis of Il1b, Cxcl1 and Nos2 (Figure 6d). IL1β and NOS2 are well-established markers for inflammatory M1 macrophages [36], indicating that AS may directly inhibit macrophage polarization toward the pro-inflammatory state. ## 4. Discussion Injection of vaccines, composed of immunogens, preservatives, adjuvants, and by-products, can elicit adverse skin reactions, such as localized or generalized eczema vaccinia, in susceptible individuals [37,38]. Conflicting results have been reported regarding the relationship between vaccination and the development of atopic diseases [39,40,41], but none of these studies investigated the specific immunomodulatory effects of the individual component of the vaccines. Aluminum salts are widely used as adjuvants in preventive vaccines, enhancing their immunogenicity and effectiveness by stimulating a type 2 immune response [11,12]. Theoretically, AS could increase the risk of type 2 lymphocyte-mediated allergic and hyper-responsive diseases, such as atopic dermatitis. On the other hand, AS application may have a therapeutic effect against psoriasis, dominated by Th1 and Th17 cells, which are antagonistic to Th2 cells. To rest this theory, in the present study, we investigated the immuno-modulatory effect of aluminum salt in two distinct murine models of inflammatory skin diseases. We found that the injection of aluminum salt into the IMQ-induced psoriasis mouse model promoted T cell polarization from the Th1/Th17 to Th2 phenotype, suppressed neutrophil recruitment and macrophage activation, and therefore suppressed the development of psoriasis-like skin inflammation. In contrast, injection of aluminum salt promoted the development of the Th2 immune response and clinical phenotype of dermatitis in the MC903-induced atopic dermatitis-like mouse model. Our results indicate that AS application may inhibit the Th1/Th17-mediated activation of autoimmunity but enhance the Th2-mediated activation of the allergic immune response in the skin. In line with our results, it has been reported that patients receiving subcutaneous allergen-specific immunotherapy with aluminum adjuvants are associated with a lower risk of autoimmune diseases, including psoriasis [42]. In contrast, several reports have shown that although rare, patients can develop delayed hypersensitivity or allergic cutaneous reactions to vaccines containing aluminum salts [43,44,45,46,47]. By in vitro primary culture, we showed that the addition of aluminum salt promoted Th2 differentiation and inhibited the Th1 differentiation of naïve lymphocytes. To our knowledge, this is the first study investigating the direct effect of aluminum salt on naive T cell differentiation/polarization. Furthermore, we found that the addition of aluminum salt inhibited macrophage polarization toward the pro-inflammatory state but had no direct effect on neutrophil activation. AS-mediated differential effects on neutrophil recruitment in the IMQ- or MC903-induced dermatitis models were likely indirectly mediated by AS-dependent changes in T cell effector immune responses. It has been shown that aluminum hydroxide has a high adsorption capacity for endotoxins and LPS (283 μg/mg of Al), whereas endotoxins are electrostatically repelled by aluminum phosphate [48]. As a result, the adsorption capacities of phosphate-treated aluminum hydroxide or aluminum phosphate are only 23 and 3 μg endotoxin/mg of Al, respectively [48]. Here we report that the AS solution (which is a mixture of aluminum hydroxide and aluminum phosphate) can inhibit the inflammatory activity of LPS (0.5 μg/mL) even at low concentration (~1.7 μg AL3+/mL), which should only absorb 5~29 ng/mL LPS in theory. Therefore, AS-mediated absorption/neutralization of LPS may contribute to the inhibitory effect of AS against LPS, but it is unlikely the major mechanism for this inhibition. Future studies are still needed to determine the mechanism underlying the anti-inflammatory effect of AS against the LPS-mediated inflammatory response in macrophages. A limitation of our study is that we administered aluminum salt as a single intraperitoneal injection instead of multiple intramuscular injections as performed in routine vaccinations in clinics. In addition, we investigated the immunomodulatory effect of only one high dose of AS in mouse dermatitis models, although this dose was chosen based on our and others’ previous studies in the mouse potency assay of human vaccines with aluminum adjuvants [15,16,17,27,28]. Our study offers a clue for the immunomodulatory role of aluminum, not an effective therapy for human disease. Future studies are needed to determine the optimal injection method and dose to observe the immunomodulatory function of AS in animal dermatitis models. Together, our results provide new mechanisms underlying the immunomodulatory effect of aluminum salt on immune cell activation. Systemic injection of a high dose of aluminum adjuvant may inhibit or promote skin inflammation, depending on the involvement of specific effector T cells and/or myeloid cells during disease pathogenesis. ## References 1. Sun L., Liu W., Zhang L.J.. The Role of Toll-Like Receptors in Skin Host Defense, Psoriasis, and Atopic Dermatitis*. J. Immunol. Res.* (2019) 2019 1824624. DOI: 10.1155/2019/1824624 2. Zhang L.J.. Recent progress at the psoriasis and atopic dermatitis research front: An experimental dermatology perspective. Exp. Derm. (2021) 30 756-764. DOI: 10.1111/exd.14388 3. Inui N., Chida K., Suda T., Nakamura H.. TH1/TH2 and TC1/TC2 profiles in peripheral blood and bronchoalveolar lavage fluid cells in pulmonary sarcoidosis. J. Allergy Clin. Immunol. (2001) 102 337-344. DOI: 10.1067/mai.2001.112273 4. Leung S., Liu X., Fang L., Chen X., Guo T., Zhang J.. The cytokine milieu in the interplay of pathogenic Th1/Th17 cells and regulatory T cells in autoimmune disease. Cell. Mol. Immunol. (2010) 7 182-189. DOI: 10.1038/cmi.2010.22 5. Eyerich S., Onken A.T., Weidinger S., Franke A., Nasorri A., Pennino D., Grosber M., Pfab F., Schmidt-Weber C.B., Mempel M.. Mutual antagonism of T cells causing psoriasis and atopic eczema. N. Engl. J. Med. (2011) 365 231-238. DOI: 10.1056/NEJMoa1104200 6. Guo Z., Yang Y., Liao Y., Shi Y., Zhang L.-j.. Emerging Roles of Adipose Tissue in the Pathogenesis of Psoriasis and Atopic Dermatitis in Obesity. JID Innov. (2022) 2 100064. DOI: 10.1016/j.xjidi.2021.100064 7. Sabat R., Wolk K., Loyal L., Döcke W.-D., Ghoreschi K.. T cell pathology in skin inflammation. Seminar Immunopathology (2019) Volume 41 359-377 8. Ho A.W., Kupper T.S.. T cells and the skin: From protective immunity to inflammatory skin disorders. Nat. Rev. Immunol. (2019) 19 490-502. DOI: 10.1038/s41577-019-0162-3 9. Volk V.K., Bunney W.E.. Diphtheria Immunization with fluid toxoid and alum-precipitated toxoid. Am. J. Public Health Nations Health (1942) 32 690-699. DOI: 10.2105/AJPH.32.7.690 10. Jones F.G., Moss J.M.. Studies on Tetanus Toxoid: I. The Antitoxic Titer of Human Subjects Following Immunization with Tetanus Toxoid and Tetanus Alum Precipitated Toxoid. J. Immunol. (1936) 30 115-125. DOI: 10.4049/jimmunol.30.2.115 11. Sokolovska A., Hem S.L., HogenEsch H.. Activation of dendritic cells and induction of CD4(+) T cell differentiation by aluminum-containing adjuvants. Vaccine (2007) 25 4575-4585. DOI: 10.1016/j.vaccine.2007.03.045 12. Cunningham A.F., Serre K., Toellner K.-M., Khan M., Alexander J., Brombacher F., MacLennan I.C.M.. Pinpointing IL-4-independent acquisition and IL-4-influenced maintenance of Th2 activity by CD4 T cells. Eur. J. Immunol. (2004) 34 686-694. DOI: 10.1002/eji.200324510 13. Comoy E.E., Capron A., Thyphronitis G.. In vivo induction of type 1 and 2 immune responses against protein antigens. Int. Immunol. (1997) 9 523-531. DOI: 10.1093/intimm/9.4.523 14. Brewer J.M., Conacher M., Satoskar A., Bluethmann H., Alexander J.. In interleukin-4-deficient mice, alum not only generates T helper 1 responses equivalent to freund’s complete adjuvant, but continues to induce T helper 2 cytokine production. Eur. J. Immunol. (1996) 26 2062-2066. DOI: 10.1002/eji.1830260915 15. Nie M., Liu Y., Huang X., Zhang Z., Zhao Q.. Microtiter Plate-Based Differential Scanning Fluorimetry: A High-Throughput Method for Efficient Formulation Development. J. Pharm. Sci. (2022) 111 2397-2403. DOI: 10.1016/j.xphs.2022.05.015 16. Wu Y., Huang X., Yuan L., Wang S., Zhang Y., Xiong H., Chen R., Ma J., Qi R., Nie M.. A recombinant spike protein subunit vaccine confers protective immunity against SARS-CoV-2 infection and transmission in hamsters. Sci. Transl. Med. (2021) 13 eabg1143. DOI: 10.1126/scitranslmed.abg1143 17. Zhang Y., Li M., Yang M., Li Y., Zheng Z., Zhang X., Lin Q., Wang Y., Li S., Xia N.. Comparable quality attributes of hepatitis E vaccine antigen with and without adjuvant adsorption-dissolution treatment. Hum. Vaccines Immunother. (2015) 11 1129-1139. DOI: 10.1080/21645515.2015.1009343 18. Zong J., Cheng J., Fu Y., Song J., Pan W., Yang L., Zhang T., Zhou M.. Serum Metabolomic Profiling Reveals the Amelioration Effect of Methotrexate on Imiquimod-Induced Psoriasis in Mouse. Front. Pharm. (2020) 11 558629. DOI: 10.3389/fphar.2020.558629 19. Neu S.D., Strzepa A., Martin S., Sorci-Thomas M.G., Pritchard K.A., Dittel B.N.. Myeloperoxidase Inhibition Ameliorates Plaque Psoriasis in Mice. Antioxidants (2021) 10. DOI: 10.3390/antiox10091338 20. Langley R.G., Ellis C.N.. Evaluating psoriasis with Psoriasis Area and Severity Index, Psoriasis Global Assessment, and Lattice System Physician’s Global Assessment. J. Am. Acad. Derm. (2004) 51 563-569. DOI: 10.1016/j.jaad.2004.04.012 21. Livak K.J., Schmittgen T.D.. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods (2001) 25 402-408. DOI: 10.1006/meth.2001.1262 22. Shun Z., Silverberg A., Chang C.K., Ouyang P.. Dunnett’s many-to-one test and least square means. J. Biopharm. Stat. (2003) 13 17-28. DOI: 10.1081/BIP-120017723 23. Armstrong R.A.. When to use the Bonferroni correction. Ophthalmic. Physiol. Opt. (2014) 34 502-508. DOI: 10.1111/opo.12131 24. Benjamini Y., Krieger A.M., Yekutieli D.. Adaptive linear step-up procedures that control the false discovery rate. Biometrika (2006) 93 491-507. DOI: 10.1093/biomet/93.3.491 25. Ishizaki M., Akimoto T., Muromoto R., Yokoyama M., Ohshiro Y., Sekine Y., Maeda H., Shimoda K., Oritani K., Matsuda T.. Involvement of tyrosine kinase-2 in both the IL-12/Th1 and IL-23/Th17 axes in vivo. J. Immunol. (2011) 187 181-189. DOI: 10.4049/jimmunol.1003244 26. Baylor N.W., Egan W., Richman P.. Aluminum salts in vaccines--US perspective. Vaccine (2002) 20 S18-S23. DOI: 10.1016/S0264-410X(02)00166-4 27. Wang R., Sun C., Ma J., Yu C., Kong D., Chen M., Liu X., Zhao D., Gao S., Kou S.. A Bivalent COVID-19 Vaccine Based on Alpha and Beta Variants Elicits Potent and Broad Immune Responses in Mice against SARS-CoV-2 Variants. Vaccines (2022) 10. DOI: 10.3390/vaccines10050702 28. Mao Q., Dong C., Li X., Gao Q., Guo Z., Yao X., Wang Y., Gao F., Li F., Xu M.. Comparative analysis of the immunogenicity and protective effects of inactivated EV71 vaccines in mice. PLoS ONE (2012) 7. DOI: 10.1371/journal.pone.0046043 29. Niyonsaba F., Ushio H., Nakano N., Ng W., Sayama K., Hashimoto K., Nagaoka I., Okumura K., Ogawa H.. Antimicrobial peptides human beta-defensins stimulate epidermal keratinocyte migration, proliferation and production of proinflammatory cytokines and chemokines. J. Investig. Derm. (2007) 127 594-604. DOI: 10.1038/sj.jid.5700599 30. Qi C., Wang Y., Li P.. Gamma Delta T Cells and Their Pathogenic Role in Psoriasis. Front. Immunol. (2021) 12 627139. DOI: 10.3389/fimmu.2021.627139 31. Deng Y., Chang C., Lu Q.. The Inflammatory Response in Psoriasis: A Comprehensive Review. Clin. Rev. Allergy Immunol. (2016) 50 377-389. DOI: 10.1007/s12016-016-8535-x 32. Rahman K., Vengrenyuk Y., Loke P., Fisher E.A.. Inflammatory Ly6Chi monocytes and their conversion to M2 macrophages drive atherosclerosis regression. J. Clin. Investig. (2017) 127 2904-2915. DOI: 10.1172/JCI75005 33. Cai Y., Xue F., Quan C., Qu M., Liu N., Zhang Y., Fleming C., Hu X., Zhang H.-G., Weuchselbaum R.. A Critical Role of the IL-1beta-IL-1R Signaling Pathway in Skin Inflammation and Psoriasis Pathogenesis. J. Investig. Derm. (2019) 139 146-156. DOI: 10.1016/j.jid.2018.07.025 34. Metzemaekers M., Gouwy M., Proost P.. Neutrophil chemoattractant receptors in health and disease: Double-edged swords. Cell. Mol. Immunol. (2020) 17 433-450. DOI: 10.1038/s41423-020-0412-0 35. Leyva-Castillo J.M., Hener P., Michea P., Karasuyama H., Chan S., Soulmelis V., Li M.. Skin thymic stromal lymphopoietin initiates Th2 responses through an orchestrated immune cascade. Nat. Commun. (2013) 4 2847. DOI: 10.1038/ncomms3847 36. Orecchioni M., Ghosheh Y., Pramod A.B., Ley K.. Corrigendum: Macrophage Polarization: Different Gene Signatures in M1(LPS+) vs. Classically and M2(LPS-) vs. Alternatively Activated Macrophages. Front. Immunol. (2020) 11 234. DOI: 10.3389/fimmu.2020.00234 37. Leventhal J.S., Berger E.M., Brauer J.A., Cohen D.E.. Hypersensitivity reactions to vaccine constituents: A case series and review of the literature. Dermatitis (2012) 23 102-109. DOI: 10.1097/DER.0b013e31825228cf 38. Engler R.J., Kenner J., Leung D.Y.. Smallpox vaccination: Risk considerations for patients with atopic dermatitis. J. Allergy Clin. Immunol. (2022) 110 357-365. DOI: 10.1067/mai.2002.128052 39. Ayasse M., Ahmed A., McCullum C., Espinosa M.L., Paller A.S., Silverberg J.I.. Vaccines do not cause atopic dermatitis: A systematic review and meta-analysis. Vaccine (2021) 39 1805-1811. DOI: 10.1016/j.vaccine.2021.02.036 40. Enriquez R., Addington W., Davis F., Freels S., Park C.L., Hershow R.S., Persky V.. The relationship between vaccine refusal and self-report of atopic disease in children. J. Allergy Clin. Immunol. (2005) 115 737-744. DOI: 10.1016/j.jaci.2004.12.1128 41. Koppen S., de Groot R., Neijens H.J., Nagelkerke N., van Eden W., Rümke H.C.. No epidemiological evidence for infant vaccinations to cause allergic disease. Vaccine (2004) 22 3375-3385. DOI: 10.1016/j.vaccine.2004.02.033 42. Linneberg A., Jacobsen R.K., Jespersen L., Abildstrøm S.Z.. Association of subcutaneous allergen-specific immunotherapy with incidence of autoimmune disease, ischemic heart disease, and mortality. J. Allergy Clin. Immunol. (2012) 129 413-419. DOI: 10.1016/j.jaci.2011.09.007 43. Mistry B.D., DeKoven J.G.. Widespread cutaneous eruption after aluminum-containing vaccination: A case report and review of current literature. Pediatr. Derm. (2021) 38 872-874. DOI: 10.1111/pde.14613 44. Aquino M.R., Bingerman T.A., Nanda A., Maples K.M.. Delayed allergic skin reactions to vaccines. Allergy and Asthma Proceedings (2022) Volume 43 20-29 45. Hoffmann S.S., Thiesson E.M., Johansen J.D., Hviid A.. Risk factors for granulomas in children following immunization with aluminium-adsorbed vaccines: A Danish population-based cohort study. Contact Dermat. (2022) 87 430-438. DOI: 10.1111/cod.14180 46. Gordon S.C., Bartenstein D.W., Tajmir S.H., Song J.S., Hawryluk E.B.. Delayed-type hypersensitivity to vaccine aluminum adjuvant causing subcutaneous leg mass and urticaria in a child. Pediatr. Derm. (2018) 35 234-236. DOI: 10.1111/pde.13390 47. Thierry-Carstensen B., Stellfeld M.. Itching nodules and hypersensitivity to aluminium after the use of adsorbed vaccines from SSI. Vaccine (2004) 22 1845. DOI: 10.1016/j.vaccine.2003.11.048 48. Shi Y., HogenEsch H., Regnier F.E., Hem S.L.. Detoxification of endotoxin by aluminum hydroxide adjuvant. Vaccine (2001) 19 1747-1752. DOI: 10.1016/S0264-410X(00)00394-7
--- title: Comparison of the Activity of Fecal Enzymes and Concentration of SCFA in Healthy and Overweight Children authors: - Katarzyna Śliżewska - Michał Włodarczyk - Martyna Sobczak - Renata Barczyńska - Janusz Kapuśniak - Piotr Socha - Aldona Wierzbicka-Rucińska - Aneta Kotowska journal: Nutrients year: 2023 pmcid: PMC9966664 doi: 10.3390/nu15040987 license: CC BY 4.0 --- # Comparison of the Activity of Fecal Enzymes and Concentration of SCFA in Healthy and Overweight Children ## Abstract In modern societies obesity has become a serious issue which must be urgently addressed. The health implications of neglected obesity are substantial, as not only does it affect individuals’ everyday lives, but it also leads to significantly increased mortality due to the development of several disorders such as type-2 diabetes, cardiovascular diseases, cancers, and depression. The objective of this research was to investigate the alterations in selected health markers caused by overweight and obesity in children. The measured parameters were the activity of the fecal enzymes, the concentration of short-chain fatty acids (SCFAs), and the concentration of branched-chain fatty acids (BCFAs). The activity of the fecal enzymes, specifically α-glucosidase, α-galactosidase, β-glucosidase, β-galactosidase, and β-glucuronidase, was determined using spectrophotometry at a wavelength of 400 nm. Furthermore, concentrations of lactic acid, SCFAs (formic, acetic, propionic, butyric, and valeric acids), and BCFAs (isobutyric and isovaleric acids) were determined using the HPLC method. The obtained results reveal that obese children have different fecal enzyme activity and a different profile of fatty acids from children of normal weight. The group of obese children, when compared to children of normal weight, had increased concentrations of BCFAs ($p \leq 0.05$) and higher activity of potentially harmful enzymes such as β-glucosidase and β-glucuronidase ($p \leq 0.05$). In comparison, children of normal weight exhibited significantly increased concentrations of lactic acid and SCFAs (especially formic and butyric acids) ($p \leq 0.05$). Furthermore, their α-glucosidase and α-galactosidase activity were higher when compared to the group of obese children ($p \leq 0.05$). These results suggest that the prevalence of obesity has a significant impact on metabolites produced in the gastrointestinal tract, which might result in a higher chance of developing serious diseases. ## 1. Introduction Obesity has become an epidemic of the 21st century, which is confirmed by the alarming data from the WHO on the NCD-RisC indicating a rapid increase in the prevalence of metabolic disorders in well-developed countries [1,2]. Records show that in 2016 more than 1.9 billion adults were overweight, of whom $13\%$ were obese, meaning they had a BMI higher than 30 [3]. Similar findings concern children and adolescents. It was estimated by the WHO that over 340 million children aged 5–19 years were overweight in 2016, and over $30\%$ of these were obese [4]. This is alarming given the fact that obese children have a significantly higher chance of becoming obese adults than non-obese children, which will directly impact the number of people suffering from obesity-related health implications [5]. Since then, there has been no record of improvement in the situation. According to several predictions, by the year 2030 the number of obese children may double and further increase with time [2,6,7]. There are various factors that promote the development of metabolic disorders, such as lifestyle changes that favor sitting or less active occupations, high stress levels, rush, and ill-considered dietary decisions [8]. It is well-established knowledge that diet influences the gastrointestinal microbiota, which is reflected in several studies, i.e., comparing the microbiota of normal-weight and obese people or examining how certain types of food impact the gut microorganisms [9,10,11,12,13]. Apart from contributing to energy metabolism, gut microorganisms produce a variety of bioactive compounds, such as vitamins, short-chain and branched-chain fatty acids (SCFAs, BCFAs), and fecal enzymes, which participate in several metabolic pathways where they exhibit anti-inflammatory, anticarcinogenic, and antioxidative effects (genera Lactobacillus, Bifidobacterium) [14,15,16]. In contrast, a dysbiosis of gut microorganisms, or their low diversity favoring the domination of certain genera (Bacteroides, Clostridium, Escherichia, Enterococcus), may cause reduced synthesis of bioactive compounds and promote synthesis with the possible accumulation of potentially harmful ones [17,18]. SCFAs have a variety of beneficial effects on the host organism, i.e., regulation of energy metabolism, immunoregulation, and stabilization of the integrity of the intestinal barrier [8,19]. BCFAs, however, are often associated with the process of protein fermentation, which may lead to the accumulation of potentially harmful bioactive compounds and can increase the risk of developing colonic cancer [20]. Microbial diversity in the gastrointestinal tract is similarly a factor that influences enzymatic activity in the colon. Bacteria produce a variety of enzymes that help to digest food ingredients while producing several bioactive compounds as byproducts. Reductases and hydrolases, which are the main classes of enzymes secreted by the intestinal bacteria, can help to produce beneficial SCFAs, but may likewise be responsible for the formation of toxic or carcinogenic compounds [21,22]. Due to the high impact of the above-mentioned bioactive compounds on the human organism, they can be considered health markers, which help to assess the physical status of a person. Based on these facts, the aim of this study was to further confirm differences in the activities of fecal enzymes and the profiles of fatty acids in fecal samples between a group of healthy overweight or obese children and a group of children of normal weight. Accordingly, concentrations of SCFAs, namely formic, acetic, propionic, butyric, and valeric acids, together with concentrations of lactic acid were determined using the HPLC method. Furthermore, concentrations of BCFAs such as isobutyric and isovaleric acids were measured. The next part of the study involved the determination of activity of fecal enzymes, namely α-glucosidase, α-galactosidase, β-glucosidase, β-galactosidase, and β-glucuronidase, in the same fecal samples. The outcome of the experimental work allowed for the evaluation whether the group of obese individuals without dietary supervision has significantly different health markers (fecal enzyme activity and concentration of fatty acids). The obtained results were subject to statistical analysis to verify possible correlations among the mentioned health markers and elucidate more comprehensive conclusions. ## 2.1. Biological Material The fecal samples were collected from 97 overweight and obese children (participants of the PreSTFibre4kids research program) and 26 children of normal weight (both males and females) aged 6–10 years (Table 1). Patients from the PreSTFibre4kids program were children with one of the following health implications: simple obesity and overweight without organ complications, fatty liver disease, and high blood pressure. Volunteers were recruited in Poland. Every participant had voluntarily signed a consent form that confirmed his/her involvement in the project. Exclusion criteria from the study included organ failure, metformin treatment, and allergy to any component of the prebiotic preparation that would be used in further phases of the research. The exclusion criteria did not include a medical history of taking antibiotics or the usage of dietary supplements. The overweight status of children was assigned according to the WHO’s BMI classification charts [23]. Participants were given general dietary advice by a trained professional. Fecal samples were transferred into sterile containers and frozen immediately (−20 °C) prior to transport to the laboratory. The project was accepted by the Research Ethics Committee of the Children’s Memorial Health Institute in Warsaw, Poland (18/KBE/2021). ## 2.2. Activity of Fecal Enzymes Prior to the enzymatic assays, the samples were prepared according to the following protocol. Firstly, 0.7 g of sample was suspended in 3.5 mL of 0.2 M phosphate buffer and vortexed thoroughly (Vortex RS-VA 10, Phoenix Instrument, Garbsen, Germany). Secondly, the samples underwent sonification (Time = 2 min, Amplitude = 60, Pulse = 6 s, Cole–Parmer Instrument Co., Vernon Hills, IL, USA). Afterwards, the samples were centrifuged (12,000 rpm, Time = 20 min, Centrifuge MPW-251, MPW, Warszawa, Poland) and the supernatant was transferred to the sterile Eppendorf tubes. With the use of spectrophotometric methods, the activity of the fecal enzymes α-glucosidase, α-galactosidase, β-glucosidase, β-galactosidase, and β-glucuronidase was determined. The protocols used in the study were based on the reaction of α-glucosidase, β-glucosidase, α-galactosidase, β-galactosidase, and β-glucuronidase with 4-nitrophenyl α-D-glucopyranoside (TCI, Tokyo, Japan), 4-nitrophenyl β-D-glucopyranoside (TCI, Tokyo, Japan), 4-nitrophenyl α-D-galactopyranoside (TCI, Tokyo, Japan), 4-nitrophenyl β-D-galactopyranoside (TCI, Tokyo, Japan), and 4-nitrophenyl β-D-glucuronide (TCI, Tokyo, Japan), respectively. The used substrates were specific to the respective enzymes present in the fecal sample. The reaction mixture contained 0.5 mL of phosphate buffer (pH = 7, 0.02 M), 0.05 mL of substrate solution (20 mM), and 0.25 mL of sample. Incubation was performed at 37 °C for 15 min (α-glucosidase, α-galactosidase and β-glucuronidase) or 60 min (β-glucosidase and β-galactosidase). An observed hue shift in the sample to yellow proved the reaction had taken place. The intensity of the color was directly proportional to the amount of p-nitrophenol released. The reactions were inhibited using 0.25 M sodium carbonate after the given reaction time. The absorbance of the samples was measured using the spectrophotometer Rayleigh UV-2601 (BFRL, Beijing, China) at a wavelength of λ = 400 nm. The unit of enzyme activity refers to the amount of p-nitrophenol (expressed in µM) which was released during 1 h of reaction for 1 mg of protein in 1 mL of sample [µMh·mg−1]. ## 2.3. Concentration of Lactic Acid, SCFAs and BCFAs Prior to the HPLC analysis the samples were prepared according to the following protocol. Firstly, 0.5 g of sample was suspended in 3 mL of demineralized water and vortexed thoroughly (Vortex RS-VA 10, Phoenix Instrument, Garbsen, Germany). Afterwards, samples were centrifuged (12,000 rpm, Time = 20 min, Centrifuge MPW-251, MPW, Poland) and the supernatant was filtered (0.22 µm filters, ALWSCI Technologies, Shaoxing, China) and transferred to the sterile autosampler vials. HPLC with the Surveyor liquid chromatography system (Termo Scientifc, Waltham, MA, USA) was employed in the experiment. The following parameters of the process were used: Aminex HPX-87H column (300 × 7.8 mm), UV detector, 0.005 mL−1 sulphuric acid as eluent, flow rate 0.6 μLmin−1, single sample analysis time 40 min. ## 2.4. Statistical Analysis The normality of the distribution of variables was examined with the Shapiro–Wilk test, whereas the homogeneity of variances was assessed with Bartlett’s test. After the confirmation of normality and equal variance, the results were analyzed using the one-way ANOVA test and Tukey’s post hoc test. Python was used for statistical testing, where the p-value of 0.05 was considered significant. The data is presented in a mean ± standard deviation (SD) format. ## 3.1. Analysis of Metabolites The main objective of the analysis was to determine the influence of excess weight on the concentrations of lactic acid, SCFAs, and BCFAs, as described in the previous section. The primary hypothesis was that the obese subjects would have altered concentrations of fatty acids due to the alterations in intestinal microbiota commonly seen in such patients. Specifically, it was suspected that the concentrations of lactic acid and SCFAs (formic, acetic, propionic, butyric, and valeric acids) would be lower in obese patients, while the concentrations of BCFAs (isobutyric and isovaleric acids) might be higher. It was determined that the concentrations of lactic acid and most of the SCFAs were significantly higher ($p \leq 0.05$) in the group of children of normal weight, whereas the concentration of BCFAs was lower (Figure 1). In the case of lactic acid, the concentration in the obese subjects was $24.9\%$ lower in comparison to the children of normal weight (Table 2). Similarly, the concentrations of formic and butyric acids were lower by $27.3\%$ and $29.0\%$, respectively, which was a clear confirmation of the hypothesis. The concentrations of acetic, propionic, and valeric acids also maintained that trend, as they were lower in the obese children by, respectively, $12.7\%$, $12.2\%$, and $10.1\%$. Conversely, the concentration of isobutyric acid was noticeably elevated in the group of obese children ($15.3\%$ higher in comparison to the children of normal weight). In the case of isovaleric acid, the difference in concentration was not significant; however, it was slightly higher in the group of obese children (Table 2). In addition to the main objective of the study, the differences in fatty acid profiles were also assessed between male and female participants to evaluate whether sex could possibly have an influence. According to the obtained results for the total population of study participants, the differences in concentrations between the tested groups for lactic, formic, and isobutyric acids were statistically significant ($p \leq 0.05$). As portrayed in Table 3, the concentration of lactic and valeric acids were significantly higher in the case of males, by $24.3\%$ and $35.1\%$ ($p \leq 0.05$), respectively. On the contrary, the concentrations of formic, isobutyric, and isovaleric acids were lower in the male group by $18.5\%$, $21.4\%$ and $19.1\%$ ($p \leq 0.05$), respectively. In the case of acetic, propionic, and butyric acids, the concentrations were not significantly different between groups. When the comparison was made within the group of children of normal weight and with obese children, the differences became even more apparent, with significantly increased mean concentrations of formic, acetic, propionic, butyric, valeric, isobutyric, and isovaleric acids in females by $61.9\%$, $19.7\%$, $40.9\%$, $22.8\%$, $34.3\%$, $51.7\%$, and $30.6\%$, respectively (Figure 2A). The opposite was observed in the case of lactic acid, which was lower in the group of females of normal weight by $18.7\%$. On the contrary, the gender differences within the group of obese children were noticeably less significant and more diverse (Figure 2B). Concentrations of lactic acetic, propionic, and valeric acids were increased in the male children by $35.9\%$, $19.4\%$, and $60.1\%$, respectively. The opposite was observed for formic, isobutyric, and isovaleric acids, where concentrations were decreased in male children by $12.2\%$, $21.6\%$, and $21.6\%$, respectively. Concentrations of acetic and butyric acids were not significantly different between males and females in the tested group. ## 3.2. Activities of Fecal Enzymes The aim of this experiment was to evaluate whether the activity of fecal enzymes differed significantly between the group of obese children and the group of children of normal weight. The hypothesis was similar to the one described in the previous section, given that dietary habits and alterations in gut microbiota can potentially influence the enzymatic activity of bacteria in the large intestine. Accordingly, it was suspected that activity of fecal enzymes closely related to the presence of less desired microbiota (associated with obesity) would cause increased activity of potentially harmful enzymes such as β-glucosidase and β-glucuronidase. Correspondingly to the concentrations of fatty acids, the calculated activity of fecal enzymes was in most cases significantly different ($p \leq 0.05$) in the group of children of normal weight when compared to the group of obese children (Figure 3). In the case of α-glucosidase and α-galactosidase, the activity in the obese subjects was, respectively, $25.8\%$ and $35.1\%$ lower than in the case of children with normal weight (Table 4). In the case of potentially harmful enzymes (β-glucosidase and β-glucuronidase), the activity was influenced differently. The activity of β-glucuronidase was significantly higher in the group of obese children (by $16.9\%$), whereas the activity of β-glucosidase was only slightly lower; however, the result was not statistically different ($p \leq 0.05$). Similarly, the activity of β-galactosidase was not affected by the BMI of the children (Table 4). Additionally, the comparison of enzymatic activity was conducted between male and female children (Figure 4). Surprisingly, the results for the total population of study participants (displayed in Table 5) demonstrated no significant differences between the groups, which suggests that the relationship between the enzymatic activity and metabolic state of patients is different from in the case of fatty acids. On the other hand, when the same comparison is made between the groups of obese male and female children, and between normal-weight male and female children (Figure 4A,B), it becomes apparent that there are noticeable gender differences in the group of children of normal weight in the case of α-glucosidase and α-galactosidase (both are increased in the female group by 9.2 and $12.8\%$, respectively). The group of obese children showed no differentiation between genders (Figure 4B). The mean values for the activity of each enzyme in the group of overweight children and the control group of children of normal weight indicate different patterns of enzyme activity. In the overweight children group, α-glucosidase had the highest mean value (11,453 μMh/mg), followed by α-galactosidase (14,115 μMh/mg) and β-glucuronidase (10,177 μMh/mg). The lowest mean value was observed for β-galactosidase [4103]. The ratio of α-glucosidase to β-galactosidase in this group was 2.79:1. In the control group of children of normal weight, α-galactosidase had the highest mean value (17,474 μMh/mg), followed by α-glucosidase (13,015 μMh/mg) and β-glucuronidase (7674 μMh/mg). The lowest mean value was observed for β-glucosidase (2845 μMh/mg). The ratio of α-glucosidase to β-glucosidase in this group was 4.56:1. Comparing the ratios of the enzymes within each group, it appears that the group of overweight children had a lower ratio of α-glucosidase to β-galactosidase compared to the control group. This suggests that the group of overweight children might have had a lower carbohydrate digestion rate and a lower capacity for breaking down galactose, which could contribute to their overweight status. Furthermore, the ratios between the two groups show that the control group had a higher ratio of α-glucosidase to β-glucosidase (4.56:1) compared to the group of overweight children (2.79:1). This indicates that the control group might have had a higher carbohydrate digestion rate, which could help to maintain their normal weight. In the overweight children, the ratio of β-glucuronidase to α-glucosidase was approximately 0.9, while in the control group it was approximately 0.6. This suggests that the activity of β-glucuronidase is relatively higher compared to α-glucosidase in overweight children. Similarly, the ratio of β-glucuronidase to β-glucosidase was approximately 2.9 in the overweight children and 2.7 in the control group, indicating that the activity of β-glucuronidase is also relatively higher compared to β-glucosidase in both groups. In accordance, the ratios of β-glucuronidase to α-galactosidase and β-galactosidase were lower in the overweight children compared to the control group. The ratio of β-glucuronidase to α-galactosidase was approximately 0.7 in the overweight children and 0.6 in the control group, while the ratio of β-glucuronidase to β-galactosidase was approximately 2.5 in the overweight children and 2.1 in the control group. When comparing the ratios of all enzymes between overweight female children and normal-weight female children (Figure 4A,B), it appears that the normal-weight female children had a higher mean activity for each of the five enzymes compared to the overweight female children. For example, the mean activity of α-glucosidase in normal-weight female children was 13,542 μMh/mg, which was higher than the mean activity of 9625 μMh/mg in overweight female children. This pattern can also be seen in the other enzymes. When comparing the ratios of all enzymes between overweight male children and normal-weight male children, it appears that the normal-weight male children had a higher mean activity for each of the five enzymes compared to the overweight male children. For example, the mean activity of α-glucosidase in normal-weight male children was 12,489 μMh/mg, which was higher than the mean activity of 9770 μMh/mg in overweight male children. When comparing the ratios of α-galactosidase in overweight female children and normal weight female children, it appears that the normal-weight female children had a higher mean activity of the enzyme compared to the overweight female children. The mean activity of α-galactosidase in normal-weight female children was 18,506 μMh/mg, which was higher than the mean activity of 11,425 μMh/mg in overweight female children. When comparing the ratios of β-galactosidase in overweight female children and normal-weight female children, it appears that the normal-weight female children had a higher mean activity of the enzyme compared to the overweight female children. The mean activity of β-galactosidase in normal-weight female children was 3551 μMh/mg, which was lower than the mean activity of 3725 μMh/mg in overweight female children. When comparing the ratios of β-glucuronidase in overweight female children and normal-weight female children, it appears that the normal weight female children had a slightly higher mean activity of the enzyme compared to the overweight female children. The mean activity of β-glucuronidase in normal-weight female children was 7610 μMh/mg, which was lower the mean activity of 9111 μMh/mg in overweight female children. A similar pattern can be seen when comparing the ratios of the enzymes in overweight male children and normal-weight male children. In general, normal-weight male children have higher mean activities of the neutral fecal enzymes and lower activities of potentially harmful fecal enzymes compared to overweight male children. When comparing the ratios of the five enzymes between male and female children of normal weight (Figure 4A), it appears that male children had lower mean activities of α-glucosidase and β-glucosidase compared to female children. The mean activity of α-glucosidase in male children of normal weight was 12,489 μMh/mg, which was lower than the mean activity of 13,542 μMh/mg in female children of normal weight. Furthermore, the mean activity of β-glucosidase in male children of normal weight was 2953 μMh/mg, which was higher than the mean activity of 2737 μMh/mg in female children of normal weight. However, when it comes to α-galactosidase and β-galactosidase, female children of normal weight had higher mean activities of the enzymes compared to male children of normal weight. The mean activity of α-galactosidase in female children of normal weight was 18,506 μMh/mg, which was higher than the mean activity of 16,442 μMh/mg in male children of normal weight. The mean activity of β-galactosidase in female children of normal weight was 3551 μMh/mg, which was lower than the mean activity of 3705 μMh/mg in male children of normal weight. In terms of β-glucuronidase, there was no significant difference in the mean activity between male and female children of normal weight. The mean activity of β-glucuronidase in male children of normal weight was 7739 μMh/mg, which was only slightly higher than the mean activity of 7610 μMh/mg in female children of normal weight. When comparing the ratios of the five enzymes between overweight male and overweight female children, it appears that there is little difference in mean activities between the two groups (Figure 4B). ## 3.3. Correlation Analysis In order to obtain further insight and explore possible trends, different heatmaps were created to determine the possible correlations between tested concentrations of fatty acids and activities of fecal enzymes (Figure 5, Figure 6 and Figure 7). As demonstrated in Figure 5, the activity of most enzymes has rather insignificant correlations with the concentrations of fatty acids. Even the most relevant correlations discovered in this study (between α-glucosidase and propionic acid, or β-glucuronidase and isovaleric acid) have relatively low values, which suggests that, while they exist, they are not strong. Significantly stronger correlations were observed between different fatty acids. It was noted that there is relatively strong correlation between SCFAs and equally strong correlation between two tested BCFAs. The strongest correlations were calculated for the concentrations of acetic and propionic acids (Figure 6). Nonetheless, the concentration of lactic acid was also strongly correlated with those of the SCFAs, especially formic, acetic, and propionic acids. Another strong link was found between the concentrations of isovaleric and isobutyric acids, which further confirms that they are likely produced by the bacteria associated with obesity (Figure 6). In the case of enzymatic activity in fecal samples, the strongest correlation was discovered between the activity of α-glucosidase and that of α-galactosidase. Surprisingly, there is likewise a correlation between the activity of α-glucosidase, α-galactosidase, and β-glucuronidase (Figure 7.). It is possible that this is because, in normal-weight patients, the general abundance of bacteria was higher, which resulted in a higher count of bacteria producing β-glucuronidase. Nonetheless, the results clearly indicate that, even in this particular case, the activity of α-glucosidase and α-galactosidase was significantly higher in the group of children with normal weight. In most cases, the outcomes produced may be rationalized by the fact that, due to the BMI disparity, differences in diet and the ratio of probiotic bacteria to less desired genera, the concentrations and/or activity of different metabolites are influenced, favoring either the beneficial or potentially harmful ones. The pH in the large intestine also changes, which can potentially influence the activity of various enzymes. These results support the hypothesis that obesity has significant consequences regarding the production of metabolites in the large intestine of children. ## 4. Discussion In the presented study, the main objective was to evaluate whether obesity has a significant effect on the metabolites present in the large intestine of children, namely lactic acid, SCFAs (formic, acetic, propionic, butyric, and valeric acid), BCFAs (isobutyric and isovaleric acid), and selected fecal enzymes (α-glucosidase, β-glucosidase, α-galactosidase, β-galactosidase, β-glucuronidase). Additionally, these parameters were also compared by gender. The obtained results confirmed the stated hypothesis, that obese children have noticeably different activity of fecal enzymes and profiles of fatty acids. The observed disparity is mostly negative, as obese children when compared to children of normal weight had increased concentrations of BCFAs and increased activity of potentially harmful enzymes such as β-glucosidase and β-glucuronidase. Contrastingly, children of normal weight exhibited significantly increased concentrations of lactic acid and SCFAs (especially formic and butyric acids). Nonetheless, there was no linear correlation found between the increase in obesity status (value of BMI) and shifts in the activity of fecal enzymes and concentrations of investigated metabolites. The changes were observed only among two compared groups and among certain individuals, where participants with the highest BMI also exhibited extreme values for tested metabolites. This finding might suggest that other factors are very likely also affecting the state of metabolic activity in the large intestine. Furthermore, the activity of beneficial fecal enzymes (α-glucosidase and α-galactosidase) was drastically higher when compared to the group of obese children. Such an observation should be considered positive given that these enzymes are beneficial for the host. α-glucosidase allows individuals to digest more fiber, thus producing more SCFAs, while α-galactosidase contributes to the digestion of dairy products which are of high importance for developing children. In comparison, significantly elevated activity of β-glucuronidase may promote the formation of colon cancer [24,25,26]. Moreover, β-glucuronidase (similarly to β-glucosidase) has the ability to convert heterocyclic aromatic amines, polycyclic aromatic hydrocarbons, and some bile acids into harmful compounds (the products of such conversion are carcinogenic aglycons) [27,28,29]. Furthermore, it was found by Li et al. [ 30] that by inhibition of β-glucosidase it is possible to increase the effectiveness of chemotherapy against colorectal cancer and even to suppress the growth of cancer. Another important aspect to consider is that, although α-glucosidase is a beneficial fecal enzyme, the α-glucosidase inhibitors are the major antidiabetic agents. A-glucosidase inhibitors are a class of drugs that are used to manage hyperglycemia, or high blood sugar levels, in people with type 2 diabetes. These drugs work by blocking the action of the α-glucosidase enzyme in the gut, which is responsible for breaking down carbohydrates into simple sugars. This results in a slower digestion of carbohydrates and a slower release of glucose into the bloodstream. Studies have shown that α-glucosidase inhibitors can have an effect on the gut microbiota and fecal enzyme activity [31,32,33]. In general, these drugs have been shown to cause alterations in the gut microbial community, including changes in the abundance of certain bacterial species and the overall diversity of the gut microbiome. In terms of fecal enzyme activity, α-glucosidase inhibitors can decrease the production of certain enzymes, including α-glucosidase, which is the target of these drugs. This decrease in enzyme activity can affect the ability of the gut microbiome to digest and utilize carbohydrates, potentially leading to changes in the composition of the gut microbiota. It is important to note that the effects of α-glucosidase inhibitors on the gut microbiota and fecal enzyme activity may vary between individuals, and more research is needed to fully understand the implications of these changes. Additionally, the effects of these drugs on the gut microbiome may depend on the dose and duration of treatment, as well as other factors such as diet and lifestyle. Overall, the effects of α-glucosidase inhibitors on the gut microbiota and fecal enzyme activity are an area of active research, and it is important to monitor and understand these effects in order to optimize the use of these drugs in managing hyperglycemia in type 2 diabetes. Further studies are needed to determine the long-term implications of these changes on human health, and to develop strategies for mitigating any negative effects. It is also important to note that elevated glucose levels in the bloodstream can contribute to the development of insulin resistance, which is a hallmark of type 2 diabetes [34]. High α-glucosidase and β-glucosidase activity can increase glucose levels by breaking down complex carbohydrates into simple sugars that are rapidly absorbed into the bloodstream. In response to elevated glucose levels, the pancreas secretes insulin to bring glucose levels back to normal. However, if this process occurs repeatedly over time, the cells in the body may become resistant to insulin, leading to higher glucose levels that are harder to control [35]. Another important aspect to consider is the role of incretins in regulating glucose levels. Incretins are hormones that are released in response to food intake and stimulate insulin secretion to help regulate glucose homeostasis by increasing insulin secretion and decreasing glucagon secretion in response to a meal. High α-glucosidase activity can increase incretin secretion and enhance insulin secretion, leading to improved glucose control [36]. This enzyme is also important for the metabolism of dietary fibers, which are indigestible carbohydrates that are fermented in the large intestine by the gut microbiota. Increased activity of these enzymes can lead to an increase in the production of short-chain fatty acids (SCFAs), which can stimulate the release of incretins such as glucagon-like peptide 1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP) [37,38]. This highlights the interplay between fecal enzyme activity, glucose levels, insulin secretion, and incretin levels, and underscores the importance of considering all these factors when trying to understand and manage glucose metabolism. It is also worth noting that fecal enzyme activity can vary between individuals and can be influenced by a variety of factors, including diet, genetics, and gut microbiome composition. For example, diets high in complex carbohydrates can increase the demand for α-glucosidase and β-glucosidase, while diets high in simple sugars can lead to decreased demand for these enzymes. Understanding the factors that influence fecal enzyme activity can help us better understand the factors that contribute to glucose metabolism and inform the development of new therapeutic approaches for metabolic disorders. In summary, the effects of fecal enzyme activity on glucose, insulin, and incretin levels are complex and multifaceted, and a better understanding of these effects can induce the development of new approaches to manage glucose metabolism and metabolic disorders. An interesting finding was discovered after a comparison of profiles of metabolites by gender. These results clearly indicated that fatty acids were affected by gender, in several cases by a significant amount in both groups of children (obese and with healthy weight). Nonetheless, this phenomenon was observed to a lesser extent in the case of the activity of fecal enzymes, where differences between genders were found only within group of children of normal weight. Furthermore, the in-depth analysis of the results suggested that there may be some differences in the ratios of fecal enzymes between male and female children, with male children generally having higher mean activities of α-glucosidase and β-glucosidase, while female children generally having higher mean activities of α-galactosidase and β-galactosidase. However, these differences may not be significant, and further research is needed to fully understand the impact of gender on fecal enzyme activities. Nonetheless, when comparing the ratios of the enzyme activities within each group, it is important to keep in mind that the ratios will vary depending on the specific activity of each enzyme. The ratios can provide insight into the relative proportions of the different enzymes present in each group, but they should not be used to make direct comparisons between the groups. When comparing the enzyme activity ratios between the two groups, it is important to note that the control group had higher levels of all five enzymes studied, which suggests that the gut microbiota in the control group may be more diverse and healthier compared to the gut microbiota in the overweight group. In conclusion, the results of this study suggest that there are differences in the gut microbiota and fecal enzyme activity levels between overweight children and children of normal weight, but further research is needed to better understand the underlying mechanisms and the potential implications of these differences for health and disease. All these differences can be explained by shifts of balance in the gut microbiota or changes in its diversity caused by unhealthy (or gender-influenced) diet or associated with the occurrence of obesity, which can have a major impact on the secretion of active compounds such as lactic acid, SCFAs, and BCFAs and the activity of fecal enzymes [14,39,40]. However, further studies on the subject are necessary to assess the differences in dietary habits between children of different genders, and to link these habits with the occurrence of specific strains or makeups of microbiota present in the large intestine. It may be suspected that specific genders, even at young age, tend to follow a different dietary trend either imposed by parents, society, or their own preferences, which finally lead to differences in microbiota composition. Another factor might be connected with hormonal development or differences in children, which may also influence the composition of fecal microbiota, which was demonstrated in several studies [41,42,43]. Fatty acids are produced in the large intestine as byproducts of the fermentation of dietary fibers and resistant starches by several species of bacteria residing in this environment. As mentioned before, SCFAs have many confirmed positive effects on the human body, while BCFAs are instead associated with negative impacts. Nonetheless, many studies have shown their beneficial effects, i.e., anti-inflammatory, anticarcinogenic, and anti-obesity, which suggests that their concentration might play a crucial role in the properties they exhibit [44,45]. Even though the topic of links between the anthropometric status of patients and the metabolites of their microbiota has been raised in the scientific community, to the author’s current knowledge there are not many studies that directly compare these differences. An interesting study by Covian et al. [ 46] revealed quite similar findings to our study. It was shown that concentrations of BCFAs (isobutyrate and isovalerate) were directly correlated with the BMI. The higher the BMI, the higher the BCFA concentration. Nonetheless, contrary to our findings, the overall effect on fatty acids in the above-mentioned study was considered non-significant. Moreover, several studies have indicated the association of specific genera of gut bacteria with either health benefits or the development of diseases. Bacteria from genera Clostridium, Bacteroides, Enterococcus, and Escherichia are usually presented as non-desired microbiota and are related to increased activity of β-glucuronidase (EC 3.2.1.31) and β-glucosidase (EC 3.2.1.21), which both have a significant correlation with the development of colorectal cancer [27,47,48]. In contrast, genera such as Bifidobacterium and Lactobacillus present a positive impact on the enzymatic activity in the gastrointestinal tract. Various studies have demonstrated their anticarcinogenic properties and ability to decrease the activity of carcinogen-metabolizing enzymes [49,50]. First and foremost, these species are also directly connected with the prevalence of obesity. Bacteria from the genera Clostridium, Bacteroides, Enterococcus, and Escherichia are more prevalent in obese patients, while species from the genera Bifidobacterium and Lactobacillus are often considered probiotics and are associated with a healthy and balanced diet rich in fibers and vitamins [8,51,52]. Furthermore, the gut microbiota and its metabolites might be affected by various other factors such as the amount and bioavailability of amino acids in the intestines, different metabolic pathways not necessarily directly linked to the gut, or several diseases such as depression and anorexia [46,53,54,55]. Hence, the studied metabolites can be treated as health markers, but with serious consideration and caution. Further limitations of this study include that the determination of metabolites, both fatty acids and fecal enzymes, was performed in fecal samples, which does not necessarily reflect the actual concentrations and activity of these metabolites in different parts of the colon, but are mainly derived from the end of the gastrointestinal tract. Additionally, the tested metabolites are in constant fluctuation due to their absorption in the large intestine. Therefore, the true concentrations or activities of these metabolites might vary slightly. A further important factor that was not considered in this study was the water content of feces, which could also influence the results. Nonetheless, this problem was partially prevented by the selection process in which children with gastric problems were excluded from the study. In light of these findings and the limitations, further studies on the comparison of profiles of metabolites produced by the gut microbiota in patients with different anthropometric statuses should be conducted to acquire further insight in the field of relationships in the gastrointestinal environment of humans. ## 5. Conclusions In conclusion, the difference between the investigated metabolites present in the fecal samples of children with obesity and with healthy BMIs is considerable. Obese children exhibited significantly higher concentrations of BCFAs, which are associated with gut dysbiosis. Moreover, the results showed that their levels of lactic acid and SCFAs were lower in comparison with the children of healthy weight. Similar findings concerned the activity of fecal enzymes. The activity of beneficial α-glucosidase and α-galactosidase was significantly higher in the group of children of normal weight and lower in the obese children. In comparison, the activity of β-glucuronidase (similarly to β-glucosidase) increased in the group of children with higher BMI. These findings further confirm the negative effects linked with the prevalence of obesity, which can lead to the development of severe diseases. Interestingly, differences were found in the profiles of fatty acids with gender as the division factor. This might designate the direction for future research on either the social level of diet composition for different genders, or on possible differences in the composition of the microbiota. However, gender had no significant implications for the activity of fecal enzymes. Further studies are necessary to shed more light on the vast network of interactions between the gut microbiota, their environment, and anthropometric factors of humans. ## References 1. 1. World Health Organization WHO Methods and Data Sources for Country-Level Causes of Death 2000–2016WHOGeneva, Switzerland2018. WHO Methods and Data Sources for Country-Level Causes of Death 2000–2016 (2018.0) 2. Rodriguez-Martinez A., Zhou B., Sophiea M.K., Bentham J., Paciorek C.J., Iurilli M.L., Carrillo-Larco R.M., E Bennett J., Di Cesare M., Taddei C.. Height and body-mass index trajectories of school-aged children and adolescents from 1985 to 2019 in 200 countries and territories: A pooled analysis of 2181 population-based studies with 65 million participants. Lancet (2020.0) 396 1511-1524. DOI: 10.1016/S0140-6736(20)31859-6 3. Vaamonde J.G., Álvarez-Món M.. Obesidad y sobrepeso. Med. Programa De Form. Médica Contin. Acreditado (2020.0) 13 767-776. DOI: 10.1016/j.med.2020.07.010 4. Obesity and Overweight, June 2021 5. Simmonds M., Llewellyn A., Owen C.G., Woolacott N.. Predicting adult obesity from childhood obesity: A systematic review and meta-analysis. Obes. Rev. (2015.0) 17 95-107. DOI: 10.1111/obr.12334 6. Janssen F., Bardoutsos A., Vidra N.. Obesity Prevalence in the Long-Term Future in 18 European Countries and in the USA. Obes. Facts (2020.0) 13 514-527. DOI: 10.1159/000511023 7. Finkelstein E.A., Khavjou O.A., Thompson H., Trogdon J.G., Pan L., Sherry B., Dietz W.. Obesity and Severe Obesity Forecasts Through 2030. Am. J. Prev. Med. (2012.0) 42 563-570. DOI: 10.1016/j.amepre.2011.10.026 8. Włodarczyk M., Śliżewska K.. Obesity as the 21st Century’s major disease: The role of probiotics and prebiotics in prevention and treatment. Food Biosci. (2021.0) 42 101115. DOI: 10.1016/j.fbio.2021.101115 9. Li W.-Z., Stirling K., Yang J.-J., Zhang L.. Gut microbiota and diabetes: From correlation to causality and mechanism. World J. Diabetes (2020.0) 11 293-308. DOI: 10.4239/wjd.v11.i7.293 10. Brusaferro A., Cozzali R., Orabona C., Biscarini A., Farinelli E., Cavalli E., Grohmann U., Principi N., Esposito S.. Is It Time to Use Probiotics to Prevent or Treat Obesity?. Nutrients (2018.0) 10. DOI: 10.3390/nu10111613 11. Tokarek J., Gadzinowska J., Młynarska E., Franczyk B., Rysz J.. What Is the Role of Gut Microbiota in Obesity Prevalence? A Few Words about Gut Microbiota and Its Association with Obesity and Related Diseases. Microorganisms (2021.0) 10. DOI: 10.3390/microorganisms10010052 12. Pinart M., Dötsch A., Schlicht K., Laudes M., Bouwman J., Forslund S.K., Pischon T., Nimptsch K.. Gut Microbiome Composition in Obese and Non-Obese Persons: A Systematic Review and Meta-Analysis. Nutrients (2021.0) 14. DOI: 10.3390/nu14010012 13. Duan M., Wang Y., Zhang Q., Zou R., Guo M., Zheng H.. Characteristics of gut microbiota in people with obesity. PLoS ONE (2021.0) 16. DOI: 10.1371/journal.pone.0255446 14. Frampton J., Murphy K.G., Frost G., Chambers E.S.. Short-chain fatty acids as potential regulators of skeletal muscle metabolism and function. Nat. Metab. (2020.0) 2 840-848. DOI: 10.1038/s42255-020-0188-7 15. Butel M.-J.. Probiotics, gut microbiota and health. Med. Mal. Infect. (2014.0) 44 1-8. DOI: 10.1016/j.medmal.2013.10.002 16. Markowiak P., Śliżewska K.. Effects of Probiotics, Prebiotics, and Synbiotics on Human Health. Nutrients (2017.0) 9. DOI: 10.3390/nu9091021 17. Palmas V., Pisanu S., Madau V., Casula E., Deledda A., Cusano R., Uva P., Vascellari S., Loviselli A., Manzin A.. Gut microbiota markers associated with obesity and overweight in Italian adults. Sci. Rep. (2021.0) 11 5532. DOI: 10.1038/s41598-021-84928-w 18. Cerdó T., García-Santos J.A., Bermúdez M.G., Campoy C.. The Role of Probiotics and Prebiotics in the Prevention and Treatment of Obesity. Nutrients (2019.0) 11. DOI: 10.3390/nu11030635 19. Sivaprakasam S., Bhutia Y.D., Yang S., Ganapathy V.. Short-Chain Fatty Acid Transporters: Role in Colonic Homeostasis. Comprehensive Physiology (2017.0) Volume 8 299-314. DOI: 10.1002/cphy.c170014 20. Mäkeläinen H., Mäkivuokko H., Salminen S., Rautonen N., Ouwehand A.. The Effects of Polydextrose and Xylitol on Microbial Community and Activity in a 4-Stage Colon Simulator. J. Food Sci. (2007.0) 72 M153-M159. DOI: 10.1111/j.1750-3841.2007.00350.x 21. Rowland I., Gibson G., Heinken A., Scott K., Swann J., Thiele I., Tuohy K.. Gut microbiota functions: Metabolism of nutrients and other food components. Eur. J. Nutr. (2018.0) 57 1-24. DOI: 10.1007/s00394-017-1445-8 22. Rendic S., Guengerich F.P.. Contributions of Human Enzymes in Carcinogen Metabolism. Chem. Res. Toxicol. (2012.0) 25 1316-1383. DOI: 10.1021/tx300132k 23. De Onis M., Onyango A.W., Borghi E., Siyam A., Nishida C., Siekmann J.. Development of a WHO growth reference for school-aged children and adolescents. Bull. World Health Organ. (2007.0) 85 660-667. DOI: 10.2471/BLT.07.043497 24. Mroczyńska M., Galecka M., Szachta P., Kamoda D., Libudzisz Z., Roszak D.. Beta-glucuronidase and Beta-glucosidase activity in stool specimens of children with inflammatory bowel disease. Pol. J. Microbiol. (2013.0) 62 319-325. DOI: 10.33073/pjm-2013-043 25. van den Heuvel E.G.H.M., Wils D., Pasman W.J., Saniez M.-H., Kardinaal A.F.M.. Dietary supplementation of different doses of NUTRIOSE. Eur. J. Nutr. (2005.0) 44 445-451. DOI: 10.1007/s00394-005-0552-0 26. Włodarczyk M., Śliżewska K., Barczyńska R., Kapuśniak J.. Effects of Resistant Dextrin from Potato Starch on the Growth Dynamics of Selected Co-Cultured Strains of Gastrointestinal Bacteria and the Activity of Fecal Enzymes. Nutrients (2022.0) 14. DOI: 10.3390/nu14102158 27. Śliżewska K., Markowiak-Kopeć P., Śliżewska W.. The Role of Probiotics in Cancer Prevention. Cancers (2020.0) 13. DOI: 10.3390/cancers13010020 28. Azcarate-Peril M.A., Sikes M., Bruno-Bárcena J.M.. The intestinal microbiota, gastrointestinal environment and colorectal cancer: A putative role for probiotics in prevention of colorectal cancer?. Am. J. Physiol. Liver Physiol. (2011.0) 301 G401-G424. DOI: 10.1152/ajpgi.00110.2011 29. Zhu Q., Gao R., Wu W., Qin H.. The role of gut microbiota in the pathogenesis of colorectal cancer. Tumor Biol. (2013.0) 34 1285-1300. DOI: 10.1007/s13277-013-0684-4 30. Li Z., Xu D., Tong X., Shan C.. Inhibition of β-glucosidase overcomes gastric cancer chemoresistance through inducing lysosomal dysfunction. Clin. Res. Hepatol. Gastroenterol. (2020.0) 45 101456. DOI: 10.1016/j.clinre.2020.04.020 31. Kyriachenko Y., Falalyeyeva T., Korotkyi O., Molochek N., Kobyliak N.. Crosstalk between Gut Microbiota and Antidiabetic Drug Action. World J. Diabetes (2019.0) 10 154. DOI: 10.4239/wjd.v10.i3.154 32. Zhao Y., Wang Y., Lou H., Shan L.. Alpha-Glucosidase Inhibitors and Risk of Cancer in Patients with Diabetes Mellitus: A Systematic Review and Meta-Analysis. Oncotarget (2017.0) 8 81027. DOI: 10.18632/oncotarget.17515 33. Zhang L., Song P., Zhang X., Metea C., Schleisman M., Karstens L., Leung E., Zhang J., Xu Q., Liu Y.. Alpha-Glucosidase Inhibitors Alter Gut Microbiota and Ameliorate Collagen-Induced Arthritis. Front. Pharmacol. (2020.0) 10 1684. DOI: 10.3389/fphar.2019.01684 34. Ormazabal V., Nair S., Elfeky O., Aguayo C., Salomon C., Zuñiga F.A.. Association between Insulin Resistance and the Development of Cardiovascular Disease. Cardiovasc. Diabetol. (2018.0) 17 122. DOI: 10.1186/s12933-018-0762-4 35. Röder P.V., Wu B., Liu Y., Han W.. Pancreatic Regulation of Glucose Homeostasis. Exp. Mol. Med. (2016.0) 48 e219. DOI: 10.1038/emm.2016.6 36. Joshi S.R., Standl E., Tong N., Shah P., Kalra S., Rathod R.. Therapeutic Potential of α-Glucosidase Inhibitors in Type 2 Diabetes Mellitus: An Evidence-Based Review. Expert Opin. Pharmacother. (2015.0) 16 1959-1981. DOI: 10.1517/14656566.2015.1070827 37. Pais R., Gribble F.M., Reimann F.. Stimulation of Incretin Secreting Cells. Ther. Adv. Endocrinol. Metab. (2016.0) 7 24-42. DOI: 10.1177/2042018815618177 38. Yan J., Zhao J., Yang R., Zhao W.. Bioactive Peptides with Antidiabetic Properties: A Review. Int. J. Food Sci. Technol. (2019.0) 54 1909-1919. DOI: 10.1111/ijfs.14090 39. Peredo-Lovillo A., Romero-Luna H., Jiménez-Fernández M.. Health promoting microbial metabolites produced by gut microbiota after prebiotics metabolism. Food Res. Int. (2020.0) 136 109473. DOI: 10.1016/j.foodres.2020.109473 40. Gomes A.C., Hoffmann C., Mota J.F.. Gut microbiota is associated with adiposity markers and probiotics may impact specific genera. Eur. J. Nutr. (2019.0) 59 1751-1762. DOI: 10.1007/s00394-019-02034-0 41. Martin A.M., Sun E.W., Rogers G.B., Keating D.J.. The Influence of the Gut Microbiome on Host Metabolism through the Regulation of Gut Hormone Release. Front. Physiol. (2019.0) 10 428. DOI: 10.3389/fphys.2019.00428 42. Teitelbaum A.A., Gareau M.G., Jury J., Yang P.C., Perdue M.H.. Chronic Peripheral Administration of Corticotropin-Releasing Factor Causes Colonic Barrier Dysfunction Similar to Psychological Stress. Am. J. Physiol. Liver Physiol. (2008.0) 295 G452-G459. DOI: 10.1152/ajpgi.90210.2008 43. Song Q., Zhang X.. The Role of Gut–Liver Axis in Gut Microbiome Dysbiosis Associated NAFLD and NAFLD-HCC. Biomedicines (2022.0) 10. DOI: 10.3390/biomedicines10030524 44. Taormina V.M., Unger A.L., Schiksnis M.R., Torres-Gonzalez M., Kraft J.. Branched-Chain Fatty Acids—An Underexplored Class of Dairy-Derived Fatty Acids. Nutrients (2020.0) 12. DOI: 10.3390/nu12092875 45. Mika A., Stepnowski P., Kaska L., Proczko M., Wisniewski P., Sledzinski M., Sledzinski T.. A comprehensive study of serum odd- and branched-chain fatty acids in patients with excess weight. Obesity (2016.0) 24 1669-1676. DOI: 10.1002/oby.21560 46. Rios-Covian D., González S., Nogacka A.M., Arboleya S., Salazar N., Gueimonde M., de Los Reyes-Gavilán C.G.. An Overview on Fecal Branched Short-Chain Fatty Acids Along Human Life and as Related With Body Mass Index: Associated Dietary and Anthropometric Factors. Front. Microbiol. (2020.0) 11 973. DOI: 10.3389/fmicb.2020.00973 47. Singh R.K., Chang H.-W., Yan D., Lee K.M., Ucmak D., Wong K., Abrouk M., Farahnik B., Nakamura M., Zhu T.H.. Influence of diet on the gut microbiome and implications for human health. J. Transl. Med. (2017.0) 15 73. DOI: 10.1186/s12967-017-1175-y 48. Dabek M., McCrae S.I., Stevens V.J., Duncan S., Louis P.. Distribution of Î²-glucosidase and Î²-glucuronidase activity and of Î²-glucuronidase gene gus in human colonic bacteria. FEMS Microbiol. Ecol. (2008.0) 66 487-495. DOI: 10.1111/j.1574-6941.2008.00520.x 49. Flint H.J., Scott K.P., Duncan S.H., Louis P., Forano E.. Microbial degradation of complex carbohydrates in the gut. Gut Microbes (2012.0) 3 289-306. DOI: 10.4161/gmic.19897 50. Bedada T.L., Feto T.K., Awoke K.S., Garedew A.D., Yifat F.T., Birri D.J.. Probiotics for cancer alternative prevention and treatment. Biomed. Pharmacother. (2020.0) 129 110409. DOI: 10.1016/j.biopha.2020.110409 51. Kasai C., Sugimoto K., Moritani I., Tanaka J., Oya Y., Inoue H., Tameda M., Shiraki K., Ito M., Takei Y.. Comparison of the gut microbiota composition between obese and non-obese individuals in a Japanese population, as analyzed by terminal restriction fragment length polymorphism and next-generation sequencing. BMC Gastroenterol. (2015.0) 15. DOI: 10.1186/s12876-015-0330-2 52. Marchesi J.R., Adams D.H., Fava F., Hermes G.D.A., Hirschfield G.M., Hold G., Quraishi M.N., Kinross J., Smidt H., Tuohy K.M.. The gut microbiota and host health: A new clinical frontier. Gut (2016.0) 65 330-339. DOI: 10.1136/gutjnl-2015-309990 53. Szczesniak O., Hestad K.A., Hanssen J.F., Rudi K.. Isovaleric acid in stool correlates with human depression. Nutr. Neurosci. (2015.0) 19 279-283. DOI: 10.1179/1476830515Y.0000000007 54. Skonieczna-Żydecka K., Grochans E., Maciejewska D., Szkup M., Schneider-Matyka D., Jurczak A., Łoniewski I., Kaczmarczyk M., Marlicz W., Czerwińska-Rogowska M.. Faecal Short Chain Fatty Acids Profile is Changed in Polish Depressive Women. Nutrients (2018.0) 10. DOI: 10.3390/nu10121939 55. de Clercq N.C., Frissen M.N., Davids M., Groen A.K., Nieuwdorp M.. Weight Gain after Fecal Microbiota Transplantation in a Patient with Recurrent Underweight following Clinical Recovery from Anorexia Nervosa. Psychother. Psychosom. (2019.0) 88 58-60. DOI: 10.1159/000495044
--- title: Clinical Profiles and Prognoses of Adult Patients with Full-Frequency Sudden Sensorineural Hearing Loss in Combination Therapy authors: - Yuanping Zhu - Sihai He - Kang Liao - Meihua Li - Zhibin Zhao - Hongyan Jiang journal: Journal of Clinical Medicine year: 2023 pmcid: PMC9966669 doi: 10.3390/jcm12041478 license: CC BY 4.0 --- # Clinical Profiles and Prognoses of Adult Patients with Full-Frequency Sudden Sensorineural Hearing Loss in Combination Therapy ## Abstract We aimed to characterize the clinical profiles and short-term outcomes of adult patients with full-frequency idiopathic sudden sensorineural hearing loss (ISSNHL) treated uniformly with combination therapy, and to determine the prognostic predictors for the combination therapy. A total of 131 eligible cases hospitalized in our department from January 2018 to June 2021 were retrospectively reviewed. All enrolled cases received a standardized combination therapy employing intravenous methylprednisolone, batroxobin, and Ginkgo biloba extract during the 12 days of hospitalization. The clinical and audiometric profiles were compared between recovered patients and their unrecovered counterparts. The overall recovery rate was $57.3\%$ in the study. Accompanying vertigo (odds ratio = 0.360, $$p \leq 0.006$$) and body mass index (BMI, odds ratio = 1.158, $$p \leq 0.016$$) were two independent predictors of hearing outcomes of the therapy. The male gender and cigarette-smoking history were marginally associated with good hearing prognosis ($$p \leq 0.051$$ and 0.070, respectively). Patients with BMI ≥ 22.4 kg/m2 had a better chance of hearing recovery ($$p \leq 0.02$$). Conclusions: Accompanying vertigo and low BMI (<22.4 kg/m2) were independently associated with poor prognosis for full-frequency ISSNHL in combination therapy. Male gender and cigarette-smoking history might be considered positive effects on hearing prognosis. ## 1. Introduction Sudden sensorineural hearing loss (SSNHL), defined as a rapid onset of sensorineural hearing loss of no less than 30 dB affecting at least three consecutive frequencies on an audiogram between 250 and 8000 Hz within 72 h [1], is a typical otolaryngological emergency requiring prompt diagnosis and treatment. The prevalence is estimated at 5 to 27 per 100,000 in the United States [1], and is rising in China. A recent investigation estimated that the annual incidence of SSNHL in mainland China is 19 per 100,000 [2]. More than $90\%$ of SSNHL is considered idiopathic with no identifiable causes. Several mechanisms have been proposed to address the pathogenesis of idiopathic SSNHL (ISSNHL), including viral attack, cochlear ischemia, and autoimmune disorders [3]. Corticosteroids are universally accepted and prescribed as the first-line treatment for ISSNHL [1,4,5]. Other therapies, such as hyperbaric oxygen therapy (HBOT), antivirals, vasoactive agents including prostaglandin E1, Ginkgo biloba extract, and nerve growth factors (NGFs), are also selectively used among institutions [6,7,8,9,10]. Unfortunately, none of these therapies has been proven to be effective for all cases. Since ISSNHL is a multifactorial disorder, the ideal management of it requires an individualized treatment strategy. According to the affected frequencies on the audiogram, ISSNHL can be classified as low-frequency (250, 500, and 1000 Hz), high-frequency (2000, 4000, and 8000 Hz), and full-frequency (250, 500, 1000, 2000, 4000, and 8000 Hz), with different underlying etiology [4]. Vascular occlusion has been reckoned to be a significant cause of full-frequency ISSNHL. The cochlea is fed only by a single tenuous end artery and has no collateral circulation. Thus, it is very susceptible to damage by vascular blockage, resulting in cochlear ischemia and full-frequency hearing loss [11]. In China, combination therapy employing corticosteroids, batroxobin, and Ginkgo biloba extract has been recommended as the first-line option in treating full-frequency ISSNHL [4]. Corticosteroids have powerful inhibitory effects on the inflammatory cell-death cascade in ISSNHL. The extract of Ginkgo biloba 761 (EGb 761), containing approximately $24\%$ flavonoid and $6\%$ terpenes lactones, could improve cochlear blood flow by promoting vasodilation. It also acts as an antioxidant to attenuate oxidative stress provoked in ISSNHL. Additionally, batroxobin is a thrombin-like serine protease isolated from the venom of the snakes Bothrops atrox and B. moojeni. It is applied in clinical treatment to the management of thrombotic conditions as a defibrinogenating agent [12]. However, the hearing outcomes and potential prognostic factors of the therapy remain debatable. In this study, we aimed to report the clinical profiles and short-term hearing outcomes of patients with full-frequency ISSNHL receiving combination therapy and to determine the prognostic factors associated with hearing improvements. ## 2.1. Subjects Adult patients diagnosed with unilateral full-frequency ISSNHL admitted to our department from January 2018 to June 2021 were enrolled in the study. The medical records were retrospectively investigated. Eligible subjects must not have had either a prior episode of SSNHL or a history of Ménière’s disease or other cochlear lesions, and must have received prompt treatment within two weeks of symptom onset. Otoscopy and audiometric examinations were performed in all involved patients, including pure tone and speech audiometry, tympanometry, distortion product otoacoustic emission (DPOAE), and auditory brainstem response (ABR), to confirm the diagnosis of ISSNHL. Magnetic resonance imaging (MRI) of the brain and internal auditory canals was also conducted to evaluate the presence of retrocochlear pathology. Those with middle ear diseases or abnormal MRI findings, such as labyrinthine hemorrhage, vestibular schwannoma, or other cerebellopontine angle tumors, were excluded. Patients with comorbidities of hypertension or diabetes, or taking any lipid-lowering drugs, anticoagulants, antiplatelet or fibrinolytic agents, were also excluded. Clinical data of all enrolled patients were collected and analyzed, including age, sex, laterality, the time to initial treatment, the severity of hearing loss, accompanying tinnitus, vertigo, and aural fullness, and cigarette-smoking history. Peripheral blood samples were obtained on admission for basic blood tests and biochemistry studies. ## 2.2. Treatment All enrolled patients received a standardized combination therapy during the 12 days of hospitalization. Methylprednisolone was dripped intravenously at 120 mg/d for the first four days, followed by a 40 mg taper every four days. Up to five shots of alternate-day batroxobin were given intravenously at 10 Bu for the first shot, then reduced to 5 Bu for the following shots. The plasma fibrinogen (Fg) level was closely monitored before each shot. Once the Fg level fell below 1.0 g/L, batroxobin injection was skipped to the next day to avoid the risk of hemorrhage. Additional Ginkgo biloba extract EGb 761 was infused intravenously at 87.5 mg/d. ## 2.3. Audiometric Evaluation Hearing thresholds were determined by measuring air conduction pure tone average (PTA) at 250, 500, 1000, 2000, 4000, and 8000 Hz. According to the revised grading system recently released by World Health Organization (WHO) [13], the severity of hearing loss was graded as mild (20 to <35 dB HL), moderate (35 to <50 dB HL), moderately severe (50 to <65 dB HL), severe (65 to <80 dB HL), profound (80 to <95 dB HL), or complete (≥95 dB HL) based on PTA. Hearing gains were assessed by comparing the initial PTA on admission with the follow-up PTA 2 weeks after discharge. Hearing outcomes were evaluated according to the guideline proposed by the Chinese Society of Otorhinolaryngology-Head and Neck Surgery (CSOHNS) [4]. Complete recovery required a hearing level within 25 dB, or equivalent to that of the unaffected ear. Marked recovery was defined as a hearing gain of 30 dB or more. A hearing gain of 15 to <30 dB was considered partial recovery. Any hearing gain of less than 15 dB was classified as no recovery. The overall recovery rate was calculated as the sum of all complete, marked, and partial recovery rates. For statistical analysis, patients were subclassified into two groups based on their hearing outcomes: one consisting of those with complete, marked, and partial recovery, and the other with no recovery. ## 2.4. Statistical Analysis Statistical analysis was performed using SPSS 22.0 software. The independent samples t-test was carried out to compare the means of metric variables between two groups when complying with a normal distribution. Otherwise, non-parametric tests were used. The Pearson χ2 test was conducted to compare frequencies of categorical variables between different groups. Spearman’s rank correlation coefficient 𝜌 was used when one variable had a continuous normal distribution and the other was non-normally distributed. Kendall’s 𝜏 was calculated to measure correlations between two non-normally distributed variables. In addition, multivariate logistic regression was employed to determine the independent prognostic factors on early hearing outcomes. The receiver operating characteristic (ROC) curve was further used to calculate the optimal cut-off points (also called the Youden Index) in the continuous variables predicting the hearing prognosis. Any p-value less than 0.05 was considered to be of statistical significance. ## 3. Results A total of 131 eligible patients were enrolled in the study, consisting of 74 ($56.5\%$) males and 57 ($43.5\%$) females, aged between 19 and 64 years old, with a median age of 47. Their clinical profiles are presented in Table 1. Over $90\%$ of enrolled patients had severe to complete hearing loss regarding their initial PTA. No adverse events were observed during treatment. Hearing outcomes were presented as follows according to CSOHNS criteria: 9 ($6.9\%$) patients made a complete recovery; 23 ($17.6\%$), a marked recovery; 43 ($32.8\%$), a partial recovery; and 56 ($42.7\%$), no recovery. The overall recovery rate of the enrolled patients was $57.3\%$ in our study. The linear-by-linear association indicated no significant trend for hearing recovery with the alleviating severity of hearing loss (χ2 = 0.593, df = 1, $$p \leq 0.441$$). Univariate analysis demonstrated a higher recovery rate in male patients compared to female counterparts (χ2 = 5.584, df = 1, $$p \leq 0.018$$). Similarly, patients with cigarette-smoking history revealed a better hearing prognosis (χ2 = 4.330, df = 1, $$p \leq 0.037$$). Furthermore, accompanying vertigo significantly reduced the hearing recovery rate (χ2 = 8.134, df = 1, $$p \leq 0.004$$). Other factors, including age, affected side, time to initial treatment, initial PTA, and presence of associated tinnitus and aural fullness seemed to have little impact on the hearing prognosis ($p \leq 0.05$). Nearly $90\%$ of involved patients demonstrated a normal value (2.0 to 4.0 g/L) of baseline plasma fibrinogen levels (Fg1) before the first shot of 10-Bu batroxobin (Figure 1a). The average Fg1 levels in hearing-recovered patients were not significantly different from those in their unrecovered counterparts (T1 = 66.36, n1 = 56; T2 = 65.73, n2 = 75; $$p \leq 0.926$$, Figure 1c). As shown in Figure 1b, the fibrinogen levels measured before the second batroxobin injection (Fg3) fell below the critical value (<1.0 g/L) in $38.7\%$ of patients with recovery and $44.6\%$ without hearing recovery (χ2 = 0.473, df = 1, $$p \leq 0.492$$). The Mann–Whitney U test indicated a similar reduction in the fibrinogen levels between the two groups (T1 = 69.17, n1 = 56; T2 = 63.63, n2 = 75; $$p \leq 0.409$$, Figure 1c). No statistical correlation was found between hearing gains and the reduction of fibrinogen levels (Fg1–Fg3) in this study (Kendall’s τ = −0.019, $$p \leq 0.755$$; Figure 1d). Figure 2a compares average BMI points between patients with opposite hearing outcomes, showing statistically greater points in the recovered group than in the unrecovered one ($t = 2.601$, df = 129, $$p \leq 0.010$$). Furthermore, the BMI points were positively correlated with hearing gains in the presented study (Spearman’s ρ = 0.235, $$p \leq 0.007$$; Figure 2b). Patients’ serum lipid levels are shown in Figure 3. The Shapiro–Wilk test indicated that only triglyceride (TG) concentrations were not normally distributed ($p \leq 0.001$). Recovered patients were similar to their unrecovered counterparts in terms of serum TG levels (T1 = 62.58, n1 = 56; T2 = 68.55, n2 = 75; $$p \leq 0.373$$, Figure 3a). In contrast, the serum concentrations of total cholesterol (TC) and low-density lipoprotein cholesterol (LDL-C) in patients with hearing recovery were significantly higher than in those without recovery ($$p \leq 0.026$$ and 0.037, respectively; Figure 3b,d). There was no statistical difference in serum high-density lipoprotein cholesterol (HDL-C) levels between the two groups (t = −0.378, df = 129, $$p \leq 0.706$$; Figure 3c). Correlation analysis further revealed that higher serum TC concentrations correlated with greater hearing gains (Spearman’s ρ = 0.225, $$p \leq 0.010$$; shown in Figure 4b). Similarly, a significantly positive correlation of serum LDL-C levels was presented with hearing gains (Spearman’s ρ = 0.185, $$p \leq 0.034$$; shown in Figure 4d). In order to determine which noticeable variables independently predict the hearing outcome of the combination therapy, the multivariate logistic regression model was conducted, including parameters that achieved $p \leq 0.05$ in the univariate analysis (gender, presence of associated vertigo, cigarette-smoking, BMI, TC, and LDL-C; details shown in Table 2). The result indicated that accompanying vertigo and BMI were significant predictors of early hearing outcomes of the therapy ($$p \leq 0.006$$ and 0.016, respectively). The male gender and cigarette-smoking history also seemed to be associated with a good hearing prognosis; however, their statistical significance levels were marginal ($$p \leq 0.051$$ and 0.070, respectively). As shown in Figure 5a, the odds of patients with concurrent vertigo achieving hearing recovery were 0.360 times those of their counterparts without vertigo. For each unit increase in patients’ BMI, the chances of hearing recovery were 1.158 times higher. The ROC curve illustrated that the optimal cut-off point (the Youden Index) in BMI was 22.4 km/m2, predicting that patients with BMI ≥ 22.4 km/m2 tended to have hearing recovery after the therapy (shown in Figure 5b). Furthermore, we examined how well these two independent prognostic factors predict hearing outcomes. Based on their status of vertigo (V+, presence of vertigo; V-, absence of vertigo) and BMI (B+, BMI ≥ 22.4 kg/m2; B-, BMI < 22.4 kg/m2), all the enrolled patients were subdivided into four groups: V+B+, V+B-, V-B+, and V-B-. The overall recovery rate reached $75\%$ ($\frac{27}{36}$) in the group of V-B+, but decreased to $62.5\%$ ($\frac{20}{32}$) and $51.5\%$ ($\frac{17}{33}$), respectively, in V-B- and V+B+. The lowest recovery rate of $36.7\%$ ($\frac{11}{30}$) was found in V+B- (shown in Figure 6a). The Pearson χ2 test indicated significant differences among patients with disparate predictors for hearing prognosis (χ2 = 10.632, df = 3, $$p \leq 0.014$$). In parallel with this, V-B+ patients had the greatest hearing gain of 22.5 [36] dB, while the least hearing gain of 9.0 [15] dB was observed in V+B- counterparts ($H = 13.201$, df = 3, $$p \leq 0.004$$; shown in Figure 6b). ## 4. Discussion Several therapeutic options have been tried based on the various mechanisms proposed for the pathogenesis of ISSNHL, but none guarantee effectiveness. In this study, we paid particular attention to the triad combination therapy employing corticosteroid, defibrinogenation agent batroxobin, and Ginkgo biloba extract, which is recommended in China for full-frequency cases [4]. The short-term hearing outcomes and prognostic factors of the treatment were investigated. About $57.3\%$ of the enrolled patients achieved hearing recovery to some extent. Our study’s overall recovery rate makes sense compared to other studies. Sun et al. [ 14] assessed the hearing outcomes one week after the initial treatment and reported a total recovery rate of $68.9\%$ ($\frac{73}{106}$) in adult patients aged between 18 and 65 years. By contrast, Lee et al. [ 15] reported an even lower hearing recovery rate ($40\%$) two to three months after onset in patients with profound hearing loss. The following reasons should be considered to explain the difference in hearing recovery rates among studies. It has been well-documented that hearing recovery negatively correlates with the severity of initial hearing loss [1,16,17,18]. Correspondingly, over $90\%$ ($\frac{118}{131}$) of the enrolled patients in our study suffered from severe to complete deafness before treatment, which reduced the chance of hearing recovery. Additionally, the definition of hearing recovery varies among studies [1,14,15], which makes the comparison of effective rates less meaningful. All enrolled patients in the study were between 19 and 64 years old, with a similar number of men and women. Univariate analysis suggested that gender, cigarette-smoking history, presence of associated vertigo, BMI, and serum TC and LDL-C levels may interfere with early hearing outcomes of the combination therapy, which, with the exception of vertigo, are also certified risk factors for arteriosclerotic cardiovascular disease (ASCVD) [19,20]. The multivariate logistic regression model further confirmed concurrent vertigo and BMI as two independent predictors for hearing prognosis in the study. It also implied that the male gender and cigarette-smoking history were marginally associated with hearing recovery. The majority of patients ($59.5\%$) were middle-aged (46 to 65 years old); thus, age had little impact on the hearing outcomes in this study. Interestingly, a significantly higher hearing recovery rate was found in male patients than in female ones, and there was a more significant proportion of smoking patients in the recovered group. Multivariate analysis further revealed that both the male gender and cigarette-smoking history were marginally associated with a good hearing prognosis. Such findings have seldom been reported in the literature. There was an equal prevalence of ISSNHL between the left and right ears. No significant association was found between hearing recovery and affected sides, which aligns with the literature [16,21]. All the enrolled patients received prompt therapy within 14 d of symptom onset, ensuring the intervals between symptom onset and initial treatment barely affected the hearing outcomes in the study, which is in accordance with preceding reports [16,17]. The presence of vertigo at the time of symptom onset was proven to be an independent predictor for poor hearing prognosis in the presented study, which matches the finding from previous studies. Possible theories include occlusion of the internal auditory artery, labyrinthine membrane rupture, and concurrence of vestibular neuritis. Tinnitus was nearly a universal accompanying complaint among patients, and about half of the patients reported a feeling of ear fullness. There was no significant association of hearing recovery with either tinnitus or aural fullness. All these findings were already well-known [16,22]. Prior studies have stated that high BMI and serum TC concentration correlate with poor hearing recovery in corticosteroid therapy [23,24]. However, contrary findings were observed in our study. Univariate analysis revealed significantly larger BMI and higher TC and LDL-C levels in patients with hearing recovery than in their unrecovered counterparts when treated using the combination regimen. Furthermore, hearing improvement was positively correlated with BMI, TC, and LDL-C. However, the multivariate analysis only identified BMI as an independent predictor for hearing prognosis. We theorize that the relatively small sample size, slight difference, and wide variation weakened the statistical power of the other two variables, TC and LDL-C. The discrepancy unveiled in our study could be partly explained by the cochlear microcirculation thrombosis theory. Theoretically, atherosclerosis, which induces vascular occlusion of coronary and cerebral arteries resulting in angina, myocardial infarction, and stroke, could play a similar role in blocking the terminal arteries supplying the cochlea, thus interrupting cochlear perfusion and eventually resulting in full-frequency ISSNHL [25]. Obesity has been reckoned as an additional ASCVD risk factor by the American Association of Clinical Endocrinologists (AACE). Furthermore, high TC and LDL-C levels are two independent major risk factors for ASCVD [19]. Hence ISSNHL patients with larger BMI or higher TC/LDL-C levels have a greater chance of atherosclerotic thrombogenesis in the cochlea. Fibrinogen has also been found to be essential in developing atherosclerosis [26]. Previous in vivo animal experiments revealed that reducing fibrinogen levels significantly improved cochlear microcirculation and hearing loss [27]. In the presented study, the extra defibrinogenation agent batroxobin was employed in the combination therapy, and could effectively reduce atherosclerosis by remarkably lowering the fibrinogen levels, resulting in cochlear microcirculation improvement. Therefore, the larger the BMI, the greater the possibility of cochlear atherosclerosis, and the more influential the combination therapy. Similar explanations apply to our findings, mentioned above, that the male gender and cigarette-smoking history had a borderline positive impact on hearing outcomes in the study. Both gender and smoking are explicitly listed as risk factors for ASCVD in the latest guidelines for dyslipidemia management proposed by the European Society of Cardiology (ESC) and European Atherosclerosis Society (EAS). Male smokers are at much higher risk for ASCVD than female non-smokers under the same conditions of age, systolic blood pressure, and TC level [20]. The presented study provides further support for the hypothesis that microvascular attack may play a critical role in the pathogenesis of full-frequency ISSNHL. Although this study is limited by its retrospective design, short observational window, and relatively small sample size, it is the first clinical research addressing the correlation of the short-term hearing outcomes of the triad combination therapy for full-frequency ISSNHL with atherosclerotic risk factors in Chinese patients, and suggesting that combination therapy could be recommended for full-frequency ISSNHL patients with a high BMI (≥ 22.4 km/m2). The findings from our study provide preliminary clinical guidance for physicians to make individualized treatment decisions for patients with full-frequency ISSNHL and support the theory of cochlear microcirculation disturbance as a vital cause of ISSNHL. ## 5. Conclusions Our study identified the presence of concurrent vertigo and low BMI (<22.4 kg/m2) as two independent prognostic factors for poor hearing outcomes in combination therapy employing methylprednisolone, batroxobin, and Ginkgo biloba extract. Additionally, male patients and those with cigarette-smoking history may also tend to have better hearing outcomes after combination treatment. ## References 1. Chandrasekhar S.S., Tsai D.B., Schwartz S.R., Bontempo L.J., Faucett E.A., Finestone S.A., Hollingsworth D.B., Kelley D.M., Kmucha S.T., Moonis G.. Clinical practice guideline: Sudden hearing loss (Update). Otolaryngol. Head Neck Surg. (2019) 161 S1-S45. DOI: 10.1177/0194599819859885 2. Xie W., Dai Q., Liu J., Liu Y., Hellstrom S., Duan M.. Analysis of clinical and laboratory findings of idiopathic sudden sensorineural hearing loss. Sci. Rep. (2020) 10 6057. DOI: 10.1038/s41598-020-63046-z 3. Schreiber B.E., Agrup C., Haskard D.O., Luxon L.M.. Sudden sensorineural hearing loss. Lancet (2010) 375 1203-1211. DOI: 10.1016/S0140-6736(09)62071-7 4. Society of Otorhinolaryngology Head and Neck Surgery, Chinese Medical Association [Guideline of diagnosis and treatment of sudden deafness (2015)]. Chin. J. Otorhinolaryngol. Head Neck Surg. (2015) 50 443-447 5. Marx M., Younes E., Chandrasekhar S.S., Ito J., Plontke S., O’Leary S., Sterkers O.. International consensus (ICON) on treatment of sudden sensorineural hearing loss. Eur. Ann. Otorhinolaryngol. Head Neck Dis. (2018) 135 S23-S28. DOI: 10.1016/j.anorl.2017.12.011 6. Liu X.H., Liang F., Jia X.Y., Zhao L., Zhou Y., Yang J.. Hyperbaric oxygen treatment improves hearing level via attenuating TLR4/NF-κB mediated inflammation in sudden sensorineural hearing loss patients. Biomed. Environ. Sci. (2020) 33 331-337. PMID: 32553077 7. Reisser C.H., Weidauer H.. Acta Otolaryngol. (2001) 121 579-584. DOI: 10.1080/000164801316878854 8. Uri N., Doweck I., Cohen-Kerem R., Greenberg E.. Acyclovir in the treatment of idiopathic sudden sensorineural hearing loss. Otolaryngol. Head Neck Surg. (2003) 128 544-549. DOI: 10.1016/S0194-59980300004-4 9. Xi J., Liu H., Wang X.. Prostaglandin E1 effects on CD62p and PAC-1 in patients with sudden sensorineural hearing loss. Thromb. Res. (2020) 188 31-38. DOI: 10.1016/j.thromres.2020.01.023 10. Rauch S.D.. Clinical practice. Idiopathic sudden sensorineural hearing loss. N. Engl. J. Med. (2008) 359 833-840. DOI: 10.1056/NEJMcp0802129 11. Narula T., Rennie C., Watkinson J.C., Clarke R.W.. Idiopathic sudden sensorineural hearing loss. Scott-Brown’s Otorhinolaryngology Head and Neck Surgery: Volume 2: Paediatrics, the Ear, and Skull Base Surgery (2018) 739-751 12. Masuda H., Sato A., Shizuno T., Yokoyama K., Suzuki Y., Tokunaga M., Asahara T.. Batroxobin accelerated tissue repair via neutrophil extracellular trap regulation and defibrinogenation in a murine ischemic hindlimb model. PLoS ONE (2019) 14. DOI: 10.1371/journal.pone.0220898 13. 13. World Health Organization World Report on HearingWorld Health OrganizationGeneva, Switzerland2021. World Report on Hearing (2021) 14. Sun H., Hu J., Mao Z., Ma Z.. Efficacy of combination therapy in adolescent and adult patients with total-deafness sudden sensorineural hearing loss. Acta Otolaryngol. (2019) 139 258-262. DOI: 10.1080/00016489.2018.1554266 15. Lee H.Y., Kim D.K., Park Y.H., Cha W.W., Kim G.J., Lee S.H.. Prognostic factors for profound sudden idiopathic sensorineural hearing loss: A multicenter retrospective study. Eur. Arch. Otorhinolaryngol. (2017) 274 143-149. DOI: 10.1007/s00405-016-4276-y 16. Perez F.N.A., Da C.M.R., Dore S.J.L., Sonzzini R.D.S.L., de Oliveira P.N.. Clinical profile of patients with unilateral sudden sensorineural hearing loss: Correlation with hearing prognosis. Otolaryngol. Head Neck Surg. (2021) 165 563-570. DOI: 10.1177/0194599820986571 17. Shimanuki M.N., Shinden S., Oishi N., Suzuki N., Iwabu K., Kitama T., Kida A., Sakamoto K., Ogawa K.. Early hearing improvement predicts the prognosis of idiopathic sudden sensorineural hearing loss. Eur. Arch. Otorhinolaryngol. (2021) 278 4251-4258. DOI: 10.1007/s00405-020-06532-4 18. Bogaz E.A., Maranhao A.S., Inoue D.P., Suzuki F.A., Penido N.O.. Variables with prognostic value in the onset of idiopathic sudden sensorineural hearing loss. Braz. J. Otorhinolaryngol. (2015) 81 520-526. DOI: 10.1016/j.bjorl.2015.07.012 19. Jellinger P.S., Smith D.A., Mehta A.E., Ganda O., Handelsman Y., Rodbard H.W., Shepherd M.D., Seibel J.A.. American association of clinical endocrinologists’ guidelines for management of dyslipidemia and prevention of atherosclerosis. Endocr. Pract. (2012) 18 1-78. DOI: 10.4158/EP.18.S1.1 20. Mach F., Baigent C., Catapano A.L., Koskinas K.C., Casula M., Badimon L., Chapman M.J., De Backer G.G., Delgado V., Ference B.A.. 2019 ESC/EAS Guidelines for the management of dyslipidaemias: Lipid modification to reduce cardiovascular risk. Eur. Heart J. (2020) 41 111-188. DOI: 10.1093/eurheartj/ehz455 21. Huafeng Y., Hongqin W., Wenna Z., Yuan L., Peng X.. Clinical characteristics and prognosis of elderly patients with idiopathic sudden sensorineural hearing loss. Acta Otolaryngol. (2019) 139 866-869. DOI: 10.1080/00016489.2019.1641218 22. Weiss D., Bocker A.J., Koopmann M., Savvas E., Borowski M., Rudack C.. Predictors of hearing recovery in patients with severe sudden sensorineural hearing loss. J. Otolaryngol. Head Neck Surg. (2017) 46 27. DOI: 10.1186/s40463-017-0207-1 23. Lee J.S., Kim D.H., Lee H.J., Kim H.J., Koo J.W., Choi H.G., Park B., Hong S.K.. Lipid profiles and obesity as potential risk factors of sudden sensorineural hearing loss. PLoS ONE (2015) 10. DOI: 10.1371/journal.pone.0122496 24. Quaranta N., Squeo V., Sangineto M., Graziano G., Sabba C.. High total cholesterol in peripheral blood correlates with poorer hearing recovery in idiopathic sudden sensorineural hearing loss. PLoS ONE (2015) 10. DOI: 10.1371/journal.pone.0133300 25. Kim J.Y., Hong J.Y., Kim D.K.. Association of sudden sensorineural hearing loss with risk of cardiocerebrovascular disease: A study using data from the korea national health insurance service. JAMA Otolaryngol. Head Neck Surg. (2018) 144 129-135. DOI: 10.1001/jamaoto.2017.2569 26. Suckfull M., Wimmer C., Reichel O., Mees K., Schorn K.. Hyperfibrinogenemia as a risk factor for sudden hearing loss. Otol. Neurotol. (2002) 23 309-311. DOI: 10.1097/00129492-200205000-00013 27. Weiss B.G., Bertlich M., Bettag S.A., Desinger H., Ihler F., Canis M.. Drug-induced defibrinogenation as new treatment approach of acute hearing loss in an animal model for inner ear vascular impairment. Otol. Neurotol. (2017) 38 648-654. DOI: 10.1097/MAO.0000000000001400
--- title: 'Assessment of the Effectiveness, Socio-Economic Impact and Implementation of a Digital Solution for Patients with Advanced Chronic Diseases: The ADLIFE Study Protocol' authors: - Borja García-Lorenzo - Ania Gorostiza - Nerea González - Igor Larrañaga - Maider Mateo-Abad - Ana Ortega-Gil - Janika Bloemeke - Oliver Groene - Itziar Vergara - Javier Mar - Sarah N. Lim Choi Keung - Theodoros N. Arvanitis - Rachelle Kaye - Elinor Dahary Halevy - Baraka Nahir - Fritz Arndt - Anne Dichmann Sorknæs - Natassia Kamilla Juul - Mikael Lilja - Marie Holm Sherman - Gokce Banu Laleci Erturkmen - Mustafa Yuksel - Tim Robbins - Ioannis Kyrou - Harpal Randeva - Roma Maguire - Lisa McCann - Morven Miller - Margaret Moore - John Connaghan - Ane Fullaondo - Dolores Verdoy - Esteban de Manuel Keenoy journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9966680 doi: 10.3390/ijerph20043152 license: CC BY 4.0 --- # Assessment of the Effectiveness, Socio-Economic Impact and Implementation of a Digital Solution for Patients with Advanced Chronic Diseases: The ADLIFE Study Protocol ## Abstract Due to population ageing and medical advances, people with advanced chronic diseases (ACD) live longer. Such patients are even more likely to face either temporary or permanent reduced functional reserve, which typically further increases their healthcare resource use and the burden of care on their caregiver(s). Accordingly, these patients and their caregiver(s) may benefit from integrated supportive care provided via digitally supported interventions. This approach may either maintain or improve their quality of life, increase their independence, and optimize the healthcare resource use from early stages. ADLIFE is an EU-funded project, aiming to improve the quality of life of older people with ACD by providing integrated personalized care via a digitally enabled toolbox. Indeed, the ADLIFE toolbox is a digital solution which provides patients, caregivers, and health professionals with digitally enabled, integrated, and personalized care, supporting clinical decisions, and encouraging independence and self-management. Here we present the protocol of the ADLIFE study, which is designed to provide robust scientific evidence on the assessment of the effectiveness, socio-economic, implementation, and technology acceptance aspects of the ADLIFE intervention compared to the current standard of care (SoC) when applied in real-life settings of seven different pilot sites across six countries. A quasi-experimental trial following a multicenter, non-randomized, non-concurrent, unblinded, and controlled design will be implemented. Patients in the intervention group will receive the ADLIFE intervention, while patients in the control group will receive SoC. The assessment of the ADLIFE intervention will be conducted using a mixed-methods approach. ## 1. Rationale Due to population ageing and medical advances, people with chronic disease(s)—including advanced chronic disease(s) (ACDs)—live longer. Patients living with chronic disease may face either temporarily or permanently reduced functionalities, which typically increases their caregiver’s burden and their healthcare resource use [1,2]. Chronic Obstructive Pulmonary Disease (COPD) and Heart Failure (HF) are two of the most prevalent ACDs [3,4]. Both COPD and HF are important causes of morbidity and mortality [3,5] together with high social and economic costs [6,7]. Patients with such conditions may benefit from integrated supportive care provided via digitally supported interventions [8,9]. This approach may either maintain or improve their quality of life, increase their independence, and optimize the healthcare resource use from early stages [1,2]. However, there is limited evidence on the implementation of such approaches. The EU-funded ADLIFE project (H2020, SC1-DTH-11-2019, 875209) aims to meet the current societal and health system challenges in Europe that are created due to the increasing number of persons living with ACD, often with accompanying co-morbidity and polypharmacy. ADLIFE is designed to provide to such patients and their caregiver(s) with digitally enabled, integrated, and personalized care. The ADLIFE toolbox is a digital solution which consists of three Information and Communication Technology (ICT) components: a platform to manage and customize care plans by the interdisciplinary healthcare team with the patient and for the patient, services supporting the clinical decisions, and a platform encouraging independence and self-management of patients and caregivers. The ICT solutions will be scaled in the seven pilot sites from six countries participating in the intervention in an integrated manner with the existing ICT solutions utilized for healthcare systems. The ADLIFE intervention objective is to have an impact for three stakeholders: patients, informal caregivers, and healthcare professionals, across the seven participating international healthcare systems. Thus, the ADLIFE intervention aims at slowing down the patients’ functional deterioration ensuring their quality of life and promoting shared decision making, whilst in parallel reducing the caregiver burden, improving the healthcare professional’s working conditions, and optimizing healthcare resource use. The evaluation of such an intervention is challenging, since multiple dimensions [10,11] arise. Published results have reported notable improvements in some of the primary outcome measures [12,13,14,15], but there is also lack of evidence-based changes in comparison with SoC [16,17,18,19,20]. Therefore, there is a lack of systematic assessments of the effect of personalized integrated care initiatives on patients with ACD and limited consistent evidence on the related impact on conventional and emergency department visits, hospital admissions, and length of hospitalization stay. The aim of this research is to evaluate whether the ADLIFE intervention, when applied in real-life settings, is able to deliver appropriate targeted and timely care for patients living with ACD. Here we present the protocol of the ADLIFE study. Using a mixed-methods approach, this study will provide robust scientific evidence on the effectiveness, socio-economic implementation, and technology acceptance assessment of ADLIFE compared to the SoC, in order to provide the scientific evidence-based supporting the funding decision-making of the ADLIFE intervention. ## 2.1. Study Design The ADLIFE intervention is a quasi-experimental trial following a multicenter, A quasi-experimental, non-randomized, non-concurrent, unblinded, and controlled design. Patients belonging to the intervention group will receive the ADLIFE intervention, while patients in the control group will receive SoC [21]. The SoC description can be found in Supplementary File S1. The trial will be implemented in a real-life setting thus allowing other programs and behavior interventions can be in place. The ADLIFE intervention will be implemented in seven different pilot sites: Basque Country (Osakidetza), United Kingdom (National Health Service Lanarkshire and University Hospitals Coventry and Warwickshire National Health Service Trust), Denmark (Odense University Hospital), Germany (Gesunder Werra-Meißner Kreis), Sweden (Region Jämtland Härjedalen), and Israel (Assuta Ashdod Hospital—Maccabi Healthcare Services Southern Region) involving healthcare professionals, care services, and patients with their caregiver(s). The trial was registered in ClinicalTrials.gov, accessed on October 2022. Number of identification: NCT05575336. The study will be conducted in line with the ethical standards and the applicable European, international, and national law on ethical principles. Ethics approval will be obtained from all pilot sites’ Ethics Committee. Patients and caregivers will provide written informed consent. ## 2.1.1. Non-Randomized, Non-Concurrent, and Unblinded The non-randomized design of the ADLIFE clinical trial lies on the recruitment process, as detailed in Section 2.3. To address potential biases and guarantee comparability between the control and the intervention group caused by the afore-mentioned non-random design, the control group selection will be based on a propensity score matching with the intervention group [22]. First, the probability of all eligible individuals of being assigned to the intervention group will be estimated using the patient variables of age and sex, and the number of emergency room (ER) visits and number of hospital admissions observed during the 9 months prior to the ADLIFE intervention. Then, each intervention patient will be matched to a control subject with a similar propensity of being assigned to intervention. The control group will be retrospectively selected to guarantee that the control group receives usual care. The study will be unblinded, since all stakeholders involved in the intervention group will use the ADLIFE toolbox and will thus be aware of the intervention However, since data will be retrospectively collected from the electronic health records (EHR) for the control and the intervention group, no information bias is expected in the data collection process. The study design and the data collection flow are shown in Figure 1. Data collection is also described in Section 2.7 of this paper. ## 2.2. Study Population The study population consists of patients with ACD, their informal caregiver(s), and their healthcare professionals, meeting the following eligibility criteria. Eligible patients will have to meet the below inclusion criteria:Aged over 55;HF in functional stage III/IV according to the NYHA scale and/or stages C and D of the ACCF/AHA classification. Stable phase (at least two months without decompensation requiring hospital care);And/or COPD GOLD scale > 2 (FEV1 < 50) and/or mMRC ≥ 2 and/or CAT ≥ 10 and/or use of oxygen at home;With or without comorbidities;They are able to provide informed consent;They still live and generally plan on living in their home for the intervention duration;They or their informal caregiver(s) are able to use digital technology, communication tools, and/or networks and have access to a computer, laptop, tablet or smartphone and wifi/internet connection;They or their informal caregiver(s) understand, read, and talk the native language. The informal caregiver will be a person who provides occasional or regular support to the patient needs. Caregivers will be eligible if the patient they care for meets the inclusion criteria and is included in the study. Healthcare professionals will be eligible if they are involved in the included patients’ care, are open to new ways of working, specifically as part of a coordinated and collaborative team, and are also open to the use of new technology. Criteria for patient exclusion include an existing diagnosis of an active malignant neoplastic disease, being in any active list of transplantation, or refusing/unable to sign the informed consent. Patients who have withdrawn their participation from ADLIFE will be excluded; caregivers will not be eligible if the patients they care for meet the exclusion criteria; Healthcare professionals not caring for eligible patients will not be included. ## 2.3. Recruitment The recruitment of healthcare professional, patients and caregivers will last 7 months. Healthcare professionals will be recruited during the first 4 months by each pilot site using a convenience sampling based on their individual profiles. The healthcare professionals will be contacted via email, letter, or face-to-face meeting to provide them with initial information about the ADLIFE intervention. Healthcare professionals consenting to participate will be recruited. Once healthcare professionals have been recruited and trained in the use of the ADLIFE intervention, each pilot site will identify a tentative target patient population fulfilling the eligibility criteria from EHR. After a validation process conducted by the healthcare professionals, a final target population will be defined. The intervention and control group will be recruited from the final target population during the last 2 months. Candidates for the intervention group will be recruited by healthcare professionals or research assistants according to the inclusion/exclusion criteria. They will be contacted by email, post, phone, or face-to-face meetings, and provided with information describing the ADLIFE intervention, their expected role, and their required ICT skills. Patients participating in the intervention will sign an informed consent and will receive ADLIFE intervention training. Patients in the intervention group may appoint an informal caregiver who will participate with them in ADLIFE in order to assist them with ICT management and self-management empowerment of their disease. Informal caregivers will sign an informed consent to participate in the intervention, and a second informed consent to access their corresponding patient’s data. They will receive ADLIFE training together with the corresponding recruited patient they care for. In the event of a participant’s drop-out of the project, their previously collected data will be retained, unless otherwise stated, and analyzed under the intention-to-treat principle. Drop-out reasons will be defined as: (i) death, (ii) not interested in the intervention anymore, (iii) too time-consuming, (iv) technology issues, (v) declaring lack of help from their informal caregiver, (vi) institutionalization, which implies a change of healthcare professional, and (vii) no response. In the case of healthcare professionals, any position change will also be considered as drop-out reason. Informal caregivers will automatically drop-out of the project if their patients do so. After removing the intervention patients from the final target population, a target control population will be identified. Control group patients will be retrospectively selected based on the afore-mentioned matching techniques (see Section 2.1.1.) from the target control population. No signed informed consent will be needed since anonymous data will be extracted from EHR by the pilot sites. Intervention patients withdrawing from the intervention will not be eligible for the control group. The recruitment process is described in Figure 2. ## 2.4. The ADLIFE Intervention and Standard of Care The ADLIFE intervention consists of the deployment and use of the ADLIFE toolbox by patients, informal caregivers, and healthcare professionals in the afore-mentioned pilot settings. The ADLIFE toolbox involves two interconnected platforms. Patients will use the Patient Empowerment Platform (PEP), and healthcare professionals will be assisted by clinical decision support services within the Personalized Care Plan Management Platform (PCPMP). Patients participating in ADLIFE will have a personalized care plan, created in PCPMP, which will be developed and managed together with their healthcare professionals. PCPMP will be used in integration with the clinical sites’ ICT systems to create patient care plans based on each patient’s baseline and most recent clinical data, following clinical evidence. PEP will facilitate the patient’s independence and self-management by presenting their personalized goals, activities, and educational materials, collecting their observations and questionnaires, and providing real-time interventions tailored to the patient’s lifestyle. The main task of patients and their informal caregivers will be to use the PEP as part of their healthcare management process together with their healthcare professionals. The ADLIFE intervention will consider the health-related outcomes relevant for the patient in actual health service planning and evaluation. By identifying the outcome that will be responsive to each measure, professionals and patients will have the chance of reviewing the health-related outcomes, and of jointly choosing the activity, objective or goal that boosts the desired one. The health-related outcomes will be reflected as labels that bind every activity, goal, and/or indicator included in a care plan. The labelling mechanism has been co-created with healthcare professionals and automated to enable health-related outcome tracking over time and over a wide spectrum of patients. The control group follows the SoC according to the pilot site organizations’ criteria. Since pilot sites belong to different health care systems, information on SoC was derived from semi-structured interviews with three stakeholder groups, 5–7 persons in each group [21]. ## 2.5. Outcomes The set of health-related outcomes grouped around relevant domains for the ADLIFE target population was defined following the International Consortium for Health Outcomes Measurement (ICHOM) methodology (see online the Supplementary File S2 for details) [23,24]. The health-related outcome set provides a consensual definition of desired end results for both the outcome-based care planning and for the project evaluation. The framework groups each of the primary and secondary outcomes by the broad health-related outcome to which it is responsive. Figure 3 shows the Health-Related Outcome set for people over 55 years old with severe heart failure and/or COPD. The conceptual data framework identifies the relevant health areas (outer ring).Each health outcome area comprehends multiple dimensions (inner ring). Besides health-related outcomes, implementation-related outcomes including barriers/facilitators related to the implementation process, technology acceptance and adoption, and contextual factors for further exploitation to later scaling-up will also be assessed. To define the variables of interest, a framework for implementation assessment was developed (see Figure 4) which is based on a human, organization, and technology-fit (HOT-fit) framework as a Health Information System evaluation framework [25], and is complemented with the intervention, process, and outer setting of the Consolidated Framework for Implementation Research (CFIR) [26]. Together, these frameworks provide a structured and systematic way to identify constructs influencing the implementation of ADLIFE on different levels. ## 2.5.1. Primary The primary outcome will be the number of ER visits during the 9-month follow-up. The ER visits regularly become a crucial point in healthcare systems when care is poorly coordinated. A high number of ER visits has been associated with functional decline and mortality [27]. In this context, existing evidence suggests that frequent ED users are at-risk patients for whom interventions may improve health outcomes [28,29]. Then, this primary outcome, defined as the main effectiveness outcome, is a proxy of the appropriateness of care in real-life settings for patients with ADC, and consequently, a gain in health-related outcomes [27,28,29]. ER visits will be collected from the EHR. ## 2.5.2. Secondary The secondary outcomes will be assessed on patients, caregivers, and healthcare professionals. Patients will be assessed on:Patient-Reported Outcome Measurements (PROMs):Health-related quality of life (EQ-5D-5L) [30];Mood/emotional health (HADS—Hospital Anxiety and Depression Scale) [31];Activities of daily living (Lawton scale, Barthel Index, Kansas City Cardiomyopathy Questionnaire score, COPD assessment test score) [32,33,34,35];Complexity (Modified Medical Research Council—mMRC—Dyspnea Scale) [36]. Technology acceptance and future adoption of the ADLIFE intervention (the Unified Theory of Acceptance and Use of Technology, UTAUT) [37];Resource use and their associated costs will be also assessed on patients [38] (see data collection guide 2 in Supplementary File S4);Caregiver will be assessed on:Burden (Zarit Burden Interview, ZBI) [39];Mental well-being (Warwick-Edinburgh Mental Wellbeing Scale, WEMWBS) [40].Healthcare professionals will be qualitatively assessed on:Perceived coordination among settings;Quality of the integration of care;Decision making process;Working conditions. All stakeholders will be qualitatively assessed on:Perceived communication;Satisfaction with accessibility, security, and personalized care plans;Barriers/facilitators related to the implementation process. Within the scope of the implementation assessment, contextual factors focusing on local technological, organizational, and human factors, will also be qualitatively assessed before and one year after the start of the intervention in a sub-group of different stakeholders including physicians, general practitioners, nurses, health coordinators, IT staff, and hospital CEOs [41]. ## 2.6. Sample Size The sample size was calculated using the number of ER visits as the main primary outcome. In order to detect an effect size of 0.6 ER visits per year, with a standard deviation of 1.2 with a $5\%$ level of significance, a $90\%$ of statistical power set, assuming a conservative intra-cluster correlation coefficient of 0.06 (each pilot site defined as a cluster) and a drop-out rate of $30\%$, 1692 patients will be required (846 per branch) across pilot sites. For the qualitative approach, a purposeful sampling will be used to select the participants, recruiting those participants who might provide in-depth and detailed information about the ADLIFE intervention. The pre-intervention interviews for the implementation assessment will involve: one medical director, six healthcare professionals (two physicians, two general practitioners, and two nurses), and two IT staff for each pilot site. The post-intervention interviews will be structured in the same way, adding three to six patients and three to six informal caregivers for each site. ## 2.7. Data Collection A data collection guide including a codebook and a template will be provided to pilot sites to conduct the quantitative data collection on socio-demographic, clinical, economic, and usability variables. Variables will be observed at baseline and at endpoint, except for resource use which will be measured during the 9-months before and after the baseline, and except for the usability variables which will be measured at endline. Data collection flows are shown in Figure 1. Data will be collected from three data sources: the EHR, Qualtrics questionnaire for technology acceptance completed by participants, and the on FHIR repository [42]. For the effectiveness and socio-economic evaluation, each pilot site will provide its corresponding dataset to the evaluator site to be merged into a single data space. Data collection guides 1 and 2 in the Supplementary File S3 and S4 show comprehensive variable lists to be collected. For the technology acceptance and adoption, pilot sites will communicate the online questionnaire links to the participants and the data will be collected. A data collection guide for the technology acceptance and adoption evaluation can be found in the Supplementary File S5. For the qualitative work, semi-structured interviews will be conducted on patients, informal caregivers, healthcare professionals, managers, and IT staff. A detailed protocol with the data collection guide and templates to be used for summarizing the data interviews will be circulated across pilot sites. Interviews at each pilot site will be conducted in the corresponding national language, and will preferably take place face-to-face, or virtually if necessary (telephone or videoconference). The interviews will be audio recorded and verbally transcribed to a structured template. For effectiveness assessment, data from interviews will be collected after follow-up period, whereas for implementation assessment a pre-post approach will be followed. ## 2.8. Data Management A robust approach to data protection and data management is adopted prior to any contact with patients or their health data. This approach is described in the Data Management Plan (DMP) [43]. The DMP describes the data protection roadmap for each of the five categories of patient level data (mock healthcare data, training healthcare data, control healthcare data, intervention healthcare data and patient/HCP reported data). At a glance, the DMP presents the data processing and flows that will largely take place within each of the seven pilot sites, to identify patients who match the eligibility criteria for the ADLIFE intervention, and then how these data will be processed including pseudonymization and anonymization steps, before being made available for wider consortium use. This wider use includes the design and development of technical components, and the training and validation of artificial intelligence algorithms. The anonymization process begins with the classification of the sensitivity of those variables that are essential for the purpose of the study. The protection criteria established include the anonymization of direct identifiers using hash algorithms, and the anonymization of indirect identifiers using micro-aggregation and temporal perturbation, reducing the risk of re-identification of the information as a whole to $5\%$. It will therefore be ensured that, within reasonableness and within the procedures adopted, no personal data or datasets can be associated to any person, within a reasonable time. The DMP also focuses on the formal Data Management Plan template published by Horizon 2020. This mostly confirms the intention of the project to make available some open research data at the end of the project, and how it intends to comply with the FAIR principles. The DMP therefore, sets the principles to enable a privacy-by-design approach to data sharing by balancing the benefits of using personal health data with a range of risk-management controls. ## 2.9. Analysis The ADLIFE study will provide the assessment of the effectiveness, socio-economic, implementation, and technology acceptance aspects of the ADLIFE intervention compared to the SoC. The threefold approach is presented in the following subsections. ## 2.9.1. Effectiveness Assessment A mixed assessment strategy will be performed using quantitative and qualitative methods. For the quantitative approach, first a descriptive analysis followed by univariate statistical tests will be conducted. The effect of the intervention will then be assessed by generalized mixed models for longitudinal data. Linear models will be used for continuous variables and logistic models for dichotomous variables. Given the hierarchical structure of the data where patients are nested in pilot sites, the latter will be included as random effect to control the models by the variability across sites. In order to consider the different time of follow-up of each participant, all models will be adjusted by this factor, i.e., the time of follow-up will be included as a covariable. All models will also be adjusted for potentially confounding factors. Since healthcare services data are usually characterized by being discrete, zero-inflated counts and right-skewed, special attention will be paid to the selection of the distribution which best fits the data. In all quantitative analyses, we will use an intention-to-treat approach and set the level of significance at $p \leq 0.05.$ For the qualitative approach, a two-step analysis will be conducted. The pilot sites will first perform a content analysis in their own languages [44], based on a systematic approach to determine trends and patterns within the text, and will organize and identify topics and their relationships. For each identified topic, a comprehensive set of quotes will be transcribed to ensure comparability across regions. Then, in a second step, an aggregate cross-country analysis to assess the validity of the results, by comparing the main results among the pilot sites and the different stakeholders, will be conducted by the evaluation coordinator. The COREQ Checklist (Consolidated criteria for Reporting Qualitative research) [45] for quality assurance will be followed. ## 2.9.2. Implementation Assessment For the quantitative approach technology acceptance and adoption assessment, the questionnaire data collected at two timepoints in the intervention, will be exported from Qualtrics to a statistical software package [46]. Descriptive statistics will be used to summarize the participants’ demographics and core set of UTAUT constructs. To measure the reliability of the UTAUT model’s constructs and form correlations between them, data analysis will be performed using techniques such as structural equation modelling, a multivariate statistical analysis technique that is used to analyze structural relationships and tests the underlying factors and hypotheses. For the qualitative analysis of the contextual factors a similar two-step analysis approach is planned, as described in in the effectiveness assessment in Section 1. This includes a content analysis in the national language using a standardized coding system based on the HOT-fit framework, followed by an aggregated cross-country analysis in English to summarize the main content of all pilot sites according to the HOT-fit framework and support it with respective quotes. The COREQ Checklist [45] for quality assurance will be followed. ## 2.9.3. Socio-Economic Assessment Simulation modelling applying discrete event simulation (DES) [47] technique will be used to represent the natural history of patients with ACD and estimate the long-term socio-economic impact. A common simulation model will be estimated for all pilot sites and then, adapted to each pilot site situation. First, the conceptual model that will rule the simulation model will be agreed and defined with experts. Second, the simulation model will be built up to represent the SoC scenario and validated by contrasting it with the data observed in real life. After being calibrated and validated, in the third step the previously estimated ADLIFE effectiveness will be added to the model in order to also represent the ADLIFE scenario. Finally, for both scenarios, the budget impact analysis (BIA) will be estimated multiplying the resource consumption rate by unit costs, and then forecasted using population projection. Then, the burden of the disease under both scenarios will be obtained over time and the changes in the expenditure of the healthcare system after the adoption of the intervention addressed. ## 3. Discussion This study will conduct a holistic evaluation, not only based on a mixed-methods approach addressing the effectiveness, technology acceptance, and socio-economic impact, but also by involving the key stakeholders: patients, caregivers, and healthcare professionals. In addition, the use of the health outcome framework will shape the scientific evidence compared to the SoC. Study findings will be disseminated through peer-reviewed publication and scientific conferences [48]. However, this study is not free of limitations. The first is related to the recruitment process. The healthcare professionals/research assistants will recruit the intervention participants according to their subjective assessment, which will not allow for the participants’ randomization. Second, the design of the data collection process, where the control period information will be retrospectively collected, will not allow the quality of life to be used as a main outcome. Moreover, thirdly, the existing differences in the ADLIFE intervention and the SoC across the different health systems might lead to a bias of the ADLIFE global effect. To address this twofold concern, the ADLIFE intervention must first be defined as a pragmatic trial [49] of a complex intervention [50]. Pragmatic trials measure effectiveness in routine clinical practice and its design reflects variations across patients. While interventions should be precisely described in Randomized Control Trials (RCT), in pragmatic trials this does not imply that the same interventions are offered to each patient. Indeed, clinicians and patients’ biases are not necessarily considered as a weakness but accepted as part of the physicians and patients’ responses to the intervention, and therefore included in the overall assessment [51]. Pragmatic trials may vary across similar participants, by chance, by practitioner preference, and according to institutional policies. Furthermore, the ADLIFE intervention as a complex intervention implies working closely with local stakeholders, and considering the implementation as an iterative, recursive, and long-term process. The complex interventions allow the implementation to vary across different contexts with no loss of the core intervention components. The standardization of interventions might be rather into the underlying process and functions than on the specific form of the components delivered [51]. In the context of the ADLIFE intervention, the differences across healthcare systems becomes a challenge, but also it becomes a strength when assessing the intervention across diverse clinical real settings. Moreover, the methodological approach offers the chance to control for the afore-mentioned differences when estimating the ADLIFE intervention effect, including the site as random effect for this purpose. ## References 1. Lunney J.R., Lynn J., Foley D.J., Lipson S., Guralnik J.M.. Patterns of functional decline at the end of life. JAMA (2003.0) 289 2387-2392. DOI: 10.1001/jama.289.18.2387 2. Tanuseputro P., Wodchis W.P., Fowler R., Walker P., Bai Y.Q., Bronskill S.E., Manuel D.. The Health Care Cost of Dying: A Population-Based Retrospective Cohort Study of the Last Year of Life in Ontario, Canada. PLoS ONE (2015.0) 10 e0121759. DOI: 10.1371/journal.pone.0121759 3. Groenewegen A., Rutten F.H., Mosterd A., Hoes A.W.. Epidemiology of heart failure. Eur. J. Heart Fail. (2020.0) 22 1342-1356. DOI: 10.1002/ejhf.1858 4. Blanco I., Diego I., Bueno P., Casas-Maldonado F., Miravitlles M.. Geographic distribution of COPD prevalence in the world displayed by Geographic Information System maps. Eur. Respir. J. (2019.0) 54 1900610. DOI: 10.1183/13993003.00610-2019 5. Quaderi S.A., Hurst J.R.. The unmet global burden of COPD. Glob. Health Epidemiol. Genom. (2018.0) 3 e4. DOI: 10.1017/gheg.2018.1 6. Iheanacho I., Zhang S., King D., Rizzo M., Ismaila A.S.. Economic Burden of Chronic Obstructive Pulmonary Disease (COPD): A Systematic Literature Review. Int. J. Chronic Obstr. Pulm. Dis. (2020.0) 15 439-460. DOI: 10.2147/COPD.S234942 7. Shafie A.A., Tan Y.P., Ng C.H.. Systematic review of economic burden of heart failure. Heart Fail. Rev. (2018.0) 23 131-145. DOI: 10.1007/s10741-017-9661-0 8. Gray C.S.. Integrated Care’s New Protagonist: The Expanding Role of Digital Health. Int. J. Integr. Care (2021.0) 21 1. DOI: 10.5334/ijic.6437 9. Yadav L., Gill T.K., Taylor A., Jasper U., De Young J., Visvanathan R., Chehade M.J.. Cocreation of a digital patient health hub to enhance education and person-centred integrated care post hip fracture: A mixed-methods study protocol. BMJ Open (2019.0) 9 e033128. DOI: 10.1136/bmjopen-2019-033128 10. Baxter S., Johnson M., Chambers D., Sutton A., Goyder E., Booth A.. The effects of integrated care: A systematic review of UK and international evidence. BMC Health Serv. Res. (2018.0) 18. DOI: 10.1186/s12913-018-3161-3 11. Allen D., Rixson L.. How has the impact of ‘care pathway technologies’ on service integration in stroke care been measured and what is the strength of the evidence to support their effectiveness in this respect?. Int. J. Evid. Based Healthc. (2008.0) 6 78-110. PMID: 21631815 12. Brännström M., Boman K.. Effects of person-centred and integrated chronic heart failure and palliative home care. PREFER: A randomized controlled study. Eur. J. Heart Fail. (2014.0) 16 1142-1151. DOI: 10.1002/ejhf.151 13. Dorr D.A., Wilcox A.B., Brunker C.P., Burdon R.E., Donnelly S.M.. The Effect of Technology-Supported, Multidisease Care Management on the Mortality and Hospitalization of Seniors. J. Am. Geriatr. Soc. (2008.0) 56 2195-2202. DOI: 10.1111/j.1532-5415.2008.02005.x 14. Hogg W., Lemelin J., Dahrouge S., Liddy C., Armstrong C.D., Legault F., Dalziel B., Zhang W.. Randomized controlled trial of anticipatory and preventive multidisciplinary team care: For complex patients in a community-based primary care setting. Can. Fam. Physician (2009.0) 55 e76-e85. PMID: 20008582 15. Janse B., Huijsman R., De Kuyper R.D.M., Fabbricotti I.N.. The effects of an integrated care intervention for the frail elderly on informal caregivers: A quasi-experimental study. BMC Geriatr. (2014.0) 14. DOI: 10.1186/1471-2318-14-58 16. Boult C., Leff B., Boyd C.M., Wolff J.L., Marsteller J.A., Frick K.D., Wegener S., Reider L., Frey K., Mroz T.M.. A Matched-Pair Cluster-Randomized Trial of Guided Care for High-Risk Older Patients. J. Gen. Intern. Med. (2013.0) 28 612-621. DOI: 10.1007/s11606-012-2287-y 17. Colla C.H., Lewis V.A., Kao L.S., O’Malley A.J., Chang C.H., Fisher E.S.. Association between Medicare Accountable Care Organization Implementation and Spending among Clinically Vulnerable Beneficiaries. JAMA Intern. Med. (2016.0) 176 1167-1175. DOI: 10.1001/jamainternmed.2016.2827 18. Counsell S.R., Callahan C.M., Clark D.O., Tu W., Buttar A.B., Stump T.E., Ricketts G.D.. Geriatric Care Management for Low-Income Seniors. A Randomized Controlled Trial. JAMA (2007.0) 298 2623-2633. DOI: 10.1001/jama.298.22.2623 19. Fagan P.J., Schuster A.B., Boyd C., Marsteller J.A., Griswold M., Murphy S.M.E., Dunbar L., Forrest C.B.. Chronic Care Improvement in Primary Care: Evaluation of an Integrated Pay-for-Performance and Practice-Based Care Coordination Program among Elderly Patients with Diabetes. Health Serv. Res. (2010.0) 45 1763-1782. DOI: 10.1111/j.1475-6773.2010.01166.x 20. Gray D., Armstrong C.D., Dahrouge S., Hogg W., Zhang W.. Cost-effectiveness of Anticipatory and Preventive multidisciplinary Team Care for complex patients: Evidence from a randomized controlled trial. Can. Fam. Physician (2010.0) 56 e20-e29. PMID: 20090057 21. Recommendations for Change Management in Integrated Personalized Care Delivery. 2022. CORDIS Reports 22. Diamond A., Sekhon J.S.. Genetic Matching for Estimating Causal Effects: A General Multivariate Matching Method for Achieving Balance in Observational Studies. Rev. Econ. Stat. (2013.0) 95 932-945. DOI: 10.1162/REST_a_00318 23. Akpan A., Roberts C., Bandeen-Roche K., Batty B., Bausewein C., Bell D., Bramley D., Bynum J., Cameron I.D., Chen L.-K.. Standard set of health outcome measures for older persons. BMC Geriatr. (2018.0) 18. DOI: 10.1186/s12877-017-0701-3 24. Burns D.J., Arora J., Okunade O., Beltrame J.F., Bernardez-Pereira S., Crespo-Leiro M.G., Filippatos G.S., Hardman S., Hoes A.W., Hutchison S.. International Consortium for Health Outcomes Measurement (ICHOM): Standardized Patient-Centered Outcomes Measurement Set for Heart Failure Patients. JACC Heart Fail. (2020.0) 8 212-222. DOI: 10.1016/j.jchf.2019.09.007 25. Yusof M.M., Kuljis J., Papazafeiropoulou A., Stergioulas L.K.. An evaluation framework for Health Information Systems: Human, organization and technology-fit factors (HOT-fit). Int. J. Med. Inform. (2008.0) 77 386-398. DOI: 10.1016/j.ijmedinf.2007.08.011 26. Damschroder L.J., Aron D.C., Keith R.E., Kirsh S.R., Alexander J.A., Lowery J.C.. Fostering implementation of health services research findings into practice: A consolidated framework for advancing implementation science. Implement. Sci. (2009.0) 4 50. DOI: 10.1186/1748-5908-4-50 27. Coleman E.A., Eilertsen T.B., Kramer A.M., Magid D.J., Beck A., Conner D.. Reducing emergency visits in older adults with chronic illness. A randomized, controlled trial of group visits. Eff. Clin. Pract. ECP (2001.0) 4 49-57. PMID: 11329985 28. Moe J., Kirkland S.W., Rawe E., Ospina M.B., Vandermeer B., Campbell S., Rowe B.H.. Effectiveness of Interventions to Decrease Emergency Department Visits by Adult Frequent Users: A Systematic Review. Acad. Emerg. Med. (2017.0) 24 40-52. DOI: 10.1111/acem.13060 29. Mateo-Abad M., Fullaondo A., Merino M., Gris S., Marchet F., Avolio F., Graps E., Ravic M., Kovac M., Benkovic V.. Impact Assessment of an Innovative Integrated Care Model for Older Complex Patients with Multimorbidity: The CareWell Project. Int. J. Integr. Care (2020.0) 20 8. DOI: 10.5334/ijic.4711 30. Herdman M., Gudex C., Lloyd A., Janssen M., Kind P., Parkin D., Bonsel G., Badia X.. Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual. Life Res. (2011.0) 20 1727-1736. DOI: 10.1007/s11136-011-9903-x 31. Zigmond A.S., Snaith R.P.. The Hospital Anxiety and Depression Scale. Acta Psychiatr. Scand. (1983.0) 67 361-370. DOI: 10.1111/j.1600-0447.1983.tb09716.x 32. Graf C.. The Lawton Instrumental Activities of Daily Living Scale. AJN Am. J. Nurs. (2008.0) 108 52-62. DOI: 10.1097/01.NAJ.0000314810.46029.74 33. Mahoney F.I., Barthel D.W.. Functional evaluation: The Barthel Index. Md. State Med. J. (1965.0) 14 61-65. PMID: 14258950 34. Spertus J.A., Jones P.G., Sandhu A.T., Arnold S.V.. Interpreting the Kansas City Cardiomyopathy Questionnaire in Clinical Trials and Clinical Care. J. Am. Coll. Cardiol. (2020.0) 76 2379-2390. DOI: 10.1016/j.jacc.2020.09.542 35. Jones P.W., Harding G., Berry P., Wiklund I., Chen W.H., Kline Leidy N.. Development and first validation of the COPD Assessment Test. Eur. Respir. J. (2009.0) 34 648-654. DOI: 10.1183/09031936.00102509 36. Mahler D.A., Wells C.K.. Evaluation of Clinical Methods for Rating Dyspnea. Chest (1988.0) 93 580-586. DOI: 10.1378/chest.93.3.580 37. Venkatesh M., Davis D.. User Acceptance of Information Technology: Toward a Unified View. MIS Q. (2003.0) 27 425. DOI: 10.2307/30036540 38. Drummond M.F., Sculpher M.J., Claxton K., Stoddart G.L., Torrance G.W.. Methods for the Economic Evaluation of Health Care Programmes (2015.0) 39. Bédard M., Molloy D.W., Squire L., Dubois S., Lever J.A., O’Donnell M.. The Zarit Burden Interview. Gerontologist (2001.0) 41 652-657. DOI: 10.1093/geront/41.5.652 40. Tennant R., Hiller L., Fishwick R., Platt S., Joseph S., Weich S., Parkinson J., Secker J., Stewart-Brown S.. The Warwick-Edinburgh mental well-being scale (WEMWBS): Development and UK validation. Health Qual. Life Outcomes (2007.0) 5 63. DOI: 10.1186/1477-7525-5-63 41. Bloemeke J., Moll C., Arndt F., González N., Ortega A., Verdoy D., Groene O.. Qualitative Research Protocol as Part of the ADLIFE International Large-Scale Pilot Project 42. onFHIR Repository 43. D1.1 Data Management Plan 44. Vaismoradi M., Turunen H., Bondas T.. Content analysis and thematic analysis: Implications for conducting a qualitative descriptive study. Nurs. Health Sci. (2013.0) 15 398-405. DOI: 10.1111/nhs.12048 45. Tong A., Sainsbury P., Craig J.. Consolidated criteria for reporting qualitative research (COREQ): A 32-item checklist for interviews and focus groups. Int. J. Qual. Health Care (2007.0) 19 349-357. DOI: 10.1093/intqhc/mzm042 46. 46. StataCorp Stata Statistical Software: Release 16StataCorp. LLC.College Station, TX, USA2019. Stata Statistical Software: Release 16 (2019.0) 47. Karnon J., Stahl J., Brennan A., Caro J.J., Mar J., Möller J.. Modeling using Discrete Event Simulation: A Report of the ISPOR-SMDM Modeling Good Research Practices Task Force-4. Value Health (2012.0) 15 821-827. DOI: 10.1016/j.jval.2012.04.013 48. Communication and Dissemination Plan and Communication Material (CORDIS Reports) 49. Tunis S.R., Stryer D.B., Clancy C.M.. Practical clinical trials. JAMA (2003.0) 290 1624-1632. DOI: 10.1001/jama.290.12.1624 50. Skivington K., Matthews L., Simpson S.A., Craig P., Baird J., Blazeby J.M., Boyd K.A., Craig N., French D.P., McIntosh E.. A new framework for developing and evaluating complex interventions: Update of Medical Research Council guidance. BMJ (2021.0) 374 n2061. DOI: 10.1136/bmj.n2061 51. Hawe P., Shiell A., Riley T.. Complex interventions: How “out of control” can a randomised controlled trial be?. BMJ (2004.0) 328 1561-1563. DOI: 10.1136/bmj.328.7455.1561
--- title: The Comparison of Retinal Microvascular Findings in Acute COVID-19 and 1-Year after Hospital Discharge Assessed with Multimodal Imaging—A Prospective Longitudinal Cohort Study authors: - Kristina Jevnikar - Andrej Meglič - Luka Lapajne - Mateja Logar - Nataša Vidovič Valentinčič - Mojca Globočnik Petrovič - Polona Jaki Mekjavić journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966689 doi: 10.3390/ijms24044032 license: CC BY 4.0 --- # The Comparison of Retinal Microvascular Findings in Acute COVID-19 and 1-Year after Hospital Discharge Assessed with Multimodal Imaging—A Prospective Longitudinal Cohort Study ## Abstract This study aimed to quantify possible long-term impairment of the retinal microcirculation and microvasculature by reassessing a cohort of patients with acute COVID-19 without other known comorbidities one year after their discharge from the hospital. Thirty patients in the acute phase of COVID-19 without known systemic comorbidities were enrolled in this prospective longitudinal cohort study. Fundus photography, SS-OCT, and SS-OCTA using swept-source OCT (SS-OCT, Topcon DRI OCT Triton; Topcon Corp., Tokyo, Japan) were performed in the COVID-19 unit and 1-year after hospital discharge. The cohort’s median age was 60 years (range 28–65) and 18 ($60\%$) were male. Mean vein diameter (MVD) significantly decreased over time, from 134.8 μm in the acute phase to 112.4 μm at a 1-year follow-up ($p \leq 0.001$). A significantly reduced retinal nerve fiber layer (RNFL) thickness was observed at follow-up in the inferior quadrant of the inner ring (mean diff. 0.80 $95\%$ CI 0.01–1.60, $$p \leq 0.047$$) and inferior (mean diff. 1.56 $95\%$ CI 0.50–2.61, $p \leq 0.001$), nasal (mean diff. 2.21 $95\%$ CI 1.16–3.27, $p \leq 0.001$), and superior (mean diff. 1.69 $95\%$ CI 0.63–2.74, $p \leq 0.001$) quadrants of the outer ring. There were no statistically significant differences between the groups regarding vessel density of the superior and deep capillary plexuses. The transient dilatation of the retinal vessels in the acute phase of COVID-19, as well as RNFL thickness changes, could become a biomarker of angiopathy in patients with severe COVID-19. ## 1. Introduction Coronavirus disease 19 (COVID-19) is caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and even though several viral variants have emerged since the outbreak in December 2019, the proposed pathophysiological mechanism remains the same [1,2]. SARS-CoV-2 enters the cell by binding to angiotensin-converting enzyme 2 (ACE2), a key enzyme of the renin–angiotensin–aldosterone (RAAS) pathway and downregulates its activity [3]. Angiotensin-converting enzyme (ACE) and ACE2 are two of the key enzymes of the RAAS pathway. ACE catalyzes angiotensin-I (Ang-I) to angiotensin II (Ang-II), then is hydrolyzed to Ang 1–7 by ACE2. By binding to ACE2, SARS-CoV-2 downregulates its activity and creates an imbalance in the signaling effects of Angiotensin II (Ang II) and its receptor (angiotensin II type 1 receptor, AT1), resulting in the accumulation of Ang II. Increased serum Ang II leads to vasoconstriction, inflammation, cellular differentiation and growth, endothelial dysfunction, the formation of reactive oxidative species, and microvascular thrombosis. On the contrary, Ang II actions mediated through angiotensin II type 2 receptor (AT2) have a vasoprotective and anti-inflammatory role [4,5,6]. Nevertheless, AT2 has been proposed as an alternative entry point of SARS-CoV-2, blocking its possible protective role [7]. The resulting imbalance in the signaling effects of the RAAS pathway leads to endothelial dysfunction, which, combined with a hypercoagulable state, predisposes the patients to thromboembolic events [4,8]. ACE2 is expressed on the host cell surface of several tissues, including the retinal vascular endothelium, Müller glia and ganglion cells, and neurons in the inner nuclear layer [5]. In addition, another mechanism has been proposed, resulting in SARS-CoV-2-induced pyroptosis—an inflammatory type of programmed host cell death caused by direct infection of the retinal cells, leading to a production of cytokines, neuronal damage, and a hypercoagulable state [3,5,6,9,10,11,12,13]. The retina is especially susceptible to potential microvascular thrombosis, given its high metabolic demands and the fact that its plexuses contain terminal vessels without anastomotic connections [4,14,15]. COVID-19 retinopathy, encompassing several retinal findings, such as flame-shaped hemorrhages, cotton wool spots, dilated veins, and tortuous vessels, have been previously described. Furthermore, increased RNFL thickness and increased GCL thickness in several quadrants of the inner and outer EDTRS ring were described in the acute phase of severe COVID-19 [16]. The studies regarding the vessel density (VD) of the superficial and deep capillary plexuses have been inconclusive. While some found no differences regarding the vessel density [17] and the foveal avascular zone (FAZ) area [18], others reported decreased VD [18,19,20,21] and an increase in the FAZ area [20,22,23]. COVID-19 severity was found to affect the presence of retinopathy as a higher incidence of findings was reported in the moderate and severe course of the disease [4,17,24]. Several factors were shown to influence the severity of COVID-19, including age, male gender, and pre-existing comorbidities, especially cardiovascular disease, diabetes, and hypertension [25,26,27]. Underlying conditions enhance RAAS/Ag II imbalance; moreover, diabetes and resulting chronic hyperglycemia are known to compromise the innate immune system, which increases susceptibility to hyperinflammation and the development of the cytokine storm. This has been confirmed by increased inflammation-related biomarkers, such as C-reactive protein, serum ferritin, and IL-6 in diabetic patients with COVID-19 [8,25,28]. Furthermore, the polymorphisms in the genes of the RAAS pathway were shown to play a role in COVID-19 severity [29,30,31,32]. ACE insertion/deletion polymorphisms (rs4646994 and rs179752) were shown to affect the disease course. Individuals with a D/D genotype, which leads to an increased serum ACE concentration, were shown to have higher rates of pulmonary embolism and COVID-19-related mortality [18,25,26]. Conflicting results regarding the role of ACE2-related polymorphisms have been published [16,19,20,21]. It was hypothesized that the presence of A-allele, linked to increased gene expression of ACE2, resulting in higher ACE2 serum levels, and therefore and an increased number of viral binding sites, could increase COVID-19 severity [19,21]. Nevertheless, GG genotype and G allele carriers, associated with lower serum ACE levels, were shown to have an increased risk of SARS-CoV-2 infection and a more severe course of COVID-19 [33]. Therefore, the presence of A-allele was found to have a protective role, possibly explained by counterbalancing of the effects of increased Ang II resulting from the RAAS dysregulation [33]. Nonetheless, several studies found no association with disease susceptibility or severity [16,20,21]. An increased risk of severe COVID-19 was also associated with an AGTR2 polymorphism (rs1914711) [27]. It is noteworthy that ACE, ACE2, and AGTR2 polymorphisms have also been associated with hypertension, diabetes, coronary artery disease, and stroke [16,17,18,19,20,21], which complicates the risk estimation, as those comborbities alone are associated with an increased risk of COVID-19. Nevertheless, our previous study, focusing on patients without known comorbidities, showed an increased risk of COVID-19 retinopathy in males with the AGTR2-AA genotype of the rs1403543 polymorphism [34]. The longitudinal nature of our study is especially interesting given that up to $10\%$ of patients are expected to develop long COVID-19, a multisystemic condition encompassing new onset thrombotic cardiovascular and cerebrovascular disease, dysautonomia, and Chronic Fatigue Syndrome [35]. Several mechanisms of long-COVID-19 pathogenesis have been proposed. It has been hypothesized that persisting reservoirs of SARS-CoV-2 in tissues could affect immune dysregulation and play a role in autoimmunity. In addition, long COVID-19 could result from the reactivation of herpesviruses, including the Epstein–Barr virus and human herpesvirus 6. Moreover, virus-induced dysfunctional signaling in the vagus nerve and the brainstem has also been postulated to lead to dysautonomia. In addition, endothelial dysfunction and microvascular thromboses are believed to have a role not only in the acute phase, but also long-term [35]. Optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA) are non-invasive imaging modalities that provide high-resolution visualization of the retinal layers and the flow in its superficial and deep capillary plexuses, enabling the assessment of the microcirculatory changes in the retina [36]. This study aimed to quantify possible long-term impairment of the retinal microcirculation and microvasculature by reassessing the cohort of patients with acute COVID-19 without other known comorbidities one year after their discharge from the hospital. ## 2. Results The cohort consisted of 30 consecutive patients in the acute phase of COVID-19 without known systemic comorbidities who were reassessed a year after their discharge from the hospital. Baseline demographic and clinical characteristics are presented in Table 1. The cohort’s median age was 60 years (range 28–65) and 18 ($60\%$) were male. Mean systolic and diastolic pressures were within normal limits. The most frequent symptom was fever ($86.7\%$), followed by cough ($76.7\%$), dyspnea ($60\%$), chest pain ($33.3\%$), headache ($10\%$), anosmia ($10\%$), and diarrhea ($10\%$). The course was complicated by deep vein thrombosis and pulmonary embolism in one patient ($3.3\%$). Patients were treated with dexamethasone, remdesivir, and oxygen, according to the guidelines. Twenty patients ($66.7\%$) required oxygen supplementation. None of the patients presented with any signs or ocular symptoms, such as itching, photophobia, foreign body, conjunctivitis, or diminished visual acuity or exhibited them at 1-year follow-up. None of the patients in our cohort exhibited any of the long-COVID-19 symptoms. MVD significantly decreased over time, from 134.8 μm in the acute phase to 112.4 μm at 1-year follow-up ($p \leq 0.001$). MAD values were 96.1 μm and 93.1 μm, respectively, and even though the difference did not reach a statistical significance ($$p \leq 0.181$$), a tendency toward a decrease in the diameter can be observed (Figure 1). The comparison of the OCT parameters of patients in the acute phase of COVID-19 compared with 1-year follow-up is presented in Table 2, Figure 2. A significantly decreased RNFL was observed at follow-up in the inferior quadrant of the inner ring (mean diff. 0.80 $95\%$ CI 0.01–1.60, $$p \leq 0.047$$) and inferior (mean diff. 1.56 $95\%$ CI 0.50–2.61, $p \leq 0.001$), nasal (mean diff. 2.21 $95\%$ CI 1.16–3.27, $p \leq 0.001$), and superior (mean diff. 1.69 $95\%$ CI 0.63–2.74, $p \leq 0.001$) quadrants of the outer ring. There were no statistically significant differences between the groups regarding the vessel density of the SCP and DCP. Moreover, no significant differences were observed in the FAZ parameters, including FAZ area, FAZ perimeter, FAZ circularity, and axial ratio (Table 3 and Figure 3). ## 3. Discussion In this study, we evaluated the evolution of retinal microvascular alterations assessed with multimodal imaging in patients without known ocular or systemic comorbidities in the acute phase of COVID-19 and at one-year follow-up. We found a significantly decreased MVD and a tendency toward a decrease in MAD at 1-year follow-up. These results are especially interesting, given that the significantly increased mean vessel diameters of both veins and arteries were present in hospitalized patients with severe COVID-19 in our previous study [16]. Similar results were reported, where a significantly decreased MVD and MAD were present after a 6-month follow-up; however, their baseline measurements were not performed in the acute phase of COVID-19 [37,38]. Our results can be explained by several mechanisms. First, the transient dilatation of the acute phase could reflect the vasogenic response to hypoxic and hypercapnic conditions of acute COVID-19 [15,37]. However, none of the confounding variables, including treatment with oxygen and inflammatory parameters, were shown to affect the vessel diameter in the acute phase [16]. Therefore, the hypothesis of SARS-CoV-2 induced RAAS dysregulation and the resulting endothelial dysfunction and injury seem more likely [9,39]. This is further supported by the findings from autopsy studies, where microvascular alterations such as small vessel thickening, vascular remodeling, and micro thrombosis were reported [40,41]. Hence, the longitudinal results and gradual decrease in the vessel diameters could reflect the regeneration of the vascular endothelium over time, a biphasic process where cellular proliferation and the resulting hypertrophy of the early phase are followed by normalization of cell density by pruning of excess cells in the circulation [42]. In addition, a significantly thinner RNFL in the outer ring of inferior, nasal, and superior quadrants was observed at follow-up. Given that the superficial layer of the retinal capillaries lies in the RNFL layer, this thinning possibly reflects the aforementioned decrease in the MVD. Similar results of thinner RNFL and GCL were reported over time [20,22,43,44,45]; however, vessel diameters were not assessed. The strength of our study is the simultaneous evaluation of those parameters, as the observed thinning seems to reflect regeneration rather than progressive SARS-CoV-2-induced RNFL atrophy. Moreover, no significant differences were found between the two groups in the VD of the SCP and DCP or any FAZ parameters, including the FAZ area, perimeter, circularity, and axial ratio. Notably, we also found no differences when comparing acute COVID-19 patients with a healthy control group [16]. While some of the previous studies are in line with our results [17,18], a decreased VD and an enlarged FAZ area were reported [18,19,20,21,46] and the observed differences were either stable [47] or more pronounced after a 3–8 month follow-up [21,21,22,23,23]. However, none of the studies reported any macroscopic changes on OCTA images, suggestive of either thromboembolisms or ischemia. It is noteworthy that only patients without known comorbidities that are known to affect the retina and retinal microvasculature, such as diabetes and hypertension, were included in our study. While we initially hypothesized that our results reflect the subtlety of acute changes, the long-term results may reflect the capability of the retinal intrinsic autoregulatory mechanisms to enable sufficient oxygenation to prevent ischemic damage even in the state of acute COVID-19 infection [14,15]. We acknowledge the limitations of our study. First, the baseline imaging was performed in the acute phase in the COVID-19 unit and was limited to patients admitted to the hospital during the study period. Moreover, only a proportion of initially enrolled patients were willing to attend the follow-up imaging. Second, OCT and OCTA have a limited reach; therefore, possible peripheral alterations could not be evaluated. ## 4.1. Study Design A prospective longitudinal cohort study was conducted at the University Medical Center Ljubljana (UMCL) between December 2020 and May 2022. The study was approved by the Slovenian medical ethics committee (protocol ID number: 0120-$\frac{553}{2020}$/3) and adhered to the tenets of the Declaration of Helsinki. Written informed consent was obtained from all participants enrolled in the study. ## 4.2. Patient Selection, Inclusion, and Exclusion Criteria The patient cohort consisted of consecutive patients aged 18–65 with PCR-confirmed SARS-CoV-2 admitted to the COVID-19 unit of the department of infectious diseases UMCL. The exclusion criteria were as follows: systemic comorbidities (diabetes, arterial hypertension, hyperlipidemia, coronary artery disease, history of stroke), concomitant infectious diseases (HIV, HSV, VZV, CMV), systemic treatment linked to retinal toxicity, smoking, pre-existing ocular pathology (age-related macular degeneration and other retinal diseases, a history of glaucoma, high myopia (>−6)), and other conditions that could have affected the retinal morphology. The cohort was reassessed one year after their hospital discharge to confirm that they continued to meet the inclusion criteria. ## 4.3. Study Protocol The study was conducted in two locations; patients in the acute phase of COVID-19 underwent imaging in the COVID-19 unit of the Department of Infectious Diseases, UMCL, whereas the follow-up imaging of the same cohort a year later was performed at the Department of Ophthalmology, UMCL. All enrolled subjects were asked about the presence of the following ocular symptoms and signs: conjunctivitis, photophobia, itching, and diminished visual acuity. After dilating the pupils ($1\%$ tropicamide), fundus images, OCT, and OCTA were obtained using swept-source OCT (SS-OCT, Topcon DRI OCT Triton; Topcon Corp., Tokyo, Japan). The study protocol consisted of 4 images per eye: 2 color fundus images (one centered on the fovea, one on the optic disc), OCT centered on the fovea using the 7 × 7 mm scanning protocol, and OCTA centered on the fovea using the 3 × 3 mm scanning protocol. All the images were obtained by two doctors (KJ, LL). The appropriate full-body protective gown with the FPP 3 mask was worn in the COVID-19 unit. Patients’ electronic medical records were reviewed during hospitalization to collect the following demographic, clinical, and laboratory parameters: age, sex, presence of comorbidities (diabetes, arterial hypertension, hyperlipidemia, coronary artery disease, history of stroke), history of smoking, alcohol consumption, concomitant infectious diseases (HIV, HSV, VZV, CMV), time from the symptom’s onset or positive PCR to the day of fundus imaging, the presence of COVID-19-related symptoms, the need for oxygen, COVID-19-related treatment, and outcome. Laboratory parameters included lactate dehydrogenase (LDH), ferritin, CRP, procalcitonin, white blood cells, red cell distribution width (RDW), platelets, lymphocytes, D-dimer, and 25-OH-D3. ## 4.4. Image Analysis The eye with a better signal strength index was included in the analysis. Only images with a signal strength index above 60 were analyzed. All obtained imaging was independently reviewed by three researchers (KJ, AM, and PJM). Fundus photographs were screened for the presence of signs synonymous with COVID-19 retinopathy. Mean vein (MVD) and mean artery (MAD) diameters were assessed with the Automated Retinal Image Analyser (ARIA, V1-09-12-11) using a previously described method, in which the vessel diameters of the four main veins and four main arteries between 0.5 and 1 disc diameter from the optic disc margin were used to calculate the mean vein diameter (MVD) and mean artery diameter (MAD) [37]. OCT and OCTA images were automatically segmented by the built-in software (Topcon Corp., Tokyo, Japan), reviewed for the presence of abnormalities, checked for correct auto-segmentation, and manually readjusted if necessary. The thicknesses of the retinal nerve fiber layer (RNFL), ganglion cell layer (GCL), and retina in the four quadrants of the inner and the outer ring of the early treatment diabetic retinopathy (ETDRS) grid were exported using OCT Data Collector software (Topcon Inc., Tokyo, Japan). OCTA images of the superficial capillary plexus (SCP) and deep capillary plexus (DCP) were processed using MATLAB (MathWorks Inc., Natick, MA, USA), which was also used to analyze the foveal avascular zone (FAZ) [48]. The OCTA image was pre-processed using top-hat transformation to increase the vessel intensity. Extraction of vessel edges with a Canny edge detector was followed by morphological closure, image inversion, and removal of small elements to reduce the number of potential FAZ candidates. From the remaining candidates, the final FAZ was identified based on predefined area and eccentricity limits. Since the use of morphological operators reduces the FAZ segmentation precision, region growing was applied to the eroded FAZ. Pixels within $30\%$ of the original region intensity were added. In cases where automatic segmentation did not produce satisfactory results, FAZ areas were outlined manually. FAZ area, perimeter, and circularity index were calculated. Once the center of FAZ was defined, ETDRS chart was superimposed to calculate the parafoveal vessel density (VD) in the four quadrants within the 3 mm circle of the center of FAZ. Each quadrant was separately binarized using the Otsu method. VD was expressed in percentage derived from the ratio of the total vessel area (white pixels) to the total area of the analyzed region (number of pixels in quadrant), a method previously described by Nicoló et al. [ 49]. The average vessel densities of the SCP and DCP were used for quantitative analysis. ## 4.5. Statistical Analysis Values are reported as mean (standard deviation, SD) or median (interquartile range, IQR) for numerical variables and as frequency (%) for descriptive variables. The normality was examined with the Shapiro–Wilk test. The differences in OCTA parameters before–after were evaluated as appropriate by dependent t-test or Wilcoxon rank-sum test. Linear mixed-effects regression was used to compare the OCT parameters between the groups. The subject was included as a random intercept to account for multiple measurements in each subject (N, S, T, I). P values of all pairwise comparisons were adjusted using the Benjamini–Hochberg method. Statistical analysis was performed with R statistical software (version 4.1.3, Vienna, Austria). ## 5. Conclusions In conclusion, this is the first study to show the evolution of the retinal microvasculature and structure from the acute phase of COVID-19 to one-year follow-up. The transient dilatation of the retinal vessels in the acute phase of COVID-19 could become a biomarker of angiopathy in patients with severe COVID-19. Further longitudinal studies are warranted to assess possible long-term complications. ## References 1. Guan W.-J., Ni Z.-Y., Hu Y., Liang W.-H., Ou C.-Q., He J.-X., Liu L., Shan H., Lei C.-L., Hui D.S.C.. Clinical Characteristics of Coronavirus Disease 2019 in China. N. Engl. J. Med. (2020.0) 382 1708-1720. DOI: 10.1056/NEJMoa2002032 2. Huang C., Wang Y., Li X., Ren L., Zhao J., Hu Y., Zhang L., Fan G., Xu J., Gu X.. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet (2020.0) 395 497-506. DOI: 10.1016/S0140-6736(20)30183-5 3. Hoffmann M., Kleine-Weber H., Schroeder S., Krüger N., Herrler T., Erichsen S., Schiergens T.S., Herrler G., Wu N.H., Nitsche A.. SARS-CoV-2 Cell Entry Depends on ACE2 and TMPRSS2 and Is Blocked by a Clinically Proven Protease Inhibitor. Cell (2020.0) 181 271-280.e278. DOI: 10.1016/j.cell.2020.02.052 4. Jevnikar K., Jaki Mekjavic P., Vidovic Valentincic N., Petrovski G., Globocnik Petrovic M.. An Update on COVID-19 Related Ophthalmic Manifestations. Ocul. Immunol. Inflamm. (2021.0) 29 684-689. DOI: 10.1080/09273948.2021.1896008 5. Choudhary R., Kapoor M.S., Singh A., Bodakhe S.H.. Therapeutic targets of renin-angiotensin system in ocular disorders. J. Curr. Ophthalmol. (2017.0) 29 7-16. DOI: 10.1016/j.joco.2016.09.009 6. Kanberg N., Ashton N.J., Andersson L.M., Yilmaz A., Lindh M., Nilsson S., Price R.W., Blennow K., Zetterberg H., Gisslén M.. Neurochemical evidence of astrocytic and neuronal injury commonly found in COVID-19. Neurology (2020.0) 95 e1754-e1759. DOI: 10.1212/WNL.0000000000010111 7. Panel C.-T.G.. Coronavirus Disease 2019 (COVID-19) Treatment Guidelines 8. Callender L.A., Curran M., Bates S.M., Mairesse M., Weigandt J., Betts C.J.. The Impact of Pre-existing Comorbidities and Therapeutic Interventions on COVID-19. Front. Immunol. (2020.0) 11 1991. DOI: 10.3389/fimmu.2020.01991 9. Connors J.M., Levy J.H.. Thromboinflammation and the hypercoagulability of COVID-19. J. Thromb. Haemost. (2020.0) 18 1559-1561. DOI: 10.1111/jth.14849 10. Pereira L.A., Soares L.C.M., Nascimento P.A., Cirillo L.R.N., Sakuma H.T., Veiga G.L.d., Fonseca F.L.A., Lima V.L., Abucham-Neto J.Z.. Retinal findings in hospitalised patients with severe COVID-19. Br. J. Ophthalmol. (2020.0) 106 102-105. DOI: 10.1136/bjophthalmol-2020-317576 11. Hussman J.P.. Cellular and Molecular Pathways of COVID-19 and Potential Points of Therapeutic Intervention. Front. Pharmacol. (2020.0) 11 1169. DOI: 10.3389/fphar.2020.01169 12. Biswas S., Thakur V., Kaur P., Khan A., Kulshrestha S., Kumar P.. Blood clots in COVID-19 patients: Simplifying the curious mystery. Med. Hypotheses (2020.0) 146 110371. DOI: 10.1016/j.mehy.2020.110371 13. Senanayake P.d., Drazba J., Shadrach K., Milsted A., Rungger-Brandle E., Nishiyama K., Miura S.-I., Karnik S., Sears J.E., Hollyfield J.G.. Angiotensin II and Its Receptor Subtypes in the Human Retina. Investig. Ophthalmol. Vis. Sci. (2007.0) 48 3301-3311. DOI: 10.1167/iovs.06-1024 14. Wong T.Y., Islam F.M.A., Klein R., Klein B.E.K., Cotch M.F., Castro C., Sharrett A.R., Shahar E.. Retinal vascular caliber, cardiovascular risk factors, and inflammation: The multi-ethnic study of atherosclerosis (MESA). Investig. Ophthalmol. Vis. Sci. (2006.0) 47 2341-2350. DOI: 10.1167/iovs.05-1539 15. Jaki Mekjavic P., Lenassi E., Eiken O., Mekjavic I.B.. Effect of acute hypercapnia during 10-day hypoxic bed rest on posterior eye structures. J. Appl. Physiol. (1985) (2016.0) 120 1241-1248. DOI: 10.1152/japplphysiol.00930.2015 16. Jevnikar K., Meglič A., Lapajne L., Logar M., Vidovič Valentinčič N., Globočnik Petrovič M., Jaki Mekjavić P.. The impact of acute COVID-19 on the retinal microvasculature assessed with multimodal imaging. Graefes Arch. Clin. Exp. Ophthalmol. (2022.0) 1-11. DOI: 10.1007/s00417-022-05887-0 17. Savastano M.C., Gambini G., Cozzupoli G.M., Crincoli E., Savastano A., De Vico U., Culiersi C., Falsini B., Martelli F., Minnella A.M.. Retinal capillary involvement in early post-COVID-19 patients: A healthy controlled study. Graefe’s Arch. Clin. Exp. Ophthalmol. = Albrecht Von Graefes Arch. Fur. Klin. Exp. Ophthalmol. (2021.0) 259 2157-2165. DOI: 10.1007/s00417-020-05070-3 18. Hazar L., Karahan M., Vural E., Ava S., Erdem S., Dursun M.E., Keklikçi U.. Macular vessel density in patients recovered from COVID 19. Photodiagnosis Photodyn. (2021.0) 34. DOI: 10.1016/j.pdpdt.2021.102267 19. Zapata M.Á., Banderas García S., Sánchez-Moltalvá A., Falcó A., Otero-Romero S., Arcos G., Velazquez-Villoria D., García-Arumí J.. Retinal microvascular abnormalities in patients after COVID-19 depending on disease severity. Br. J. Ophthalmol. (2020.0) 106 559-563. DOI: 10.1136/bjophthalmol-2020-317953 20. Abrishami M., Emamverdian Z., Shoeibi N., Omidtabrizi A., Daneshvar R., Saeidi Rezvani T., Saeedian N., Eslami S., Mazloumi M., Sadda S.. Optical coherence tomography angiography analysis of the retina in patients recovered from COVID-19: A case-control study. Can. J. Ophthalmol. (2021.0) 56 24-30. DOI: 10.1016/j.jcjo.2020.11.006 21. Cennamo G., Reibaldi M., Montorio D., D’Andrea L., Fallico M., Triassi M.. Optical Coherence Tomography Angiography Features in Post-COVID-19 Pneumonia Patients: A Pilot Study. Am. J. Ophthalmol. (2021.0) 227 182-190. DOI: 10.1016/j.ajo.2021.03.015 22. Bilbao-Malavé V., González-Zamora J., Saenz de Viteri M., de la Puente M., Gándara E., Casablanca-Piñera A., Boquera-Ventosa C., Zarranz-Ventura J., Landecho M.F., García-Layana A.. Persistent Retinal Microvascular Impairment in COVID-19 Bilateral Pneumonia at 6-Months Follow-Up Assessed by Optical Coherence Tomography Angiography. Biomedicines (2021.0) 9. DOI: 10.3390/biomedicines9050502 23. Abrishami M., Hassanpour K., Hosseini S., Emamverdian Z., Ansari-Astaneh M.R., Zamani G., Gharib B., Abrishami M.. Macular vessel density reduction in patients recovered from COVID-19: A longitudinal optical coherence tomography angiography study. Graefes Arch. Clin. Exp. Ophthalmol. (2021.0) 260 771-779. DOI: 10.1007/s00417-021-05429-0 24. Teo K.Y., Invernizzi A., Staurenghi G., Cheung C.M.G.. COVID-19 related retinal micro-vasculopathy—A review of current evidence: COVID-19 related retinal micro-vasculopathy. Am. J. Ophthalmol. (2021.0) 235 98-110. DOI: 10.1016/j.ajo.2021.09.019 25. de Almeida-Pititto B., Dualib P.M., Zajdenverg L., Dantas J.R., de Souza F.D., Rodacki M., Bertoluci M.C., Brazilian Diabetes Society Study G.. Severity and mortality of COVID 19 in patients with diabetes, hypertension and cardiovascular disease: A meta-analysis. Diabetol. Metab. Syndr. (2020.0) 12 75. DOI: 10.1186/s13098-020-00586-4 26. Matsushita K., Ding N., Kou M., Hu X., Chen M., Gao Y., Honda Y., Zhao D., Dowdy D., Mok Y.. The Relationship of COVID-19 Severity with Cardiovascular Disease and Its Traditional Risk Factors: A Systematic Review and Meta-Analysis. Glob. Heart (2020.0) 15 64. DOI: 10.5334/gh.814 27. Bae S., Kim S.R., Kim M.-N., Shim W.J., Park S.-M.. Impact of cardiovascular disease and risk factors on fatal outcomes in patients with COVID-19 according to age: A systematic review and meta-analysis. Heart (2020.0) 107 373-380. DOI: 10.1136/heartjnl-2020-317901 28. Apicella M., Campopiano M.C., Mantuano M., Mazoni L., Coppelli A., Del Prato S.. COVID-19 in people with diabetes: Understanding the reasons for worse outcomes. Lancet Diabetes Endocrinol. (2020.0) 8 782-792. DOI: 10.1016/S2213-8587(20)30238-2 29. Karakaş Çelik S., Çakmak Genç G., Pişkin N., Açikgöz B., Altinsoy B., Kurucu İşsiz B., Dursun A.. Polymorphisms of ACE (I/D) and ACE2 receptor gene (Rs2106809, Rs2285666) are not related to the clinical course of COVID-19: A case study. J. Med. Virol. (2021.0) 93 5947-5952. DOI: 10.1002/jmv.27160 30. Zmorzynski S., Szudy-Szczyrek A., Popek-Marciniec S., Korszen-Pilecka I., Wojcierowska-Litwin M., Luterek M., Chocholska S., Styk W., Swiderska-Kołacz G., Januszewska J.. ACE Insertion/Deletion Polymorphism (rs4646994) Is Associated With the Increased Risk of Multiple Myeloma. Front. Oncol. (2019.0) 9 44. DOI: 10.3389/fonc.2019.00044 31. Calabrese C., Annunziata A., Coppola A., Pafundi P.C., Guarino S., Di Spirito V., Maddaloni V., Pepe N., Fiorentino G.. ACE Gene I/D Polymorphism and Acute Pulmonary Embolism in COVID19 Pneumonia: A Potential Predisposing Role. Front. Med. (2020.0) 7 631148. DOI: 10.3389/fmed.2020.631148 32. Möhlendick B., Schönfelder K., Breuckmann K., Elsner C., Babel N., Balfanz P., Dahl E., Dreher M., Fistera D., Herbstreit F.. ACE2 polymorphism and susceptibility for SARS-CoV-2 infection and severity of COVID-19. Pharm. Genom. (2021.0) 31 165-171. DOI: 10.1097/FPC.0000000000000436 33. 33. R Core Team R: A Language and Environment for Statistical ComputingR Foundation for Statistical ComputingVienna, Austria2022. R: A Language and Environment for Statistical Computing (2022.0) 34. Jevnikar K., Lapajne L., Petrovič D., Meglič A., Logar M., Vidovič Valentinčič N., Globočnik Petrovič M., Cilenšek I., Mekjavić P.J.. The Role of ACE, ACE2, and AGTR2 Polymorphisms in COVID-19 Severity and the Presence of COVID-19-Related Retinopathy. Genes (2022.0) 13. DOI: 10.3390/genes13071111 35. Davis H.E., McCorkell L., Vogel J.M., Topol E.J.. Long COVID: Major findings, mechanisms and recommendations. Nat. Rev. Microbiol. (2023.0) 21 133-146. DOI: 10.1038/s41579-022-00846-2 36. Kim K., Kim E.S., Yu S.Y.. Optical coherence tomography angiography analysis of foveal microvascular changes and inner retinal layer thinning in patients with diabetes. Br. J. Ophthalmol. (2018.0) 102 1226-1231. DOI: 10.1136/bjophthalmol-2017-311149 37. Invernizzi A., Torre A., Parrulli S., Zicarelli F., Schiuma M., Colombo V., Giacomelli A., Cigada M., Milazzo L., Ridolfo A.. Retinal findings in patients with COVID-19: Results from the SERPICO-19 study. eClinicalMedicine (2020.0) 27 100550. DOI: 10.1016/j.eclinm.2020.100550 38. Invernizzi A., Schiuma M., Parrulli S., Torre A., Zicarelli F., Colombo V., Marini S., Villella E., Bertoni A., Antinori S.. Retinal vessels modifications in acute and post-COVID-19. Sci. Rep. (2021.0) 11 19373. DOI: 10.1038/s41598-021-98873-1 39. Ho D., Low R., Tong L., Gupta V., Veeraraghavan A., Agrawal R.. COVID-19 and the Ocular Surface: A Review of Transmission and Manifestations. Ocul. Immunol. Inflamm. (2020.0) 28 726-734. DOI: 10.1080/09273948.2020.1772313 40. Lang M., Som A., Carey D., Reid N., Mendoza D.P., Flores E.J., Li M.D., Shepard J.-A.O., Little B.P.. Pulmonary Vascular Manifestations of COVID-19 Pneumonia. Radiol. Cardiothorac. Imaging (2020.0) 2 e200277. DOI: 10.1148/ryct.2020200277 41. Penkava J., Muenchhoff M., Badell I., Osterman A., Delbridge C., Niederbuchner F., Soliman S., Rudelius M., Graf A., Krebs S.. Detection of SARS-CoV-2-RNA in post-mortem samples of human eyes. Graefes Arch. Clin. Exp. Ophthalmol. (2021.0) 260 1789-1797. DOI: 10.1007/s00417-021-05529-x 42. McDonald A.I., Shirali A.S., Aragón R., Ma F., Hernandez G., Vaughn D.A., Mack J.J., Lim T.Y., Sunshine H., Zhao P.. Endothelial Regeneration of Large Vessels Is a Biphasic Process Driven by Local Cells with Distinct Proliferative Capacities. Cell Stem. Cell (2018.0) 23 210-225.e216. DOI: 10.1016/j.stem.2018.07.011 43. González-Zamora J., Bilbao-Malavé V., Gándara E., Casablanca-Piñera A., Boquera-Ventosa C., Landecho M.F., Zarranz-Ventura J., García-Layana A.. Retinal Microvascular Impairment in COVID-19 Bilateral Pneumonia Assessed by Optical Coherence Tomography Angiography. Biomedicines (2021.0) 9. DOI: 10.3390/biomedicines9030247 44. Burgos-Blasco B., Güemes-Villahoz N., Vidal-Villegas B., Martinez-de-la-Casa J.M., Donate-Lopez J., Martín-Sánchez F.J., González-Armengol J.J., Porta-Etessam J., Martin J.L.R., Garcia-Feijoo J.. Optic nerve and macular optical coherence tomography in recovered COVID-19 patients. Eur. J. Ophthalmol. (2021.0) 32 628-636. DOI: 10.1177/11206721211001019 45. Oren B., Aksoy Aydemır G., Aydemır E., Atesoglu H.I., Goker Y.S., Kızıltoprak H., Ozcelık K.C.. Quantitative assessment of retinal changes in COVID-19 patients. Clin. Exp. Optom. (2021.0) 104 717-722. DOI: 10.1080/08164622.2021.1916389 46. Cetinkaya T., Kurt M.M., Akpolat C.. Analysis of swept-source optical coherence tomography angiography measurement alterations in adult patients recovered from COVID-19. Clin. Exp. Optom. (2021.0) 105 848-852. DOI: 10.1080/08164622.2021.1999770 47. Banderas García S., Aragón D., Azarfane B., Trejo F., Garrell-Salat X., Sánchez-Montalvá A., Otero-Romero S., Garcia-Arumi J., Zapata M.A.. Persistent reduction of retinal microvascular vessel density in patients with moderate and severe COVID-19 disease. BMJ Open Ophthalmol. (2022.0) 7 e000867. DOI: 10.1136/bmjophth-2021-000867 48. Díaz M., Novo J., Cutrín P., Gómez-Ulla F., Penedo M.G., Ortega M.. Automatic segmentation of the foveal avascular zone in ophthalmological OCT-A images. PLoS ONE (2019.0) 14. DOI: 10.1371/journal.pone.0212364 49. Nicolò M., Rosa R., Musetti D., Musolino M., Saccheggiani M., Traverso C.E.. Choroidal Vascular Flow Area in Central Serous Chorioretinopathy Using Swept-Source Optical Coherence Tomography Angiography. Investig. Ophthalmol. Vis. Sci. (2017.0) 58 2002-2010. DOI: 10.1167/iovs.17-21417
--- title: Ticagrelor Can Regulate the Ion Channel Characteristics of Superior Cervical Ganglion Neurons after Myocardial Infarction authors: - Lijun Cheng - Lin Yu - Xiaoping Zhan - Gary Tse - Tong Liu - Huaying Fu - Guangping Li journal: Journal of Cardiovascular Development and Disease year: 2023 pmcid: PMC9966694 doi: 10.3390/jcdd10020071 license: CC BY 4.0 --- # Ticagrelor Can Regulate the Ion Channel Characteristics of Superior Cervical Ganglion Neurons after Myocardial Infarction ## Abstract Background: The superior cervical ganglion (SCG) plays a key role in cardiovascular diseases. The aim of this study was to determine the changes in the ion channel characteristics of the SCG following myocardial infarction (MI) and the role of pretreatment with the P2Y12 receptor antagonist ticagrelor (TIC). Methods: A total of 18 male rabbits were randomly divided into a control group, MI group, and P2Y12 receptor antagonist (TIC) group (abbreviated as the TIC group). Rabbit MI was performed via two abdominal subcutaneous injections of 150 mg·kg−1·d−1 of isoproterenol (ISO) with an interval of 24 h. TIC pretreatment at 20 mg·kg−1·d−1 was administered via gavage for two consecutive days. The cardiac function of each group was evaluated with echocardiography. ADP receptor P2Y12 expressions in SCGs were determined using RT-PCR and immunofluorescence staining. Ion channel characteristics of SCG neurons were measured using a whole-cell patch clamp. Intracellular calcium concentrations for SCG neurons were measured using confocal microscopy. Results: Cardiac function was reduced in the rabbits of the MI group, the sympathetic nerve activity of SCGs was increased, and the current amplitude of the neuron ion channel was increased. MI led to alterations in the activation and inactivation characteristics of INa channels accompanied by increased expression of P2Y12 in SCGs. Most of these abnormalities were prevented by TIC pretreatment in the TIC group. Conclusions: TIC pretreatment could attenuate the increase in P2Y12 expression in SCGs and the changes to the ion channel characteristics of SCG neurons after MI. This may be the mechanism underlying the cardiac protective effects of TIC. ## 1. Introduction Myocardial infarction (MI), as one of the common cardiovascular diseases, is a significant problem for middle-aged and elderly individuals. Cardiac sympathetic nerves are over-excited after MI [1,2], and the accumulation of neurotransmitters is an important cause and mechanism of arrhythmia after MI [3,4]. The superior cervical ganglion (SCG) is an important extracardiac ganglion that participates in the regulation of heart function. Previous studies have demonstrated that many components of SCGs are involved in the regulation of cardiac function after MI [5,6]. Our previous results showed that the electrophysiological levels of extracardiac ganglion neurons increased significantly after myocardial ischemia and infarction [7,8,9,10]. These changes in SCGs were closely related to the occurrence of ventricular arrhythmias after MI. Therefore, attenuating the activity of SCGs in MI patients can prevent abnormal cardiac electrophysiological function. Ion channels mediate electrical activity in neuronal cells, and these channels for SCG neurons provide the electrophysiological basis of sympathetic nerve excitability. Our prior studies have shown that, after myocardial ischemia, the characteristics of the ion channels of the SCG neurons changed significantly, and the channel current amplitude increased [8]. Reducing changes to the electrical properties of ion channels helped to reduce the increase in sympathetic nerve activity after MI. P2Y12, a P2Y metabotropic G-protein-coupled purinergic receptor, is expressed in the sympathetic ganglia and can regulate neuronal function [11]. P2Y12 was found to be involved in the enhanced excitatory synaptic responses in the substantia gelatinosa neurons after nerve injury [12]. After MI, the expression of P2Y12 receptors in some sympathetic ganglia increased significantly [11]. Moreover, P2Y12 affects ion channel characteristics [13,14]. P2Y12 receptor inhibitors can improve ion channel function [15] and protect against ischemia-induced neural injury [16]. Therefore, P2Y12 may be involved in the process of changing the electrical properties of SCG neurons’ ion channels after MI. The use of P2Y12 receptor inhibitors may be beneficial for the recovery of sympathetic nerves after MI, but the specific mechanism is not clear. The aim of this study was to investigate the role of the P2Y12 receptor antagonist ticagrelor (TIC) in electrophysiological remodeling of SCG neurons in a rabbit model of MI. ## 2.1. Animals and Experimental Design This study was approved by the Animal Ethical and Welfare Committee of the Chinese Academy Medical Sciences Institute of Radiation Medicine. The animals’ use, their groupings, and the MI preparation methods for experimental animals are briefly introduced as follows: 18 male rabbits weighing 300–500 g (Tianjin Yuda Experimental Animal Co., Ltd., Tianjin, China) were randomly and equally divided into a control group, MI group, and P2Y12 receptor antagonist (TIC) group (abbreviated as the TIC group). Rabbit MI was performed via two abdominal subcutaneous injections of 150 mg·kg−1·d−1 of isoproterenol (ISO) with an interval of 24 h. The successful establishment of the MI model was verified in our previous study [7]. Rabbits in the TIC groups were administered 20 mg·kg−1·d−1 TIC (on two consecutive days) via gavage, then MI was induced via ISO. Rabbits in the control group were administered a normal saline injection subcutaneously. After 24 h of treatment, SCG tissues and blood were collected from the three groups of rabbits for the following experiments. The successful establishment of the MI model was determined through the assessment of myocardial enzymes (creatine kinase (CK) and creatine kinase isoenzyme (CKMB)), electrocardiograms (ECGs), and echocardiography recordings. ## 2.2. Echocardiography After 24 h of the intervention, rabbits in each group were anesthetized with $3\%$ pentobarbital sodium to perform transthoracic echocardiography. Echocardiography parameters, including left atrial diameter (LAD), left ventricular end-diastolic dimension (LVDD), left ventricular ejection fraction (LVEF), and fractional shortening (FS), were recorded using a small-animal ultrasound system (Vevo 2100, VisualSonics, Toronto, ON, Canada). ## 2.3. Preparation and Electrophysiological Recordings of SG Neurons The methods for the SCG neuron acquisition and whole-cell patch-clamp recording were described in our previous reports [7,8,9,10]. In short, SCG slices were cut, digested in the enzyme solution (50–60 min, 37 °C), and dispersed gently with glass tubes into a single neuron. For the enzyme solution, 0.6–0.7 g/L pronase E (Roche, Basel, Switzerland), 1.7–1.8 g/L collagenase type II (Worthington, Lakewood, CO, USA), and 7.0–8.0 g/L bovine serum albumin (Roche, Basel, Switzerland) were added to the incubation solution (NaCl 130 mmol/L, MgCl2 1 mmol/L, KCl 5 mmol/L, glucose 10 mmol/L, CaCl2 2 mmol/L, NaH2PO4 1.5 mmol/L, NaHCO3 25 mmol/L, and HEPES 10 mmol/L). Here, we mainly recorded the characteristics of the delayed rectifier potassium channel current (IK), sodium channel current (INa), and N-type calcium channel current (ICa). The data was recorded using Axopach 200B patch-clamp systems and analyzed using Clampex 10.2 software (Axon, Scottsdale, AZ, USA). ## 2.4. Intracellular Calcium Measurement in SCG Neurons The method for recording the intracellular calcium concentrations of SCG neurons was described in detail in our previous studies [7,8]. The SCG neurons were loaded in incubation solution including 3 μM Fluo 4-AM and Pluronic F-127 (<$0.05\%$) (40 min, 37 °C). After that, the change in the fluorescence intensity in SCG neurons was monitored using confocal microscopy (FV1000, Olympus, Tokyo, Japan). F/F0 was used to reflect intracellular calcium concentrations, where F0 and F are the fluorescence intensity before and after adding 60 mmol/L KCl, respectively. ## 2.5. Immunofluorescence Staining SCG tissues were fixed in $4\%$ paraformaldehyde, embedded in paraffin, sectioned into 5 μm slices, and incubated with anti-tyrosine hydroxylase (TH, 1:300, Boster, Wuhan, China) and anti-P2Y12 (1:200, Bioss, Beijing, China) overnight at 4 °C. After washes with PBS, SCG tissue slices were incubated with CoraLite594-conjugated Goat Anti-Rabbit IgG (1:200, Proteintech, Wuhan, China) and CoraLite488-conjugated Affinipure Goat Anti-Mouse IgG (1:200, Proteintech, Wuhan, China) secondary antibodies. In order to identify the nucleus, the slices were counterstained with 4′,6-diamidino-2-phenylindole (DAPI, Solarbio, Beijing, China). Then, the fluorescence intensity of the slices was detected using a confocal microscope (FV1000). ## 2.6. RT-PCR Briefly, total RNA extraction and cDNA synthesis were completed using a kit (Tiangen, Beijing, China) following the manufacturer’s instructions. RT-PCR was performed with a Quant Gene 9600 System (Bioer Technology, Hangzhou, China). Relative gene expression was calculated with the 2-ΔΔCT method. All values obtained with the P2Y12 primers were normalized to the values obtained with the GAPDH primers. The primer sequences of GAPDH were as follows: (forward) TCGTGGATGACCTTGGCC and (reverse) GATGCTGGTGCCGAGTAC. The primer sequences of P2Y12 were as follows: (forward) ACCAGTTTGGAACCGCTAAA and (reverse) GTAGGCCCACATCAAATGCT. ## 2.7. ELISA Blood samples were collected for norepinephrine (NE) determination 24 h after MI and TIC intervention. Serum was collected and stored at −80 °C. We used a commercial NE ELISA kit (Cloud-Clone Corp, Wuhan, China) to measure the NE concentration in rabbit serum. ## 2.8. Statistical Analysis The variables are presented as means ± standard deviation (SD). The comparison between the groups was analyzed with one-way ANOVA. An LSD t-test was used for post hoc analysis. $p \leq 0.05$ was considered to be statistically significant. Data were analyzed with Origin 6.0 (OriginLab, Northampton, MA, USA), Clampex 10.2, and SPSS 17.0 software (SPSS Inc., Chicago, IL, USA). ## 3.1. MI Model Validation and Changes in Cardiac Function after MI and TIC MI was validated by (a) increased levels of myocardial enzymes, (b) ST segment elevation in ECGs, and (c) reduction in cardiac function (Figure 1A–D). The effects of MI and TIC on cardiac function were detected by echocardiography. Compared with the control group, the cardiac chamber dilation was decreased and left ventricle function reduced in the MI group, which manifested as increasing LAD and LVDD and decreasing LVEF and FS ($$n = 6$$, $p \leq 0.05$; Figure 1C,D). These changes were mostly reversed by the TIC intervention. Thus, although TIC did not completely restore cardiac function after MI, it improved cardiac function. ## 3.2. P2Y12 Expression in SCG and NE content in Serum after MI and TIC First, we examined the effects of MI and TIC treatment on P2Y12 expression in SCGs using immunofluorescence staining and RT-PCR. Double staining with antibodies against TH and P2Y12 was applied to SCG neurons. The results showed that P2Y12 was mainly expressed in the TH neurons in the SCGs of the three groups. Immunofluorescence and RT-PCR results showed that P2Y12 levels in SCGs were significantly higher in the MI group compared to the control group ($$n = 5$$, $p \leq 0.05$) and significantly lower in the TIC group compared to the MI group ($$n = 5$$, $p \leq 0.05$) (Figure 2A,B). As sympathetic activity is directly related to neurotransmitter release, we examined the effects of TIC and MI on the NE content in serum. The ELISA results showed that NE content in serum was significantly increased in MI-group rabbits compared to those in the control group ($$n = 6$$, $p \leq 0.05$) and significantly lower in the TIC group compared to the MI group ($$n = 6$$, $p \leq 0.05$) (Figure 2C). It can be seen that TIC treatment can significantly improve P2Y12 expression in the SCGs and NE content in serum after MI. Next, we examined the effects of MI and TIC on electrophysiological properties and intracellular calcium concentrations in SCG neurons. ## 3.3. Effects of MI and TIC on Activation Kinetics of IK The patch-clamp protocol and the drawing of the curves (current-voltage (I-V) curves and activation curves) for IK were carried out with reference to our previous study [7,8]. I-V curves were drawn by plotting the peak current density (the x-axis represented the different test membrane potentials; the y-axis represented the maximum current amplitude/neuron membrane capacitance). The patch-clamp protocol, curve sample, and I-V curves for IK are shown in Figure 3A,B. As seen in the I-V curve, the peak current density of IK was increased (97.76 ± 18.91 vs. 75.15 ± 19.14 pA/pF, $$n = 10$$, $p \leq 0.05$) in the MI group. TIC pretreatment slightly reduced the current amplitude of IK (86.99 ± 32.02 pA/pF), but there was no significant difference with the MI group ($$n = 10$$, $p \leq 0.05$). The patch-clamp protocol and the activation curves for IK are shown in Figure 3C. The current amplitude with different membrane potentials was converted into conductance using the equation: G = I/(V − Vrev), where Vrev is the reversal potential and V is the membrane potential. The activation curves were fitted using the Boltzmann equation: G/Gmax = 1/{1 + exp[(V$\frac{1}{2}$ − V)/k]}, where Gmax is the maximum conductance, V$\frac{1}{2}$ is the potential at half-activation, and k is the slope factor of the curves. According to the activation curve-fitting results, the V$\frac{1}{2}$ values for the three groups were 6.35 ± 0.64 mV, −2.35 ± 0.38 mV, and 0.68 ± 0.68 mV ($$n = 10$$, $p \leq 0.05$). The k values for the three groups were 25.79 ± 1.11, 16.99 ± 0.43, and 25.51 ± 1.08 ($$n = 10$$, $p \leq 0.05$). A comparison showed that there were no significant differences in the activation curves for the three groups ($$n = 10$$, $p \leq 0.05$). Therefore, TIC could improve the current amplitude of IK, but neither MI nor TIC had an effect on the activation curve of IK. ## 3.4. Effects of MI and TIC on Activation Kinetics of INa The patch-clamp protocol and the drawing of the curves (I-V curves, the activation curves, the inactivation curves, and the recovery curves) for INa were carried out with reference to our previous study [7,8]. The patch-clamp protocol, curve sample, and I-V curves for INa are shown in Figure 4A,B. I-V curves were drawn using the aforementioned method. The I-V curve showed that the peak amplitude of the INa density was increased (−143.30 ± 46.11 vs. −88.96 ± 20.65 pA/pF, $p \leq 0.05$) in the MI group. TIC pretreatment could reduce the current amplitude of INa (−108.52 ± 21.88 pA/pF, $$n = 10$$, $p \leq 0.05$). The patch-clamp protocol and activation curves for INa for the three groups are shown in Figure 4C. The activation curves were drawn using the aforementioned method. The fitting results for the activation curve showed that the V$\frac{1}{2}$ values for the three groups were −40.18 ± 0.27 mV, −47.11 ± 0.64 mV, and −44.59 ± 0.93 mV ($$n = 10$$, $p \leq 0.05$). The k values were 1.88 ± 1.78, 5.18 ± 0.57, and 1.61 ± 0.30 ($$n = 10$$, $p \leq 0.05$). The activation curves for the MI group were shifted toward negative potential compared to those of the control group, and the k values in the activation curves were increased ($$n = 10$$, $p \leq 0.05$). However, these changes were partially reversed in the TIC group. The patch-clamp protocol and inactivation curves for INa for the three groups are shown in Figure 4D. The inactivation curves were fitted with the Boltzmann equation: I/Imax = 1/{1 + exp[(V − V$\frac{1}{2}$)/k]}, where Imax is the maximum current amplitude, V$\frac{1}{2}$ is the potential at half-inactivation, and k is the slope factor. From the fitting results for the inactivation curve, the V$\frac{1}{2}$ values for the three groups were −70.63 ± 0.72 mV, −69.34 ± 0.55 mV, and −64.77 ± 0.51 mV ($$n = 10$$, $p \leq 0.05$). The k values were 11.36 ± 0.71, 8.45 ± 0.50, and 7.69 ± 0.46 ($$n = 10$$, $p \leq 0.05$). Compared to the control group, the inactivation curves for the MI group and TIC group did not change significantly ($$n = 10$$, $p \leq 0.05$). The patch-clamp protocol and the recovery curves for INa for the three groups are shown in Figure 4E. The intervals of the double pulses were 1 ms, 2 ms, 3 ms, 5 ms, 8 ms, 12 ms, 20 ms, 50 ms, and 100 ms, respectively. The recovery curves for INa were fitted with the exponential equation: I/Imax = 1 − exp(−t/τ), where Imax is the maximum current amplitude, t is the recovery time, and τ is the time constant for channel recovery. The fitting results for the recovery curve showed that the τ values of the three groups were 8.45 ± 0.60 ms, 4.30 ± 0.39 ms, and 6.09 ± 0.44 ms ($$n = 10$$, $p \leq 0.05$). The recovery curves for the MI group were shifted toward negative potential compared to those of the control group, and τ values were reduced ($$n = 10$$, $p \leq 0.05$). These changes were mostly reversed in the TIC group ($$n = 10$$, $p \leq 0.05$). Therefore, TIC could improve the changes to INa characteristics after MI, including the current amplitude, activation, and recovery curve. ## 3.5. Effects of MI and TIC on ICa and Intracellular Calcium Concentration Next, we measured the ICa characteristics and the intracellular calcium concentration of SCG neurons after MI. The patch-clamp protocol, curve sample, and I-V curves for ICa are shown in Figure 5A,B. The I-V curves were drawn using the aforementioned method. The I-V curve showed that the peak amplitude of the ICa density was increased (−12.73 ± 4.04 vs. −10.14 ± 1.79 pA/pF, $$n = 8$$, $p \leq 0.05$) in the MI group. This was reduced by TIC pretreatment (−12.23 ± 1.81 pA/pF, $$n = 8$$, $p \leq 0.05$). When the voltage was at −10 mV and 10 mV, the peak amplitude of the MI group increased significantly ($p \leq 0.05$). Although the current amplitude of the TIC group decreased compared to that of the MI group, there was no significant difference between the TIC group and MI group ($p \leq 0.05$). The intraneuronal calcium concentration was obtained using Fluo 4-AM staining and confocal microscopy. Representative recordings of fluorescence curves and F/F0 ratios ($$n = 10$$) are shown in Figure 5C. The F/F0 ratio was higher in the MI group compared to that in the control group (3.42 ± 0.55 vs. 2.86 ± 0.47; $p \leq 0.05$), and TIC pretreatment (3.18 ± 0.37) had no significant effects compared to the control and MI groups ($p \leq 0.05$). TIC pretreatment could partly attenuate the effects of MI on the current amplitude of the calcium channel and the intracellular calcium concentration. ## 4. Discussion In the present study, we investigated the effects of P2Y12 receptor antagonist (TIC) on P2Y12 expression and the electrophysiological characteristics of SCG neurons after MI. After MI, cardiac function and the NE content in serum changed, the expression of P2Y12 protein in SCGs increased, and the characteristics of neuronal ion channels were altered. TIC pretreatment produced reversals of these changes in SCGs. ## 4.1. Remodeling in SCG Neurons following MI The sympathetic nervous system (SNS) plays an important role in the regulation of cardiac function after MI [17,18]. MI can lead to excessive excitement of sympathetic nerves, leading to neurotransmitter accumulation [1,2], which is an important reason for the cardiac arrhythmias observed post-MI [3,4]. However, little is known about the molecular mechanisms regulating cardiac sympathetic innervation. The SCG and the stellate ganglion (SG) are important extracardiac ganglia of the heart. As the SG is closer to the heart, there are more studies on the SG. Our previous study showed that not only were the electrophysiological properties of SG neurons significantly altered after myocardial ischemia or infarction but the neuronal electrophysiological properties of the SCG were also significantly altered after MI [7,8,9,10]. In fact, the SCG plays a crucial role in the regulation of cardiac function. Anatomic evidence that the SCG innervates the heart has been reported for a long time, and it is known that postganglionic fibers participate in the formation of the cardiac plexus, innervating myocardial tissues. SCG blockade was also found to improve cardiac fibrosis and cardiac function, and the instability of ventricular electrophysiology was reduced [5]. Many components of SCGs are involved in the regulation of cardiac function after MI. For example, activation of the GABAergic signaling system in SCG sympathetic neurons can suppress sympathetic activity, thereby facilitating cardiac protection and making it a potential target to alleviate ventricular arrhythmias [6]. The expression of the oxytocin receptor in SCGs was enhanced after MI, suggesting that the involvement of the oxytocin receptor in SCGs may contribute to the transmission of sympathetic responses after MI [19]. ATP and the P2X7 receptor in the SCG also participate in sympathoexcitatory transmission after myocardial ischemia injury [20,21]. All these findings indicate that SCGs play important roles in pathophysiology after MI. Our previous results showed that MI led to significantly higher sympathetic neuron activity, which was confirmed by the electrophysiological activity of SCG neurons increasing significantly after myocardial ischemia and infarction [7,8,9]. The experimental results of the present study also confirmed that not only did the NE content in serum increase but the ion channel and intracellular calcium concentration of SCG neurons also changed significantly after MI. These changes reflected the activation of sympathetic nerves after MI, which then induced cardiac electrophysiological instability. Therefore, reducing the activity of SCGs following MI is helpful to prevent dysfunction in the sympathetic nerves and cardiac electrophysiological instability. ## 4.2. Ion Channels in SCGs after MI Sympathetic hyperactivity is known to contribute to the pathophysiology of a variety of cardiovascular diseases, including MI [1,2]. Sympathetic ganglion neuron ion channels play an important role in sympathetic hyperactivity. Many of these ion channels are involved in the pathophysiological processes of cardiovascular diseases, such as hypertension and ischemic heart disease [22,23,24]. Blocking the Nav1.8 channel significantly attenuated ischemia-induced ventricular arrhythmia, primarily by suppressing sympathetic ganglion activity [24]. Overactivation of Cav2.2 channels in sympathetic ganglion neurons contributed to cardiac sympathetic hyperactivation and the occurrence of ventricular arrhythmogenesis [25,26]. Ion channels can serve as potential therapeutic targets in disease treatment [27]. Similarly, ion channels are the basis of neurons’ electrical activity in sympathetic ganglia. Among these channels, it has been found that sodium channels govern AP upstroke and propagation in neurons [28], potassium channels mediate the AP repolarization process [29], and calcium channels are not only involved in the process of action-potential firing but are also closely related to intracellular calcium concentration and neurotransmitter release [30,31]. These channels are also expressed in SCG neurons, and they are the basis of sympathetic nerve excitability. Our results showed that not only did the activity of neuronal ion channels increase significantly in SCG neurons after MI but the NE content in serum also increased significantly. Reducing the activity of ion channels in SCG neurons is an important way of reducing the adverse effects of sympathetic nerve overexcitation after MI. ## 4.3. P2Y12 Expression and Roles of P2Y12 Receptor Antagonist Treatment in MI P2Y receptors are expressed ubiquitously in the body, including in the central nervous system and microglia, with physiological roles in neurotransmission, neurogenesis, and a number of peripheral pathophysiological processes [32]. P2Y12 is a P2Y metabolic G-protein-coupled purinergic receptor expressed in sympathetic ganglia that regulates neuronal function [11]. P2Y12 is involved in the enhanced excitatory synaptic responses in the substantia gelatinosa neurons after nerve injury [12]. After MI, the expression of P2Y12 receptors in some sympathetic ganglia increased significantly [11]. In some diseases, sympathetic activity can be reduced by inhibiting P2Y12 in sympathetic ganglia [33]. It can be seen that P2Y12 is involved in the pathophysiological process of increasing sympathetic activity after MI. Downregulation of the P2Y12 receptor in the SCG after MI may improve cardiac function by alleviating the sympathoexcitatory reflex [34]. In addition, many literature reports have shown that P2Y receptor affects the characteristics of neuronal ion channels and the release of neurotransmitters [32,35,36,37]. Similarly, P2Y12 also has a direct impact on ion channels [13,14], which may be involved in the regulation of SCG neuronal ion channel characteristics after MI. 32, 6T TIC is a typical P2Y12 receptor inhibitor. TIC significantly reduced the incidence of major adverse cardiovascular events in acute coronary syndrome patients [38,39]. The application of TIC in the treatment of ST-elevated acute coronary syndrome can increase the level of myocardial microcirculation perfusion and improve left heart function [40]. TIC can protect against ischemia-induced neural injury and has neuroprotective effects [16]. In addition, TIC improved ion channel function [15] and reduced calcium influx [41,42]. Our previous studies have shown that TIC attenuated the changes in the electrophysiological characteristics of SG neurons following MI [7]. Similarly, the results of this study showed that TIC attenuated the change in the electrophysiological characteristics of SCG neurons after MI, and the expression of P2Y12 in SCGs and the NE content in serum after MI were reduced, which also confirmed the neuroprotective effects of TIC from another perspective. It can be deduced that TIC prevented the development of electrophysiological abnormalities in SCG neurons after MI by improving the expression of P2Y12 in ganglia and the characteristics of the neuronal ion channel, and it further protected the cardiovascular system and sympathetic nerves. Several studies have shown that TIC is associated with bradycardia, atrioventricular blockage, and ventricular pauses [43,44]. Caution and careful monitoring are required, especially for patients with already compromised conduction systems and/or treated with AV blocking agents [45]. The mechanism whereby TIC leads to bradyarrhythmia is unclear; one possibility is a direct effect of ticagrelor on cardiac autoregulation and conduction and another may be a modulatory effect of adenosine. TIC inhibits cellular uptake and increases the plasma concentration of adenosine [46]. However, the increase in adenosine reduces cardiac sympathetic efferent nerve activity [47]. The reduction in sympathetic nerve activity may further lead to bradycardia and atrioventricular blockage, which may be one of the mechanisms whereby TIC causes bradycardia, as is also verified from the perspective of sympathetic nerves. TIC is a commonly used P2Y12 receptor inhibitor in clinical contexts. In addition, it also includes prasugrel, clopidogrel, and other inhibitors. Although our study did not examine whether other inhibitors have effects on SCG neuronal ion channels after MI, the findings suggest that other inhibitors may have the same effects. No relevant research has been published at present, and the hypothesis needs to be verified in subsequent work. In conclusion, our experimental results suggest that TIC has neuroprotective effects after myocardial infarction. However, for clinical applications, ECG and sympathetic nerve activity should be monitored in patients with acute coronary syndrome when P2Y12 receptor inhibitors are prescribed. For patients with bradycardia and low sympathetic activity, P2Y12 receptor inhibitors should be used with caution. ## 4.4. Limitations Several limitations of this study should be recognized. Firstly, ion channel protein and P2Y12 expression were not validated by Western blotting in the present study because of the small size of the SCG tissue. Secondly, our previous studies have shown that TIC can inhibit the ion channel currents of stellate ganglion neurons, so experiments demonstrating that TIC acts on the ion channel currents of SCG neurons directly were not conducted in the present study. ## 5. Conclusions P2Y12 receptor antagonist (TIC) pretreatment partly reversed P2Y12 expression and abnormal neuronal electrophysiological changes in SCGs after MI. This may represent one of the mechanisms underlying the cardiovascular protection brought about by TIC. ## References 1. Yu L., Zhou L., Cao G., Po S.S., Huang B., Zhou X., Wang M., Yuan S., Wang Z., Wang S.. Optogenetic Modulation of Cardiac Sympathetic Nerve Activity to Prevent Ventricular Arrhythmias. J. Am. Coll. Cardiol. (2017) 70 2778-2790. DOI: 10.1016/j.jacc.2017.09.1107 2. Xiong L., Liu Y., Zhou M., Wang G., Quan D., Shen C., Shuai W., Kong B., Huang C., Huang H.. Targeted ablation of cardiac sympathetic neurons improves ventricular electrical remodelling in a canine model of chronic myocardial infarction. EP Eur. (2018) 20 2036-2044. DOI: 10.1093/europace/euy090 3. Tan A.Y., Elharrif K., Cardona-Guarache R., Mankad P., Ayers O., Joslyn M., Das A., Kaszala K., Lin S.-F., Ellenbogen K.A.. Persistent Proarrhythmic Neural Remodeling Despite Recovery From Premature Ventricular Contraction-Induced Cardiomyopathy. J. Am. Coll. Cardiol. (2020) 75 1-13. DOI: 10.1016/j.jacc.2019.10.046 4. Zhang W., Zhou Q., Lu Y., Li Y., Zhang L., Zhang J., Xing Q., Lv W., Cheng X., Zhang G.. Renal Denervation Reduced Ventricular Arrhythmia After Myocardial Infarction by Inhibiting Sympathetic Activity and Remodeling. J. Am. Heart Assoc. (2018) 7 e009938. DOI: 10.1161/JAHA.118.009938 5. Wang Y., Jiang W., Chen H., Zhou H., Liu Z., Liu Z., Zhou Y., Zhou X., Yu L., Jiang H.. Sympathetic Nervous System Mediates Cardiac Remodeling After Myocardial Infarction in a Circadian Disruption Model. Front. Cardiovasc. Med. (2021) 8 668387. DOI: 10.3389/fcvm.2021.668387 6. Shi Y., Li Y., Yin J., Hu H., Xue M., Li X., Cheng W., Wang Y., Li X., Wang Y.. A novel sympathetic neuronal GABAergic signalling system regulates NE release to prevent ventricular arrhythmias after acute myocardial infarction. Acta Physiol. (2019) 227 e13315. DOI: 10.1111/apha.13315 7. Cheng L., Fu H., Wang X., Ye L., Lakhani I., Tse G., Zhang Z., Liu T., Li G.. Effects of ticagrelor pretreatment on electrophysiological properties of stellate ganglion neurons following myocardial infarction. Clin. Exp. Pharmacol. Physiol. (2020) 47 1932-1942. DOI: 10.1111/1440-1681.13385 8. Cheng L., Wang X., Liu T., Tse G., Fu H., Li G.. Modulation of Ion Channels in the Superior Cervical Ganglion Neurons by Myocardial Ischemia and Fluvastatin Treatment. Front. Physiol. (2018) 9 1157. DOI: 10.3389/fphys.2018.01157 9. Cheng L.J., Zhang Y., Li J., Li G.P.. Characteristics of sympathetic ganglion neuron channels in two myocardial infarction models: A comparative study. Chin. J. Biomed. Eng. (2017) 23 193-198 10. Cheng L.-J., Li G.-P., Li J., Chen Y., Wang X.-H.. Effects of Fluvastatin on Characteristics of Stellate Ganglion Neurons in a Rabbit Model of Myocardial Ischemia. Chin. Med. J. (2016) 129 549-556. DOI: 10.4103/0366-6999.176991 11. Zou L., Han X., Liu S., Gong Y., Wu B., Yi Z., Liu H., Zhao S., Jia T., Li L.. Baicalin Depresses the Sympathoexcitatory Reflex Induced by Myocardial Ischemia via the Dorsal Root Ganglia. Front. Physiol. (2018) 9 928. DOI: 10.3389/fphys.2018.00928 12. Yu T., Zhang X., Shi H., Tian J., Sun L., Hu X., Cui W., Du D.. P2Y12 regulates microglia activation and excitatory synaptic transmission in spinal lamina II neurons during neuropathic pain in rodents. Cell Death Dis. (2019) 10 165. DOI: 10.1038/s41419-019-1425-4 13. Swiatkowski P., Murugan M., Eyo U., Wang Y., Rangaraju S., Oh S., Wu L.-J.. Activation of microglial P2Y12 receptor is required for outward potassium currents in response to neuronal injury. Neuroscience (2016) 318 22-33. DOI: 10.1016/j.neuroscience.2016.01.008 14. Gafar H., Rodriguez M.D., Chandaka G.K., Salzer I., Boehm S., Schicker K.. Membrane coordination of receptors and channels mediating the inhibition of neuronal ion currents by ADP. Purinergic Signal. (2016) 12 497-507. DOI: 10.1007/s11302-016-9516-5 15. Olgar Y., Durak A., Degirmenci S., Tuncay E., Billur D., Ozdemir S., Turan B.. Ticagrelor alleviates high-carbohydrate intake induced altered electrical activity of ventricular cardiomyocytes by regulating sarcoplasmic reticulum–mitochondria miscommunication. Mol. Cell. Biochem. (2021) 476 3827-3844. DOI: 10.1007/s11010-021-04205-2 16. Gelosa P., Lecca D., Fumagalli M., Wypych D., Pignieri A., Cimino M., Verderio C., Enerbäck M., Nikookhesal E., Tremoli E.. Microglia is a Key Player in the Reduction of Stroke Damage Promoted by the New Antithrombotic Agent Ticagrelor. J. Cereb. Blood Flow Metab. (2014) 34 979-988. DOI: 10.1038/jcbfm.2014.45 17. Amoni M., Dries E., Ingelaere S., Vermoortele D., Roderick H.L., Claus P., Willems R., Sipido K.R.. Ventricular Arrhythmias in Ischemic Cardiomyopathy—New Avenues for Mechanism-Guided Treatment. Cells (2021) 10. DOI: 10.3390/cells10102629 18. Wu P., Vaseghi M.. The autonomic nervous system and ventricular arrhythmias in myocardial infarction and heart failure. Pacing Clin. Electrophysiol. (2020) 43 172-180. DOI: 10.1111/pace.13856 19. Yu R., Wang F., Yin J., Shi Y., Wang Y., Wen S., Hu H., Yan S.. Expression of oxytocin receptor in the rat superior cervical ganglion after myocardial infarction. Int. J. Clin. Exp. Pathol. (2018) 11 739-747. PMID: 31938160 20. Liu J., Li G., Peng H., Tu G., Kong F., Liu S., Gao Y., Xu H., Qiu S., Fan B.. Sensory–sympathetic coupling in superior cervical ganglia after myocardial ischemic injury facilitates sympathoexcitatory action via P2X7 receptor. Purinergic Signal. (2013) 9 463-479. DOI: 10.1007/s11302-013-9367-2 21. Tu G., Zou L., Liu S., Wu B., Lv Q., Wang S., Xue Y., Zhang C., Yi Z., Zhang X.. Long noncoding NONRATT021972 siRNA normalized abnormal sympathetic activity mediated by the upregulation of P2X7 receptor in superior cervical ganglia after myocardial ischemia. Purinergic Signal. (2016) 12 521-535. DOI: 10.1007/s11302-016-9518-3 22. Davis H., Herring N., Paterson D.J.. Downregulation of M Current Is Coupled to Membrane Excitability in Sympathetic Neurons Before the Onset of Hypertension. Hypertension (2020) 76 1915-1923. DOI: 10.1161/HYPERTENSIONAHA.120.15922 23. Bardsley E.N., Davis H., Ajijola O.A., Buckler K.J., Ardell J.L., Shivkumar K., Paterson D.J.. RNA Sequencing Reveals Novel Transcripts from Sympathetic Stellate Ganglia During Cardiac Sympathetic Hyperactivity. Sci. Rep. (2018) 8 8633. DOI: 10.1038/s41598-018-26651-7 24. Yu L., Wang M., Hu D., Huang B., Zhou L., Zhou X., Wang Z., Wang S., Jiang H.. Blocking the Nav1.8 channel in the left stellate ganglion suppresses ventricular arrhythmia induced by acute ischemia in a canine model. Sci. Rep. (2017) 7 534. DOI: 10.1038/s41598-017-00642-6 25. Zhang D., Tu H., Wang C., Cao L., Hu W., Hackfort B.T., Muelleman R.L., Wadman M.C., Li Y.-L.. Inhibition of N-type calcium channels in cardiac sympathetic neurons attenuates ventricular arrhythmogenesis in heart failure. Cardiovasc. Res. (2020) 117 137-148. DOI: 10.1093/cvr/cvaa018 26. Zhang D., Tu H., Cao L., Zheng H., Muelleman R.L., Wadman M.C., Li Y.. Reduced N-Type Ca. J. Am. Heart Assoc. (2018) 7 e007457. DOI: 10.1161/JAHA.117.007457 27. Hutchings C.J., Colussi P., Clark T.G.. Ion channels as therapeutic antibody targets. Mabs (2019) 11 265-296. DOI: 10.1080/19420862.2018.1548232 28. Barker B.S., Nigam A., Ottolini M., Gaykema R.P., Hargus N.J., Patel M.K.. Pro-excitatory alterations in sodium channel activity facilitate subiculum neuron hyperexcitability in temporal lobe epilepsy. Neurobiol. Dis. (2017) 108 183-194. DOI: 10.1016/j.nbd.2017.08.018 29. Zhang X., Wan J.-Q., Tong X.-P.. Potassium channel dysfunction in neurons and astrocytes in Huntington’s disease. CNS Neurosci. Ther. (2018) 24 311-318. DOI: 10.1111/cns.12804 30. Wang Q., Zhu G.-H., Xie D.-H., Wu W.-J., Hu P.. Taurine Enhances Excitability of Mouse Cochlear Neural Stem Cells by Selectively Promoting Differentiation of Glutamatergic Neurons Over GABAergic Neurons. Neurochem. Res. (2015) 40 924-931. DOI: 10.1007/s11064-015-1546-9 31. Liu J., Wang F., Wu D., Wang Y., Deng A.. Effects of Anthopleurin-Q on the Intracellular Free Ca. Drug Res. (2014) 64 436-441. DOI: 10.1055/s-0033-1361125 32. Guzman S.J., Gerevich Z.. P2Y Receptors in Synaptic Transmission and Plasticity: Therapeutic Potential in Cognitive Dysfunction. Neural Plast. (2016) 2016 1207393. DOI: 10.1155/2016/1207393 33. Sheng X., Wang J., Guo J., Xu Y., Jiang H., Zheng C., Xu Z., Zhang Y., Che H., Liang S.. Effects of Baicalin on Diabetic Cardiac Autonomic Neuropathy Mediated by the P2Y12 Receptor in Rat Stellate Ganglia. Cell. Physiol. Biochem. (2018) 46 986-998. DOI: 10.1159/000488828 34. Zou L., Gong Y., Zhao S., Yi Z., Han X., Wu B., Jia T., Li L., Yuan H., Shi L.. Downregulation of P2Y. J. Cell. Physiol. (2017) 233 3375-3383. DOI: 10.1002/jcp.26184 35. Boehm S.. P2Ys go neuronal: Modulation of Ca. Br. J. Pharmacol. (2003) 138 1-3. DOI: 10.1038/sj.bjp.0705044 36. Lechner S.G., Boehm S.. Regulation of neuronal ion channels via P2Y receptors. Purinergic Signal. (2004) 1 31-41. DOI: 10.1007/s11302-004-4746-3 37. Zaika O., Tolstykh G.P., Jaffe D.B., Shapiro M.S.. Inositol Triphosphate-Mediated Ca. J. Neurosci. (2007) 27 8914-8926. DOI: 10.1523/JNEUROSCI.1739-07.2007 38. Varenhorst C., Alström U., Braun O., Storey R., Mahaffey K.W., Bertilsson M., Cannon C.P., Scirica B.M., Himmelmann A., James S.K.. Causes of mortality with ticagrelor compared with clopidogrel in acute coronary syndromes. Hear (2014) 100 1762-1769. DOI: 10.1136/heartjnl-2014-305619 39. Lesiak M., Komosa A.. Dual antiplatelet therapy for reduction in mortality in patients with acute and chronic coronary syndromes. Adv. Interv. Cardiol. (2021) 17 340-343. DOI: 10.5114/aic.2021.112082 40. Liu Y., Ding L.-Y., Li X.-Z.. Therapy with ticagrelor for ST-elevated acute coronary syndrome accompanied by diabetes mellitus. Eur. Rev. Med. Pharmacol. Sci. (2019) 23 312-318. PMID: 31389599 41. Lariccia V., Macrì M.L., Matteucci A., Maiolino M., Amoroso S., Magi S.. Effects of ticagrelor on the sodium/calcium exchanger 1 (NCX1) in cardiac derived H9c2 cells. Eur. J. Pharmacol. (2019) 850 158-166. DOI: 10.1016/j.ejphar.2019.01.067 42. Kucuk M., Celen M., Yamasan B.E., Olgar Y., Ozdemir S.. Effects of Ticagrelor on Ionic Currents and Contractility in Rat Ventricular Myocytes. Cardiovasc. Drugs Ther. (2015) 29 419-424. DOI: 10.1007/s10557-015-6617-2 43. Scirica B.M., Cannon C.P., Emanuelsson H., Michelson E.L., Harrington R.A., Husted S., James S., Katus H., Pais P., Raev D.. The Incidence of Bradyarrhythmias and Clinical Bradyarrhythmic Events in Patients With Acute Coronary Syndromes Treated With Ticagrelor or Clopidogrel in the PLATO (Platelet Inhibition and Patient Outcomes) Trial: Results of the Continuous Electrocardiographic Assessment Substudy. J. Am. Coll. Cardiol. (2011) 57 1908-1916. DOI: 10.1016/j.jacc.2010.11.056 44. Pujade I., Perino J., Mathieu C., Arnaud M., Raschi E., Gatti M., Bezin J., Salvo F.. Risk of bradyarrhythmia related to ticagrelor: A systematic review and meta-analysis. Pharmacol. Res. (2020) 160 105089. DOI: 10.1016/j.phrs.2020.105089 45. De Maria E., Borghi A., Modonesi L., Cappelli S.. Ticagrelor therapy and atrioventricular block: Do we need to worry?. World J. Clin. Cases (2017) 5 178-182. DOI: 10.12998/wjcc.v5.i5.178 46. Goldberg A., Rosenfeld I., Nordkin I., Halabi M.. Life-threatening complete atrioventricular block associated with ticagrelor therapy. Int. J. Cardiol. (2015) 182 379-380. DOI: 10.1016/j.ijcard.2014.12.162 47. da Silva V.J.D., Gnecchi-Ruscone T., Bellina V., Oliveira M., Maciel L., de Carvalho A.C.C., Salgado H.C., Bergamaschi C.M., Tobaldini E., Porta A.. Acute adenosine increases cardiac vagal and reduces sympathetic efferent nerve activities in rats. Exp. Physiol. (2012) 97 719-729. DOI: 10.1113/expphysiol.2011.063925
--- title: Analysis of Morphological Parameters and Body Composition in Adolescents with and without Intellectual Disability authors: - Bogdan Constantin Ungurean - Adrian Cojocariu - Beatrice Aurelia Abalașei - Lucian Popescu journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9966700 doi: 10.3390/ijerph20043019 license: CC BY 4.0 --- # Analysis of Morphological Parameters and Body Composition in Adolescents with and without Intellectual Disability ## Abstract Compared to the tremendous volume of studies focusing on children and teenagers without disabilities, research regarding weight and body composition among young populations with an intellectual disability is relatively rare. Their number further decreases when we refer to specific age groups with intellectual deficits, such as children and adolescents younger than 18. In addition, studies are even scarcer when we wish to compare groups of subjects with different degrees of intellectual disability by gender. This study has a constative nature. The research sample comprises 212 subjects—girls and boys with an average age of 17.7 ± 0.2, divided into six groups by gender and type of intellectual disability. The parameters considered within the study include anthropometrical data and body composition determined using a professional device (Tanita MC 580 S). The findings of this study highlight the impact of intellectual disability on body composition in this age category. We hope it will help develop efficient strategies, recommendations, and intervention plans to ensure active participation in physical activities and categorisation within the optimal parameters of body composition indicators. ## 1. Introduction The prevalence of obesity at the paediatric age has doubled in the past 30 years among pre-schoolers and adolescents, and it has tripled in the age group of 6–11 [1]. Child obesity is a predictive factor for morbidity and mortality among adults: up to $80\%$ of obese children will be obese adults who will be highly susceptible to cancer, high blood pressure, strokes, hepatic and bile duct diseases, and osteoarthritis [2,3]. Youths with ID who are overweight or obese are also more likely to develop secondary conditions related to obesity, such as asthma, high blood cholesterol, diabetes, depression, and fatigue, compared to youths of the same population with normal weights [4]. Furthermore, children with intellectual disability (ID) can face several challenges concerning information processing (for instance, cognitive disorders, communication disorders, and limited mental function). Consequently, they have difficulties understanding and acquiring knowledge concerning health and developing healthy behaviours [5]. Recommended by the World Health Organisation, body mass index (BMI) and body composition are commonly used to measure obesity in different populations [6]. Considering these recommendations, it cannot be said whether these methods accurately measure body composition or fat distribution in populations with ID. Often, such individuals have specific anthropometry compared to those without disabilities [7]. Understanding the causes and effects of high body mass index or obesity remains essential when assessing the health states of persons with ID. For most populations with various forms of ID, BMI is a reasonable measure to identify the individuals most prone to the harmful effects of obesity. Current research [8] suggests that fat tissue is still the most detrimental to health and should be the target of any intervention measure. Several methods have been implemented to measure body composition among persons with ID, such as waist circumference, skinfold measurements, as well as bioelectrical impedance analysis (BIA) [9]. However, the findings of a recent study have indicated that skinfold measurements provide too many errors among people with ID. For instance, Waninge et al. [ 10] have concluded that it is impossible to observe the measurement conditions, requiring the measurement of skinfolds three times precisely in the same spot on a person’s body. Previous research on body composition among people with ID focused solely on assessing body fat while alternating lean mass [11]. However, total muscle mass, of which the skeletal mass is a primary component related primarily to the physical function of an individual, is considered, for the most part, the most significant compartment of the body in assessing the physiological and nutritional state. In addition, a recent study on children with intellectual disabilities has reported that bioelectrical impedance analysis (BIA) is more viable than skinfold measurements [12,13]. Because communication with people with ID and particularly with a severe intellectual disability is challenging and no other non-invasive body composition measuring tools are available, the feasibility and viability of BIA measurements can be more relevant. This study aimed to assess a series of morphological and body compositions among children with and without intellectual disability to characterise the morphofunctional normality and its disturbance. The data obtained after using the statistical–mathematical indicators will be analysed in relation to the literature. The research tasks may be summarised as collecting and studying the literature and processing the data collected based on the statistical–mathematical methods to provide an objective interpretation and to elaborate the conclusions of the research conducted. Primary assumption 1: Secondary assumptions: ## 2. Materials and Methods Ethics: All the procedures in this study conformed with the 1964 Declaration of Helsinki and its subsequent amendments. We conducted the research with the approval of the Ethics Commission of scientific research No. $\frac{10}{2020}$, and the date of approval was 7 October 2020. The research per se began a long time ago through discussions with the Physical Education and Sports teachers within the centres schooling children with intellectual disability, meetings with specialists in the field, and collecting and studying the literature. The measurements began in April 2021 and continued until November 2022, considering the pandemic context of that moment. Participants. The activities took place in the gymnasiums of the academic units, as well as in the physical therapy practices of the “Sf. Andrei” School Centre Gura Humorului, Suceava County; the “Constantin Păunescu” School Centre Iaşi; “Elisabeta Polihroniade” Inclusive Education School Centre Vaslui; “Emil Gârleanu” Special School No. 1 Galați. It is worth mentioning that the measurements were carried out in the first part of the day (in 9–13) for all the groups. These institutions educate children with different types of intellectual disabilities. The inclusion criterion in the study was that ID should not be associated with other disabilities. The subjects’ parents or tutors signed a protocol at the beginning of the school year. This study included 212 subjects of the aforementioned educational establishments, distributed into six groups by gender and type of disability, as illustrated in Table 1. Procedure. Morphological parameters and some components of body composition represent the dependent variables. TANITA PRO SOFTWARE Version 3.4.5—The Tanita PRO software pack was developed in partnership with an essential medical software developer (Medizin & Sevice GmbH, Chemnitz, Germany). The software can store and analyse the data from the Tanita MC 580 S monitor. In conformity with the EU Regulations, the software is medically approved and observes the standing Regulations [14] Eur lex. ( The Medical Devices Directives, Directive $\frac{93}{42}$/EEC of the Council of 14 June 1993 on medical devices). The use of TANITA MC580 S and TANITA PRO SOFTWARE generates many measurements; the most representative introduced within this study as dependent variables are body mass (kg), body mass index (BMI kg/h2), body fat (%), muscle mass (%), basal metabolic rate (kcal), body fat (Kg), muscle mass (Kg), and skeletal muscle mass (SMM). Statistical Analysis. MANOVA (two-way MANOVA)—The two-way multivariate analysis of variance (two-way MANOVA) is often considered an extension of the two-way ANOVA for situations in which there are two dependent variables. The primary purpose of the two-way MANOVA is to understand if there is an interaction between two independent variables on the other dependent variables combined. Considering the significant number of data (over 200 subjects), it is recommended to use a skewness statistical indicator to test the normality of data distribution, which evaluates the asymmetry degree of distribution and the kurtosis indicator. SPSS provides both tests. In [15], two z thresholds are proposed by the number of subjects tested. For a more significant number of data (over 150–200), the z threshold is 1.96 [16]. The Kruskal–Wallis H test is a rank-based non-parameter test that can determine if there are statistically significant differences between two or more groups of an independent variable on a continuous or ordinal dependent variable. The Kruskal–Wallis H test may be used when the data do not observe the unidirectional ANOVA assumptions. It occurs if (a) the data are not normally distributed or (b) there is an ordinal dependent variable. Descriptive analyses—in SPSS 20.0, through graphical and numerical synthesis, leaving some of the information out to gain relevance; The Tukey Procedure—(honestly significant difference—HSD) is a method based on q statistics, and it is preferred for group comparisons, two by two. The technique is effective for multiple group comparisons when groups are uneven. ## 3. Results To determine the tests used for data interpretation, we considered the values of skewness and kurtosis indicators concerning data distribution. After analysing the values of kurtosis (Table 2), we note that in four dependent variables (body mass—Kg, BMI (kg/m2), muscle mass %, and body fat Kg), for the nine dependent variables, there is no normal data distribution, which made us use a nonparametric test (Kruskal–Wallis H). For the other five dependent variables (height—cm, body fat %, BMR—kcal, muscle mass—Kg, and SMM) with normal data distribution, we used the Manova test. In this respect, we wished to determine the existence of a statistically significant interaction effect by interpreting the multivariate testing. It tests the null assumption according to which the covariance matrices of the dependent variables are equal between groups [17]. Consequently, we relied on Pillai’s Trace ($$p \leq 0.008$$); as shown in Table 3, there is a statistically significant interaction effect between gender and type of disability on the dependent variables combined ($p \leq 0.05$). In this situation, the assumption of the homogeneity of covariance matrices was not observed (Table 4), as assessed using Box’s M test ($p \leq 0.001$). Box’s M test is known to be very sensitive when multivariate normality is not observed, leading to a statistically significant result due to non-normality [18]. However, the MANOVA test is considered robust to the non-observance of this assumption. Hence, if the assumption of covariance equality is not observed, we may continue, regardless of whether the groups have similar sizes. Though Wilks’ Lambda test is usually recommended, Pillai’s Trace test is more robust, and it is a reliable choice when the samples are uneven, and the M matrix is present (Table 4); Box’s test—significant covariance equality ($p \leq 0.001$). As illustrated in Table 5, with the multiple comparisons by gender, several significant differences were recorded between the dependent variables with normal distribution for the groups of boys. The only dependent variable not influenced by the type of intellectual disability was body fat %. It is worth noting that we found significant differences ($p \leq 0.05$) in the other four dependent variables (height, BMR kcal, SMM, and muscle mass—kg) between the group of boys without intellectual disability and the group of boys with moderate intellectual disability (height $$p \leq 0.013$$, BMR kcal $$p \leq 0.005$$, SMM $$p \leq 0.001$$, muscle mass—kg $p \leq 0.001$) and between the group of boys without intellectual disability and the group of boys with severe intellectual disability (height $p \leq 0.001$, BMR kcal $$p \leq 0.009$$, SMM $$p \leq 0.001$$, muscle mass—kg $p \leq 0.001$). However, between the group of boys with moderate intellectual disability and the group of boys with severe intellectual disability, we found no significant differences in any dependent variable. We performed a Kruskal–Wallis H test to determine the existence of significant differences concerning the dependent variables without a normal distribution between the three groups of boys. The median body mass (kg) scores were statistically significantly different between groups, $$p \leq 0.005.$$ The median scores for IMC, body fat (kg), and muscle mass (%) were not statistically significantly different between groups p ˃ 0.05. The paired comparisons for body mass (kg) were performed using Dunn’s procedure, with a Bonferroni correction for multiple comparisons. Here, we feature the adjusted values of p. The post hoc analysis revealed significant differences in the median scores of body mass (kg) between the groups of boys with severe intellectual disability and without intellectual disability ($$p \leq 0.024$$) and between the groups of boys with moderate intellectual disability and without intellectual disability ($$p \leq 0.014$$). Between the groups of boys with severe intellectual disability and moderate intellectual disability, we found no significant differences (Appendix A). Concerning the groups of girls, as shown in Table 6, we only recorded significant differences in the dependent variable of height between the group of girls without intellectual disability and the group of girls with moderate intellectual disability ($$p \leq 0.036$$) and between the group of girls without intellectual disability and the group of girls with severe intellectual disability ($$p \leq 0.024$$). Regarding the other four dependent variables (BMR kcal, SMM, body fat %, muscle mass—kg), we found no significant differences between the three groups of girls. To analyse the data without a normal distribution in the groups of girls, we applied the Kruskal–Wallis H test. We found no significant differences between the four groups for any dependent variable (p ˃ 0.05). After analysing the data by gender, we found significant differences for the dependent variables with a normal distribution (Table 7). We identified values of $p \leq 0.05$ between almost all pairs of subjects, except for the SMM variable, between the group of girls with severe intellectual disability and the group of boys with severe intellectual disability ($$p \leq 0.137$$). For the dependent variables without a normal distribution, we recorded significant differences in three of the four dependent variables. The only dependent variable without differences by gender was body mass index (for BMI p ˃ 0.05) (Table 8). ## 4. Discussion In this study, significant differences ($p \leq 0.05$) were found in the groups of boys for five of the nine dependent variables, particularly between the group of boys without intellectual disability and the group of boys with moderate intellectual disability, as well as between the group of boys without intellectual disability and the group of boys with severe intellectual disability. However, between the groups of boys with different types of intellectual disability, no significant differences could be reported. One of the dependent variables not influenced by the type of disability or gender was body mass index. Though no statistical differences between groups were recorded, the means per group (Table 8) of the BMI exceeded the WHO guidelines [19]. Recent research shows similar findings among children without intellectual disability [20,21] and with intellectual disability [22]. The combined prevalence of overweight and obesity among European teenagers is 22–$25\%$ [23]. This figure has increased constantly in the past few decades, and now it appears to be rising faster in Western countries. Nonetheless, studies show differences in socioeconomic status and geographical position [24,25]; an increase was noted in children and adolescents with financial difficulties and those with intellectual disability [26]. Because they go through a stage of growth and development, body composition modifications are to be expected at this age. In boys, muscle mass indicators will increase, thus recording statistically significant differences by gender for all three categories ($p \leq 0.05$, Table 8), as sexual hormones lead to a substantial increase in muscle mass. Among girls, on the other hand, puberty development involves a period of fat tissue storage [27]. As illustrated in Table 8, the average values of the dependent variable (body fat—kg) were higher than the boys’ groups in all three categories; we found significant differences for $p \leq 0.05.$ This aspect is seen as a physiological preparation for birth, where extra energy is necessary to have and feed the new-born [28]. The differences can be due to multiple features of persons with ID, which may be approached individually or collectively, ascribing them to idleness, and social barriers hindering access to exercise programs, which affect the BMI and BMC alike [29]. Substantial body composition modifications among teenagers are due to puberty, the development process inherent to this stage of life, with significant differences between genders. Thus, it is challenging in longitudinal studies to distinguish between unhealthy weight gains and natural body composition modifications [30]. Unfortunately, few studies focus on the relationships between body composition in populations with different types of intellectual disability. It is even more challenging to find recent studies featuring differences in this population by gender. Against this backdrop, this study aimed to assess the relationship between body composition (assessed using a Tanita 580 S professional device relying on BIA technology) in teenagers with and without intellectual disability by gender. The substantial differences between genders concerning the body composition of adolescents argue for the presentation of our study findings by gender. This study has several limitations: the data analysis did not assess how much the subjects exercised and how many calories they burnt; the body composition using BIA technology was assessed in this research, which is both commonly used and reliable, but there was no comparison with other methods. This limitation can be considered as a future research topic to include diet and physical activity measurement. However, the study has several strong points, too: it is among the few pieces of research discussing body composition and body mass index by different types of intellectual disability and gender. ## 5. Conclusions Bioelectrical impedance analysis (BIA) is a commonly used technology in research concerning body composition because it is non-invasive and quick, and the data are highly reliable. It can be moved to various locations and is particularly easy to use for populations with different types of intellectual disability. This research has confirmed that the primary factors of body composition (body mass (Kg), body mass index (BMI kg/h2), body fat (%), muscle mass (%), basal metabolic rate (BMR kcal), body fat (Kg), muscle mass (Kg), skeletal muscle mass (SMM), and the morphological indicators (height and weight) may be influenced by both the type of disability and gender. The prevalence of overweight and obesity among people with intellectual disabilities was similar between male and female subjects. This shows an increasing trend with age. Body composition is an essential determinant of health states and nutritional indicators. Hence, body composition analysis is crucial in assessing such populations’ physiological and pathological states. Body composition evaluation has become increasingly popular in clinical practice, primarily due to the constant increase in the obesity rate. The results obtained in this study may help to develop intervention strategies for the treatment of obesity, provided that decision-makers prioritise the treatment of people with intellectual disabilities. ## References 1. Global Action Plan on Physical Activity 2018–2030: More Active People for a Healthier World 2. Hazart J., Montel F., Gentes E., Lahaye C., Pouget M., Farigon N., Miolanne M., Mulliez A., Boirie Y.. Body Mass Trajectory Affects the Long-Term Occurrence of Metabolic Syndrome in Adult Patients with Severe Obesity. Children (2023.0) 10. DOI: 10.3390/children10010027 3. Ferreira J.P., Matos R., Campos M.J., Monteiro D., Antunes R., Jacinto M.. Effects of Physical Exercise Program in Adults with Intellectual and Developmental Disabilities—A Study Protocol. J. Clin. Med. (2022.0) 11. DOI: 10.3390/jcm11247485 4. Krause S., Ware R., McPherson L., Lennox N., O’Callaghan M.. Obesity in adolescents with intellectual disability: Prevalence and associated characteristics. Obes. Res. Clin. Pract. (2016.0) 10 520-530. DOI: 10.1016/j.orcp.2015.10.006 5. Sun Y., Supriya R., Gao Y., Yu S., Wang A., Ou X., Tao D., Baker J.S.. The Relationships between Parenting Practices and Child Health-Related Behaviors in Children with Intellectual Disability: The Moderating Role of Child Body Weight Status. Nutrients (2022.0) 14. DOI: 10.3390/nu14245206 6. Morin D., Mérineau-Côté J., Ouellette-Kuntz H., Tassé M.J., Kerr M.. A comparison of the prevalence of chronic disease among people with and without intellectual disability. Am. J. Intellect. Dev. Disabil. (2012.0) 117 455-463. DOI: 10.1352/1944-7558-117.6.455 7. Farías-Valenzuela C., Ferrari G., Espoz-Lazo S., Ferrero-Hernández P., Jofré-Saldia E., Álvarez-Arangua S., Poblete-Aro C., Godoy-Cumillaf A., Cofre-Bolados C., Valdivia-Moral P.. Anthropometric Indicators of the Cardiometabolic Risk, Muscle Strength, and Functional Capacity of Schoolchildren with Intellectual Disabilities during Lockdown in Chile. Children (2022.0) 9. DOI: 10.3390/children9091315 8. Zwierzchowska A., Rosołek B., Sikora M., Celebańska D.. Forced Sedentariness and Sports Activity as Factors Differenti-ating Anthropometric Characteristics, Indices, and Body Composition in People with Disabilities. Biology (2022.0) 11. DOI: 10.3390/biology11060906 9. Muhanna R.G., Aljuraiban G.S., Almadani N.K., Alquraishi M., El-Sharkawy M.S., Abulmeaty M.M.A.. Value of Adding Bioelectrical Impedance Analysis to Anthropometric Indices in the Diagnosis of Metabolic Syndrome in 10–16 Years Old Schoolgirls. Healthcare (2022.0) 10. DOI: 10.3390/healthcare10030419 10. Waninge A., Van Der Weide W., Evenhuis I.J., Van Wijck R., Van Der Schans C.P.. Feasibility and reliability of body composition measurements in adults with severe intellectual and sensory disabilities. J. Intellect. Disabil. Res. (2009.0) 53 377-388. DOI: 10.1111/j.1365-2788.2009.01153.x 11. Kfir A., Lahav Y., Gepner Y.. Cross-Validation of a New General Population Resting Metabolic Rate Prediction Equation Based on Body Composition. Nutrients (2023.0) 15. DOI: 10.3390/nu15040805 12. Maury-Sintjago E., Rodríguez-Fernández A., Ruíz-De la Fuente M.. Predictive Equations Overestimate Resting Metabolic Rate in Young Chilean Women with Excess Body Fat. Metabolites (2023.0) 13. DOI: 10.3390/metabo13020188 13. Staśkiewicz W., Grochowska-Niedworok E., Zydek G., Grajek M., Krupa-Kotara K., Białek-Dratwa A., Jaruga-Sękowska S., Kowalski O., Kardas M.. The Assessment of Body Composition and Nutritional Awareness of Football Players According to Age. Nutrients (2023.0) 15. DOI: 10.3390/nu15030705 14. EUR-lex 15. Field A.. Discovering Statistic Using SPSS for Windows (2000.0) 16. Sava F.. Data Analysis in Psychological Research. Complementary Statistical Methods (2004.0) 17. Tabachnick B.G., Fidell L.S.. Using Multivariate Statistics (2014.0) 18. Rencher A.C., Christensen W.F.. Methods of Multivariate Analysis (2012.0) 19. 19. World Health Organization Available online: https://www.who.int/tools/growth-reference-data-for-5to19-years/indicators/bmi-for-age(accessed on 22 December 2022) 20. Müller A., Nagy Z., Kovács S., Szőke S., Bendíková E., Ráthonyi G., Ráthonyi-Ódor K., Szabados G., Gabnai Z., Bába É.B.. Correlations between Physical Fitness and Body Composition among Boys Aged 14–18—Conclusions of a Case Study to Reverse the Worsening Secular Trend in Fitness among Urban Youth Due to Sedentary Lifestyles. Int. J. Environ. Res. Public Health (2022.0) 19. DOI: 10.3390/ijerph19148765 21. Ballarin G., Licenziati M.R., Alicante P., Di Vincenzo O., Valerio G., Scalfi L.. Bioelectrical Impedance Analysis-Derived Phase Angle and Body Composition Are Predictors of Health-Related Fitness in Children and Adolescents with Obesity. Children (2022.0) 9. DOI: 10.3390/children9121943 22. Skrzypek M., Koch W., Goral K., Soczyńska K., Poźniak O., Cichoń K., Przybysz O., Czop M.. Analysis of the Diet Quality and Nutritional State of Children, Youth and Young Adults with an Intellectual Disability: A Multiple Case Study. Preliminary Polish Results. Nutrients (2021.0) 13. DOI: 10.3390/nu13093058 23. Nils A.A.. Physical Activity and Body Composition in Norwegian Adolescent. Ph.D. Thesis (2020.0) 24. Chung A., Backholer K., Wong E., Palermo C., Keating C., Peeters A.. Trends in child and adolescent obesity prevalence in economically advanced countries according to socioeconomic position: A systematic review. Obes. Rev. (2016.0) 17 276-295. DOI: 10.1111/obr.12360 25. Martinho D.V., Gouveia É.R., França C., Lopes H., Ihle A., Marques A., Rodrigues A., Alves R., Sarmento H.. Body Composition and Physical Fitness in Madeira Youth. Children (2022.0) 9. DOI: 10.3390/children9121833 26. Farías-Valenzuela C., Ferrero-Hernández P., Ferrari G., Espoz-Lazo S., Castillo-Paredes A., Álvarez-Arangua S., Valdivia-Moral P.. Reference Values of Absolute and Relative Handgrip Strength in Chilean Schoolchildren with Intellectual Disabilities. Children (2022.0) 9. DOI: 10.3390/children9121912 27. Domaradzki J.. Congruence between Physical Activity Patterns and Dietary Patterns Inferred from Analysis of Sex Differences in Lifestyle Behaviors of Late Adolescents from Poland: Cophylogenetic Approach. Nutrients (2023.0) 15. DOI: 10.3390/nu15030608 28. Kaplowitz P.B.. Link between body fat and the timing of puberty. Pediatrics (2008.0) 121 S208-S217. DOI: 10.1542/peds.2007-1813F 29. Abe T., Thiebaud R.S., Ozaki H., Yamasaki S., Loenneke J.P.. Children with Low Handgrip Strength: A Narrative Review of Possible Exercise Strategies to Improve Its Development. Children (2022.0) 9. DOI: 10.3390/children9111616 30. Rigamonti A.E., Frigerio G., Caroli D., De Col A., Cella S.G., Sartorio A., Fustinoni S.. A Metabolomics-Based Investigation of the Effects of a Short-Term Body Weight Reduction Program in a Cohort of Adolescents with Obesity: A Prospective Interventional Clinical Study. Nutrients (2023.0) 15. DOI: 10.3390/nu15030529
--- title: Novel Antioxidant Peptides Identified from Arthrospira platensis Hydrolysates Prepared by a Marine Bacterium Pseudoalteromonas sp. JS4-1 Extracellular Protease authors: - Congling Liu - Gong Chen - Hailian Rao - Xun Xiao - Yidan Chen - Cuiling Wu - Fei Bian - Hailun He journal: Marine Drugs year: 2023 pmcid: PMC9966703 doi: 10.3390/md21020133 license: CC BY 4.0 --- # Novel Antioxidant Peptides Identified from Arthrospira platensis Hydrolysates Prepared by a Marine Bacterium Pseudoalteromonas sp. JS4-1 Extracellular Protease ## Abstract Crude enzymes produced by a marine bacterium Pseudoalteromonas sp. JS4-1 were used to hydrolyze phycobiliprotein. Enzymatic productions showed good performance on DPPH radical and hydroxyl radical scavenging activities (45.14 ± $0.43\%$ and 65.11 ± $2.64\%$, respectively), especially small peptides with MWCO <3 kDa. Small peptides were fractioned to four fractions using size-exclusion chromatography and the second fraction (F2) had the highest activity in hydroxyl radical scavenging ability (62.61 ± $5.80\%$). The fraction F1 and F2 both exhibited good antioxidant activities in oxidative stress models in HUVECs and HaCaT cells. Among them, F2 could upregulate the activities of SOD and GSH-Px and reduce the lipid peroxidation degree to scavenge the ROS to protect Caenorhabditis elegans under adversity. Then, 25 peptides total were identified from F2 by LC-MS/MS, and the peptide with the new sequence of INSSDVQGKY as the most significant component was synthetized and the ORAC assay and cellular ROS scavenging assay both illustrated its excellent antioxidant property. ## 1. Introduction The microalgae Arthrospira belongs to the cyanobacteria division, the cyanophycean class, and the Oscillatoriaceae family. Arthrospira sp. is a photosynthetic cyanobacterium that has recently grown in value as a nutritional supplement and food additive due to its high protein content of around 60–$70\%$ in their biomass and various valuable natural compounds, such as pigments, β-carotenes, polysaccharides and peptides. In addition, Arthrospira sp. has a wide range of uses in the pharmaceutical, cosmetic, and nutritional industries. Because of its high protein content and nutritive value of amino acid composition, Arthrospira sp. was evaluated as a potential meat substitute and has received increased attention [1]. Phycobiliprotein is a class of proteins with antioxidant, antitumor and anti-inflammatory properties, but the heat sensitivity and allergenicity limit its wider applications. Phycobiliprotein is highly abundant in Arthrospira, accounting for about $20\%$ of its dry weight [2]. Using proteases to hydrolyze the phycobiliprotein to produce antioxidant peptides is an efficient way to reach higher bioactive potential and promote high value utilization of marine proteins while avoiding its natural limitations. Moreover, the proteins of microalgae can be converted by enzymatic hydrolysis into value-added products with better functional properties such as antioxidant peptides (APs), anti-hypertension peptides, and antibacterial peptides, etc. Therefore, more and more researchers are interested in obtaining effective bioactive peptides (BPs) from Arthrospira sp. hydrolysates. Enzymatic hydrolysis of the dietary proteins generates BPs, which are short chains of 2–15 amino acids residues, can enhance nutritional value, safety, bioactive function and reduce the allergenicity [3,4]. Numerous studies have demonstrated that BPs with antioxidant activity produced by proteolysis can achieve antioxidation by scavenging free radicals, chelating transition metals, and enhancing the activity of endogenous antioxidant enzymes such as catalase (CAT), superoxide dismutase (SOD), and glutathione peroxidase (GSH-Px). Redox homeostasis (balance) is an important cellular process that plays a vital role in maintaining the normal physiological steady state of the human body. A disturbance of balance between oxidants and antioxidants results in oxidative stress. Reactive oxygen species (ROS) are molecules derived from aerobic cellular metabolism; in the oxidative stress condition excessive production of ROS can disrupt the intracellular redox balance and cause numerous diseases such as cancer, diabetes, atherosclerosis, cardiovascular and neurodegenerative [4]. A growing number of studies have demonstrated that APs from protease hydrolysates can effectively scavenge free radicals or inhibit the production of ROS. Meanwhile, food-derived APs also performed nontoxicity, are low cost, absorb efficiency advantages, and were considered as potential substitutes for commercial synthetic antioxidants [5,6]. Since different proteases have specific hydrolyzed performance on protein substrate and can produce diverse BPs with different length, sequence, and composition, these hydrolyzes may exhibit various hydrophobicity and functionalities [7,8]. Therefore, it is necessary to find a new enzymes resource to hydrolyze Arthrospira platensis (A. platensis) and produce novel APs with high antioxidant activity and bioactive properties. Our previous study suggested using extracellular proteases secreted by a marine bacterium Pseudoalteromonas sp. JS4-1 to hydrolyze chlorella can obtain hydrolysates with high antioxidant activity [9]. In this study, proteases from Pseudoalteromonas sp. JS4-1 were used to hydrolyze A. platensis proteins and produced diverse Aps; a series of biochemical assays, cell models, and animal models were applied for testing the bioactivity of the APs. ## 2.1. Screening of Proteases for Enzymatic Hydrolysis Pseudoalteromonas is a strictly marine genus and able to produce extracellular enzymes, which have good advantages in degrading natural marine proteins [10]. In order to evaluate which enzyme is more suitable for hydrolyzing Arthrospira sp. proteins, the extracellular enzymes from eight strains named Pseudoalteromonas sp. B27-3, Pseudoalteromonas sp. B47-6, Pseudoalteromonas sp. B62-3, Pseudoalteromonas sp. WH05-1, Pseudoalteromonas sp. WH06-2, Pseudoalteromonas sp. WH16-2, Pseudoalteromonas sp. JS4-1 and Pseudoalteromonas sp. ZB23-2 were extracted and screened by zymography. The enzymatic hydrolysates were identified by SDS-PAGE and their DPPH RSA and OH RSA activities were also detected. The zymography showed with the exception of Pseudoalteromonas sp. WH06-2 and Pseudoalteromonas sp. WH16-2, all the other strains can secrete a variety of proteases, indicating that most marine bacteria could produce extracellular proteases with a variety of cleavage sites, and suitable for enzymatic hydrolysis (Figure 1A). Proteases from Pseudoalteromonas sp. B27-3 and Pseudoalteromonas sp. JS4-1 have stronger hydrolysis activity than other bacterial proteases toward Arthrospira sp. proteins. SDS-PAGE showed that the hydrolysates existed in smaller fragments compared with the control protein, and almost no protein band left on the gel, while the hydrolysates produced by other bacterial proteases still left visible bands on the gel (Figure 1B). The hydrolysates of Pseudoalteromonas sp. JS4-1 proteases have the highest DPPH RSA (45.14 ± $0.43\%$) and OH RSA (65.11 ± $2.64\%$) than other productions (Figure 1C,D). Therefore, proteases from Pseudoalteromonas sp. JS4-1 were selected as a candidate for Arthrospira sp. proteins hydrolyzing. ## 2.2. Optimization of Enzymatic Hydrolyzing Parameters Enzymatic hydrolyzation was optimized using the one-variable-at-a-time (OVAT) approach by changing one parameter while keeping other parameters constant. First, the hydrolyzing was performed at 50 °C from 1 h to 7 h with 1 h at intervals, at a E: S ratio of 1:10, to estimate the optimal hydrolyzing time. Data showed that the hydrolysis degree increased along with the hydrolyzing process, especially in the initial three hours. After five hours hydrolyzation, the hydrolysis degree was not improved significantly. DPPH RSA as an evaluation index suggested that when the E/S ratio ranged from 1:6 to 1:10, the DPPH RSA of the hydrolytes exhibited no significant difference. While the E/S ratio adjusted to 1:12 and 1:14, the DPPH RSA of the hydrolytes dropped $6.72\%$ and $15.95\%$, respectively, indicating that the substrate exceeded the catalytic capacity of the enzymes. When the hydrolyzing temperature increased, the DPPH RSA improved and reached the maximal activity at 50 °C, but when the temperature exceeded >50 °C the DPPH RSA decreased, probably due to the enzymes being inactivated at higher temperatures. After OVAT optimization, when E:S ratio was 1:10 and enzymatic hydrolyzing performed at 50 °C for 4 h, the DPPH RSA and the OH RSA of the hydrolysates achieved the maximal level of 48.51 ± $2.42\%$ and 67.16 ± $3.21\%$, respectively. The ninhydrin colorimetry was used to determine the hydrolysis degree, and data showed that about 0.67 ± 0.02 mM of •NH2 was released after hydrolyzing reaction. In order to evaluate the free RSA of small peptides, the hydrolysates were separated into 2 parts by ultrafiltration. The DPPH RSA of < 3 kDa and > 3 kDa parts were 48.26 ± $1.34\%$ and 42.60 ± $0.25\%$, respectively. The OH RSA of < 3 kDa and > 3 kDa parts were 63.30 ± $1.11\%$ and 44.30 ± $0.68\%$, respectively (Figure 1E). It has been proven that the small peptides (<3 kDa) contained more antioxidant components and were more easy to exert biological effects [11]. The bioactive peptides with small molecular weight have broad application potential in food, medicine, cosmetics and other industries due to their good absorbability and fewer side effects besides their higher bioactivity [12]. ## 2.3. Antioxidant Activities of Size-Exclusion Liquid Chromatography Fractions of Phycobiliprotein Hydrolysis Products In order to investigate the antioxidant components of the small peptides, the peptides with MWCO <3 kDa were further separated using size-exclusion liquid chromatography, and in total four fractions defined as F1 to F4 were obtained (Figure 1F). By measuring the DPPH RSA (F1 to F4 were 52.06 ± $2.13\%$, 52.56 ± $1.38\%$, 24.43 ± $0.85\%$ and 27.58 ± $2.24\%$, respectively; Figure 1G) and the OH RSA (F1 to F4 were 40.75 ± $14.35\%$, 62.61 ± $5.80\%$, 30.86 ± $10.90\%$, and 10.21 ± $7.91\%$, respectively; Figure 1H), F1 and F2 showed higher antioxidant activity than F3 and F4. We could obtain about 40 mg of F1 and 40 mg of F2 lyophilized powder after the size-exclusion liquid chromatography from 1 g of A. Platensis powder. Then, the ORAC assay was performed to compare the antioxidant activity of F1 and F2. At peptide concentration of 0.015 mg/mL, both F1 and F2 could significantly slow down the decrease rate of the fluorescence decay compared with PBS. The Trolox equivalent antioxidant capacity of F1 and F2 were 0.27 ± 0.02 mmol TE/g and 0.41 ± 0.04 mmol TE/g, respectively (Figure 1I). F2 exhibited higher antioxidant activity than F1. The hydroxyl radicals could open the circular of supercoiled DNA (SC DNA) and make it become partly opened circular DNA (OC DNA), which can slow down the DNA mobility on agarose gel. Comparing with the no damaged plasmid DNA pET-22b (Figure 1J, lane 2), when the plasmid DNA has no protection, it can be completely degraded by hydroxyl radicals into small fragments and no band observed on the gel (Figure 1J, lane 3). An amount of 0.2 mg/mL of Vitamin C can protect the plasmid DNA from oxidative damage, but the protection is limited and the OC DNA is the major form retained on the gel (Figure 1J, lane 10). F1 and F2 concentrations range from 0.2 mg/mL to 0.4 mg/mL; all can significantly protect plasmid DNA from oxidative damage and these protections were far better than Vitamin C, which make the SC DNA and OC DNA co-existing on the gel (Figure 1J, lane 4, 5, 7, 8). So, this study proved that both F1 and F2 have strong antioxidant activity to protect DNA from oxidative stress. ## 2.4. Antioxidant Activities of F1 and F2 at Cellular Levels The intracellular oxidized cell model is a common method that is used to assess the antioxidative capacity of compounds. Vascular endothelial cell injury is associated with several factors, among them oxidative stress is an important cause of cardiovascular disease. APs are able to protect vascular endothelial cell function which is a key point in the prevention and treatment for cardiovascular diseases. The skin is the first immune defense of the body and susceptible to a variety of physical and chemical stimuli. Protecting skin cells could be an effective strategy to prevent and treat dermatosis. Therefore, we chose HUVECs and HaCaT cells as cellular models to investigate the intracellular antioxidant effects of F1 and F2 [13,14,15,16]. MTT assay indicated that both F1 and F2 had no significant cytotoxicity on HUVECs and HaCaT cells; moreover, they can obviously increase the cell viability at a concentration range of 100–500 μg/mL (Figure 2A,B and Figure 3A,B). An amount of 35 mM of glucose was added to the HUVECs and 1.5 mM H2O2 was used in the HaCaT cells. Cells treated with glucose and H2O2 exhibited higher fluorescence intensity than control groups, indicating the oxidative stress models have been successfully established (Figure 2C,E and Figure 3C,E). The intracellular ROS were labelled by DCFH-DA probe. Cells treated with low concentration of F1 and F2 (25 μg/mL) displayed less fluorescence intensity than the glucose group in HUVECs, suggesting the fractions were able to scavenge intracellular ROS to protect HUVECs from hyperglycemia-induced oxidative stress (Figure 2D,F). Cells treated with a high concentration of F1 (100 μg/mL) and a low concentration of F2 (50 μg/mL) displayed less fluorescence intensity than the H2O2 group in HaCaT cells, suggesting that the fractions were able to scavenge intracellular ROS to protect HaCaT cells from H2O2 induced oxidative stress (Figure 3D,F). The antioxidant activities of peptides on a cellular level in this study were close to those reported extracts of plants [17,18,19], which suggested the peptides in our study may also have great potential applications. To further investigate the underlying mechanisms of fractions scavenging intracellular ROS, we detected the activities of antioxidant enzymes such as SOD, GSH-Px and CAT after being treated with fractions due to their role in scavenging ROS and suppressing oxidative stress [20,21]. The glucose and H2O2 reduced the activities of SOD, GSH-Px and CAT in HUVECs and HaCaT cells. The administration of F1 and F2 dose-dependently increased the activities of these enzymes at a concentration range of 25–200 μg/mL (Figure 2G–L and Figure 3G–L). ## 2.5. Effects of F2 on Resistance to Oxidative Stress in C. elegans Caenorhabditis elegans (C. elegans) have many highly conserved genes and signaling pathways involved in the regulation of aging and oxidative stress [22,23]. The genes and pathways in aging, oxidative stress and inflammation between C. elegans and human are highly homologous [24]. C. elegans as an animal model is widely used for anti-aging and anti-oxidative studies. In this study, 0.5 mg/mL of F1 and F2 had no toxicity and no significant life-extending effect on N2 (Figure 4A). The in vivo ROS assay showed that 0.5 mg/mL of F1 had no significant ROS scavenging effects, while F2 could scavenge ROS at low concentration (0.2 mg/mL, Figure 4E,F). The fluorescence accumulation assay of 80 nematodes showed similar results (Figure 4G). Then, we speculate whether the APs can enhance the stress resistance of N2. Oxidizing agent H2O2 (10 mM), pesticide paraquat (10 mM) and heat (35 °C) were chosen as stress conditions. Considering F2 has higher anti-oxidative activity than F1, in this study N2 were treated with F2 and then exposed to stress conditions. Data showed that the F2 can extend the stressed N2 survival time in all stress conditions, and the survival rate was positively correlated with the addition of F2. An amount of 0.2 mg/mL of F2 increased $35\%$ survival rate after 7 h of H2O2 stress, $50\%$ after 120 h of paraquat stress and $24\%$ under heat stress (Figure 4B–D). Although F2 did not help the C. elegans live longer, it really improved the resistant ability of C. elegans obviously to the oxidation, pesticide and heat stress. We then investigated the MDA levels and activities of antioxidant enzymes in nematodes. The MDA level is an important biomarker of lipid peroxidation [25]. F2 could significantly reduce the MDA levels in nematodes (Figure 4H). As shown in Figure 4I–K, F2 could have dose-dependently increased the activities of SOD and GSH-Px, while it had no effects on CAT. These results suggested that F2 could upregulate the activities of SOD and GSH-Px and reduce the lipid peroxidation degree to scavenge the ROS in the organisms. Activating these antioxidant enzymes contributes to maintaining the redox balance, delaying aging process and prolonging lifetime [26]. Antioxidant enzymes with high activity are able to reduce ROS levels, maintain normal physiological steady state and prevent numerous diseases such as cancer, diabetes, neurodegenerative diseases and so on [4]. ## 2.6. Antioxidant Activities of Synthesized Peptide INSSDVQGKY Since F2 has the highest antioxidant activity in all fractions, its components were investigated using LC-MS/MS. The total ion chromatography of F2 was shown in Figure S1. A total of 25 peptides with relative abundance >$1\%$ were identified, and their sequence and abundance were listed in Table 1. Among them, one peptide with the sequence of INSSDVQGKY accounting for $17.83\%$ was considered as the major component of F2. The MS/MS of INSSDVQGKY was shown in Figure S2. Sequence alignment showed that the peptide INSSDVQGKY was a part of a beta-subunit of allophycocyanin. This was a novel AP first be sequenced and proved with antioxidant activity. The structure of INSSDVQGKY was predicted and it was a linear peptide with a little helix (Figure 5A). In order to validate the antioxidant function of this peptide, it was synthetized. ORAC assay of the synthetized peptide was 0.32 ± 0.03 mmol TE/g and slightly lower than F2 (0.41 ± 0.04 mmol TE/g) (Figure 5B). The peptide had no significant cytotoxicity on HaCaT cells within 400 μg/mL (Figure 5C). The peptide dose-dependently reduced the intracellular ROS levels at a concentration range from 100 μg/mL to 400 μg/mL (Figure 5D,E). This peptide is therefore hypothesized to be the major antioxidant component of F2. The antioxidant activity of INSSDVQGKY is close to peptide PNN which was obtained from hydrolysis of phycobiliprotein by protease K [27] and may have potential applications in further study. ## 3. Discussion Phycobiliprotein is an ideal marine protein resource with antioxidant properties widely spread in Rhodophyta and Cyanophyta, which could be used in bioactive peptides preparation [28]. This study aimed to apply the bacterial extracellular protease in preparing antioxidant peptides from phycobiliprotein. To utilize phycobiliprotein more comprehensively and avoid its natural limitations, we reported bacterial extracellular protease of Pseudoalteromonas sp. JS4–1 hydrolyzed phycobiliprotein to obtain antioxidant peptides in a single process. The bioactive fractions exhibited antioxidant activities both in vitro and in vivo. The biochemical mechanisms of the antioxidant activities of the fractions at cellular and C. elegans levels may be through activating the activities of antioxidant enzymes and scavenging the oxidative stress in the organisms. The molecular mechanism underlying this biological progress still need to be further studied in the future. In this study, extracellular proteases from marine bacteria were used to hydrolyze phycobiliprotein. Marine proteases can degrade organics in the ocean naturally; thus, they have extraordinary advantages in digesting marine-sourced proteins such as fish and plants compared with land proteases [9,29,30]. Compared with commercial proteases, marine proteases have higher catalyzing efficiency and shorter hydrolyzing time. The unique cleavage sites of marine proteases also produced peptides with different amino acid sequences which is helpful to select more bioactive peptides and develop various applications of phycobiliprotein [31]. The identified peptide INSSDVQGKY exhibited antioxidant activities at a concentration of 100–200 μg/mL which was close to some newly found antioxidant peptides derived from other resources [32,33]. The cellular ROS scavenging assay is strong proof of the outstanding antioxidant activity of the peptide during biological process. ROS are important mediators of biological process. The increased formation of ROS could cause oxidative stress which is associated with oxidative damage to biomolecules [34]. Oxidative stress happens in various pathophysiological processes and is an important pathophysiological mechanism underlying many diseases. Thus, decreasing the ROS levels to inhibit the oxidative stress has become an efficient and promising way in medicine [35]. Activating endogenous antioxidant enzymes such as SOD, CAT and GSH-Px by small molecules is an effective and protective approach in diseases [36]. The enzymatic hydrolysates and newly identified peptide in this study were able to activate the antioxidant enzymes to scavenge ROS under exogenous stress such as high glucose, and H2O2. Li G et al. identified a novel peptide EDEQKFWGK with DPPH RSA $26.76\%$ and the OH RSA $32.19\%$ from porcine plasma hydrolysate, and this peptide could increase SOD, CAT, and GSH-Px activities in HepG2 cells [37]. Park YR et al. identified a novel peptide NCWPFQGVPLGFQAPP with DPPH RSA $50\%$ from clam worms and exhibited good antioxidant and anti-inflammation effects in RAW264 7 cells [38]. Compared with antioxidant peptides from other resources, phycobiliprotein hydrolysates also showed good antioxidant activities (DPPH RSA: 45.14 ± $0.43\%$ and OH RSA: 65.11 ± $2.64\%$). Further, phycobiliprotein can be easily obtained and produced massively in the industry. So, exploring bioactive peptides from phycobiliprotein is of great economic value. Conclusively, the enzymatic hydrolysis of phycobiliprotein is a promising and efficient method to produce antioxidant peptides. The peptides hydrolyzed from phycobiliprotein by marine proteases exhibited antioxidant activities both at cellular levels and in C. elegans. ## 4.1. Strains and Reagents The A. platensis powder was purchased from a local supplier in Shanghai. Strains used in this study were isolated from the sediment of the South China Sea and stored in the lab: Pseudoalteromonas sp. JS4-1 (GenBank accession number: MT116988), Pseudoalteromonas sp. B27-3, Pseudoalteromonas sp. B47-6, Pseudoalteromonas sp. B62-3, Pseudoalteromonas sp. WH05-1, Pseudoalteromonas sp. WH06-2, and Pseudoalteromonas sp. ZB23-2. Immortal human umbilical vein endothelial cell line (HUVECs) and human immortal keratinocyte line (HaCaT) were obtained from the National Collection of Authenticated Cell Cultures (Shanghai, China). Wild type C. elegans (N2) were obtained from Caenorhabditis Genetics Center (CGC). Ninhydrin, penicillin-streptomycin solution (cell culture grade), trypsin-EDTA solution, dimethyl sulfoxide (DMSO), thiazolyl blue tetrazolium bromide (MTT), and phosphate buffered saline (PBS) were purchased from Sangon Biotech (Shanghai, China). Trisodium citrate dihydrate, SnCl2, n-propanol, FeSO4, $30\%$ H2O2, cholesterol and glucose were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). In addition, 1,1-diphenyl-2-picrylhydrazyl (DPPH), 1,10-phenanthroline monohydrate (OP), fluorescein sodium salt, 2,20-azobis (2-amidinopropane) dihydrochloride (AAPH), 6-hydroxy-2,5,7,8-tetramethylchroman-2-carboxylic acid (Trolox), L-leucine, antifade mounting medium and NaN3 were purchased from Sigma-Aldrich, Ltd. (St. Louis, MO, USA). AmiconTM Ultra-15 centrifugal filter units were purchased from Millipore® (Billerica, MA, USA). Sephadex LH-20 was purchased from GE Healthcare Life Sciences (Uppsala, Sweden). Fetal bovine serum (FBS) was purchased from PAN-BiotechGibco company (Aidenbach, Germany). RPMI-1640 medium was purchased from Gibco® ThermoFisher Scientific Company (Waltham, MA, USA). Methyl viologen dichloride was purchased from Aladdin® (Shanghai, China). BCA assay kit, cell lysis buffer, ROS assay kit, catalase (CAT) assay kit, total superoxide dismutase (SOD) assay kit with NBT, cellular glutathione peroxidase (GSH-Px) assay kit with NADPH and lipid peroxidation MDA assay kit were purchased from Beyotime Biotechnology (Shanghai, China). ## 4.2.1. Extraction of Phycobiliprotein from A. platensis One gram of A. platensis powder was stirred in 100 mL of sterilized ultrapure water, then frozen in −20 °C and thawed in 37 °C-water bath repeatedly 3–4 times to help proteins released. The protein content was measured by BCA method [39]. The soluble phycobiliprotein (crude extract) was obtained by centrifugation at 12,000× g for 20 min, and stored at −20 °C for future use. We could obtain 100 mL of 3.75 mg/mL protein solution from 1 g of A. platensis powder and the yield rate was $37.5\%$. ## 4.2.2. Preparation of Crude Bacterial Extracellular Protease The activated seed inoculum was prepared by adding bacterial strains preserved in glycerin into 5 mL 2216E liquid medium (tryptone 5 g/L, yeast extract 1 g/L, Fe2(SO4)3 0.01 g/L, dissolved in artificial sea water, pH 7.8), and incubating at 16 °C for 16–24 h with 180 rpm shaking. Then, $2\%$ of seed inoculum was transformed to fresh fermentation medium (bean powder 20 g/L, corn powder 20 g/L, wheat bran 10 g/L, Na2HPO4·12H2O 1 g/L, KH2PO4 0.3 g/L, CaCl2 1 g/L, Na2CO3 1 g/L, dissolved in artificial sea water, pH 7.8) and continuously culturing at the same condition. After 5 days cultivation, the crude enzyme was obtained by collecting the supernatant of the fermentation broth at 12,000× g centrifugation for 10 min at 4 °C, and stored at −20 °C prior to use. ## 4.2.3. Substrate Immersing Zymography Substrate immersing zymography was conducted to determine the enzymatic activities against casein according to a previous study, with minor modifications [40]. The SDS-PAGE gel was immersed in the pre-warmed $0.5\%$ (w/v) casein solution and incubated at 37 °C for 1 h with 75 rpm shaking. The gel was then stained with $0.1\%$ of Coomassie Brilliant Blue dye solution for 4 h, and then decolored with $30\%$ of ethanol and $70\%$ of acetic acid (v/v) until the bands were clearly visible. ## 4.3.1. Optimization of Hydrolysis Conditions First, the enzyme/substrate (E: S) ratio (v/w, mL/mg) was designed as 1:6, 1:8, 1:10, 1:12 and 1:14 with 1 mg/mL enzyme at 50 °C, to optimize the E/S ratio during the hydrolyzing. Then, the hydrolysis reaction was carried out at 35 to 60 °C, with 5 °C at intervals, to estimate the optimal temperature for hydrolyzing. Finally, hydrolyzing was taken from 1 h to 7 h at hourly intervals at 50 °C to optimize the reaction time. The reaction was terminated by heating at 95 °C for 10 min. The reaction buffer was ultrapure water. The hydrolysate was centrifuged at 12,000× g and 4 °C for 20 min and the hydrolysis degree was measured by ninhydrin coloration method [41]. The •NH2 in hydrolysate could react with ninhydrin and have absorption peak at 570 nm. The standard curve was determined using L-leucine at a concentration of 0, 1, 1.5, 2, 2.5 and 3 mM. ## 4.3.2. Purification of Hydrolysates After digestion, the hydrolysates were added into the upper casing of the AmiconTM Ultra-15 centrifugal filter units and centrifuged at 5000× g for 30 min to separate peptides with MWCO (molecular weight cut-off) > 3 kDa (in the upper casing) and MWCO < 3 kDa (in the lower casing). The peptides with MWCO < 3 kDa was further fractionated by NGC Scout Plus fast protein liquid chromatography (FPLC) system (Bio-Rad, USA) and a XK$\frac{16}{70}$ column (1.6 × 60 cm) equipped with Sephadex LH-20 gel. One mL of sample was injected into the loading loop and the column was continuously eluted with filtered ultrapure water at a flow rate of 0.75 mL/min. The protein absorbance was monitored at 220 nm with a UV detection all the time. Fractions were lyophilized and redissolved in ultrapure water at the same concentration for further assay. ## 4.3.3. Determination of Antioxidant Activities To evaluate antioxidant activities of hydrolytes in vitro, DPPH free radical scavenging activity (DPPH RSA) assay, hydroxyl free radical (OH) scavenging activity (OH RSA) assay and oxygen radical absorbance capacity (ORAC) assay were conducted according to previous studies, with minor modifications [42,43,44]. The DPPH and OH radical scavenging rate were calculated. The ORAC was defined as Trolox equivalents (mmol TE/g sample or mmol TE/mmol sample) according to the area under the curve (AUC) and calculated using the following formula:ORAC = (AUCsample − AUCcontrol)/(AUCTrolox − AUCcontrol) × (MTrolox/Msample) [1] where AUCsample, AUCcontrol and AUCTrolox were the integral areas under the fluorescence decay curve of the sample, PBS and Trolox, and MTolox and Msample were the concentrations of the Trolox and sample, respectively. ## 4.3.4. Protective Effects on Oxidative Damage of Plasmid DNA To evaluate the protective effects of peptides on the oxidative damage of biological macromolecules (proteins and DNA), the protection effect assay was conducted according to a previous method, with minor modifications [15]. An amount of 8 μL of pET-22b DNA, 2 μL of FeSO4 (2 mM), 8 μL of sample and 2 μL of $0.1\%$ H2O2 were mixed in a sterilized tube and incubated at 37 °C for 10 min. For the blank, 12 μL of sterilized water replaced other reagents incubating with the DNA. For the negative control, sample was replaced with sterilized water. For the positive control, the sample was replaced with Vitamin C (200 μg/mL). The reaction solution was then analyzed by $1\%$ agarose gel electrophoresis under 120 V for 30 min. ## 4.3.5. Determination of Intracellular ROS Levels and the Activities of Intracellular Oxidative Enzymes HUVECs and HaCaT cells were cultured in RPMI-1640 medium with $10\%$ FBS, 100 U/mL penicillin and 100 μg/mL streptomycin (complete medium) at 37 °C in a humidified atmosphere of $5\%$ CO2. Cell viability was measured by MTT assay. The 2′,7′-dichloro-fluorescence diacetate (DCFH-DA) probe can be oxidized by ROS to produce green fluorescent materials which can be observed under fluorescence microscope. An amount of 1 × 105 of cells were plated in a 24-well plate and incubated overnight. For the HUVECs cells, the medium was replaced with RPMI-1640 medium (without FBS instead, added 35 mM of glucose) for 24 h. For the HaCaT cells, the medium was replaced with RPMI-1640 medium (without FBS, but added 1.5 mM of H2O2) for 4 h. For the blank, cells were treated with RPMI-1640 medium (without FBS). An amount of 10 μM DCFH-DA probe was added to each well and incubated for 1 h. Cells were washed with PBS for 3 times and re-immersed in 500 μL of PBS. Images were taken using an inverted fluorescence microscope (Axio vert. A1, Carl Zeiss, Oberkochen, Germany). Next, 2 × 106 of cells were plated in a 6-well plate and treated with the same condition. After incubating for 24 h, cells were washed with PBS for 3 times, then incubating with 150 μL of cell lysis buffer for 30 min on ice. The suspension was centrifuged at 12,000× g for 10 min at 4 °C. The protein concentration was measured by BCA method. The CAT, SOD and GSH-Px activities were determined according to the manufacturer’s instructions. ## Maintenance of C. elegans and Lifespan Assay N2 was cultured on nematode growth medium (NGM) plates supplied with E. coli OP50 as food at 20 °C. The age-synchronized C. elegans were prepared using the alkaline hypochlorite method according to Phaniendra et al. [ 45]. Fresh NGM plates were coated with E. coli OP50 containing a different concentration of APs (0.2 and 0.5 mg/mL) and incubated at 37 °C for 48 h to prepare the sample plates. Age-synchronized nematodes (L4 stage) were plated in blank plates and sample plates (80 worms on each plate), as Day 0 for lifespan assay. Nematodes on each plate were transferred to a new plate to avoid new-born worms and counted every day until all nematodes were dead. The survival rate was calculated. ## Determination of ROS Levels and Activities of Antioxidant Enzymes in C. elegans After 3 days of treatment on the sample plates, nematodes were collected into a M9 buffer (added 100 μM DCFH-DA probe) and incubated at 20 °C for 30 min without light. Nematodes were washed with M9 buffer 3 times. For inverted fluorescence microscope observation, nematodes were anesthetized with 1 μL of NaN3 (25 mM) and 20 μL of suspension was dropped onto the center of the slide, and 5 μL of antifade mounting medium was added for observation. The ROS accumulation was conducted according to previous studies, with minor modifications [46]. The fluorescence intensity was measured at 15 min intervals for 6 h with excitation 485 nm and emission 530 nm at 25 °C. For intracellular proteins extraction, nematodes were ultrasonicated on ice and the suspension was centrifuged at 12,000× g for 10 min at 4 °C. The protein concentration was measured by BCA method. The CAT, SOD and GSH-Px activities and MDA levels in the supernatant were determined according to the manufacturer’s instructions. ## Oxidative Stress Experiment After 3 days of treatment, nematodes were transferred onto NGM plates supplied with 10 mM H2O2 or 10 mM paraquat, and each plate with 80 worms. The number of survival nematodes was counted per 30 min intervals (the H2O2 plates) and per 12 h intervals (the paraquat plates) until all the nematodes were dead [47]. The survival rate was calculated. After 3 days of treatment, nematodes were transferred onto the fresh NGM plates (80 worms on each plate) and incubated at 35 °C for 10 h [48]. Numbers of survival nematodes were counted after the heat shock and the survival rate was calculated. ## 4.3.7. Identification and Solid Phase Synthesis of APs The sequence and composition of peptides in the active fractions were analyzed by liquid chromatography-mass spectrometry (LC-MS/MS) in Sangon Biotech Co., Ltd. (Shanghai, China). The LC-MS/MS was developed on an UltiMate 3000 RSLCnano system (Thermo, Waltham, MA, USA) coupled with a Q Exactive Plus HPLC-Mass Spectrometer (Thermal, Riverside County, CA, USA). The chromatographic separation was achieved using a C18 column (3 μm, 120 Å, 100 μm × 20 mm). The analysis was performed on a C18 analytical column (2 μm, 120 Å, 750 μm × 150 mm). Mobile phase A was $3\%$ DMSO, $0.1\%$ formic acid, and $97\%$ H2O and mobile phase B was $3\%$ DMSO, $0.1\%$ formic acid, and $97\%$ ACN. The flow rate was set as 300 nL/min. The MS data was processed with MaxQuant (V1.6.2.10) to identify detected peptides in Arthrospira platensis proteomics database. Peptide with maximal abundance was synthesized by solid phase synthesis in ChinaPeptides Co., Ltd. (Shanghai, China) according to the sequence information given by LC-MS/MS. The structure of peptide was predicted by Phyre2 (http://www.sbg.bio.ic.ac.uk/phyre2 (accessed on 20 December 2022)). ## 4.4. Statistical Analysis All the experiments and analysis were carried out in triplicate. The Normality of experimental data was analyzed by Shapiro–Wilk test. Normal experimental data were analyzed with by Student t-test. Non-parametric data were analyzed by Mann–Whitney test. Data were presented as the mean ± standard deviation (SD) and $P \leq 0.05$ was considered as a statistically significant difference. ## 5. Conclusions Algal protein resources are considered to be a huge treasury of bioactive peptides. More and more researchers are attempting to prepare novel bioactive peptides from algal protein through protease hydrolysis. In addition, novel proteases from marine bacteria also possess great potential in antioxidant peptide preparation, due to their high efficiency and low cost. In this study, phycobiliprotein was hydrolyzed by extracellular proteases from Pseudoalteromonas sp. JS4-1 to produce antioxidant peptides. A novel antioxidant peptide INSSDVQGKY was separated and identified from the hydrolysates. The hydrolysates and peptide INSSDVQGKY exhibited excellent antioxidant activities both at cellular levels and in C. elegans. ## References 1. Lupatini A.L., Colla L.M., Canan C., Colla E.. Potential application of microalga. J. Sci. Food Agric. (2017) 97 724-732. DOI: 10.1002/jsfa.7987 2. Romay C., González R., Ledón N., Remirez D., Rimbau V.. C-phycocyanin: A biliprotein with antioxidant, anti-inflammatory and neuroprotective effects. Curr. Protein Pept. Sci. (2003) 4 207-216. DOI: 10.2174/1389203033487216 3. Marthandam Asokan S., Wang T., Su W.T., Lin W.T.. Antidiabetic Effects of a Short Peptide of Potato Protein Hydrolysate in STZ-Induced Diabetic Mice. Nutrients (2019) 11. DOI: 10.3390/nu11040779 4. Qiao Q., Chen L., Li X., Lu X., Xu Q.. Roles of Dietary Bioactive Peptides in Redox Balance and Metabolic Disorders. Oxid. Med. Cell Longev. (2021) 2021 5582245. DOI: 10.1155/2021/5582245 5. Amigo L., Hernández-Ledesma B.. Current Evidence on the Bioavailability of Food Bioactive Peptides. Molecules (2020) 25. DOI: 10.3390/molecules25194479 6. Wang K., Han L., Hong H., Pan J., Liu H., Luo Y.. Purification and identification of novel antioxidant peptides from silver carp muscle hydrolysate after simulated gastrointestinal digestion and transepithelial transport. Food Chem. (2021) 342 128275. DOI: 10.1016/j.foodchem.2020.128275 7. Chalamaiah M., Dinesh Kumar B., Hemalatha R., Jyothirmayi T.. Fish protein hydrolysates: Proximate composition, amino acid composition, antioxidant activities and applications: A review. Food Chem. (2012) 135 3020-3038. DOI: 10.1016/j.foodchem.2012.06.100 8. Lu X., Zhang L., Sun Q., Song G., Huang J.. Extraction, identification and structure-activity relationship of antioxidant peptides from sesame (. Food Res. Int. (2019) 116 707-716. DOI: 10.1016/j.foodres.2018.09.001 9. Olena Z., Yang Y., TingTing Y., XiaoTao Y., HaiLian R., Xun X., Dong X., CuiLing W., HaiLun H.. Simultaneous preparation of antioxidant peptides and lipids from microalgae by pretreatment with bacterial proteases. Bioresour. Technol. (2022) 348 126759. DOI: 10.1016/j.biortech.2022.126759 10. Xu F., Cha Q.Q., Zhang Y.Z., Chen X.L.. Degradation and Utilization of Alginate by Marine Pseudoalteromonas: A Review. Appl. Environ. Microbiol. (2021) 87 e0036821. DOI: 10.1128/AEM.00368-21 11. Kou X., Gao J., Xue Z., Zhang Z., Wang H., Wang X.. Purification and identification of antioxidant peptides from chickpea (. LWT—Food Sci. Technol. (2013) 50 591-598. DOI: 10.1016/j.lwt.2012.08.002 12. Hou H., Wang J., Wang J., Tang W., Shaikh A.S., Li Y., Fu J., Lu L., Wang F., Sun F.. A Review of Bioactive Peptides: Chemical Modification, Structural Characterization and Therapeutic Applications. J. Biomed. Nanotechnol. (2020) 16 1687-1718. DOI: 10.1166/jbn.2020.3001 13. Cai S., Lu C., Liu Z., Wang W., Lu S., Sun Z., Wang G.. Derivatives of gecko cathelicidin-related antioxidant peptide facilitate skin wound healing. Eur. J. Pharmacol. (2021) 890 173649. DOI: 10.1016/j.ejphar.2020.173649 14. Hseu Y.C., Vudhya Gowrisankar Y., Wang L.W., Zhang Y.Z., Chen X.Z., Huang P.J., Yen H.R., Yang H.L.. The in vitro and in vivo depigmenting activity of pterostilbene through induction of autophagy in melanocytes and inhibition of UVA-irradiated α-MSH in keratinocytes via Nrf2-mediated antioxidant pathways. Redox Biol. (2021) 44 102007. DOI: 10.1016/j.redox.2021.102007 15. Wu R., Wu C., Liu D., Yang X., Huang J., Zhang J., Liao B., He H.. Antioxidant and anti-freezing peptides from salmon collagen hydrolysate prepared by bacterial extracellular protease. Food Chem. (2018) 248 346-352. DOI: 10.1016/j.foodchem.2017.12.035 16. Zou J., Fei Q., Xiao H., Wang H., Liu K., Liu M., Zhang H., Xiao X., Wang K., Wang N.. VEGF-A promotes angiogenesis after acute myocardial infarction through increasing ROS production and enhancing ER stress-mediated autophagy. J. Cell Physiol. (2019) 234 17690-17703. DOI: 10.1002/jcp.28395 17. Nensat C., Songjang W., Tohtong R., Suthiphongchai T., Phimsen S., Rattanasinganchan P., Metheenukul P., Kumphune S., Jiraviriyakul A.. Porcine placenta extract improves high-glucose-induced angiogenesis impairment. BMC Complement. Med. Ther. (2021) 21. DOI: 10.1186/s12906-021-03243-z 18. Rezabakhsh A., Rahbarghazi R., Malekinejad H., Fathi F., Montaseri A., Garjani A.. Quercetin alleviates high glucose-induced damage on human umbilical vein endothelial cells by promoting autophagy. Phytomedicine (2019) 56 183-193. DOI: 10.1016/j.phymed.2018.11.008 19. Son D.H., Yang D.J., Sun J.S., Kim S.K., Kang N., Kang J.Y., Choi Y.H., Lee J.H., Moh S.H., Shin D.M.. A Novel Peptide, Nicotinyl-Isoleucine-Valine-Histidine (NA-IVH), Promotes Antioxidant Gene Expression and Wound Healing in HaCaT Cells. Mar. Drugs (2018) 16. DOI: 10.3390/md16080262 20. Dong H., Liu M., Wang L., Liu Y., Lu X., Stagos D., Lin X., Liu M.. Bromophenol Bis (2,3,6-Tribromo-4,5-dihydroxybenzyl) Ether Protects HaCaT Skin Cells from Oxidative Damage via Nrf2-Mediated Pathways. Antioxidants (2021) 10. DOI: 10.3390/antiox10091436 21. Wang G., Hao M., Liu Q., Jiang Y., Huang H., Yang G., Wang C.. Protective effect of recombinant Lactobacillus plantarum against H(2)O(2)-induced oxidative stress in HUVEC cells. J. Zhejiang Univ. Sci. B (2021) 22 348-365. DOI: 10.1631/jzus.B2000441 22. Lin C., Xiao J., Xi Y., Zhang X., Zhong Q., Zheng H., Cao Y., Chen Y.. Rosmarinic acid improved antioxidant properties and healthspan via the IIS and MAPK pathways in. Biofactors (2019) 45 774-787. DOI: 10.1002/biof.1536 23. Nakagawa H., Shiozaki T., Kobatake E., Hosoya T., Moriya T., Sakai F., Taru H., Miyazaki T.. Effects and mechanisms of prolongevity induced by. Aging Cell (2016) 15 227-236. DOI: 10.1111/acel.12431 24. Ayuda-Durán B., González-Manzano S., González-Paramás A.M., Santos-Buelga C.. Caernohabditis elegans as a Model Organism to Evaluate the Antioxidant Effects of Phytochemicals. Molecules (2020) 25. DOI: 10.3390/molecules25143194 25. Tsikas D.. Assessment of lipid peroxidation by measuring malondialdehyde (MDA) and relatives in biological samples: Analytical and biological challenges. Anal. Biochem. (2017) 524 13-30. DOI: 10.1016/j.ab.2016.10.021 26. Shi Y., Meng X., Zhang J.. Multi- and trans-generational effects of N-butylpyridium chloride on reproduction, lifespan, and pro/antioxidant status in. Sci. Total Environ. (2021) 778 146371. DOI: 10.1016/j.scitotenv.2021.146371 27. Yu J., Hu Y., Xue M., Dun Y., Li S., Peng N., Liang Y., Zhao S.. Purification and Identification of Antioxidant Peptides from Enzymatic Hydrolysate of. J. Microbiol. Biotechnol. (2016) 26 1216-1223. DOI: 10.4014/jmb.1601.01033 28. Pagels F., Guedes A.C., Amaro H.M., Kijjoa A., Vasconcelos V.. Phycobiliproteins from cyanobacteria: Chemistry and biotechnological applications. Biotechnol. Adv. (2019) 37 422-443. DOI: 10.1016/j.biotechadv.2019.02.010 29. Gómez-Estaca J., Bravo L., Gómez-Guillén M.C., Alemán A., Montero P.. Antioxidant properties of tuna-skin and bovine-hide gelatin films induced by the addition of oregano and rosemary extracts. Food Chem. (2009) 112 18-25. DOI: 10.1016/j.foodchem.2008.05.034 30. Kim S.Y., Je J.Y., Kim S.K.. Purification and characterization of antioxidant peptide from hoki (. J. Nutr. Biochem. (2007) 18 31-38. DOI: 10.1016/j.jnutbio.2006.02.006 31. Wu R., Chen L., Liu D., Huang J., Zhang J., Xiao X., Lei M., Chen Y., He H.. Preparation of Antioxidant Peptides from Salmon Byproducts with Bacterial Extracellular Proteases. Mar. Drugs (2017) 15. DOI: 10.3390/md15010004 32. Fontoura R., Daroit D.J., Corrêa A.P.F., Moresco K.S., Santi L., Beys-da-Silva W.O., Yates J.R., Moreira J.C.F., Brandelli A.. Characterization of a novel antioxidant peptide from feather keratin hydrolysates. New Biotechnol. (2019) 49 71-76. DOI: 10.1016/j.nbt.2018.09.003 33. Ma Y., Huang K., Wu Y.. In Vivo/In Vitro Properties of Novel Antioxidant Peptide from. J. Microbiol. Biotechnol. (2021) 31 33-42. DOI: 10.4014/jmb.2006.06002 34. Li R., Jia Z., Trush M.A.. Defining ROS in Biology and Medicine. React. Oxyg Species (2016) 1 9-21. DOI: 10.20455/ros.2016.803 35. He L., He T., Farrar S., Ji L., Liu T., Ma X.. Antioxidants Maintain Cellular Redox Homeostasis by Elimination of Reactive Oxygen Species. Cell Physiol. Biochem. (2017) 44 532-553. DOI: 10.1159/000485089 36. Christofidou-Solomidou M., Muzykantov V.R.. Antioxidant strategies in respiratory medicine. Treat Respir. Med. (2006) 5 47-78. DOI: 10.2165/00151829-200605010-00004 37. Li G., Zhan J., Hu L., Yuan C., Ying X., Hu Y.. Identification of novel antioxidant peptide from porcine plasma hydrolysate and its effect in in vitro digestion/HepG2 cells model. J. Food Biochem. (2022) 46 e13853. DOI: 10.1111/jfbc.13853 38. Park Y.R., Park C.I., Soh Y.. Antioxidant and Anti-Inflammatory Effects of NCW Peptide from Clam Worm (. J. Microbiol. Biotechnol. (2020) 30 1387-1394. DOI: 10.4014/jmb.2003.03050 39. Walker J.M.. The bicinchoninic acid (BCA) assay for protein quantitation. Methods Mol. Biol. (1994) 32 5-8. PMID: 7951748 40. Liu D., Yang X., Huang J., Wu R., Wu C., He H., Li H.. In situ demonstration and characteristic analysis of the protease components from marine bacteria using substrate immersing zymography. Appl. Biochem. Biotechnol. (2015) 175 489-501. DOI: 10.1007/s12010-014-1287-2 41. Leach A.A.. The determination of the substitution achieved at the alpha-amino, epsilon-amino and imidazole groups of proteins with special reference to derivatives of gelatin. Biochem. J. (1966) 98 506-512. DOI: 10.1042/bj0980506 42. Dávalos A., Gómez-Cordovés C., Bartolomé B.. Extending applicability of the oxygen radical absorbance capacity (ORAC-fluorescein) assay. J. Agric. Food Chem. (2004) 52 48-54. DOI: 10.1021/jf0305231 43. Shimada K., Fujikawa K., Yahara K., Nakamura T.. Antioxidative properties of xanthan on the autoxidation of soybean oil in cyclodextrin emulsion. J. Agric. Food Chem. (1992) 40 945-948. DOI: 10.1021/jf00018a005 44. Wang B., Li Z.R., Chi C.F., Zhang Q.H., Luo H.Y.. Preparation and evaluation of antioxidant peptides from ethanol-soluble proteins hydrolysate of Sphyrna lewini muscle. Peptides (2012) 36 240-250. DOI: 10.1016/j.peptides.2012.05.013 45. Alugoju P., Janardhanshetty S.S., Subaramanian S., Periyasamy L., Dyavaiah M.. Quercetin Protects Yeast Saccharomyces cerevisiae pep4 Mutant from Oxidative and Apoptotic Stress and Extends Chronological Lifespan. Curr. Microbiol. (2018) 75 519-530. DOI: 10.1007/s00284-017-1412-x 46. Lin C., Su Z., Luo J., Jiang L., Shen S., Zheng W., Gu W., Cao Y., Chen Y.. Polysaccharide extracted from the leaves of. Int. J. Biol. Macromol. (2020) 143 243-254. DOI: 10.1016/j.ijbiomac.2019.12.023 47. Sakamoto T., Imai H.. Hydrogen peroxide produced by superoxide dismutase SOD-2 activates sperm in. J. Biol. Chem. (2017) 292 14804-14813. DOI: 10.1074/jbc.M117.788901 48. Yasuda K., Kubo Y., Murata H., Sakamoto K.. Cortisol promotes stress tolerance via DAF-16 in. Biochem. Biophys. Rep. (2021) 26 100961. DOI: 10.1016/j.bbrep.2021.100961
--- title: Transcriptomics Dissection of Calorie Restriction and Exercise Training in Brown Adipose Tissue and Skeletal Muscle authors: - Yonghao Feng - Zhicheng Cui - Xiaodan Lu - Hongyu Gong - Xiaoyu Liu - Hui Wang - Haoyu Cheng - Huanqing Gao - Xiaohong Shi - Yiming Li - Hongying Ye - Qiongyue Zhang - Xingxing Kong journal: Nutrients year: 2023 pmcid: PMC9966723 doi: 10.3390/nu15041047 license: CC BY 4.0 --- # Transcriptomics Dissection of Calorie Restriction and Exercise Training in Brown Adipose Tissue and Skeletal Muscle ## Abstract Calorie restriction (CR) and exercise training (EX) are two critical lifestyle interventions for the prevention and treatment of metabolic diseases, such as obesity and diabetes. Brown adipose tissue (BAT) and skeletal muscle are two important organs for the generation of heat. Here, we undertook detailed transcriptional profiling of these two thermogenic tissues from mice treated subjected to CR and/or EX. We found transcriptional reprogramming of BAT and skeletal muscle as a result of CR but little from EX. Consistent with this, CR induced alterations in the expression of genes encoding adipokines and myokines in BAT and skeletal muscle, respectively. Deconvolution analysis showed differences in the subpopulations of myogenic cells, mesothelial cells and endogenic cells in BAT and in the subpopulations of satellite cells, immune cells and endothelial cells in skeletal muscle as a result of CR or EX. NicheNet analysis, exploring potential inter-organ communication, indicated that BAT and skeletal muscle could mutually regulate their fatty acid metabolism and thermogenesis through ligands and receptors. These data comprise an extensive resource for the study of thermogenic tissue molecular responses to CR and/or EX in a healthy state. ## 1. Introduction Skeletal muscle and brown adipocytes have been found to share a common Pax7+/Myf5+ lineage, and transcriptional profiles of their precursor cells showed similarities [1,2]. Both skeletal muscle and brown adipose tissue (BAT) use energy substrates, such as glucose and fatty acids, to generate heat and to maintain core body temperature [3]. Consistent with this, skeletal muscle and BAT have been shown to have synergistic effects in response to cold challenge [3]. Skeletal muscle has multiple mechanisms of heat production involving shivering and non-shivering thermogenesis [4,5]. BAT encodes uncoupling proteins to dissipate the mitochondrial proton gradient and produce heat [6,7]. Nevertheless, the transcriptomic adaptations of these two tissues to various physiological treatments (e.g., calorie restriction (CR) and exercise) are not fully understood. CR, a classical example of negative energy balance, extends healthy lifespan from yeast to mammals [8,9,10]. Though the precise mechanisms of reaction to CR are still not fully defined, the metabolic benefits have been studied in different tissues. For example, short-term CR feeding enhances skeletal muscle stem cell function, mitochondrial function and muscle repair and, eventually, improves muscle insulin sensitivity [11]. It is also evident that CR can improve liver lipid metabolism and reduce systemic inflammation [12,13]. Given these findings, it is clear that adaptation to CR leads to progressive recruitment of metabolic tissues, but effects on BAT are still poorly explored. While a previous study demonstrated that CR led to browning of white adipose tissue [14], another study indicated that CR during pregnancy diminished thermogenic capacity in the offspring, including impaired BAT sympathetic innervation and thyroid hormone signaling [15]. It would thus be interesting to investigate the molecular changes in BAT in response to CR. Exercise training (EX) causes many adaptations in the body that contribute to beneficial effects on health, including enhancing insulin sensitivity and reducing circulating lipid concentrations, primarily via adaptations to skeletal muscle [16,17,18]. Recently, studies have begun to address exercise-induced adaptations in adipose tissue [19]. While exercise is reported to induce white adipose tissue browning [20,21,22], the effects of EX on BAT are controversial. Different results have been obtained in various studies investigating BAT mitochondrial activity and gene expression, glucose uptake, the lipidome of BAT and the thermogenesis of BAT after acute and chronic exercise. Several rodent studies determined that EX increases mitochondrial activity and thermogenesis in rodent BAT [18,23,24,25], while others showed decreased mitochondrial gene expression and thermogenesis [26,27]. In humans, studies have shown that exercise reduces glucose uptake in BAT [27,28]. Given that the biological analysis is debated, the molecular pathway analysis still needs to be clarified. Collectively, CR, EX, or a combination of both are accepted as effective strategies in obesity prevention and treatment. Enhanced skeletal muscle or BAT activities play essential roles in the treatment of obesity. Skeletal muscle can produce cytokines, including proteins, peptides, enzymes and metabolites, and these cytokines can contribute to the weight loss induced by EX [29,30]. In addition, evidence has also shown that skeletal muscle-induced cytokines can induce browning of WAT in response to both CR and EX [14,31]. Therefore, we hypothesized that CR and EX may yield greater benefits for weight loss treatments. Although the effect of CR on changes in metabolites in BAT has been reported [32], the signaling pathway at the transcriptome level for skeletal muscle and BAT in response to CR and EX is unclear. Thus, in this study, we carried out RNA-seq of BAT and skeletal muscle to investigate the molecular and pathway alterations in response to CR and/or exercise interventions. To explore whether CR and EX have synergistic benefits, we restricted the EX mice to $70\%$ calorie intake. ## 2.1. Experiment Model and Subject Details Four-week-old healthy male C57BL/6J mice were purchased from GemPharmatech Co., Ltd. All mice were housed under standard laboratory conditions (12 h on/off) and in a temperature-controlled environment (22–24 °C) in the SPF Animal Research Center of Shanghai University of Sports (SYXK 2014-0002). The mice were given ad libitum access to food and water or fed a CR diet beginning at 8 weeks of age. The glucose fed to mice was tested three days before they were sacrificed. Mice were sacrificed 24 h after the last bout of EX. Mice in all groups were fasted for 6 h prior to tissue harvesting. Tissues were then rapidly dissected and processed or stored for analysis. ## 2.2. CR Compared to the baseline food intake, the food intake of the mice was reduced by $10\%$ per week to a final level of $70\%$ of the ad libitum food intake at the 8th week. ## 2.3. Chronic Treadmill Exercise Training Eight-week-old male C57BL/6J wild-type mice with and without CRs were randomly divided into the sedentary and EX groups. The treadmill treatment included 2 days of adaptive training. For the adaptive EX, the treadmill was set at a 10° incline and began with a 5 min 0 m/min acclimation period, followed by 10 m/min for 10 min and 14 m/min for another 10 min. Mice started to run on the third day (5° incline, 12 m/min for 1 h, 5 days a week for 8 weeks). Mice were fasted for 3 h prior to EX. ## 2.4. Body Composition Measurement A MiniSpec MQ10 nuclear magnetic resonance analyzer (Bruker) was used to measure the body composition of the mice according to the manufacturer’s instructions. Briefly, mice were put in a nuclear magnetic resonance tube and loaded in the machine. The body composition was measured automatically by the machine. ## 2.5. H&E Staining BAT tissues were fixed in a $4\%$ paraformaldehyde solution for at least 24 h and embedded in paraffin. Hematoxylin and eosin (H&E) (hematoxylin—E607317-0500, eosin—E607321-0100; Sangon Biotech, Shanghai, China) staining was undertaken. The tissues were processed as per routine for paraffin embedding, and 5 μm thick sections were cut and placed on glass slides. The paraffin-embedded sections were dewaxed with xylene, washed with a gradient of ethanol to water and then incubated with hematoxylin and eosin (Servicebio, Wuhan, China) for 4 min and sealed after conventional ethanol dehydration. Finally, the sections were analyzed under a Nikon light microscope at the indicated magnification. ImageJ software was used to calculate the areas of droplets in H&E images. ## 2.6. Bulk mRNA Sequencing A total of 15 BAT samples, including 3 from the sedentary group (control), 4 from the CR group, 4 from the EX group (EX) and 4 from the CR + EX group (CREX), and 14 gastrocnemius muscle tissue samples, including 4 from the control group, 3 from the CR group, 4 from the EX group and 3 from the CREX group, were used to perform transcriptome analysis. The Trizol method was used to extract total RNA according to the manufacturer’s instructions. RNA quality was measured using an Agilent 2100 Bioanalyzer (RNA 6000 Nano Kit; Agilent Technologies, Santa Clara, CA, USA). A previously reported method was utilized to construct cDNA libraries for each sample [33]. Libraries were sequenced on a BGIseq500 platform (BGI-Shenzhen, China) using 150 bp paired-end reads aimed at 30 million reads per sample. ## 2.7. Sequence Alignment and Gene Expression Analysis The raw sequencing data were filtered with trim-galore (v0.6.7) by removing reads containing sequencing adapters and reads with low-quality bases. The clean reads were mapped to the reference genome (mm10) using STAR (v2.7.10a). Quantification of gene expression was calculated using rsem (v2.0.1) with the following command: (rsem-calculate-expression --paired-end -p 20 –alignments samples.bam mm10). We conducted a differently expressed (DE) genes analysis for the two groups. Subsequently, differential expression analysis for the groups was carried out using DESeq2 (v1.38.0) with an adjusted p value < 0.05 and \|log2FC\| > log2(1.5). Then, GO and KEGG pathway analyses of the DEGs were carried out using the enrichGO and enrichKEGG functions in the R package clusterProfiler (v4.6.0). ## 2.8. Adipokine and Myokine Analysis Based on previous secretome profiling data from adipose tissue and skeletal muscle obtained with the LC-MS/MS method, we determined the DEGs encoding adipokines and myokines. Then, we performed GO and KEGG analyses to explore the roles of these DEGs using the enrichGO and enrichKEGG functions in the R package clusterProfiler (v4.6.0) ## 2.9. Deconvolution We input gene counts from the bulk RNA-seq data from BAT and GAS, as well as the multi-subject single-cell profiles from previously published single-cell RNA sequencing (scRNA-seq) or single-nuclei RNA sequencing (snRNA-seq) data (skeletal muscle: GSE183288; BAT: GSE207705), into the MuSiC2 algorithm in R4.2.2. Cell types from the scRNA-seq and snRNA-seq were based on prior categorizations. Genes in the bulk RNA-seq data and single-cell and single-nuclei profiles used gene symbols as their identifiers. MuSiC2 used cell-type-specific gene expression from scRNA-seq data to characterize cell type compositions in the bulk RNA-seq. The estimated proportions were normalized to sum to 1 across the included cell types. The Wilcoxon test was used to calculate comparisons between group levels. ## 2.10. BAT and Skeletal Muscle Tissue Crosstalk To analysis ligand activity in the two tissues, we first defined the set of potentially active ligands in the ‘‘sender’’ tissue (i.e., skeletal muscle or BAT). Ligand–receptor interactions and weighted-networks data were downloaded from https://zenodo.org/record/3260758 (accessed on 8 January 2023). Ligands regulated by CR with or without EX for which at least one specific receptor was expressed (average mean expression over all conditions > 1 tag/kilobase (kb)) in the receiver tissue (i.e., BAT or skeletal muscle) were used for the next analyses. Then, NicheNet (v1.1.1) was used to rank the ligands based on how well they predicted whether a gene was linked with a gene set of interest compared to the background gene set. The gene sets of interest for the BAT and skeletal muscle were defined as metabolic-related DEGs. All other genes expressed in the receiver tissue (average mean expression over all conditions > 1 tag/kb) were considered background. Ligand activity scores were calculated as the Pearson correlation coefficient for the ligand–target regulatory potential scores of each selected ligand and the target indicator vectors, which indicated whether a gene belonged to the gene set of interest or not. For the top five ligands with the highest ligand activity, the corresponding receptors were exhibited in a ligand–receptor heatmap. Furthermore, the most prominent target genes for the top five ligands were chosen according to the regulatory potential score, with the genes presented relating to the indicated gene set of interested and being among the 5000 most strongly predicted targets of at least one of the top five ligands. Ligand–target gene interactions were illustrated in a circle plot using the R-package circlize (v0.4.15). ## 2.11. Quantification and Statistical Analysis Unpaired one-tailed t tests or Wilcoxon rank sum and signed rank tests were performed to compare the two groups. One-way ANOVA and post hoc Tukey’s multiple comparison tests were performed for intergroup comparisons of more than two groups. ## 3.1. Phenotypic Response to CR and/or EX and Profiling of Two Metabolic Tissues We studied 8-week-old C57BL/6J male mice treated with CR and/or treadmill running exercise training (EX) interventions for 8 weeks (Figure 1A; $$n = 38$$ across four groups) and then collected BAT and gastrocnemius muscle for transcriptomic profiling. Phenotypically, CR decreased body weight (BW) gain (Figure 1B). The reduced weight was because of the lower fat mass in the CR group compared to the control group, with no differences in lean mass (Figure 1C,D). The BAT weight was also lower in CR feeding mice (Figure 1E). EX had no significant effects on BW gain or body mass (Figure 1B–D and Figure S1A,B). The level of fed glucose was lower in the CR and/or EX groups compared to the control group (Figure 1F). Moreover, CR decreased the size of lipid droplets in BAT, whereas EX enlarged the size (Figure 1G,H). Nonetheless, both CR and EX increased browning, showing synergistic effects in the browning of inguinal white adipose tissue, for which adipocytes are usually multilocular, and decreasing lipid droplet size (Figure S1C,D). ## 3.2. BAT-Level Gene and Pathway Alterations upon CR with or without EX To investigate the molecular alterations, we conducted transcriptomic analysis of BAT under the CR and/or EX conditions. The principal component analysis plots of the different interventions are shown in Figure 2A,B. A total of 1362 differentially expressed genes (DEGs) were identified in the CR group compared to the control group, including 499 upregulated and 863 downregulated genes (Figure 2C and Figure S2A). Additionally, a total of 781 DEGs were identified in the CREX group compared to the EX group, including 310 upregulated and 471 downregulated genes (Figure 2D and Figure S2B). To gain further insights into these DEGs, we conducted gene ontology (GO) term and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses. Upregulated DEGs induced by CR with or without EX were mainly enriched in lipid and small molecule metabolism, whereas downregulated DEGs were engaged in general functions and pathways related to the extracellular matrix (Figure 2E–H). These results indicated that CR and CREX might have similar effects on BAT (Figure 2I). Next, we evaluated the similarities and differences between CR and CREX with regard to transcriptomic alterations. There were 222 upregulated and 341 downregulated genes overlapping in the CR and CREX comparisons. While 522 downregulated and 277 upregulated genes were only found in the CR comparison, 130 downregulated and 88 upregulated genes were found in the CREX comparison (Figure 2J). Analysis of pathways revealed that co-upregulated DEGs were mainly enriched in functions associated with the lipid metabolic process, while co-downregulated DEGs not only participated in the metabolic process but were also involved in cell motility (Figure S2C,D). In CR-specific upregulated DEGs, the metabolic process and biosynthetic process were significantly enriched, and downregulated DEGs were primarily involved in cytoskeleton and immune-related functions (Figure S2E,F). Upregulated DEGs in the CREX group were related to biosynthesis of amino acids and carbon metabolism, and downregulated DEGs in the CREX group related to general functions concerning the extracellular matrix (ECM) (Figure S2G,H). ## 3.3. BAT-Level Gene Alteration upon EX with or without CR Transcriptome analysis demonstrated that EX caused a few genes to change, which was consistent with the lack of changes in BW. PCA plots of the EX vs. control groups and CREX vs. CR groups are displayed in Figure S3A,B. In total, 13 upregulated and 24 downregulated genes were found in the EX group compared to control (Figure S3C,E), and 25 upregulated and 5 downregulated genes were found in the CREX group vs. the CR group (Figure S3D,F). ## 3.4. Analysis of DEGs Encoding Adipokines upon CR with or without EX BAT can secrete various adipokines executing autocrine, paracrine and endocrine regulatory functions [34]. We explored the effects of CR and/or EX on genes encoding adipokines [35,36,37]. In total, 111 different adipokine genes were profiled in the CR group compared to the control group, including 67 upregulated and 44 downregulated genes (Figure 3A), and 76 different adipokine genes were identified in the CREX group compared to the EX group, including 40 upregulated and 36 downregulated genes (Figure 3B). There were 21 upregulated and 32 downregulated adipokines overlapping in the two comparisons (Figure 3C). Pathway analysis showed that co-upregulated adipokines were mainly enriched in functions associated with amino acids, fatty acids and small molecule metabolic processes, while co-downregulated adipokines primarily participated in the cell cytoskeleton, cell migration, ECM–receptor interactions and protein digestion and absorption (Figure 3D,E). Moreover, CR-specific upregulated adipokines were related to small molecular metabolic processes, while CR-specific downregulated adipokines were associated with general functions related to the cell cytoskeleton, migration and immune response (Figure 3F,G). As only a few CREX-specific adipokines were identified, we did not analyze the CREX-specific pathways. ## 3.5. Skeletal Muscle-Level Gene and Pathway Alterations upon CR with or without EX To profile transcriptional alterations in skeletal muscle response to CR and/or EX, bulk RAN-seq of skeletal muscle was performed. The principal component analysis plot for the four-group intervention is shown in Figure 4A,B. A total of 1308 DEGs were identified in the CR group compared to the control group, including 593 upregulated and 715 downregulated genes (Figure 4C and Figure S4A). Additionally, a total of 1454 DEGs were determined in the CREX group compared to the EX group, including 588 upregulated and 866 downregulated genes (Figure 4D and Figure S4B). Pathway analysis demonstrated that all upregulated DEGs induced by CR were mainly enriched in the cholesterol metabolism function, whereas downregulated DEGs were engaged in general functions and pathways related to the ECM and ECM–receptor interactions (Figure 4E,F), similar to the findings for BAT. Upregulated DEGs induced by CREX were mainly enriched in the general function associated with the ribosome, whereas downregulated DEGs were engaged in general functions and pathways related to the ECM, ECM–receptor interaction and oxidative phosphorylation (Figure 4G,H). Next, we assessed the similarities and differences in transcriptomic alterations caused by CR and CREX (Figure 4I). There were 290 commonly upregulated and 434 commonly downregulated genes in the CR and CREX comparisons (Figure 4J). Pathway analysis revealed that co-upregulated DEGs were enriched in functions relating to the AMPK signaling pathway, neuroactive ligand–receptor interactions and cellular senescence, while co-downregulated DEGs were involved in functions relating to the inner mitochondrial membrane protein complex, cell differentiation, cell adhesion and cell migration (Figure S4C,D). Then, we analyzed the role of CR-specific regulated DEGs. Upregulated DEGs participated in the lipid metabolic process and the biosynthesis of amino acids, and downregulated DEGs were primarily involved in ECM–receptor interactions, insulin resistance and lipid metabolism (Figure S4E,F). CREX-specific upregulated DEGs were related to the ribosome, and CREX-specific downregulated DEGs were related to general functions concerning channel activity and lipid metabolism (Figure S4G,H). ## 3.6. Skeletal Muscle-Level Gene and Pathway Alterations upon EX with or without CR Similarly to the BAT transcriptome, a few DEGs were identified in skeletal muscle upon EX with or without CR. The principal component analysis plots of the EX versus control groups and CREX versus CR displayed no significant differences (Figure S5A,B). DEG analysis distinguished 48 DEGs in the EX group versus the control group, encompassing 20 upregulated genes and 28 downregulated genes (Figure S5C,E), and 15 DEGs in the CREX group versus the CR group, consisting of 6 upregulated genes and 9 downregulated genes (Figure S5D,H). GO and pathway analyses verified that the upregulated DEGs induced by EX were related to circadian rhythm, whereas the downregulated DEGs were related to the PPAR signaling pathway (Figure S5F,G). ## 3.7. Analysis of DEGs Encoding Myokines upon CR with or without EX Skeletal muscle has been reported to secret myokines to communicate with other tissues [38,39,40,41]. Therefore, we investigated the effects of CR and/or EX on genes encoding myokines. In total, 100 different myokine genes were profiled in the CR compared to control groups, including 35 upregulated and 65 downregulated myokines (Figure 5A), and 112 different myokine genes were determined in the CREX group compared to the EX group, including 30 upregulated and 82 downregulated myokines (Figure 5B). There were 9 upregulated and 44 downregulated adipokines overlapping in the two comparisons (Figure 5C). GO and pathway analyses showed that co-upregulated myokines were mainly enriched in functions associated with calcium ion transport and the ECM, while co-downregulated DE myokines were primarily related to cell growth, cell migration, cellular response to environmental stimuli and cancer-related pathways (Figure 5D,E). CR-specific upregulated DE myokines were related to vitamin binding, lipid location, carbon metabolism and the ECM, and downregulated DE myokines were linked to general functions concerning skeletal system development, cell motility, immune response and the MAPK signaling pathway (Figure 5F,G). CREX-specific upregulated DE myokines were associated with general functions concerning the ribosome, and downregulated DE myokines were related to various metabolic processes, immune response and the lipid and atherosclerosis signaling pathway (Figure 5H,I). As EX had little effect on transcriptome alterations, it only caused tiny changes in genes encoding myokines. ## 3.8. Cell Proportion Alterations resulting from CR and/or EX across the Two Tissues BAT and skeletal muscle are heterogeneous tissues with different cell types. We analyzed the cell proportions in BAT and skeletal muscle. To recognize the cell population compositions of the two tissues, we performed deconvolution analysis using published single-cell or single-nuclei RNA sequencing data (BAT, GSE207706; skeletal muscle, GSE183288). Statistical comparisons were not performed for proportions of cell subgroups close to or equal to zero. For the BAT, five cell populations—adipose cells, myogenic cells, mesothelial cells, endothelial cells and smooth muscle cells—were analyzed (Figure 6A). Only mesothelial cell proportions were decreased in the CR group compared to control. The other four cell proportions did not show significantly different responses to CR without or with EX. In addition, EX decreased the adipose cell proportions and increased the myogenic and endogenic cell proportions. For skeletal muscle, deconvolution analysis identified ten cell populations: muscle fiber cells, satellite cells, fibro-adipogenic progenitor (FAP) cells, B cells, dendritic cells, monocyte cells, neutrophil cells, smooth muscle cells, endothelial cells and glial cells (Figure 6B). CR with or without EX had no significant effects on muscle fiber cell proportions. CR probably increased the satellite and endothelial cell proportions but decreased immune cell proportions, such as dendritic and neutrophil cell proportions (Figure 6B). While endothelial cell proportions were increased by EX, immune cell (neutrophil and dendritic cells) proportions were decreased. Both CR and EX decreased smooth muscle cell proportions. CR had no effects on the proportions of the other three cell populations with or without EX. ## 3.9. The Crosstalk between BAT and Muscle upon CR with or without EX BAT and skeletal muscle originate from the same precursors in the somites and display multifaceted interactions. Our group previously reported that BAT could communicate with skeletal muscle via myostatin [1]. Others have shown that skeletal muscle secreted myokines to regulate the metabolism of adipose tissue [20,21,22]. We analyzed the networks between BAT and skeletal muscle. Gene networks linking selected DEGs from the two tissues encoding interacting proteins were constructed. Biologically meaningful modules of interacting proteins linked with metabolism were shown (Figure 7A). Considering that the main cell compositions of BAT and skeletal muscle were adipocyte and muscle fiber cells, respectively, we performed NicheNet analysis to explore potential inter-tissue communication between BAT and skeletal muscle. We found that CR-regulated, BAT-derived ligands had the potential to modulate skeletal muscle-selective gene programs related to fatty acid metabolism and thermogenesis. The five BAT ligands with the highest predicted activity in CR were Apoe, Bmp5, Nptn, Col5a3 and Inhbb. All of the five had corresponding receptors with high interaction potential in skeletal muscle (Figure 7B). These receptors could regulate several target genes in skeletal muscle correlated with fatty acid metabolism and thermogenesis (Figure 7C). Of the top five potential ligands in BAT, Nptn was upregulated, and Inhbb, Col5a3, Apoe and Bmp5 were downregulated (Figure 7D). A total of 15 target genes regulated by these receptors were identified, of which Ehhadh, Acsl1, Hadha and Hadhb were downregulated; Hsd17b7 was upregulated; and Acadvl, Hadh, Acaa2, Acadl, Ndufs1, Dhcr24, Cox5a, Ndufb2, Atp5h and Fdft1 were not significantly different in the CR group compared to the control group (Figure 7E). Next, we investigated whether CR-regulated, skeletal muscle-derived ligands had the potential to regulate BAT-selective gene programs related to fatty acid metabolism and thermogenesis (Figure 7F). The top five skeletal muscle ligands with the highest predicted activity in the CR group were Apoe, Col5a3, Gpi1, Angpt1 and Col4a1, and all had corresponding receptors with high interaction potential in BAT (Figure 7F). These receptors could regulate multiple target genes in BAT correlated with fatty acid metabolism and thermogenesis (Figure 7G). Of the top five potential ligands in skeletal muscle, Apoe and Gpi1 were upregulated and Angpt1, Col4a1 and Col5a3 were downregulated (Figure 7H). A total of seven target genes regulated by these receptors were identified, of which Hsd17b7, Dhcr24 and Fdft1were upregulated and Ndufs1, Cox5a, Atp5h and Ndufs8 were not significantly different in the CR group compared to the control group (Figure 7I). Furthermore, we performed NicheNet analyses of the CREX and EX conditions. The top five BAT-derived ligands with the highest predicted activity in the CR with EX condition were Apoe, Bmp5, Col5a3, Pdgfd and Sfrp2, and all had several receptors with high interaction potential in skeletal muscle (Figure S6A). These receptors could regulate multiple target genes in skeletal muscle correlated with fatty acid metabolism and thermogenesis (Figure S6B). The five skeletal muscle-derived ligands with the highest predicted activity in the CREX group were Gpi1, Col4a1, Icam2, Cx3cl1 and Nampt, and all had several receptors with high interaction potential in BAT (Figure S6C). These receptors could regulate multiple target genes in BAT correlated with fatty acid metabolism and thermogenesis (Figure S6D). Taken together, the findings suggested that Apoe and Col5a3 were mainly involved in interactions under the different conditions, indicating that Apoe and Col5a3 might play vital roles in thermogenesis and fatty acid metabolism. ## 4. Discussion Skeletal muscle and BAT are two well-described thermogenic sites that utilize distinct mechanisms for heat production. CR and EX are two lifestyle interventions that aim to produce improvements in metabolic health. The comparative effects of EX vs. CR on transcriptomics in BAT vs. skeletal muscle have not previously been reported. In the present study, we compared the effects of 8 weeks of CR with/without EX on mRNA expression and signaling pathways relating to thermogenesis, metabolism and ECM; in addition, we examined the cell populations and potential ligand–receptor communications between BAT and muscle. Our data showed that CR and CREX can significantly reduce body weight and fat mass; however, we did not observe significant differences between the EX and control groups. CR, EX and CREX could decrease fed glucose; however, there was a greater decrease in the CREX group compared to the EX group. Our findings highlight that strategies targeting obesity should combine both CR and EX. White adipose tissue, accounting for the largest volume of adipose tissue in humans, is critical for energy storage, endocrine communication and glucose homeostasis [42]. Browning is a process that plays a critical role in white adipose tissue regulation and leads to increased thermogenesis [43]. CR has been reported to result in browning of white adipose tissue in lean male C57BL/6 and BALB/c mice after only 1 week of restriction [14], which is consistent with our results. In a previous study, CR enhanced the thermogenesis of BAT by affecting the tricarboxylic acid cycle and fatty acid degradation [32]. Inhibition of the ECM and fatty acid metabolism have been reported to be linked to thermogenesis of BAT [44,45]. In our study, CR showed a decline in the ECM but enhanced fatty acid biosynthesis in BAT, which also indicated that CR activates thermogenesis in BAT. It is possible that, during CR, BAT switches function to store lipids, as shown in 9 month old rats subjected to 6 months of CR and ECM remodeling [46,47]. Unlike BAT, CR induced cholesterol metabolism and decreased ECM in skeletal muscle. The gene expression level of the ECM, which includes many collagen genes, coincides with changes in muscle size, increasing during muscle hypertrophy [48] and decreasing during experimental muscle atrophy [49]. Alternatively, changes in ECM gene expression may simply reflect changes in protein turnover. EX and CR can reduce weight loss by increasing skeletal muscle and BAT burning; however, evidence has shown that exercise alone without dietary restrictions cannot reduce body weight significantly, which indicates that exercise alone is not effective for weight loss, especially for those patients with cardiometabolic diseases [50]. Our study aimed to investigate the joint effects of CR and EX on RNA transcription in skeletal muscle and BAT. Our data showed that the targeted pathway of weight loss in skeletal muscle and BAT was the metabolism of fatty acid, which is consistent with the function of skeletal muscle and BAT. Our data highlighted that strategies for weight loss should consider combining CR and EX simultaneously. In line with a previous study [11], our data showed that satellite population was increased, especially in the CREX group, indicating that CREX-induced metabolic factors play an important role in regulating stem cell function. In addition, consistent with Cara’s study [32], our results showed that, compared with EX, the amino acid biosynthesis pathway was upregulated in CREX, which indicated that CR plays a critical role in amino acid metabolism. Unexpectedly, the present study did not detect significance in the ability of EX to alter mRNA levels of markers of lipid or carbohydrate metabolism in BAT or skeletal muscle. Several reasons may account for this. First, the time of tissue collection: mice were sacrificed 24 h after the last bout of running when the mRNA profiles were probably back to normal. Second, the intensity of exercise: we trained the mice with a low-intensity recipe, which might not have been sufficient to cause gene expression alterations. Third, the duration of running: the mice only ran 1 h per day. Longer duration and higher intensity for the exercise are required in further investigations. Our results in Figure 2 and Figure S3 (BAT) and Figure 4 and Figure S5 (skeletal muscle) showed that the difference in genes alteration between the CR and control groups was substantially stronger than the difference between the EX and control groups, which suggested that CR and EX may exert distinct, nonoverlapping and frequently additive effects on BAT and skeletal muscle. Further work is needed to systematically isolate targetable pathways and address the resulting interactions to develop optimal strategies to promote healthy benefits. ## References 1. Kong X., Yao T., Zhou P., Kazak L., Tenen D., Lyubetskaya A., Dawes B.A., Tsai L., Kahn B.B., Spiegelman B.M.. Brown Adipose Tissue Controls Skeletal Muscle Function via the Secretion of Myostatin. Cell Metab. (2018) 28 631-643.e3. DOI: 10.1016/j.cmet.2018.07.004 2. Timmons J.A., Wennmalm K., Larsson O., Walden T.B., Lassmann T., Petrovic N., Hamilton D.L., Gimeno R.E., Wahlestedt C., Baar K.. Myogenic gene expression signature establishes that brown and white adipocytes originate from distinct cell lineages. Proc. Natl. Acad. Sci. USA (2007) 104 4401-4406. DOI: 10.1073/pnas.0610615104 3. Bal N.C., Maurya S.K., Pani S., Sethy C., Banerjee A., Das S., Patnaik S., Kundu C.N.. Mild cold induced thermogenesis: Are BAT and skeletal muscle synergistic partners?. Biosci. Rep. (2017) 37 BSR20171087. DOI: 10.1042/BSR20171087 4. Aydin J., Shabalina I.G., Place N., Reiken S., Zhang S.J., Bellinger A.M., Nedergaard J., Cannon B., Marks A.R., Bruton J.D.. Nonshivering thermogenesis protects against defective calcium handling in muscle. FASEB J. (2008) 22 3919-3924. DOI: 10.1096/fj.08-113712 5. Dawkins M.J., Scopes J.W.. Non-shivering thermogenesis and brown adipose tissue in the human new-born infant. Nature (1965) 206 201-202. DOI: 10.1038/206201b0 6. Enerback S., Jacobsson A., Simpson E.M., Guerra C., Yamashita H., Harper M.E., Kozak L.P.. Mice lacking mitochondrial uncoupling protein are cold-sensitive but not obese. Nature (1997) 387 90-94. DOI: 10.1038/387090a0 7. Kong X., Banks A., Liu T., Kazak L., Rao R.R., Cohen P., Wang X., Yu S., Lo J.C., Tseng Y.H.. IRF4 is a key thermogenic transcriptional partner of PGC-1alpha. Cell (2014) 158 69-83. DOI: 10.1016/j.cell.2014.04.049 8. Colman R.J., Beasley T.M., Kemnitz J.W., Johnson S.C., Weindruch R., Anderson R.M.. Caloric restriction reduces age-related and all-cause mortality in rhesus monkeys. Nat. Commun. (2014) 5 3557. DOI: 10.1038/ncomms4557 9. Wei M., Fabrizio P., Hu J., Ge H., Cheng C., Li L., Longo V.D.. Life span extension by calorie restriction depends on Rim15 and transcription factors downstream of Ras/PKA, Tor, and Sch9. PLoS Genet. (2008) 4. DOI: 10.1371/journal.pgen.0040013 10. Bordone L., Guarente L.. Calorie restriction, SIRT1 and metabolism: Understanding longevity. Nat. Rev. Mol. Cell Biol. (2005) 6 298-305. DOI: 10.1038/nrm1616 11. Cerletti M., Jang Y.C., Finley L.W., Haigis M.C., Wagers A.J.. Short-term calorie restriction enhances skeletal muscle stem cell function. Cell Stem Cell (2012) 10 515-519. DOI: 10.1016/j.stem.2012.04.002 12. Fan J.G., Zhong L., Xu Z.J., Tia L.Y., Ding X.D., Li M.S., Wang G.L.. Effects of low-calorie diet on steatohepatitis in rats with obesity and hyperlipidemia. World J. Gastroenterol. (2003) 9 2045-2049. DOI: 10.3748/wjg.v9.i9.2045 13. Park C.Y., Park S., Kim M.S., Kim H.K., Han S.N.. Effects of mild calorie restriction on lipid metabolism and inflammation in liver and adipose tissue. Biochem. Biophys. Res. Commun. (2017) 490 636-642. DOI: 10.1016/j.bbrc.2017.06.090 14. Fabbiano S., Suarez-Zamorano N., Rigo D., Veyrat-Durebex C., Stevanovic Dokic A., Colin D.J., Trajkovski M.. Caloric Restriction Leads to Browning of White Adipose Tissue through Type 2 Immune Signaling. Cell Metab. (2016) 24 434-446. DOI: 10.1016/j.cmet.2016.07.023 15. Palou M., Priego T., Romero M., Szostaczuk N., Konieczna J., Cabrer C., Remesar X., Palou A., Pico C.. Moderate calorie restriction during gestation programs offspring for lower BAT thermogenic capacity driven by thyroid and sympathetic signaling. Int. J. Obes. (2015) 39 339-345. DOI: 10.1038/ijo.2014.56 16. Bonadonna R.C., Del Prato S., Saccomani M.P., Bonora E., Gulli G., Ferrannini E., Bier D., Cobelli C., DeFronzo R.A.. Transmembrane glucose transport in skeletal muscle of patients with non-insulin-dependent diabetes. J. Clin. Investig. (1993) 92 486-494. DOI: 10.1172/JCI116592 17. Goodyear L.J., Kahn B.B.. Exercise, glucose transport, and insulin sensitivity. Annu. Rev. Med. (1998) 49 235-261. DOI: 10.1146/annurev.med.49.1.235 18. Dewal R.S., Stanford K.I.. Effects of exercise on brown and beige adipocytes. Biochim. Biophys. Acta Mol. Cell Biol. Lipids (2019) 1864 71-78. DOI: 10.1016/j.bbalip.2018.04.013 19. Stanford K.I., Middelbeek R.J., Goodyear L.J.. Exercise Effects on White Adipose Tissue: Beiging and Metabolic Adaptations. Diabetes (2015) 64 2361-2368. DOI: 10.2337/db15-0227 20. Bostrom P., Wu J., Jedrychowski M.P., Korde A., Ye L., Lo J.C., Rasbach K.A., Bostrom E.A., Choi J.H., Long J.Z.. A PGC1-alpha-dependent myokine that drives brown-fat-like development of white fat and thermogenesis. Nature (2012) 481 463-468. DOI: 10.1038/nature10777 21. Rao R.R., Long J.Z., White J.P., Svensson K.J., Lou J., Lokurkar I., Jedrychowski M.P., Ruas J.L., Wrann C.D., Lo J.C.. Meteorin-like is a hormone that regulates immune-adipose interactions to increase beige fat thermogenesis. Cell (2014) 157 1279-1291. DOI: 10.1016/j.cell.2014.03.065 22. Roberts L.D., Bostrom P., O’Sullivan J.F., Schinzel R.T., Lewis G.D., Dejam A., Lee Y.K., Palma M.J., Calhoun S., Georgiadi A.. beta-Aminoisobutyric acid induces browning of white fat and hepatic beta-oxidation and is inversely correlated with cardiometabolic risk factors. Cell Metab. (2014) 19 96-108. DOI: 10.1016/j.cmet.2013.12.003 23. Yoshioka K., Yoshida T., Wakabayashi Y., Nishioka H., Kondo M.. Effects of exercise training on brown adipose tissue thermogenesis in ovariectomized obese rats. Endocrinol. Jpn. (1989) 36 403-408. DOI: 10.1507/endocrj1954.36.403 24. Stanford K.I., Goodyear L.J.. Exercise regulation of adipose tissue. Adipocyte (2016) 5 153-162. DOI: 10.1080/21623945.2016.1191307 25. Ignacio D.L., Fortunato R.S., Neto R.A., da Silva Silvestre D.H., Nigro M., Frankenfeld T.G., Werneck-de-Castro J.P., Carvalho D.P.. Blunted response of pituitary type 1 and brown adipose tissue type 2 deiodinases to swimming training in ovariectomized rats. Horm. Metab. Res. (2012) 44 797-803. DOI: 10.1055/s-0032-1314875 26. Wu M.V., Bikopoulos G., Hung S., Ceddia R.B.. Thermogenic capacity is antagonistically regulated in classical brown and white subcutaneous fat depots by high fat diet and endurance training in rats: Impact on whole-body energy expenditure. J. Biol. Chem. (2014) 289 34129-34140. DOI: 10.1074/jbc.M114.591008 27. Vosselman M.J., Hoeks J., Brans B., Pallubinsky H., Nascimento E.B., van der Lans A.A., Broeders E.P., Mottaghy F.M., Schrauwen P., van Marken Lichtenbelt W.D.. Low brown adipose tissue activity in endurance-trained compared with lean sedentary men. Int. J. Obes. (2015) 39 1696-1702. DOI: 10.1038/ijo.2015.130 28. Motiani P., Virtanen K.A., Motiani K.K., Eskelinen J.J., Middelbeek R.J., Goodyear L.J., Savolainen A.M., Kemppainen J., Jensen J., Din M.U.. Decreased insulin-stimulated brown adipose tissue glucose uptake after short-term exercise training in healthy middle-aged men. Diabetes Obes. Metab. (2017) 19 1379-1388. DOI: 10.1111/dom.12947 29. Hargreaves M., Spriet L.L.. Skeletal muscle energy metabolism during exercise. Nat. Metab. (2020) 2 817-828. DOI: 10.1038/s42255-020-0251-4 30. Pedersen B.K., Febbraio M.A.. Muscles, exercise and obesity: Skeletal muscle as a secretory organ. Nat. Rev. Endocrinol. (2012) 8 457-465. DOI: 10.1038/nrendo.2012.49 31. Aldiss P., Betts J., Sale C., Pope M., Budge H., Symonds M.E.. Exercise-induced ‘browning’ of adipose tissues. Metabolism (2018) 81 63-70. DOI: 10.1016/j.metabol.2017.11.009 32. Green C.L., Mitchell S.E., Derous D., Wang Y., Chen L., Han J.J., Promislow D.E.L., Lusseau D., Douglas A., Speakman J.R.. The Effects of Graded Levels of Calorie Restriction: XIV. Global Metabolomics Screen Reveals Brown Adipose Tissue Changes in Amino Acids, Catecholamines, and Antioxidants After Short-Term Restriction in C57BL/6 Mice. J. Gerontol. A Biol. Sci. Med. Sci. (2020) 75 218-229. DOI: 10.1093/gerona/glz023 33. Yao T., Yan H., Zhu X., Zhang Q., Kong X., Guo S., Feng Y., Wang H., Hua Y., Zhang J.. Obese Skeletal Muscle-Expressed Interferon Regulatory Factor 4 Transcriptionally Regulates Mitochondrial Branched-Chain Aminotransferase Reprogramming Metabolome. Diabetes (2022) 71 2256-2271. DOI: 10.2337/db22-0260 34. Villarroya F., Cereijo R., Villarroya J., Giralt M.. Brown adipose tissue as a secretory organ. Nat. Rev. Endocrinol. (2017) 13 26-35. DOI: 10.1038/nrendo.2016.136 35. Chen X., Cushman S.W., Pannell L.K., Hess S.. Quantitative proteomic analysis of the secretory proteins from rat adipose cells using a 2D liquid chromatography-MS/MS approach. J. Proteome Res. (2005) 4 570-577. DOI: 10.1021/pr049772a 36. Wang P., Mariman E., Keijer J., Bouwman F., Noben J.P., Robben J., Renes J.. Profiling of the secreted proteins during 3T3-L1 adipocyte differentiation leads to the identification of novel adipokines. Cell. Mol. Life Sci. (2004) 61 2405-2417. DOI: 10.1007/s00018-004-4256-z 37. Lehr S., Hartwig S., Lamers D., Famulla S., Muller S., Hanisch F.G., Cuvelier C., Ruige J., Eckardt K., Ouwens D.M.. Identification and validation of novel adipokines released from primary human adipocytes. Mol. Cell. Proteom. (2012) 11 M111.010504. DOI: 10.1074/mcp.M111.010504 38. Hartwig S., Raschke S., Knebel B., Scheler M., Irmler M., Passlack W., Muller S., Hanisch F.G., Franz T., Li X.. Secretome profiling of primary human skeletal muscle cells. Biochim. Biophys. Acta (2014) 1844 1011-1017. DOI: 10.1016/j.bbapap.2013.08.004 39. Henningsen J., Rigbolt K.T., Blagoev B., Pedersen B.K., Kratchmarova I.. Dynamics of the skeletal muscle secretome during myoblast differentiation. Mol. Cell. Proteom. (2010) 9 2482-2496. DOI: 10.1074/mcp.M110.002113 40. Bortoluzzi S., Scannapieco P., Cestaro A., Danieli G.A., Schiaffino S.. Computational reconstruction of the human skeletal muscle secretome. Proteins (2006) 62 776-792. DOI: 10.1002/prot.20803 41. Le Bihan M.C., Bigot A., Jensen S.S., Dennis J.L., Rogowska-Wrzesinska A., Laine J., Gache V., Furling D., Jensen O.N., Voit T.. In-depth analysis of the secretome identifies three major independent secretory pathways in differentiating human myoblasts. J. Proteom. (2012) 77 344-356. DOI: 10.1016/j.jprot.2012.09.008 42. Heinonen S., Jokinen R., Rissanen A., Pietiläinen K.H.. White adipose tissue mitochondrial metabolism in health and in obesity. Obes. Rev. (2020) 21 e12958. DOI: 10.1111/obr.12958 43. Herz C.T., Kiefer F.W.. Adipose tissue browning in mice and humans. J. Endocrinol. (2019) 241 R97-R109. DOI: 10.1530/JOE-18-0598 44. Yudasaka M., Okamatsu-Ogura Y., Tanaka T., Saeki K., Kataura H.. Cold-induced Conversion of Connective Tissue Skeleton in Brown Adipose Tissues. Acta Histochem. Cytochem. (2021) 54 131-141. DOI: 10.1267/ahc.21-00030 45. Heeren J., Scheja L.. Brown adipose tissue and lipid metabolism. Curr. Opin. Lipidol. (2018) 29 180-185. DOI: 10.1097/MOL.0000000000000504 46. Okita N., Hayashida Y., Kojima Y., Fukushima M., Yuguchi K., Mikami K., Yamauchi A., Watanabe K., Noguchi M., Nakamura M.. Differential responses of white adipose tissue and brown adipose tissue to caloric restriction in rats. Mech. Ageing Dev. (2012) 133 255-266. DOI: 10.1016/j.mad.2012.02.003 47. Ham D.J., Borsch A., Chojnowska K., Lin S., Leuchtmann A.B., Ham A.S., Thurkauf M., Delezie J., Furrer R., Burri D.. Distinct and additive effects of calorie restriction and rapamycin in aging skeletal muscle. Nat. Commun. (2022) 13 2025. DOI: 10.1038/s41467-022-29714-6 48. Gordon P.M., Liu D., Sartor M.A., IglayReger H.B., Pistilli E.E., Gutmann L., Nader G.A., Hoffman E.P.. Resistance exercise training influences skeletal muscle immune activation: A microarray analysis. J. Appl. Physiol. (2012) 112 443-453. DOI: 10.1152/japplphysiol.00860.2011 49. Lecker S.H., Jagoe R.T., Gilbert A., Gomes M., Baracos V., Bailey J., Price S.R., Mitch W.E., Goldberg A.L.. Multiple types of skeletal muscle atrophy involve a common program of changes in gene expression. FASEB J. (2004) 18 39-51. DOI: 10.1096/fj.03-0610com 50. Magkos F., Hjorth M.F., Astrup A.. Diet and exercise in the prevention and treatment of type 2 diabetes mellitus. Nat. Rev. Endocrinol. (2020) 16 545-555. DOI: 10.1038/s41574-020-0381-5
--- title: Feasibility Study of Lactobacillus Plantarum 299v Probiotic Supplementation in an Urban Academic Facility among Diverse Pregnant Individuals authors: - Nefertiti OjiNjideka Hemphill - Lacey Pezley - Alana Steffen - Gloria Elam - Michelle A. Kominiarek - Angela Odoms-Young - Nicollette Kessee - Alyshia Hamm - Lisa Tussing-Humphreys - Mary Dawn Koenig journal: Nutrients year: 2023 pmcid: PMC9966742 doi: 10.3390/nu15040875 license: CC BY 4.0 --- # Feasibility Study of Lactobacillus Plantarum 299v Probiotic Supplementation in an Urban Academic Facility among Diverse Pregnant Individuals ## Abstract [1] Background: Despite iron intake recommendations, over a quarter of pregnant individuals have iron deficiency. Lactobacillus plantarum 299v (LP299V®) enhances iron absorption in non-pregnant populations and may have positive effects in pregnancy among those with sufficient iron stores; however, no studies have evaluated the effect of LP299V® on maternal and neonatal iron status among individuals at risk for iron deficiency anemia in pregnancy. Thus, this study aims to assess the feasibility and preliminary efficacy of daily oral LP299V® maternal supplementation among diverse pregnant individuals. [ 2] Methods: In this double-blind placebo-controlled randomized supplementation feasibility study, participants were randomized to probiotic LP299V® + prenatal vitamin with iron or placebo + prenatal vitamin with iron from 15–20 weeks of gestation through delivery. [ 3] Results: Of the 20 enrolled and randomized participants, $58\%$ ($\frac{7}{12}$) from the LP299V® group and $75\%$ ($\frac{6}{8}$) from the placebo group were retained. Adherence to supplementation was $72\%$ for LP299V®/placebo and $73\%$ for the prenatal vitamin. A slower decline in maternal hematological and iron parameters across pregnancy was observed in the LP299V® group compared to placebo. [ 4] Conclusions: LP299V® may be a tolerable therapy during pregnancy and has the potential to affect maternal and neonatal hematological and iron status. ## 1. Introduction The most prevalent micronutrient deficiency in the United States (U.S.) is iron; a large majority of cases of iron deficiency (ID) and iron deficiency anemia (IDA) occur among pregnant women [1,2]. During pregnancy, maternal iron stores are used for the growing fetus, maternal red blood cell (RBC) expansion, and placental growth and development [3], thus increasing the risk for ID and IDA. Across all trimesters of pregnancy in the U.S., it is estimated that $18\%$ of individuals have ID and $5\%$ have IDA, and within the third trimester, the prevalence of ID exceeds $27\%$. Prevalence of IDA is even greater among those who identify as Black or low-income [2]. Maternal ID and IDA are associated with increased risk of preterm birth, low infant birth weight, maternal and fetal mortality, and irreversible infant neurocognitive defects [4,5]. To meet this increasing requirement for iron and to optimize maternal iron nutrition, the Recommended Dietary Allowance for pregnancy is 27 mg/day of iron [6]. However, given the continued high rates of maternal ID and IDA and only modest adherence to daily prenatal vitamins containing iron [7], alternative approaches to optimizing iron nutrition in pregnancy are needed. Research has shown that one-time or short-term dosing of the probiotic Lactobacillus plantarum 299v (LP299V®) enhances iron absorption in non-pregnant populations [8,9,10,11]. However, few studies have examined the effect of long-term supplementation on body iron stores. While probiotics are considered safe to consume in pregnancy, only one LP299V® supplementation trial has been conducted during the gestational period to evaluate its effects on maternal iron stores and risk of IDA [12]. This study, among iron-sufficient pregnant Swedish women, showed a significantly lower decline in iron stores and a significantly lower prevalence of IDA in the third trimester among those randomized to LP299V® compared to standard care control [12]. These results offer potential positive effects for the role of LP299V® in maintaining maternal iron status among those starting pregnancy with sufficient iron stores and who receive care in a decentralized publicly funded healthcare system [12]. However, no studies have evaluated the effect of LP299V® on maternal iron status among individuals at risk for IDA in pregnancy in the U.S., nor have studies extended findings to neonatal iron status. Moreover, it is unknown if positive feasibility and preliminary efficacy would persist in a U.S.-based health care setting with racially, ethnically, and socioeconomically diverse pregnant individuals. Therefore, the objectives of this study were as follows. First and foremost, we examined the feasibility of daily oral LP299V® maternal supplementation taken from the early second trimester through birth. Second, we explored the preliminary efficacy of LP299V® intake on maternal (at-risk for IDA defined as hemoglobin (Hb) between 10.0–12.0 g/dL) and neonatal cord hematological and iron status parameters compared to controls in an urban U.S. academic medical center with a racially, ethnically, and socioeconomically diverse patient population. ## 2. Materials and Methods Study design. This was a two-arm double-blind placebo-controlled randomized supplementation feasibility study conducted at an urban Midwest academic medical center. All participants provided written informed consent (IRB #2016-0662) prior to study participation. The study is registered at clinicaltrials.gov (NCT03646487). Research participants. Clinic schedules were reviewed daily to identify potentially eligible women. Initial eligibility was further assessed via electronic health records (EHR). Women identified as potentially eligible were approached in clinic or called to assess interest. Inclusion and exclusion criteria are presented in Table 1. Treatment groups. Women were randomly assigned to one of two treatment groups: probiotic LP299V® + prenatal vitamin with iron (PNVI), or placebo + PNVI. All participants were instructed to consume the probiotic or placebo and PNVI daily from 15–20 weeks of gestation (WG) up through admittance for delivery. Participants were asked to consume supplements with a cool or room temperature beverage at the same time each day. Participants were able to choose what time of day to consume their supplements; however, it was suggested that the probiotic/placebo and the PNVI be taken together at bedtime or at the same time each day. We used NatureMade Digestive Probiotics, Daily Balance, containing hypromellous capsule material, potato starch, magnesium stearate, and 1010 CFUs LP299V® sourced from ProbiAB, which has documented efficacy for enhancing iron absorption in non-pregnant reproductive-age women and for optimizing iron status among pregnant women [8,9,12]. The PNVI was from NatureMade and provided the % daily value for pregnant and lactating women in one daily tablet. Each PNVI tablet contained 27 mg of iron in the form of ferrous fumarate. The placebo was produced by the University of Illinois at Chicago (UIC) Investigational Drug Service (IDS). The placebo consisted of a $100\%$ gelatin capsule containing cellulose Microcryst PH-102, a non-soluble fine white powder selected to resemble the LP299V® supplement, while being non-absorbable. The placebo was indistinguishable from the probiotic to both participants and researchers. Both the probiotic/placebo capsules and PNVI were packaged in smart medication bottles (Pillsy, Seattle, WA, USA) and generically labeled as “Probiotic or Placebo” and “Prenatal Vitamin.” The medical director of Obstetrics and Gynecology provided a prescription for the supplements to be dispensed by UIC IDS. Women received approximately 32 pills at a time, or approximately one month’s supply. Refills occurred approximately every four weeks and were coupled with standard care clinic visits. In response to the COVID-19 pandemic, refills were shipped directly to the participant’s home. Randomization. Women were randomized at the baseline visit following a 2:1 allocation ratio using the Research Electronic Data Capture (REDCap®) randomization module. The biostatistician provided the randomization scheme to the UIC IDS, the party responsible for filling all supplement prescriptions, to ensure that the study recruitment and data collection staff remained blinded. Data collection. Data was collected at baseline (15–20 WG), 24–28 WG, 34–36 WG, Labor and Delivery and during monthly pill refill visits that fell outside of the main study visits. Most data collection was conducted in person with modifications made at the onset of the COVID-19 pandemic as described below. Study feasibility. The electronic health record (EHR) was screened daily to identify eligible women scheduled for a “New Obstetrics” visit. Women were tracked, and the number of women approached (by phone & in person) for enrollment and the number of women who declined was documented. Once a woman was enrolled in the study, attendance at study visits, completeness of data, and overall and treatment specific loss to follow-up/withdrawal was closely monitored and documented. To track progress of participants through the study, the Consolidated Standards of Reporting Trials (CONSORT) subject flow diagram was utilized [13]. Study feasibility was defined a priori as recruitment ≥ $50\%$ of those eligible, participants completing ≥$80\%$ of planned study visits, and retaining ≥$80\%$ of participants in both treatment arms through admittance for delivery. Health events. Health events including medication changes, iron supplementation, pregnancy-related conditions, and gastrointestinal symptoms were assessed monthly by survey and through viewing provider notes in the participant’s EHR. Study tolerability was defined a priori as no serious health events and the rate of non-serious health events being similar between study arms. Adherence monitoring. Adherence to the supplements was assessed using a Bluetooth-enabled smart pill bottle (Pillsy, Seattle, WA) and standard hand pill counts. At baseline, the smart pill bottles were paired with the participant’s smartphone by a member of the research staff, through the Pillsy phone application. Once paired, the application sent daily alerts to the subject when it was time to take the supplements (i.e., the pill bottle rang and lit up, and a secondary reminder was sent through the phone application). Data was also transmitted to a HIPAA-compliant research platform that tracked the number of times the bottle was opened with a date and timestamp. The technology allowed members of the research staff to receive daily dose-compliance information and allowed for two-way text and call communication between participants and study staff if data from the smart bottle were not transmitted. To proactively account for technological challenges related to the smart bottles, UIC IDS performed standard hand pill counts at each bottle return and recorded the number of pills taken and remaining in each bottle in a spreadsheet that was shared with research staff when the study was completed. Maternal and Cord Blood Collection and Processing. At each data collection visit, maternal venous whole blood was obtained, with a portion processed for serum, to assess maternal iron, inflammatory and hematological parameters. At delivery, a venous umbilical cord blood sample was obtained at the bedside following delivery of the placenta. All maternal and cord blood samples were processed and stored at −80 °C or sent immediately to a commercial laboratory (Quest Diagnostics, Wood Dale, IL, USA) for analysis. Complete Blood Count (CBC). CBC with differential was measured in whole maternal or cord blood by electronic cell sizing/counting/cytometry/microscopy by a local commercial lab (Quest Diagnostics, Wood Dale, IL, USA). Hb, obtained from the CBC, was used to define trimester-specific maternal IDA, with a downward correction of 0.8 g/dL for Black women [6]. At the time of the study, it was recommended to use a race-adjusted cut-point for IDA. However, this race-adjusted cut-point was recently determined to be unfounded [14]. IDA ranges included ≤11 g/dL for the first trimester, ≤10.5 g/dL for the second trimester, and ≤11 g/dL for the third trimester. IDA ranges with the correction for Black women are ≤10.2 g/dL for the first trimester, ≤9.7 g/dL for the second trimester, and ≤10.2 g/dL for the third trimester [15]. Hb is the primary iron-related outcome for pilot trials, given that it is the most common clinical marker for iron/hematological status in pregnancy. Iron Status Parameters. Serum ferritin and iron were measured from maternal and cord serum by immunoassay and spectrophotometry by a local commercial lab (Quest Diagnostics, Wood Dale, IL, USA). Normal ranges for maternal ferritin are: 2nd trimester 2–230 ng/mL and 3rd trimester 0–116 ng/mL [16,17,18]. For serum iron, normal trimester-specific levels are: 1st trimester 72–143 μg/dL, 2nd trimester 44–178 μg/dL, and 3rd trimester 30–193 μg/dL [16]. For transferrin saturation, normal levels are: 1st not reported, 2nd trimester 10–$44\%$, and 3rd trimester 5–$37\%$ [16]. C-reactive Protein. High-sensitivity CRP (hs-CRP), a common clinical marker of systemic inflammation, was measured via nephelometry from maternal and cord serum by a local commercial lab (Quest Diagnostics, Wood Dale, IL, USA). Survey Interviews. Interview-administered surveys included a socio-demographics and health history questionnaire and a 24-h diet recall. The 24-h diet recall was conducted using the Nutrition Data System for Research (NDSR) software (University of Minnesota, Minneapolis, MN, USA) using a multiple-pass interview approach, to standardize collection of the data [19]. The consumption of dietary supplements was evaluated in concurrence with 24-h diet recall using the Dietary Supplement Assessment Module included in NDSR [20]. Data from the recall was used to quantify the average intake of dietary (from food) and supplemental iron across the gestational period. Maternal Anthropometrics. Maternal height was measured using a fixed stadiometer (baseline only) and weight at each gestational data collection visit using a calibrated digital scale (or via EHR during the COVID-19 pandemic) and at admittance for labor and delivery (from EHR). BMI was calculated as kg/m2. Pre-pregnancy weight was self-reported. Neonatal Characteristics. There is evidence that fetal sex, neonatal weight at delivery and gestational age at delivery can affect maternal and neonatal iron status and hematological parameters [21,22]. We obtained these variables from the participant’s EHR. Data Management. Research staff entered all data directly into a REDCap (Vanderbilt University, Nashville, TN, USA) data structure. Standard checks for outliers, duplicates, and other errors associated with data entry were conducted. Power and Statistical Analysis. We aimed to recruit 24 individuals, with the assumption of $80\%$ retention, resulting in a final recruitment sample of 20. With a 2:1 allocation ratio distribution between groups, our sample size is adequate to estimate parameters (e.g., standard deviation, mean change) needed to inform the design of a future efficacy trial. Confidence intervals for feasibility proportions such as retention and percent adherence have a maximum width of 0.27 if proportions are 0.5, and 0.22 with a proportion of 0.8 [23,24]. Statistical analysis was performed using Stata/BE (versions 15 and 17, College Station, TX, USA). Stata variables not normally distributed were transformed using natural log transformation, and data were presented as geometric means with confidence intervals. For data that could not be normalized with transformation, non-parametric tests were utilized. Given the feasibility nature of the study, the analysis was largely descriptive. For continuous variables, means, confidence intervals, medians, interquartile ranges, and standard deviations were reported. For categorical variables, frequencies and percentages were reported. The feasibility and supplement adherence data were reported by all randomized participants and by per protocol analysis (≥$80\%$ supplement adherence, completed the study). The maternal and cord blood hematological markers presented were from the per protocol analysis (≥$80\%$ supplement adherence, completed the study). For the maternal data, we visualized changes in the hematological and iron status parameters using “spaghetti” plots to provide a demonstration of marker changes over the gestational period. Differences between treatment groups (LP299V® + standard PNVI versus placebo + standard PNVI) for continuous data were compared by t-test or the Wilcoxon rank-sum test for non-parametric data and via Fisher’s exact test for categorical data. Maternal iron-related parameters and inflammation at baseline, 24–28 WG, 34–36 WG and labor and delivery, were evaluated by t-tests. Infant iron status and hematological parameters at delivery were compared between mother and infant pairs via t-tests. Mean changes in scores for maternal iron-related parameters were calculated from baseline for each time point, 24–28 WG, 34–36 WG, and labor and delivery. All p values were based on a two-sided test of statistical significance accepted at the level of $p \leq 0.05.$ Significance testing for within- and between-group differences in hematological and iron status markers was assessed; however, this feasibility study was not powered for these types of analyses. Therefore, our data were presented descriptively. ## 3.1. Eligibility Screening and Recruitment During the time of our study recruitment (January 2019–March 2020), a total of 1505 pregnant individuals scheduled their initial prenatal care visits at the University of Illinois Hospital and Health Sciences (UI Health) Center for Women’s Health. We screened all of these individuals’ electronic medical records to assess preliminary eligibility. Of those screened, $67\%$ ($\frac{1013}{1505}$) were ineligible and the remaining $33\%$ ($\frac{492}{1505}$) were preliminarily eligible and were approached. Of those willing to participate in the study ($$n = 195$$), 72 went on to meet full eligibility criteria (initial prenatal care visit Hb 10.0–12.0 g/dL). Of these, 21 individuals completed the informed consent process and were successfully enrolled into the study. Of those enrolled, 12 participants were randomized to the LP299V® group and eight to the placebo group. Of the 20 enrolled and randomized participants, $58\%$ ($\frac{7}{12}$) from the LP299V® group and $75\%$ ($\frac{6}{8}$) from the placebo group were retained. Completion of planned assessments for LP299V® and placebo groups, respectively, were as follows: $100\%$ ($\frac{12}{12}$) and $100\%$ ($\frac{8}{8}$) at baseline, $75\%$ ($\frac{9}{12}$) and $75\%$ ($\frac{6}{8}$) at 24–28 WG, $58\%$ ($\frac{7}{12}$) and $63\%$ ($\frac{5}{8}$) at 34–36 WG, and $58\%$ ($\frac{7}{12}$) and $75\%$ ($\frac{6}{8}$) at delivery. Further details are provided in Figure 1. ## 3.2. Participants A total of 21 pregnant individuals were enrolled in the study. One participant withdrew prior to completing the baseline visit due to discomfort with study blood-draw activities, resulting in a total of 20 randomized participants. The mean maternal age was 28.9 years (SD 6.5) with a mean gestational age of 13.4 WG (SD 4.1) at baseline and 38.8 WG (SD 0.7) at delivery. The study population was comprised primarily of individuals who identified as non-Hispanic ($\frac{16}{20}$, $80\%$) or Black ($\frac{15}{20}$, $75\%$) women. Most individuals were either single and not living with a significant other ($\frac{7}{20}$, $35\%$) or married ($\frac{7}{20}$, $35\%$). More than half the participants had public health insurance (i.e., Medicaid) ($\frac{11}{20}$, $55\%$) and some high school or college education ($\frac{15}{20}$, $75\%$). Most participants reported a household income of less than or equal to $30,000 ($\frac{14}{20}$, $70\%$) and greater than half ($\frac{11}{20}$, $55\%$) were enrolled in the Supplemental Nutrition Assistance Program (SNAP). The mean pre-pregnancy BMI was 31.4 kg/m2 (SD 7.5) with over half the participants having pre-pregnancy obesity ($\frac{11}{20}$, $55\%$). At baseline, mean maternal BMI was 32.2 kg/m2 (SD 6.7) and indicative of a high degree of maternal obesity ($\frac{12}{20}$, $60\%$). A parity of 0 was reported by $40\%$ ($\frac{8}{20}$) of participants with the remaining population divided between a parity of 1 ($\frac{6}{20}$, $30\%$) and 2 or more ($\frac{6}{20}$, $30\%$). The median intake of food iron was 10.5 mg per 1000 kcal (IQR 8.78) and total iron, food, and supplement together was 36.2 mg per 1000 kcal (IQR 22.1). The participants in the intervention groups differed for household income of ≤$30,000 ($$p \leq 0.005$$). All other baseline characteristics were similar (Table 2). ## 3.3. Attrition Following baseline, two ($\frac{2}{8}$) participants withdrew from the placebo group and five ($\frac{5}{12}$) from the probiotic group. Five participants withdrew during the 24–28 WG assessment period (three probiotic and two placebo) and two probiotic participants withdrew at delivery. The reasons for withdrawal differed for all seven participants and included bereavement/miscarriage, hospital change, did not wish to continue, green stool that was perceived to be linked to study supplement, unable to complete study activities; one participant was lost to follow-up and one withdrawn by the study staff for non-compliance. Baseline characteristics were largely similar between the completed and withdrawn groups, although dietary intake of iron was higher among the participants that withdrew from the study (data not shown). ## 3.4. Supplement Adherence Supplement adherence (Table 3) was measured using hand pill counts and pill bottle dosage monitoring software through Pillsy smart pill bottles. On average, participants took $72\%$ of the LP299V®/placebo provided and $73\%$ of the PNVI provided. Adherence was similar by randomization group. We also examined those who completed the study and observed no differences in adherence by treatment group. A difference was observed, however, for adherence by completed versus withdrawal status for LP299V®/placebo ($$p \leq 0.002$$) and PNVI ($$p \leq 0.003$$), with those remaining in the study being more adherent. ## 3.5. General Adverse Events Adverse events (AEs) were captured using the Maternal Adherence Form beginning with the first pill refill visit through delivery (Table 4). By randomization group, more upper respiratory conditions were observed in the placebo group compared to the LP299V® group ($$p \leq 0.04$$). When stratified by withdrawal versus completed or completed by treatment group, reported AEs were similar. ## 3.6. Adverse Pregnancy Conditions At every research visit following baseline and excluding delivery, the use of antibiotics, iron supplements, the development of GDM, and the receipt of an intravenous iron infusion was monitored and recorded (Table 5). GI symptoms were also recorded using the GI Symptoms Checklist. The overall GI symptoms mean score was 42 (SD 24.2), and by randomization group, 41 (SD 29.2) for the LP299V® group and 44 (SD 16.6) for the placebo group. When comparing participants by withdrawal group, those that completed the study had a higher ($$p \leq 0.04$$) mean GI symptoms score of 48 (SD 24.9), as compared to 24 (SD 7.8) for participants that withdrew. ## 3.7. Maternal Hematological and Iron Status Markers We analyzed participants who completed the study with ≥$80\%$ adherence to examine changes in maternal hematological and iron status parameters by randomization group (Table 6). In this per protocol analysis, mean baseline concentrations of Hb, Hct, SI, TIBC, SF, and TSAT were the lowest and hs-CRP the highest among the LP299V® group, a pattern that continued at all subsequent time points. Regarding the development of IDA, no participants in either treatment group displayed a Hb level indicative of IDA at 24–28 WG. However, at 34–36 WG, one participant in the LP299V® group had IDA, and at delivery, one participant among the LP299V® group and two participants in the placebo group had IDA. For this per protocol analysis, changes observed between treatment groups in absolute mean values and mean change values from baseline for all time points were similar. ## 3.8. Neonatal Cord Hematological and Iron Status Markers In the per protocol analysis, the neonatal hematological and iron status parameters were examined for the newborns of mothers with ≥$80\%$ adherence (Table 7). Gestational age at delivery, neonatal weight at delivery, and infant sex were not statistically different between treatment groups. We observed lower cord Hb, Hct, TIBC, similar SF, and higher SI and TSAT among neonates from the LP299V® group. Cord hs-CRP concentrations were non-detectable. ## 4. Discussion This study was a double-blind placebo-controlled feasibility RCT designed to examine the feasibility and preliminary efficacy of LP299V® supplementation in pregnant individuals at-risk for IDA beginning ≤20 WG through delivery. Feasibility was the primary objective of this work given that, to our knowledge, this is the first study of its kind conducted in the U.S. and focused on pregnant individuals at risk for iron deficiency anemia and their neonates. Recruitment and Enrollment. Primary measures of feasibility were recruitment and enrollment, retention, and adherence to the intervention. Recruitment feasibility was defined as recruiting ≥$50\%$ of those eligible; however, only $15\%$ ($\frac{21}{141}$) of fully eligible individuals were successfully enrolled into the study. Many ($\frac{41}{141}$, $29\%$) of these individuals were unreachable following their initial expression of interest and phone screening. While it is difficult to know why contact ceased, the OB patient population at UI Health largely identifies as low-income and minority individuals, and it is well documented that such circumstances involve barriers to participation in research including transportation, childcare, lack of time from other priorities of greater personal importance than study activities, and general distrust of the medical community [25,26,27]. Recruiting minority pregnant individuals can be especially challenging, as obstacles inherent to pregnancy (e.g., spousal approval and pregnancy-related problems) may be compounded by cultural and economic factors [28,29]. To proactively prepare for these potential issues, our team implemented recommended strategies [28,30], including race/ethnicity matching research staff to the target population when possible, using non-clinical language to explain the study, providing flexible scheduling based around prenatal visits, providing transportation for study visits as needed, offering limited child care with toys and activities for accompanying children, and providing financial incentives with the completion of study activities. We also worked with clinic staff and leadership to optimize recruitment and to integrate study activities into existing clinic processes. Many studies found the incorporation of primary care providers essential to the success of clinic-based perinatal research studies [28,31,32]. While clinic staff were willing to assist with in-facility logistics, our partnership fell short of true clinician buy-in, whereby procedures were not in place for clinical care providers to endorse the study or introduce the study staff member as a member of the health care team. This shortcoming likely contributed to the stunting of our recruitment numbers, as research has shown that people are more receptive to studies that are introduced by their provider [33]. One potential approach to elicit provider support for a future study would be to recruit clinicians interested in research during the development of the study, incorporate their feedback into the study design, and organize recruitment around their patient referrals. Retention. Our goal was to retain ≥$80\%$ of enrolled participants who would then complete ≥$80\%$ of planned assessments; however, only $58\%$ ($\frac{7}{12}$) of participants from the LP299V® group and $75\%$ ($\frac{6}{8}$) from the placebo group were retained and no study visits had assessment completion rates of at least $80\%$ for either group. Participants were diverse from an ethno-racial, education and economic perspective. Our findings showed a higher percentage of less educated and single women among those who were lost to follow-up or withdrawn from the study. This is similar to the $79\%$ retention observed in Project DC-HOPE, a behavioral RCT designed to reduce smoking, depression, and intimate partner violence during pregnancy, which found that retention was lowest among women who were less educated and reported single relationship status [25]. The aforementioned study also recommended financial incentives, well-trained research staff, and consistent contact with participants as essential to longitudinal study retention [25]. All of these strategies were employed within our research protocol. Specific to our participants, we had difficulty scheduling and contacting two individuals with unstable housing, and we also had difficulty keeping engaged individuals with reported high family demands (e.g., caring for a child with disabilities and caring for multiple young children). Future studies might consider a screening period prior to randomization to gauge study engagement and commitment. This might include having the participants take a PNVI for a week to observe their compliance with the intervention. With transportation as another barrier and in response to the COVID-19 pandemic, we made several changes to study procedures: 1, remote study visits; 2, pill distribution via FedEx delivery; 3, pill adherence counts completed with study staff via video conference; 4, blood collection tubes were sent by mail and venous blood samples were collected during routine clinical phlebotomy visits, in place of the clinical research visit. These changes proved more convenient for participants and staff alike and appeared to ease participant study burden. Adherence. The a priori goal for adherence to the intervention was consumption of ≥$80\%$ of the supplements provided. Overall, adherence was $72\%$ for the LP299V®/placebo and $73\%$ for the PNVI. Among completers versus those that were withdrawn, adherence was $85\%$ for the LP299V®/placebo and $86\%$ for the PNVI. Similar adherence has been observed in other probiotic supplementation interventions conducted in pregnancy [34,35]. In a recent longitudinal RCT with 20 healthy Black and White pregnant individuals examining the feasibility of consuming the probiotic Florajen3 for prevention of group B streptococcus in pregnancy, adherence was $86\%$ [36]. The authors used the electronic cap monitoring system MEMS to track subject supplement use, similar to the Pillsy electronic pill bottle system used in this study [36]. Unfortunately, half of the MEMS cap data was unusable, and two caps were not returned. Instead, the investigators had to rely on hand pill counts to calculate adherence, a less desirable method due to the potential for pill dumping before refill visits or unrecorded make-up doses consumed in response to missed doses [36]. Similar to their experience, we also had complications with the Pillsy technology. In our study, Pillsy smart caps were used in combination with pharmacy and participant hand pill counts (shifted to participant counts during the COVID-19 pandemic). While a small subset of participants experienced intermittent connectivity issues with Pillsy technology, the larger challenge was forgetting their assigned Pillsy bottles at pill refill visits. In this case, hand pill counts were used to rectify discrepancies in electronic data. Future studies should select a more portable bottle for participant convenience and provide a more in-depth Pillsy interface training at baseline to help eliminate usability issues. In the Swedish supplementation trial evaluating the effect of LP299V® on iron status in healthy, iron-replete pregnant women, the authors reported $95\%$ adherence among LP299V® and $94\%$ among placebo-treated groups [12]. However, it is unclear if the counts were subject to inconsistencies observed in other studies, as the authors did not report use of a secondary method to corroborate adherence results [12]. A more objective method, such as quantifying an increase in the probiotic strain in a fecal sample, should be considered. This approach was used in a probiotic supplementation study in pregnant Australian women with obesity to reduce the risk of GDM [37]. The measurement of probiotic in stool was used in conjunction with subject self-reporting of intake, and the results showed an adherence of $79\%$ for the fecal samples and $90\%$ for self-report [37]. This highlights the importance of a multi-method monitoring approach to accurately determine intervention adherence. Adverse Events. We hypothesized that AEs would be similar between the treatment arms. However, the placebo group had higher reports of upper respiratory infections (URIs). In pregnancy, hormonal changes often induce hyperemia, excess blood vessels in the sinus and nasal mucosa, and increased nasal cavity secretions, increasing risk of URIs [38]. Consistent with the only other LP299V® supplementation trial in pregnancy [12], we observed no differences between groups for GI symptoms. However, when stratified by those who withdrew and completers, a higher GI symptom score was observed among study completers ($$p \leq 0.04$$). This was likely an artifact of increased GI discomfort associated with advanced gestation, given that completers reached full-term during the study as compared to participants that withdrew from the study earlier in the gestational period. Maternal Iron Status Biomarkers. Among participants included in the per protocol analysis, a slower decline in hematological and iron parameters across pregnancy was observed in the LP299V® group compared to placebo. Although not powered to reliably detect small significant differences or associations, a positive pattern was observed in mean Hb changes from baseline across all study time points; baseline Hb levels were lower among the LP299V® group but increased by delivery as compared to a decrease over time observed among the placebo group. Within the LP299V® group, one participant presented with IDA at 34–36 WG and at delivery, while among the placebo group, there were no cases of IDA at 34–36 WG, and two participants presented with IDA at delivery. A similar observation was reported in the Axling study, where the prevalence of IDA was significantly lower among the LP299V® group compared to placebo at 35 WG [12]. An important distinction to note is that the Axling study supplemented the gravida twice daily with LP299V® and 12 mg ascorbic acid, 4.2 mg of iron and 30 µg folic acid [12]. Providing smaller boluses of vitamins and minerals provides a dietary absorption advantage over large one-time boluses. It is possible that supplementation of LP299V® twice daily in conjunction with smaller amounts of iron and ascorbic acid (which facilitates iron absorption) could have potentially contributed to the positive findings observed among the pregnant women in the Axling trial [12]. Together, this suggests that regular consumption of LP299V® may have positive effects on maternal iron nutrition and occurrence of IDA among those with iron-sufficiency or among those with subclinical ID early in pregnancy. However, a larger efficacy trial is required to confirm the positive pattern observed among the high-risk group targeted in our small feasibility study. Inflammation plays a pivotal role in iron metabolism, as its presence has the capacity to downregulate dietary uptake and initiate iron sequestration [39,40]. In the present study and the Axling trial, systemic inflammatory levels were captured using the acute-phase protein hs-CRP [12]. Participants in our study had higher hs-CRP levels compared to those in the Axling trial, likely due to the differences in BMI and participant sociodemographic characteristics between the two populations. The mean BMI in the Axling study was within normal range, while the participants in the current study had a high prevalence of obesity [12]. Higher levels of adiposity have been correlated with increased levels of inflammation and decreased iron bioavailability [41]. Moreover, our cohort was largely low-income Black women, who have been shown to have higher chronic psychosocial stressors in pregnancy compared with White women [42]. One proposed mechanism for this stress-related disparity is centered around the conceptualization of minority status as a chronic stressor [43]. Stress is directly correlated with inflammation [44,45] and likely more pervasive among our largely obese, minority population, as compared to that of the homogenous Swedish population. Although LP299V® has been shown to have anti-inflammatory effects on the immune system [46], the possibility remains that elevated inflammatory levels can potentially complicate increased dietary iron absorption, presenting a greater challenge to the evaluation of LP299V® efficacy in populations that are more inflamed. Given the beneficial effects of probiotic use in different clinical settings and patient populations with shared inflammatory and stress-related mechanisms, further exploration of the probiotic LP299V® is warranted [47,48,49]. Neonatal Iron Status Biomarkers. Although not statistically significant, we observed higher SI and TSAT levels among neonates from the LP299V® group compared to placebo. We observed similar SF concentrations between groups; however, all other biomarkers of iron (Hb, Hct, and TIBC) were higher among the placebo group. In pregnancy, neonatal iron needs increase with gestational age to establish fetal iron stores that are essential to ex utero neurodevelopment in the early months of life [3]. Fetal iron endowment relies exclusively on maternal iron transfer [50]. Previous research suggests that maternal-fetal iron trafficking is largely dictated by maternal signaling [51]. Moreover, a recent study among murine and in vitro human models indicates the placenta may also play a vital role in maternal-fetal iron transfer, responding to changes in maternal iron status [52]. These findings suggest that maintaining sufficient maternal iron status has important effects on placental functions and ultimately adequate iron transfer to the developing fetus [52]. As such, the results from our study are promising, given that the LP299V® group began the study more iron-deficient than the placebo group yet improved in several hematological parameters over the course of pregnancy and delivered neonates with higher SI and TSAT levels than those in the placebo group. Strengths and Limitations. This study has several strengths. It is the first to assess the preliminary efficacy of LP299V® in pregnant individuals with subclinical ID, a group at high risk for IDA, and among a racially, ethnically, socioeconomically diverse, U.S.-based population. It is also the first study to extend the evaluation of LP299V® supplementation during pregnancy into neonatal hematological parameters and iron status at delivery. The study successfully followed participants from ≤20 WG through delivery with four data collection time points: baseline, 24–28 WG, 34–36 WG, and delivery. In addition, the collection of multiple iron and hematological parameters also helped to expand the examination of the preliminary effects of LP299V® on iron nutrition in pregnancy. Lastly, the use of two-factor adherence monitoring using both the Pillsy smart bottle technology and pharmacy pill counts is a strength in regard to the accuracy of adherence measurement and subsequent reliability of the results. There are several limitations of this study. This was a feasibility study in which the sample was too small to balance characteristics using randomization and to use multivariate methods to address imbalance. Additionally, due to the small sample size, this study was not powered to detect statistically significant changes in hematological and iron status parameters and limited our ability to conclusively determine study acceptability. Data collection challenges and protocol adaptions adopted in reaction to COVID-19 resulted in missing data. There were differences in Hb levels by group at baseline, a characteristic that we recommend be included in participant stratification for future projects. Although we excluded individuals using antibiotics in the past two months prior to enrollment, we did not withdraw participants if they started using antibiotics during the study, which could have affected the viability of LP299V®. Dietary data collected via self-reported 24-h recall is inherently subject to recall bias. Dose timing between meals should be considered to optimize iron absorption, and consumption with milk and other dairy products should be prohibited to limit chelation. Multi-day doses with smaller boluses of supplemental iron should be provided in place of one large daily bolus of iron from a PNVI. Additionally, key hematological markers of iron metabolism including hepcidin, sTfR, ERFE and EPO were not measured. Inclusion of sTfR would have provided unvarying insight into changes in plasma iron availability, as levels do not fluctuate in response to inflammation or pregnancy. Moreover, maternal-placental-fetal iron trafficking was not objectively measured and would have provided results on the direct effects of LP299V® on this critical system. ## 5. Conclusions The results of this study suggest the need for continued adjustments in the methods of recruitment and retention for probiotic supplementation among an urban U.S.-based health care setting. Once individuals were engaged in the research, there was strong adherence to the intervention and relatively few adverse events, indicating LP299V® as a low cost and tolerable therapy during pregnancy. Preliminary findings suggest LP299V® has the potential to affect several maternal and neonatal hematological and iron related parameters, and these findings should be further explored. ## References 1. Gupta P.M., Hamner H.C., Suchdev P.S., Flores-Ayala R., Mei Z.. Iron status of toddlers, nonpregnant females, and pregnant females in the United States. Am. J. Clin. Nutr. (2017) 106 1640S-1646S. DOI: 10.3945/ajcn.117.155978 2. Adebisi O.Y., Strayhorn G.. Anemia in pregnancy and race in the United States: Blacks at risk. Fam. Med. (2005) 37 655-662. PMID: 16193419 3. Bothwell T.H.. Iron requirements in pregnancy and strategies to meet them. Am. J. Clin. Nutr. (2000) 72 257S-264S. DOI: 10.1093/ajcn/72.1.257S 4. Lozoff B., Beard J., Connor J., Felt B., Georgieff M., Schallert T.. Long-Lasting Neural and Behavioral Effects of Iron Deficiency in Infancy. Nutr. Rev. (2006) 64 34-43. DOI: 10.1301/nr.2006.may.S34-S43 5. Allen L.H.. Anemia and iron deficiency: Effects on pregnancy outcome. Am. J. Clin. Nutr. (2000) 71 1280S-1284S. DOI: 10.1093/ajcn/71.5.1280s 6. 6. Institute of Medicine Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and ZincNational Academies PressWashington, DC, USA2001. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc (2001) 7. Cogswell M.E., Parvanta I., Ickes L., Yip R., Brittenham G.M.. Iron supplementation during pregnancy, anemia, and birth weight: A randomized controlled trial. Am. J. Clin. Nutr. (2003) 78 773-781. DOI: 10.1093/ajcn/78.4.773 8. Hoppe M., Önning G., Hulthén L.. Freeze-dried Lactobacillus plantarum 299v increases iron absorption in young females—Double isotope sequential single-blind studies in menstruating women. PLoS ONE (2017) 12. DOI: 10.1371/journal.pone.0189141 9. Hoppe M., Önning G., Berggren A., Hulthén L.. Probiotic strain Lactobacillus plantarum 299v increases iron absorption from an iron-supplemented fruit drink: A double-isotope cross-over single-blind study in women of reproductive age. Br. J. Nutr. (2015) 114 1195-1202. DOI: 10.1017/S000711451500241X 10. Bering S., Suchdev S., Sjoltav L., Berggren A., Tetens I., Bukhave K.. A lactic acid-fermented oat gruel increases non-haem iron absorption from a phytate-rich meal in healthy women of childbearing age. Br. J. Nutr. (2006) 96 80-85. DOI: 10.1079/BJN20061683 11. Scheers N., Rossander-Hulthen L., Torsdottir I., Sandberg A.S.. Increased iron bioavailability from lactic-fermented vegetables is likely an effect of promoting the formation of ferric iron (Fe. Eur. J. Nutr. (2016) 55 373-382. DOI: 10.1007/s00394-015-0857-6 12. Axling U., Önning G., Combs M.A., Bogale A., Högström M., Svensson M.. The Effect of Lactobacillus plantarum 299v on Iron Status and Physical Performance in Female Iron-Deficient Athletes: A Randomized Controlled Trial. Nutrients (2020) 12. DOI: 10.3390/nu12051279 13. Rennie D.. CONSORT Revised—Improving the Reporting of Randomized Trials. JAMA (2001) 285 2006-2007. DOI: 10.1001/jama.285.15.2006 14. Hamm R.F., Wang E.Y., Levine L.D., Srinivas S.K.. Association Between Race and Hemoglobin at Delivery or Need for Transfusion When Using Race-Based Definitions for Treatment of Antepartum Anemia. Obstet. Gynecol. (2021) 138 108. DOI: 10.1097/AOG.0000000000004439 15. 15. Committee on the Prevention and Management of Iron Deficiency Anemia among U.S. Children and Women of Childbearing Age Iron Deficiency Anemia: Recommended Guidelines for the Prevention, Detection, and Management among U.S. Children and Women of Childbearing AgeNational Academies PressWashington, DC, USA1994. Iron Deficiency Anemia: Recommended Guidelines for the Prevention, Detection, and Management among U.S. Children and Women of Childbearing Age (1994) 16. Abbassi-Ghanavati M., Greer L.G., Cunningham F.G.. Pregnancy and laboratory studies: A reference table for clinicians. Obstet. Gynecol. (2009) 114 1326-1331. DOI: 10.1097/AOG.0b013e3181c2bde8 17. Assessing the Iron Status of Populations: Including Literature Reviews. World Health Organization (2007) 18. Næss-Andresen M.-L., Eggemoen Å.R., Berg J.P., Falk R.S., Jenum A.K.. Serum ferritin, soluble transferrin receptor, and total body iron for the detection of iron deficiency in early pregnancy: A multiethnic population-based study with low use of iron supplements. Am. J. Clin. Nutr. (2019) 109 566-575. DOI: 10.1093/ajcn/nqy366 19. Dennis B., Ernst N., Hjortland M., Tillotson J., Grambsch V.. The NHLBI nutrition data system. J. Am. Diet. Assoc. (1980) 77 641. DOI: 10.1016/S1094-7159(21)03585-6 20. Harnack L., Stevens M., Van Heel N., Schakel S., Dwyer J.T., Himes J.. A computer-based approach for assessing dietary supplement use in conjunction with dietary recalls. J. Food Compos. Anal. (2008) 21 S78-S82. DOI: 10.1016/j.jfca.2007.05.004 21. Phillips A.K., Roy S.C., Lundberg R., Guilbert T.W., Auger A.P., Blohowiak S.E., Coe C.L., Kling P.J.. Neonatal iron status is impaired by maternal obesity and excessive weight gain during pregnancy. J. Perinatol. (2014) 34 513-518. DOI: 10.1038/jp.2014.42 22. Viteri F.E.. Iron endowment at birth: Maternal iron status and other influences. Nutr. Rev. (2011) 69 S3-S16. DOI: 10.1111/j.1753-4887.2011.00449.x 23. Newcombe R.G.. Two-sided confidence intervals for the single proportion: Comparison of seven methods. Stat. Med. (1998) 17 857-872. DOI: 10.1002/(SICI)1097-0258(19980430)17:8<857::AID-SIM777>3.0.CO;2-E 24. Fleiss J.L.. Statistical Methods for Rates and Proportions (2003) 25. El-Khorazaty M.N., Johnson A.A., Kiely M., El-Mohandes A.A., Subramanian S., Laryea H.A., Murray K.B., Thornberry J.S., Joseph J.G.. Recruitment and retention of low-income minority women in a behavioral intervention to reduce smoking, depression, and intimate partner violence during pregnancy. BMC Public Health (2007) 7. DOI: 10.1186/1471-2458-7-233 26. Shavers-Hornaday V.L., Lynch C.F., Burmeister L.F., Torner J.C.. Why are African Americans under-represented in medical research studies? Impediments to participation. Ethn. Health (1997) 2 31-45. DOI: 10.1080/13557858.1997.9961813 27. Shavers V.L., Lynch C.F., Burmeister L.F.. Racial differences in factors that influence the willingness to participate in medical research studies. Ann. Epidemiol. (2002) 12 248-256. DOI: 10.1016/S1047-2797(01)00265-4 28. Goff S.L., Youssef Y., Pekow P.S., White K.O., Guhn-Knight H., Lagu T., Mazor K.M., Lindenauer P.K.. Successful strategies for practice-based recruitment of racial and ethnic minority pregnant women in a randomized controlled trial: The IDEAS for a healthy baby study. J. Racial Ethn. Health Disparities (2016) 3 731-737. DOI: 10.1007/s40615-015-0192-x 29. Gamble A., Beech B.M., Blackshear C., Cranston K.L., Herring S.J., Moore J.B., Welsch M.A.. Recruitment planning for clinical trials with a vulnerable perinatal adolescent population using the Clinical Trials Transformative Initiative framework and principles of partner and community engagement. Contemp. Clin. Trials (2021) 104 106363. DOI: 10.1016/j.cct.2021.106363 30. Coleman-Phox K., Laraia B.A., Adler N., Vieten C., Thomas M., Epel E.. Recruitment and retention of pregnant women for a behavioral intervention: Lessons from the maternal adiposity, metabolism, and stress (MAMAS) study. Prev. Chronic Dis. (2013) 10 1-7. DOI: 10.5888/pcd10.120096 31. Frew P.M., Saint-Victor D.S., Isaacs M.B., Kim S., Swamy G.K., Sheffield J.S., Edwards K.M., Villafana T., Kamagate O., Ault K.. Recruitment and retention of pregnant women into clinical research trials: An overview of challenges, facilitators, and best practices. Clin. Infect. Dis. (2014) 59 S400-S407. DOI: 10.1093/cid/ciu726 32. Kerver J.M., Elliott M.R., Norman G.S., Sokol R.J., Keating D.P., Copeland G.E., Johnson C.C., Cislo K.K., Alcser K.H., Kruger-Ndiaye S.R.. Pregnancy recruitment for population research: The national children’s study vanguard experience in Wayne County, Michigan. Paediatr. Perinat. Epidemiol. (2013) 27 303-311. DOI: 10.1111/ppe.12047 33. Newington L., Metcalfe A.. Factors influencing recruitment to research: Qualitative study of the experiences and perceptions of research teams. BMC Med. Res. Methodol. (2014) 14. DOI: 10.1186/1471-2288-14-10 34. Davidson S.J., Barrett H.L., Price S.A., Callaway L.K., Dekker Nitert M.. Probiotics for preventing gestational diabetes. Cochrane Database Syst. Rev. (2021) 2021 CD009951. DOI: 10.1002/14651858.CD009951.pub3 35. Lindsay K.L., Brennan L., McAuliffe F.M.. Acceptability of and compliance with a probiotic capsule intervention in pregnancy. Int. J. Gynecol. Obstet. (2014) 125 279-280. DOI: 10.1016/j.ijgo.2014.01.004 36. Hanson L., Vandevusse L., Duster M., Warrack S., Safdar N.. Feasibility of oral prenatal probiotics against maternal group B streptococcus vaginal and rectal colonization. JOGNN—J. Obstet. Gynecol. Neonatal Nurs. (2014) 43 294-304. DOI: 10.1111/1552-6909.12308 37. Callaway L.K., McIntyre H.D., Barrett H.L., Foxcroft K., Tremellen A., Lingwood B.E., Tobin J.M., Wilkinson S., Kothari A., Morrison M.. Probiotics for the prevention of gestational diabetes mellitus in overweight and obese women: Findings from the SPRING double-blind randomized controlled trial. Diabetes Care (2019) 42 364-371. DOI: 10.2337/dc18-2248 38. Rosenberger K.D., Cibulka N.J., Barron M.L.. Guidelines for Nurse Practitioners in Ambulatory Obstetric Settings (2022). DOI: 10.1891/9780826148544 39. Ganz T., Nemeth E.. Iron homeostasis in host defence and inflammation. Nat. Rev. Immunol. (2015) 15 500-510. DOI: 10.1038/nri3863 40. Papanikolaou G., Pantopoulos K.. Systemic iron homeostasis and erythropoiesis. IUBMB Life (2017) 69 399-413. DOI: 10.1002/iub.1629 41. Flores-Quijano M.E., Vega-Sánchez R., Tolentino-Dolores M.C., López-Alarcón M.G., Flores-Urrutia M.C., López-Olvera A.D., Talavera J.O.. Obesity Is Associated with Changes in Iron Nutrition Status and Its Homeostatic Regulation in Pregnancy. Nutrients (2019) 11. DOI: 10.3390/nu11030693 42. Gillespie S.L., Porter K., Christian L.M.. Adaptation of the inflammatory immune response across pregnancy and postpartum in Black and White women. J. Reprod. Immunol. (2016) 114 27-31. DOI: 10.1016/j.jri.2016.02.001 43. Christian L.M., Glaser R., Porter K., Iams J.D.. Stress-induced inflammatory responses in women: Effects of race and pregnancy. Psychosom. Med. (2013) 75 658-669. DOI: 10.1097/PSY.0b013e31829bbc89 44. Slavich G.M., Irwin M.R.. From stress to inflammation and major depressive disorder: A social signal transduction theory of depression. Psychol. Bull. (2014) 140 774-815. DOI: 10.1037/a0035302 45. Redpath N., Rackers H.S., Kimmel M.C.. The Relationship Between Perinatal Mental Health and Stress: A Review of the Microbiome. Curr. Psychiatry Rep. (2019) 21 1-9. DOI: 10.1007/s11920-019-0998-z 46. Wells J.. Immunomodulatory mechanisms of lactobacilli. Microb. Cell Factories (2011) 10 S17. DOI: 10.1186/1475-2859-10-S1-S17 47. Castelli V., D’Angelo M., Quintiliani M., Benedetti E., Cifone M.G., Cimini A.. The emerging role of probiotics in neurodegenerative diseases: New hope for Parkinson’s disease?. Neural Regen. Res. (2021) 16 628-634. DOI: 10.4103/1673-5374.295270 48. Wu H., Chiou J.. Potential benefits of probiotics and prebiotics for coronary heart disease and stroke. Nutrients (2021) 13. DOI: 10.3390/nu13082878 49. Vivarelli S., Falzone L., Basile M., Nicolosi D., Genovese C., Libra M., Salmeri M.. Benefits of using probiotics as adjuvants in anticancer therapy (Review). World Acad. Sci. J. (2019) 1 125-135. DOI: 10.3892/wasj.2019.13 50. Cao C., Fleming M.D.. The placenta: The forgotten essential organ of iron transport. Nutr. Rev. (2016) 74 421-431. DOI: 10.1093/nutrit/nuw009 51. Dao M.C., Sen S., Iyer C., Klebenov D., Meydani S.N.. Obesity during pregnancy and fetal iron status: Is Hepcidin the link?. J. Perinatol. (2013) 33 177-181. DOI: 10.1038/jp.2012.81 52. Sangkhae V., Fisher A.L., Wong S., Koenig M.D., Tussing-Humphreys L., Chu A., Lelić M., Ganz T., Nemeth E.. Effects of maternal iron status on placental and fetal iron homeostasis. J. Clin. Investig. (2020) 130 625-640. DOI: 10.1172/JCI127341
--- title: 'Genetic Variations in Angiotensinogen Gene and Risk of Preeclampsia: A Pilot Study' authors: - Dong He - Xianglan Peng - Hongkai Xie - Rui Peng - Qixuan Li - Yitong Guo - Wei Wang - Hong He - Yang Chen journal: Journal of Clinical Medicine year: 2023 pmcid: PMC9966751 doi: 10.3390/jcm12041509 license: CC BY 4.0 --- # Genetic Variations in Angiotensinogen Gene and Risk of Preeclampsia: A Pilot Study ## Abstract Preeclampsia (PE) is a typical hypertensive disorders of pregnancy (HDP) which can cause substantial morbidity and mortality in both pregnant women and fetuses. The renin-angiotensin system (RAS) genes are the main HDP-causing genes, and Angiotensinogen (AGT) as the initial substrate can directly reflect the activity of the entire RAS. However, the association between AGT SNPs and PE risk has rarely been confirmed. This study was carried out to determine whether AGT SNPs could affect the risk of PE in 228 cases and 358 controls. The genotyping result revealed that the AGT rs7079 TT carrier was related to increased PE risk. Further stratified analysis illustrated that the rs7079 TT genotype significantly increased the PE risk in subgroups of Age < 35, BMI < 25, Albumin (ALB) ≥ 30 and Aspartate aminotransferase (AST) < 30. These findings demonstrated that the rs7079 might be a promising candidate SNP strongly associated with PE susceptibility. ## 1. Introduction Hypertensive disorders of pregnancy (HDP) have long been major causes of morbidity and mortality in pregnant women, and the incidence of HDP keeps rising worldwide [1,2]. These pregnancy disorders include gestational hypertension, preeclampsia (PE), and eclampsia, which are characterized by elevated blood pressure and multiple organ disturbances, ranging from mild to severe [3,4]. The PE complicates 3–$5\%$ of all pregnancies and is estimated to result in a large number of maternal and fetal deaths globally every year [5,6]. Although the majority of PE usually resolve after delivery or in the early postpartum period, there is increasing evidence that PE confers a noteworthy increase in risks for future long-term health [7,8]. It is irrefutable that a history of PE increases the risk of hypertension, peripheral arterial disease, coronary artery disease and cerebrovascular disease in the future [9,10]. Furthermore, the severity of PE has been shown to accelerate the progression of cardiovascular disease [11]. To date, there is no stable and reliable treatment for PE. Therefore, elucidating the pathogenesis of PE is particularly critical for the diagnosis and treatment of this disease. With the rapid development of molecular biology technology, the genetic basis of HDP has been initially revealed [12,13]. Previous studies have shown that renalase (RNLS) gene polymorphisms were associated with many diseases, such as essential hypertension, PE [14] and gestational diabetes [15]. A recent study displayed that the interleukin 1 receptor type 1 (IL1R1) rs2071374G variant could lead to an increased risk of PE [16]. The egl-9 family hypoxia inducible factor 1 (EGLN1) rs479200 may have the potential to become a marker to evaluate the genetic predisposition to PE [17]. However, the identified genetic mutation or SNPs can only explain the etiology of PE in a small sample of cases. Therefore, the identification and characterization of more functional mutation sites are critical to fully reveal the pathologic mechanism of PE. The renin-angiotensin system (RAS) plays a crucial role in maintaining the homeostasis of cardiovascular function, especially the regulation of arterial blood pressure [18,19]. Its dysfunction can lead to a series of cardiovascular diseases [20,21]. Angiotensinogen (AGT), as the initial substrate in the RAS, directly reflects the activity of the entire RAS [22]. So far, the investigation that directly evaluated the effect of AGT SNP on PE is relatively limited. Therefore, the current study aimed to assess the association between AGT gene SNPs and PE risk among Chinese pregnant women. ## 2. Materials and Methods All patients and controls in this case-control study were from the Department of Obstetrics and Gynecology, The Third Affiliated Hospital of Guangzhou Medical University, during the period from January 2016 to December 2021. 228 PE patients constituted the case group, and 358 normotensive pregnant women constituted the control group. All PE patients were diagnosed according to the diagnostic criteria of the American College of Obstetrics and Gynecology published in 2013 (systolic and/or diastolic blood pressure ≥ $\frac{140}{90}$ mmHg after 20th week of pregnancy, plus proteinuria ≥ 300 mg per 24-h urine collection or ≥ 1+ urine dipstick; in the absence of proteinuria, hypertension in pregnancy with any of the following features: pulmnary edema, platelets count ≤ 10 × 1010/L, impaired liver function, systolic and/or diastolic blood pressure ≥ $\frac{160}{110}$ mmHg, renal failure and visual disturbances) [23]. The samples of the control group and the case group were non-probability continuous and random. Normotensive pregnant women without a history of chronic hypertension and complication during pregnancy were defined as the control population. Clinical characteristics and biochemical indicators of all participants were collected for stratified analysis. The three SNPs to be verified were selected according to the following criteria: The selected SNPs were preferentially located in 3′ and 5′ untranslated regions and exons of AGT gene to maximize the probability that these SNPs were functional; The minor allele frequencies of these selected SNPs should be $5\%$ or greater. Furthermore, any two SNPs should have low linkage disequilibrium (R2 < 0.8). DNA of each subject for genotyping was extracted from a peripheral venous blood sample (200 µL) according to the protocol of the TIANamp Genomic DNA Kit (Tiangen Biotech, Beijing, China). The Genotyping, performed according to the instructions (Applied Biosystems, Waltham, MA, USA), was carried out in a total volume of 10 µL containing the DNA template (1 µL, 2 ng/µL), TaqMan® SNP Genotyping Assay (0.06 µL, 40X), TaqPath ProAm Master Mix (3 µL) and DEPC H2O (add to 10 µL). The Genotyping conditions were as follows: 60 °C for 30 s, 95 °C for 5 min, followed by 40 cycles of 95 °C for 15 s and 60 °C for 1 min. The Genotyping was operated with optical 96 or 384-well plate in an ABI PRISM 7500 PCR instrument (Applied Biosystems). In addition, $5\%$ of the samples were randomly re-tested to make sure that these re-tested samples attain $100\%$ recurrence rate. The student t-test for continuous variables was employed to evaluate clinical variables differences in the case and control groups. The χ2 test served to know whether these three AGT SNPs in the controls were in Hardy-Weinberg equilibrium (HWE). Crude ratios with respective $95\%$ confidence intervals (CIs) and logistic regression analysis were used to assess the association between HDP risk and AGT SNPs. SPSS software was used to analyze all data in this plot study. The p-value less than 0.05 indicated that the statistical result was defined as significant. ## 3.1. Clinical Characteristics of Cases and Controls The clinical and demographic characteristics of all participants were shown in Table 1. The statistical results indicated that the body mass index (BMI), maternal age, systolic blood pressure (SBP), diastolic blood pressure (DBP), Aspartate aminotransferase (AST), Alanine aminotransferase (ALT), creatinine (CREA) and uricacid (UA) in the case group were significantly higher than that of the control group (all $p \leq 0.01$). On the other hand, gestational age, fetal birth weight, Albumin (ALB) and platelet count (PLT) level were significantly lower in the preeclamptic patients compared to the control subjects (all $p \leq 0.01$). ## 3.2. Effect of AGT Gene SNPs on PE The genotype distributions for three selected AGT SNPs showed that there was a significant association between rs7079 TT genotype and PE risk (OR = 3.804, $95\%$ CI = 1.100–13.156, $$p \leq 0.035$$) (Table 2). These results indicated that the rs7079 TT carrier shared a significantly increased risk of PE. Moreover, the AGT rs7079 TT genotype was associated with increased preeclampsia risk in the recessive model (OR = 4.054, $95\%$ CI = 1.178–13.945, $$p \leq 0.026$$) (Table 2). However, we did not detect significant differences in SNP rs4762 and rs5050 in any analysis model. At last, the genotype distribution frequencies of three AGT SNPs of the investigated samples were in accordance with the Hardy-Weinberg equilibrium (HWE) in the control group (Table 3). ## 3.3. Stratification Analysis In order to further clarify the role of the three SNPs in different clinical subgroups, a stratified analysis based on age, BMI and clinical parameters was performed. As presented in Table 4, the rs7079 TT genotype carriers shared significantly increased PE risk in subgroups of Age < 35 (OR = 10.988, $95\%$ CI = 2.342–51.555, $$p \leq 0.001$$), BMI < 25 (OR = 5.153, $95\%$ CI = 1.314–20.212, $$p \leq 0.024$$), ALB ≥ 30 (OR = 5.029, $95\%$ CI = 1.392–18.167, $$p \leq 0.019$$) and AST < 30 (OR = 5.088, $95\%$ CI = 1.573–16.456, $$p \leq 0.007$$). ## 3.4. The Relevance of rs7079 G>T to AGT Expression To determine whether rs7079 could lead to change of AGT gene expression, the GTEx database was used to verify the correlation between rs7079 G>T and AGT mRNA expression. The results showed that the rs7079 T allele was significantly related to the decrease of AGT expression level in cultured fibroblasts ($$p \leq 0.000099$$) (Figure 1). ## 4. Discussion To date, the AGT gene has been confirmed to be closely related to the occurrence and development of various types of diseases, including cardiovascular disease [24] and colorectal cancer [25]. A recent experimental study pointed out that loss of AGT protein in specific cells can improve high-fat diet-induced insulin tolerance [26]. And Yilmaz et al. reported that urinary AGT levels in 35-week gestational women with PE were significantly elevated compared to normal pregnancies and non-pregnant women [27]. Moreover, consistent evidence found by clinical research showed that plasma-derived oxidized AGT in pregnant women with PE retained a dominant level compared to normotensive controls [28]. These above results indicated the important biological function of AGT in the process of diseases occurrence and development. Contributions of AGT SNPs to cardiovascular diseases have been widely accepted in recent years [29,30]. In patients with peripheral arterial disease, the association between AGT rs699 CC genotype and high-density lipoprotein (HDL) levels was proved to be significant [31]. The non-alcoholic fatty liver disease (NAFLD) patients with the AGT rs5051 TC + CC genotype had a significantly increased risk of coronary heart disease (CHD) in the northern Chinese Han population [32]. Nonetheless, the association between these three AGT SNPs and PE has not been well established. Through genotyping and stratified analysis of 228 cases and 358 controls, the present study successfully clarified the association between AGT SNPs and PE risk. The rs4762 accompanied by the mutation from threonine to methionine at residence 174 is a typical missense mutation of AGT gene [33]. Previous study confirmed that the rs4762 showed a significant risk for the diabetes mellitus in transplant patients [34]. Another clinical study in Egyptians pointed out the rs4762 variant may increased the risk for end-stage renal failure risk [35]. Compared with rs4762, the AGT rs5050 leads to the transformation from adenine to cytosine in the promoter region of gene. The AGT rs5050 GG carriers with astrocytoma were more likely to have poor prognosis [36]. In children with Kawasaki Disease, rs5050 T>G was associated with the risk of coronary artery aneurysm [37]. However, our findings indicated that there was no statistically significant association between these two SNPs (rs4762 and rs050) and PE susceptibility (Table 2). We speculated that the two functional SNP sites rs4762 and rs050 would not affect the dynamic balance of RAS, so that the susceptibility of PE did not change. The AGT rs7079 located in the 3′UTR is the direct binding site of miR-31 and miR-584 regulating AGT gene expression [38,39]. Moreover, the genotyping results in this study showed that the rs7079 TT genotype was related to increased PE risk (Table 2). The stratified analysis further indicated that the rs7079 TT genotype carriers shared significantly increased PE risk in subgroups of Age < 35, BMI < 25, ALB ≥ 30 and AST < 30 (Table 4). Moreover, we found that the rs7079 T allele resulted in the decrease of AGT expression level in the GTEx portal (Figure 1). Based on the above findings, we speculated that the rs7079 T allele may increase PE risk by disturbing the dynamic equilibrium of the RAS. There are some deficiencies as follows that could be improved in future research. First, this study only verified the function of three AGT SNPs. More AGT SNPs should be tested to comprehensively evaluate the association between AGT SNPs and PE risk. Second, in consideration of this study’s relatively limited sample size, larger sample sizes are needed to confirm the impact of AGT SNPs on PE risk. Third, the non-genetic risk factors, including environmental factors, life style, and health care, are also considered as the main nosogenesis of PE [40,41]. Therefore, the non-genetic risk factors merit further research in combination with the analysis between the PE risk and SNPs. Finally, the genetic diversity among different ethnic groups also affects the susceptibility to PE. In the subsequent evaluation of the association between AGT gene polymorphisms and susceptibility to PE, it will be a great challenge to exclude functional genetic diversity that may potentially affect the risk of PE. ## References 1. ACOG Practice Bulletin No. 202: Gestational Hypertension and Preeclampsia. Obstet Gynecol. (2019) 133 1. DOI: 10.1097/aog.0000000000003018 2. Garovic V.D., Dechend R., Easterling T., Karumanchi S.A., McMurtry Baird S., Magee L.A., Rana S., Vermunt J.V., August P.. Hypertension in Pregnancy: Diagnosis, Blood Pressure Goals, and Pharmacotherapy: A Scientific Statement from the American Heart Association. Hypertension (2022) 79 e21-e41. DOI: 10.1161/HYP.0000000000000208 3. Brown M.A., Magee L.A., Kenny L.C., Karumanchi S.A., McCarthy F.P., Saito S., Hall D.R., Warren C.E., Adoyi G., Ishaku S.. The hypertensive disorders of pregnancy: ISSHP classification, diagnosis & management recommendations for international practice. Pregnancy Hypertens. (2018) 13 291-310. DOI: 10.1016/j.preghy.2018.05.004 4. Nobles C.J., Mendola P., Mumford S.L., Silver R.M., Kim K., Andriessen V.C., Connell M., Sjaarda L., Perkins N.J., Schisterman E.F.. Preconception Blood Pressure and Its Change into Early Pregnancy: Early Risk Factors for Preeclampsia and Gestational Hypertension. Hypertension (2020) 76 922-929. DOI: 10.1161/HYPERTENSIONAHA.120.14875 5. Chappell L.C., Cluver C.A., Kingdom J., Tong S.. Pre-eclampsia. Lancet (2021) 398 341-354. DOI: 10.1016/S0140-6736(20)32335-7 6. Abalos E., Cuesta C., Grosso A.L., Chou D., Say L.. Global and regional estimates of preeclampsia and eclampsia: A systematic review. Eur. J. Obstet. Gynecol. Reprod. Biol. (2013) 170 1-7. DOI: 10.1016/j.ejogrb.2013.05.005 7. Ambrožič J., Lučovnik M., Cvijić M.. Evolution of cardiac geometry and function in women with severe preeclampsia from immediately post-delivery to 1 year postpartum. Int. J. Cardiovasc. Imaging (2021) 37 2217-2225. DOI: 10.1007/s10554-021-02210-6 8. Khosla K., Heimberger S., Nieman K.M., Tung A., Shahul S., Staff A.C., Rana S.. Long-Term Cardiovascular Disease Risk in Women After Hypertensive Disorders of Pregnancy: Recent Advances in Hypertension. Hypertension (2021) 78 927-935. DOI: 10.1161/HYPERTENSIONAHA.121.16506 9. Brouwers L., van der Meiden-van Roest A.J., Savelkoul C., Vogelvang T.E., Lely A.T., Franx A., van Rijn B.B.. Recurrence of pre-eclampsia and the risk of future hypertension and cardiovascular disease: A systematic review and meta-analysis. Br. J. Obstet. Gynaecol. (2018) 125 1642-1654. DOI: 10.1111/1471-0528.15394 10. Berends A.L., de Groot C.J., Sijbrands E.J., Sie M.P., Benneheij S.H., Pal R., Heydanus R., Oostra B.A., van Duijn C.M., Steegers E.A.. Shared constitutional risks for maternal vascular-related pregnancy complications and future cardiovascular disease. Hypertension (2008) 51 1034-1041. DOI: 10.1161/HYPERTENSIONAHA.107.101873 11. Craici I., Wagner S., Garovic V.D.. Preeclampsia and future cardiovascular risk: Formal risk factor or failed stress test?. Ther. Adv. Cardiovasc. Dis. (2008) 2 249-259. DOI: 10.1177/1753944708094227 12. Gray K.J., Saxena R., Karumanchi S.A.. Genetic predisposition to preeclampsia is conferred by fetal DNA variants near FLT1, a gene involved in the regulation of angiogenesis. Am. J. Obstet. Gynecol. (2018) 218 211-218. DOI: 10.1016/j.ajog.2017.11.562 13. Ardissino M., Slob E.A.W., Millar O., Reddy R.K., Lazzari L., Patel K.H.K., Ryan D., Johnson M.R., Gill D., Ng F.S.. Maternal Hypertension Increases Risk of Preeclampsia and Low Fetal Birthweight: Genetic Evidence from a Mendelian Randomization Study. Hypertension (2022) 79 588-598. DOI: 10.1161/HYPERTENSIONAHA.121.18617 14. Bagci B., Karakus S., Bagci G., Sancakdar E.. Renalase gene polymorphism is associated with increased blood pressure in preeclampsia. Pregnancy Hypertens. (2016) 6 115-120. DOI: 10.1016/j.preghy.2016.04.002 15. Fatima S.S., Jamil Z., Alam F., Malik H.Z., Madhani S.I., Ahmad M.S., Shabbir T., Rehmani M.N., Rabbani A.. Polymorphism of the renalase gene in gestational diabetes mellitus. Endocrine (2017) 55 124-129. DOI: 10.1007/s12020-016-1058-7 16. Sivaraj N., Rachel K.V., Suvvari T.K., Prasad S., Boppana S.H., Vegi P.K.. Association of IL1R1 gene (SNP rs2071374) with the risk of preeclampsia. J. Reprod. Immunol. (2022) 149 103463. DOI: 10.1016/j.jri.2021.103463 17. Kumar K.S.P., Arcot M., Munisamaiah M., Balakrishna S.. Novel association of SNP rs479200 in EGLN1 gene with predisposition to preeclampsia. Gene (2019) 705 1-4. DOI: 10.1016/j.gene.2019.04.049 18. Schmieder R.E., Hilgers K.F., Schlaich M.P., Schmidt B.M.. Renin-angiotensin system and cardiovascular risk. Lancet (2007) 369 1208-1219. DOI: 10.1016/S0140-6736(07)60242-6 19. Paz Ocaranza M., Riquelme J.A., García L., Jalil J.E., Chiong M., Santos R.A.S., Lavandero S.. Counter-regulatory renin-angiotensin system in cardiovascular disease. Nat. Rev. Cardiol. (2020) 17 116-129. DOI: 10.1038/s41569-019-0244-8 20. Takimoto-Ohnishi E., Murakami K.. Renin-angiotensin system research: From molecules to the whole body. J. Physiol. Sci. (2019) 69 581-587. DOI: 10.1007/s12576-019-00679-4 21. Wu C.H., Mohammadmoradi S., Chen J.Z., Sawada H., Daugherty A., Lu H.S.. Renin-Angiotensin System and Cardiovascular Functions. Arterioscler. Thromb. Vasc. Biol. (2018) 38 e108-e116. DOI: 10.1161/ATVBAHA.118.311282 22. Wu C., Lu H., Cassis L.A., Daugherty A.. Molecular and Pathophysiological Features of Angiotensinogen: A Mini Review. N Am. J. Med. Sci. (2011) 4 183-190. DOI: 10.7156/v4i4p183 23. Hypertension in pregnancy. Report of the American College of Obstetricians and Gynecologists’ Task Force on Hypertension in Pregnancy. Obstet. Gynecol. (2013) 122 1122-1131. DOI: 10.1097/01.Aog.0000437382.03963.88 24. Xu Y., Rong J., Zhang Z.. The emerging role of angiotensinogen in cardiovascular diseases. J. Cell. Physiol. (2021) 236 68-78. DOI: 10.1002/jcp.29889 25. Chen W., Chen Y., Zhang K., Yang W., Li X., Zhao J., Liu K., Dong Z., Lu J.. AGT serves as a potential biomarker and drives tumor progression in colorectal carcinoma. Int. Immunopharmacol. (2021) 101 108225. DOI: 10.1016/j.intimp.2021.108225 26. Tao X.R., Rong J.B., Lu H.S., Daugherty A., Shi P., Ke C.L., Zhang Z.C., Xu Y.C., Wang J.A.. Angiotensinogen in hepatocytes contributes to Western diet-induced liver steatosis. J. Lipid. Res. (2019) 60 1983-1995. DOI: 10.1194/jlr.M093252 27. Yilmaz Z., Yildirim T., Yilmaz R., Aybal-Kutlugun A., Altun B., Kucukozkan T., Erdem Y.. Association between urinary angiotensinogen, hypertension and proteinuria in pregnant women with preeclampsia. J. Renin Angiotensin Aldosterone Syst. (2015) 16 514-520. DOI: 10.1177/1470320313510585 28. Dahabiyeh L.A., Tooth D., Kurlak L.O., Mistry H.D., Pipkin F.B., Barrett D.A.. A pilot study of alterations in oxidized angiotensinogen and antioxidants in pre-eclamptic pregnancy. Sci. Rep. (2020) 10 1956. DOI: 10.1038/s41598-020-58930-7 29. Watkins W.S., Hunt S.C., Williams G.H., Tolpinrud W., Jeunemaitre X., Lalouel J.M., Jorde L.B.. Genotype-phenotype analysis of angiotensinogen polymorphisms and essential hypertension: The importance of haplotypes. J. Hypertens. (2010) 28 65-75. DOI: 10.1097/HJH.0b013e328332031a 30. Mabhida S.E., Mashatola L., Kaur M., Sharma J.R., Apalata T., Muhamed B., Benjeddou M., Johnson R.. Hypertension in African Populations: Review and Computational Insights. Genes (2021) 12. DOI: 10.3390/genes12040532 31. Junusbekov Y., Bayoglu B., Cengiz M., Dirican A., Arslan C.. AGT rs699 and AGTR1 rs5186 gene variants are associated with cardiovascular-related phenotypes in atherosclerotic peripheral arterial obstructive disease. Ir. J. Med. Sci. (2020) 189 885-894. DOI: 10.1007/s11845-019-02166-6 32. Dong M.Z., Lin Z.H., Liu S.S., Xin Y.N., Xuan S.Y.. AGT rs5051 gene polymorphism increases the risk of coronary heart disease in patients with non-alcoholic fatty liver disease in the Han Chinese population. Zhonghua Gan Zang Bing Za Zhi (2021) 29 1095-1100. DOI: 10.3760/cma.j.cn501113-20210106-00008 33. Lin R., Lei Y., Yuan Z., Ju H., Li D.. Angiotensinogen gene M235T and T174M polymorphisms and susceptibility of pre-eclampsia: A meta-analysis. Ann. Hum. Genet (2012) 76 377-386. DOI: 10.1111/j.1469-1809.2012.00722.x 34. Lee S.R., Moon J.Y., Lee S.H., Ihm C.G., Lee T.W., Kim S.K., Chung J.H., Kang S.W., Kim T.H., Park S.J.. Angiotensinogen Polymorphisms and Post-Transplantation Diabetes Mellitus in Korean Renal Transplant Subjects. Kidney Blood Press. Res. (2013) 37 95-102. DOI: 10.1159/000343404 35. El-Garawani I.M., Shaheen E.M., El-Seedi H.R., Khalifa S.A.M., Mersal G.A.M., Emara M.M., Kasemy Z.A.. Angiotensinogen Gene Missense Polymorphisms (rs699 and rs4762): The Association of End-Stage Renal Failure Risk with Type 2 Diabetes and Hypertension in Egyptians. Genes (2021) 12. DOI: 10.3390/genes12030339 36. Perdomo-Pantoja A., Mejía-Pérez S.I., Reynoso-Noverón N., Gómez-Flores-Ramos L., Soto-Reyes E., Sánchez-Correa T.E., Guerra-Calderas L., Castro-Hernandez C., Vidal-Millán S., Sánchez-Corona J.. Angiotensinogen rs5050 germline genetic variant as potential biomarker of poor prognosis in astrocytoma. PLoS ONE (2018) 13. DOI: 10.1371/journal.pone.0206590 37. Liu Y., Fu L., Pi L., Che D., Xu Y., Zheng H., Long H., Zeng L., Huang P., Zhang L.. An Angiotensinogen Gene Polymorphism (rs5050) Is Associated with the Risk of Coronary Artery Aneurysm in Southern Chinese Children with Kawasaki Disease. Dis. Markers (2019) 2849695. DOI: 10.1155/2019/2849695 38. Mopidevi B., Ponnala M., Kumar A.. Human angiotensinogen +11525 C/A polymorphism modulates its gene expression through microRNA binding. Physiol. Genom. (2013) 45 901-906. DOI: 10.1152/physiolgenomics.00056.2013 39. Wu Y., Wang M., Zhang J., Sun N., Li C.. A new model of the mechanism underlying lead poisoning: SNP in miRNA target region influence the AGT expression level. Hereditas (2019) 156 6. DOI: 10.1186/s41065-019-0084-x 40. Yang Y., Le Ray I., Zhu J., Zhang J., Hua J., Reilly M.. Preeclampsia Prevalence, Risk Factors, and Pregnancy Outcomes in Sweden and China. JAMA Netw. Open (2021) 4 e218401. DOI: 10.1001/jamanetworkopen.2021.8401 41. Zhang N., Tan J., Yang H., Khalil R.A.. Comparative risks and predictors of preeclamptic pregnancy in the Eastern, Western and developing world. Biochem. Pharmacol. (2020) 182 114247. DOI: 10.1016/j.bcp.2020.114247
--- title: Berberine Alleviates Doxorubicin-Induced Myocardial Injury and Fibrosis by Eliminating Oxidative Stress and Mitochondrial Damage via Promoting Nrf-2 Pathway Activation authors: - Yiyang Wang - Jia Liao - Yuanliang Luo - Mengsi Li - Xingyu Su - Bo Yu - Jiashuo Teng - Huadong Wang - Xiuxiu Lv journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966753 doi: 10.3390/ijms24043257 license: CC BY 4.0 --- # Berberine Alleviates Doxorubicin-Induced Myocardial Injury and Fibrosis by Eliminating Oxidative Stress and Mitochondrial Damage via Promoting Nrf-2 Pathway Activation ## Abstract Doxorubicin (DOX)-related cardiotoxicity has been recognized as a serious complication of cancer chemotherapy. Effective targeted strategies for myocardial protection in addition to DOX treatment are urgently needed. The purpose of this paper was to determine the therapeutic effect of berberine (Ber) on DOX-triggered cardiomyopathy and explore the underlying mechanism. Our data showed that Ber markedly prevented cardiac diastolic dysfunction and fibrosis, reduced cardiac malondialdehyde (MDA) level and increased antioxidant superoxide dismutase (SOD) activity in DOX-treated rats. Moreover, Ber effectively rescued the DOX-induced production of reactive oxygen species (ROS) and MDA, mitochondrial morphological damage and membrane potential loss in neonatal rat cardiac myocytes and fibroblasts. This effect was mediated by increases in the nuclear accumulation of nuclear erythroid factor 2-related factor 2 (Nrf2) and levels of heme oxygenase-1 (HO-1) and mitochondrial transcription factor A (TFAM). We also found that Ber suppressed the differentiation of cardiac fibroblasts (CFs) into myofibroblasts, as indicated by decreased expression of α-smooth muscle actin (α-SMA), collagen I and collagen III in DOX-treated CFs. Pretreatment with Ber inhibited ROS and MDA production and increased SOD activity and the mitochondrial membrane potential in DOX-challenged CFs. Further investigation indicated that the Nrf2 inhibitor trigonelline reversed the protective effect of Ber on both cardiomyocytes and CFs after DOX stimulation. Taken together, these findings demonstrated that Ber effectively alleviated DOX-induced oxidative stress and mitochondrial damage by activating the Nrf2-mediated pathway, thereby leading to the prevention of myocardial injury and fibrosis. The current study suggests that Ber is a potential therapeutic agent for DOX-induced cardiotoxicity that exerts its effects by activating Nrf2. ## 1. Introduction Doxorubicin (DOX) is an anticancer drug that has been widely used in treating several solid tumors [1,2]. Unfortunately, DOX treatment is associated with adverse effects such as cardiomyopathy and heart failure [3]. This cardiotoxicity is frequently detected years after the cessation of chemotherapy and cannot be effectively treated [4,5]. Although lowering the dose of DOX could reduce DOX-induced cardiotoxicity, this would lead to lower success rates of chemotherapy, which creates a dilemma for oncologists and cardiologists [6,7]. Dexrazoxane is currently the only cardioprotective agent approved by the United States Food and Drug Administration for use in patients receiving anthracycline chemotherapy. However, it can induce bone marrow suppression, liver toxicity and the development of secondary malignant tumors [8]. Therefore, it is of clinical importance to explore safe and effective adjuvant therapies for DOX-induced cardiotoxicity. The molecular mechanism of DOX-induced cardiotoxicity has been extensively investigated. Although many studies have revealed that this condition may be related to inflammation, topoisomerase 2β, Ca2+ handling abnormalities and autophagy dysfunction [9,10,11], oxidative stress has been described to be most commonly associated with the complex pathophysiology of DOX-induced cardiotoxicity [12,13,14]. Oxidative stress is defined as the production of excessive reactive oxygen species (ROS) that cannot be quenched by the antioxidant defense system and may contribute to the pathophysiology of cardiac remodeling and heart failure [15]. Excessive ROS lead to oxidative damage to biological macromolecules and disrupt the integrity and function of the cell membrane, further leading to mitochondrial dysfunction and eventual cardiomyocyte apoptosis [16]. Oxidative stress, which is considered to be an important manifestation of DOX-induced myocardial remodeling [17], is also implicated in the pathogenesis of cardiac fibrosis both directly and via its involvement in cytokine and growth-factor signaling [18,19]. Cardiomyocytes are more susceptible to ROS because there are lower levels of antioxidant enzymes in the heart than in other organs [20]. Therefore, improving the endogenous cardiac antioxidant defense system is an attractive strategy to prevent DOX-induced cardiotoxicity. Nuclear factor erythroid 2-related factor 2 (Nrf2), a redox-sensitive transcription factor, is the central regulator of the antioxidative stress defense system [21]. In response to oxidants, Nrf2 translocates to the nucleus, where it binds to antioxidant responsive elements to activate the transcription of downstream genes encoding cytoprotective enzymes such as heme oxygenase-1 (HO-1) [22]. Although aberrant HO-1 upregulation has been reported in several diseases such as Alzheimer disease and cancers [23,24], HO-1 protects against oxidative injury, regulates apoptosis, and modulates inflammation in normal cells [25,26]. It has been reported that activation of Nrf2/HO-1 signaling favors the enhancement of cell survival upon chemotherapy [27]. The phytochemicals such as curcumin [28] and formononetin [29] induce Nrf2/HO-1 signaling in reducing toxicity of oxaliplatin against liver and brain cells. These studies clearly demonstrate that Nrf2 signaling is of importance for alleviation of chemotherapy-mediated side effects. Furthermore, Nrf2 activates mitochondrial transcription factor A (TFAM), which leads to increased mitochondrial protein synthesis and mitochondrial mass, improved mitochondrial respiration, and depolarization of the mitochondrial membrane potential [30]. Activation of the Nrf2 pathway may attenuate DOX-induced oxidative stress and cardiac damage. Berberine (Ber) is an isoquinoline alkaloid originally isolated from Chinese goldthread (Coptis chinensis) and possesses a wide range of biological activities, including antitumor, anti-inflammatory, cardiovascular-protective, and antioxidant activities, allowing it to reduce ROS production in various tissues [31,32,33]. Ber has been found to protect against oxidative stress injury in neuronal cells by enhancing Nrf2/HO-1 signaling [34]. We previously showed that Ber inhibits DOX-induced cardiomyocyte apoptosis in vitro and in vivo [35]; however, its mechanisms need to be explored in depth. We hypothesize that Ber may play an antioxidant role by promoting the activation of the Nrf2 signaling pathway, resulting in the prevention of DOX-induced cardiotoxicity. In this study, we aimed to investigate whether Ber could protect against DOX-induced myocardial injury and fibrosis by preventing oxidative stress and mitochondrial damage via regulation of the Nrf2-mediated pathway. ## 2.1. The Effect of Ber on Preventing DOX-Induced Cardiac Diastolic Dysfunction and Fibrosis Is Associated with Alleviating Oxidative Stress Two days after the last administration, echocardiography was performed to study the effects of Ber and/or DOX treatment on left ventricular function. Echocardiographic examination demonstrated that DOX treatment markedly reduced CO and the E/A ratio, an index used to reflect diastolic function. In contrast, CO and the E/A ratio were higher in the Ber (60 mg/kg) + DOX group than in the DOX group. Moreover, LVEDD and LVESD decreased after DOX challenge. LVEDD and LVESD were increased in the Ber-pretreated group compared with the DOX group (Figure 1A,B). Usually, cardiac injury is accompanied by the release of enzymes such as creatine kinase (CK) into the blood [36]. In our study, as shown in Figure 1C, DOX increased serum CK levels from 1378.8 ± 225.1 U to 2640 ± 443.8 U. Pretreatment with Ber markedly reduced the elevated CK content to 1586.5 ± 170.8 U. To further study the mechanism of Ber against DOX-initiated myocardial injury, cardiac fibrosis in rats treated with DOX and/or Ber was detected by Masson’s trichrome staining. The results of Masson’s staining in the four groups are shown in Figure 1D,E. After Masson’s staining, cardiomyocytes were stained red, whereas collagen fibers were stained blue. More collagen deposition was observed in the hearts of rats administered DOX than in the hearts of control rats. The ratio of collagen area to total area was measured. The fractional area of cardiac fibrosis was significantly larger in the DOX group than in the control group. Fibrosis of the heart was markedly reduced in the DOX + Ber group compared with the DOX group. Furthermore, we measured MDA and SOD levels in the myocardium of rats treated with DOX with or without Ber. The MDA content in cardiac tissue was increased from 0.92 ± 0.025 nmol/mg protein to 1.54 ± 0.01 nmol/mg protein after DOX injection. However, the SOD level in cardiac tissues was decreased from 12.28 ± 0.61 U/mg to 6.23 ± 0.47 U/mg. Ber obviously decreased the elevated MDA content to 1.44 ± 0.028 nmol/mg and increased the levels of SOD to 10.64 ± 0.88 nmol/mg protein. This suggests that the effect of Ber in protecting against DOX-induced cardiac diastolic dysfunction and fibrosis is related to inhibiting oxidative stress. ## 2.2. Ber Inhibited Cardiomyocyte Apoptosis, Oxidative Stress, Mitochondrial Injury and Activated Nrf-2-Mediated Signaling Pathway in Cardiomyocytes after DOX Stimulation We found that Ber inhibits DOX-induced cardiomyocyte apoptosis in vitro and in vivo in our previous study [35], but the mechanisms involved have not been thoroughly explored. Here, the TUNEL assay results again confirmed that Ber reduced the apoptosis of neonatal rat cardiomyocytes after DOX administration (Figure 2A,B). The assessment of DOX-induced cytotoxicity in cardiomyocytes was performed using the cTnI test with a culture medium of neonatal rat cardiomyocytes. The cells were preincubated with Ber for 30 min and treated with DOX for 24 h. The results of the cTnI release assay showed that DOX significantly increased cTnI content in the culture medium of cardiomyocytes. Pretreatment with Ber considerably decreased cTnI content in the culture medium of cardiomyocytes treated with DOX (Figure 2C). To further ascertain the effects of Ber on DOX-induced oxidative stress in neonatal rat cardiomyocytes, we examined ROS production and MDA and SOD levels in cardiomyocytes. Consistent with the results of the in vivo experiments, as shown in Figure 2D–F, DOX increased ROS production and MDA content but decreased SOD levels in cardiomyocytes, and this effect was reversed by Ber. It is well known that mitochondria have important physiological functions, such as oxidative phosphorylation, electron transfer and energy metabolism, and are the source of intracellular oxidative stress [37]. Many studies have found that DOX induces the production of excess ROS in mitochondria and leads to mitochondrial dysfunction [38,39]. We previously found that exposure of neonatal rat cardiomyocytes to DOX induces a marked decrease in the mitochondrial membrane potential, an effect that is reduced by Ber [35]. In the current research, we examined the changes in mitochondrial morphology in DOX- and/or Ber-treated cardiomyocytes using transmission electron microscopy. As shown in Figure 2G, mitochondria in the control group were regular with distinct cristae. After treatment with DOX for 2 h, the mitochondrial structure was not distinct, the mitochondrial cristae were swollen or disordered, and vacuolization was visible. We also observed partial mitochondrial membrane rupture. Interestingly, irregular mitochondria, vacuolization, and partial mitochondrial membrane rupture were alleviated in the Ber pretreatment group. The changes in mitochondrial ultrastructure indicate that Ber notably reduced DOX-induced mitochondrial damage in cardiomyocytes. To clarify the mechanism underlying the protective effects of Ber against DOX-induced oxidative stress and mitochondrial damage, we examined the protein levels of nuclear/cytosolic Nrf2, HO-1, and the mitochondrial marker TFAM in cardiomyocytes 2 h after DOX treatment. The Western blot results showed that DOX induced a modest increase in the levels of nuclear Nrf2 protein in neonatal rat cardiomyocytes. Compared with the DOX treatment, Ber pretreatment elevated nuclear Nrf2 levels in cells. Additionally, the protein levels of HO-1 and TFAM were significantly lower in DOX-treated cells than in control cells, and this effect that was partially reversed by Ber pretreatment (Figure 2H–J). These results demonstrate that the protective effect of Ber against DOX-induced cardiomyocyte injury is most likely related to the complete inhibition of oxidative stress and mitochondrial damage through activation of the Nrf2-mediated pathway. ## 2.3. The Effect of Ber on Protecting Cardiomyocytes against DOX-Induced Injury, Mitochondrial Damage and Oxidative Stress Was Dependent on Nrf2 Activation To confirm the role of Nrf2 activation in the mechanism underlying the protective effect of Ber against DOX-induced cardiomyocyte apoptosis and injury, we treated cardiomyocytes with an inhibitor of Nrf2, trigonelline (TRI), 30 min prior to Ber and DOX stimulation. As depicted in Figure 3A, the Western blot results showed that the effect of Ber in increasing HO-1 and TFAM levels in DOX-treated cardiomyocytes was completely blocked by TRI. Moreover, inhibition of Nrf2 activation with 1 μM TRI reversed the inhibitory effect of Ber on cardiomyocyte apoptosis and the increase in cTnI levels induced by DOX but did not affect apoptosis or cTnI levels in cardiomyocytes treated with DOX alone (Figure 3B,C). Pretreatment with TRI also suppressed the effect of Ber in maintaining the mitochondrial membrane potential (MMP) of cardiomyocytes stimulated with DOX but had no effect on cardiomyocytes challenged with DOX alone (Figure 3D). Furthermore, we confirmed the role of Nrf2 activation in the protective effect of Ber against DOX-initiated oxidative stress in cardiomyocytes. ROS and MDA levels showed a marked decrease in cardiomyocytes costimulated with Ber and DOX for 24 h compared with those treated with DOX alone. TRI reversed the inhibitory effects of Ber on DOX-induced increases in ROS and MDA levels (Figure 3E,F). In addition, the SOD content was obviously increased in the DOX + Ber + TRI group compared with the DOX- and Ber-treated cardiomyocytes not pretreated with TRI (Figure 3G). The administration of TRI did not markedly alter the levels of HO-1, ROS, MDA or SOD in cardiomyocytes stimulated by DOX alone. These findings demonstrate that the suppressive effect of Ber on DOX-induced cardiomyocyte injury, mitochondrial damage and oxidative stress was dependent on Nrf2 activation. ## 2.4. Ber Suppressed DOX-Induced Differentiation of Cardiac Fibroblasts (CFs) into Myofibroblasts Cardiac fibrosis was observed after DOX treatment, suggesting that DOX is a potent inducer of fibroblast-to-myofibroblast differentiation characterized by high expression of α-smooth muscle actin (α-SMA) [40]. To determine the effect of DOX and/or Ber on the differentiation of CFs into myofibroblasts, first, the level of α-SMA in DOX-administered CFs with or without Ber was assessed by immunofluorescence staining. As expected, compared with control CFs, the CFs treated with DOX for 24 h exhibited an obvious increase in α-SMA expression. We observed a drastic decrease in α-SMA levels in the DOX + Ber-treated CFs compared with the CFs treated with DOX alone. Ber slightly increased the expression of α-SMA in CFs not stimulated with DOX, but the increase was not statistically significant (Figure 4A). The Western blot results showed that the protein expression levels of α-SMA, collagen I and collagen III in DOX-treated cells were significantly increased compared with those in control cells. Ber completely inhibited the DOX-induced increases in α-SMA, collagen I and collagen III levels in CFs (Figure 4B). This means that treatment with Ber prior to DOX administration inhibits the differentiation of CFs. ## 2.5. Ber Alleviated DOX-Induced Oxidative Stress and MMP Loss and Promoted Activation of the Nrf2 Pathway in CFs To evaluate the effect of Ber on oxidative stress in CFs stimulated with DOX, we measured the levels of ROS, MDA and SOD in CFs treated with DOX and/or Ber for 24 h. The results showed that DOX obviously increased the levels of ROS and MDA and decreased the SOD content in the CFs. Similar to the results in cardiomyocytes, Ber significantly inhibited DOX-induced alterations in ROS, MDA and SOD levels in CFs (Figure 5A–C). We further determined the effect of Ber on the MMP in DOX-treated CFs stained with JC-1. When the MMP is normal, JC-1 forms J-aggregates in the mitochondrial matrix, exhibiting red fluorescence. Conversely, JC-1 exists as a monomer that fluoresces green when the MMP is lost. Thus, a change in florescence from red to green indicates a decrease in the MMP. As shown in Figure 5D, the ratio of red fluorescence intensity to green fluorescence intensity was markedly lower in the group treated with DOX for 12 h than in the control group. In contrast, the intensity of red/green fluorescence decreased in DOX-stimulated CFs pretreated with Ber. As determined by Western blotting, Ber markedly increased the levels of Nrf2, HO-1 and TFAM in CFs 2 h after DOX stimulation (Figure 5E–G). These findings indicate that Ber promotes the activation of the Nrf2 signaling pathway, which may alleviate DOX-mediated CF differentiation by suppressing oxidative stress and mitochondrial injury. ## 2.6. An Inhibitor of Nrf2 Abolished the Inhibitory Effect of Ber on CF Differentiation into Myofibroblasts Induced by DOX Furthermore, we explored whether the effect of Ber in inhibiting DOX-initiated CF differentiation was achieved by activation of Nrf2. CFs were preincubated with TRI for 30 min before treatment with LPS and DOX. The Western blot results are shown in Figure 6A, Ber upregulated the expression of HO-1 and TFAM and reduced α-SMA and collagen I/III expression in DOX-challenged CFs. These effects of Ber were reversed by TRI. Stimulation with TRI had no effect on the differentiation of DOX-treated CFs. We also tested the effect of TRI on the MMP and oxidative stress in CFs after DOX and Ber treatment. Consistent with the results of the cardiomyocyte experiments, the inhibitory action of Ber on DOX-induced MMP loss in CFs was markedly prevented by TRI (Figure 6B). Moreover, TRI completely inhibited the effect of Ber on the downregulation of ROS and MDA and the increase in SOD levels in CFs after DOX stimulation (Figure 6C–E). These data indicate that the effect of Ber in preventing DOX-induced differentiation of CFs is mediated by activation of the Nrf2 signaling pathway, which suppresses oxidative stress and mitochondrial injury in CFs. ## 3. Discussion Although Doxorubicin is an effective chemotherapeutic drug for treating multiple cancers worldwide, the risk of DOX-induced cardiotoxicity, such as myocyte destruction, left ventricular dysfunction and cardiac remodeling, has been noted [41,42]. The pathological mechanisms of DOX-induced cardiac injury are complicated and have not been fully elucidated. It has been suggested that increased oxidative stress, such as the production of ROS, and compromise of the antioxidant system play pivotal roles in the disturbance of cardiac homeostasis in DOX-induced cardiotoxicity [43,44]. Mitochondria are the main sites of ROS generation and the key targets of DOX [45]. Therefore, exploring ways to inhibit oxidative stress and protect mitochondria is an attractive approach to reduce cardiac injury after DOX treatment. Ber hydrochloride is an effective antioxidant and free radical scavenger that prevents ROS formation [46]. In this research, we identified Ber as a potential therapeutic drug in treating DOX-induced cardiotoxicity, as indicated by its ability to inhibit myocardial injury and fibrosis and improve cardiac diastolic function. Mechanistically, we revealed that Ber completely inhibited oxidative stress and mitochondrial injury by inducing Nrf2-mediated upregulation of HO-1 and TFAM expression in cardiomyocytes and CFs, thereby leading to reduced cardiomyocyte apoptosis and CF differentiation; these effects were responsible for the beneficial roles of Ber in DOX-induced cardiotoxicity (Figure 7). These results indicate that Ber may be a potential therapeutic drug for the prevention of DOX-induced heart failure. The evidence shows that left ventricular dysfunction and cardiac fibrosis have crucial roles in the pathogenesis of DOX-induced cardiotoxicity [47,48]. Our previous research showed that DOX and/or Ber have no obvious effect on left ventricular ejection fraction and fractional shortening, suggesting that DOX may only cause mild impairment of left ventricular systolic function. In this study, we confirmed that DOX reduced CO, the E/A ratio, LVEDD, and LVESD, and these effects were reversed by pretreatment with Ber. This means that DOX treatment caused obvious impairment of cardiac diastolic function, as reflected by a decreased E/A ratio, and this effect was significantly prevented by Ber. CK is one of the most important markers of cardiotoxicity. In this study, Ber markedly inhibited the increase in serum CK activity induced by DOX. In addition, our results further confirmed that Ber suppressed DOX-induced cardiac fibrosis, as indicated by Masson’s staining. Oxidative stress causes damage to the lipid membrane and other cellular components due to an imbalance between the oxidant and antioxidant enzyme systems, which may be an important factor in DOX-induced cardiotoxicity [49]. Several studies have reported that DOX tends to induce the generation of ROS during metabolism [50,51]. Compared with other tissues, the myocardium is more vulnerable to oxidative stress, possibly due to low levels of antioxidant enzymes that scavenge ROS [52,53]. Consistent with previous studies, our results confirmed that DOX promoted ROS production in the myocardium of rats. As high levels of ROS cause lipid peroxidation, we measured MDA content and the activity of SOD, a major antioxidant enzyme. We found that Ber decreased the production of MDA and increased SOD levels after DOX challenge in rat cardiac tissue. Based on the above results, we conclude that Ber alleviated DOX-induced myocardial oxidative stress damage, which may be an important mechanism underlying its protective effect against cardiac diastolic dysfunction and fibrosis. Some studies have suggested that cardiomyocyte apoptosis is associated with irreversible heart failure. We previously found that Ber protects against DOX-induced cardiomyocyte apoptosis through the mitochondria-mediated apoptotic pathway. In this study, we more deeply explored the mechanisms by which Ber protects cardiomyocytes against DOX-induced injury in vitro. The results showed that 1 μM Ber not only inhibited cardiomyocyte apoptosis but also decreased ROS and MDA production and reduced SOD levels in cardiomyocytes treated with DOX. SOD is the first line of defense against oxygen-derived free radicals, and mitochondria are rich in SOD and are the source of intracellular ROS [54]. DOX is retained in the inner mitochondrial membrane during complexation with cardiolipin [55]. Given that Ber mainly enhanced SOD activity and reduced ROS production, we speculated that the key players in the effects of Ber on DOX-induced oxidative stress in cardiomyocytes may be mitochondria. Therefore, we examined the integrity of mitochondria in primary cultured cardiomyocytes. Interestingly, we observed that the morphology of mitochondria was markedly changed after 2 h of DOX treatment. TEM showed that Ber reduced the swelling of mitochondrial cristae and the disorder and rupture of mitochondria induced by DOX. Combined with our previous finding that Ber reverses DOX-induced MMP loss in cardiomyocytes, these findings indicate that Ber plays an important role in maintaining mitochondrial integrity in cardiomyocytes upon DOX stimulation. CFs, the most abundant interstitial cells in the adult mammalian heart, are a key source of extracellular matrix proteins, such as collagen I and III, which surround myocytes [56]. Under pathological conditions, injurious factors lead to CF differentiation into cardiac myofibroblasts, which highly express α-SMA and produce more collagen I and III, which contribute to the development of cardiac fibrosis [57]. Cardiac compliance and diastolic function decline accompanied by aggravation of myocardial fibrosis, ultimately leading to cardiac insufficiency and even heart failure [58]. In this study, CFs differentiated into myofibroblasts after DOX treatment in vitro, which is consistent with a previous report [19]. Interestingly, Ber abolished DOX-induced CF differentiation, as indicated by the downregulation of α-SMA and collagen I/III. Importantly, these effects of Ber were accompanied by alleviated MMP loss and decreased intracellular oxidative stress, as indicated by ROS accumulation and reduced levels of the antioxidant enzyme SOD. These studies suggest that the inhibitory impact of Ber on the effects of DOX mediates the fibroblast–myofibroblast transition and is associated with complete inhibition of oxidative stress and maintenance of mitochondrial function. Nrf2 is a transcription factor that plays a key role in cellular defense against oxidative stress by initiating the transcription of antioxidant genes, including HO-1 [59]. In addition, Nrf2 acts on many genes whose products are imported into mitochondria, such as TFAM [60]. TFAM is necessary for maintaining mitochondrial function because it protects mitochondrial DNA against ROS damage and has been found to play an important role in the Nrf2-mediated protective effect on mitochondrial bioenergetics and biogenesis [61,62]. Numerous studies support the notion that downregulation of Nrf2 is associated with various oxidative stress-related cardiomyopathies, such as DOX-induced cardiotoxicity [63,64,65]. However, some studies have shown that activation of the Nrf2 pathway accelerates ferroptosis and promotes DOX-induced cardiotoxicity [66,67]. This discrepancy indicates that the effect of Nrf2 on DOX-induced cardiotoxicity requires further investigation. The results of this study showed that DOX disrupted the nuclear translocation of Nrf2 in both cardiomyocytes and CFs, which was completely reversed by Ber. Activation of Nrf2 by Ber was found to be beneficial for preserving the intracellular activity of antioxidant enzymes and mitochondrial biogenesis during DOX stimulation due to an increase in HO-1 and TFAM protein synthesis. Moreover, we further confirmed that activated Nrf2 participates in the suppressive effects of Ber on DOX-induced cardiomyocyte injury and CF differentiation by using the Nrf2 inhibitor TRI. The data showed that TRI inhibited the Ber-mediated suppression of cardiomyocyte apoptosis, CF differentiation, MMP loss and oxidative stress in cardiomyocytes and CFs challenged with DOX. Furthermore, the Ber-induced upregulation of HO-1 and TFAM in DOX-treated cardiomyocytes and CFs was completely suppressed by Nrf2 inhibition. It has been reported that stimulation of Nrf2 signaling protects H9c2 cardiomyocytes from damage caused by DOX [68,69]. Pharmacological agents with naturally occurring compounds as the most common have been used for inducing Nrf2 signaling in DOX amelioration [70,71]. In this study, our findings provide strong evidence that Ber inhibits DOX-induced cardiomyocyte injury and CF differentiation by activating Nrf2 to alleviate oxidative stress and protect mitochondria. This study is the first to show that Ber inhibits chronic DOX-induced cardiotoxicity through activating the Nrf2-mediated pathway. However, further studies must be carried out in clinical trials to confirm how to use Ber in the treatment of DOX-triggered heart failure. Meanwhile, we only studied the effect of Ber on activation of Nrf2 in DOX-treated cardiomyocytes and CFs, and the mechanisms need to be studied in depth. ## 4.1. Materials Doxorubicin hydrochloride (D1515) and berberine hemisulfate salt (B3412) were purchased from Sigma Aldrich (St. Louis, MO, USA). Trigonelline (HY-N0414) was obtained from MedChemExpress (Monmouth Junction, NJ, USA). Anti-α-SMA (ab7817), anti-Nrf2 (ab137550), anti-TFAM (ab131607), and anti-voltage-dependent anion channel (VDAC, ab15895) antibodies were obtained from Abcam (Cambridge, UK). Antibodies against vimentin (#5741), lamin B (#13435) and GAPDH (#2118) were purchased from Cell Signaling Technology, Inc. (Beverly, MA, USA). Antibodies against collagen I (WL0088) and collagen III (WL03186) were obtained from Wanleibio (Shenyang, Liaoning, China). Alexa Fluor 647-labeled goat anti-rabbit IgG (A32733), Alexa Fluor 488-labeled goat anti-mouse IgG (A11001), goat anti-rabbit IgG (#31462) and goat anti-mouse IgG (#31438) were obtained from Thermo Fisher Scientific (Logan, UT, USA). Reactive oxygen species (ROS, S0033S), malondialdehyde (MDA, S0131S) and superoxide dismutase (SOD, S0101S) assay kits were purchased from Beyotime Biotechnology (Nantong, China). ## 4.2. Animals and Treatment Procedures Wild-type male Sprague-Dawley rats (8 to 10 weeks old) were obtained from the Medical Laboratory Animal Center of Guangdong Province (Guangzhou, China) and housed under standard conditions (12 h light/12 h dark cycle, 24 °C, and 50–$70\%$ humidity) with free access to food and water. The rats were randomly divided into four groups: the control group (received saline as vehicle); the DOX group (received 2.5 mg/kg DOX three times per week); the Ber group (treated with Ber at dose of 60 mg/kg via intragastric administration); and the DOX + Ber group (treated with Ber 30 min before each DOX injection). There were eight rats in each group. The rats were returned to their original cages after drug treatment and provided free access to food and water. After three weeks, the rats were anesthetized with diethyl ether and sacrificed. Then, the left ventricle of the heart was obtained for Western blotting and Masson’s staining. All animal experiments were conducted according to the guidelines for the Care and Use of Laboratory Animals of the National Institutes of Health and approved by the Animal Care and Use Committee at Jinan University School of Medicine. ## 4.3. Echocardiography Two days after the last drug treatment, cardiac function was assessed with the VisualSonicsR Vevo 770TM High-Resolution In Vivo Micro-Imaging System (VisualSonics, Inc., Toronto, ON, Canada), as described previously [72]. Rats were anesthetized with $1.5\%$ isoflurane (Rhodia UK Ltd., Avonmouth, Bristol, UK) and imaged in the supine position using a 17.5-MHz-centered frequency RMV 707 scanhead. Two-dimensional B-mode and M-mode images were acquired by a technician who was blinded to the study design. The left ventricular end-diastolic diameter (LVEDD) and left ventricular end-systolic diameter (LVESD) were measured. The peak early filling (E wave) and late diastolic filling (A wave) velocities were measured on the M-mode parasternal short-axis tracing at the papillary muscle level. Cardiac output (CO) and the E/A ratio were calculated with Vevo770TM imaging system software, and data from at least three consecutive cardiac cycles were averaged. ## 4.4. Masson’s Trichrome Staining and Biochemical Analyses After echocardiographic assessment, the rats were immediately anesthetized with pentobarbital sodium (50 mg/kg, intraperitoneally). An appropriate depth of anesthesia was confirmed by the disappearance of the corneal reflex, loss of the pedal reflex, and failure to respond to a skin incision. Cardiac tissues were fixed in $10\%$ buffered formalin and embedded in paraffin. Paraffin-embedded cardiac tissues were cut into 5 μm-thick sections with standard techniques. For evaluation of cardiac fibrosis, the cardiac tissue sections were stained with Masson’s trichrome according to the manufacturer’s instructions and then examined under light microscopy. Fibrosis was assessed in 10 randomly selected fields per section at high magnification. ImageJ software (v1.53) was used to quantify the fractional area of cardiac fibrosis. Blood samples were collected in tubes and allowed to clot at room temperature. Serum was separated by centrifugation at 400× g for 15 min at 4 °C for the determination of creatine kinase (CK) activity. ## 4.5. Neonatal Rat Cardiomyocyte and Cardiac Fibroblast Culture and Cytotoxicity Assay Neonatal Sprague-Dawley rats obtained from the Medical Laboratory Animal Center of Guangdong Province (Guangzhou, China) were deeply anesthetized with pentobarbital sodium (100 mg/kg). The hearts were excised, and left ventricular cardiomyocytes and fibroblasts were enzymatically dissociated. The digested cells were cultured in Dulbecco’s modified Eagle medium supplemented with $10\%$ fetal bovine serum, 100 U/mL penicillin, and 100 µg/mL streptomycin at 37 °C for 48 h in a humidified atmosphere containing $5\%$ CO2. The cells were treated with 1.0 μM Ber or vehicle for 20 min and then exposed to 1.0 μM DOX for the indicated period. Cardiomyocyte cytotoxicity was assessed by measuring the release of cardiac troponin I (cTnI, LifeSpan BioSciences, New York, NY, USA). ## 4.6. Terminal Deoxynucleotidyl Transferase dUTP Nick End Labeling (TUNEL) Assay Cardiomyocyte apoptosis was determined using an in situ cell death detection kit (Roche Applied Science, Indianapolis, IN, USA), as previously described [35]. Briefly, cardiomyocytes were fixed with $4\%$ paraformaldehyde in PBS for 40 min at room temperature, washed and permeabilized with $0.1\%$Triton X-100 for another 5 min at 4 °C. The fixed cells were subjected to TUNEL staining according to the manufacturer’s instructions. Finally, the cells were incubated with 4′,6-diamidino-2-phenylindole (DAPI) for 15 min. Fluorescence images of ten random fields from each sample were taken using a fluorescence microscope. ## 4.7. Estimation of Oxidative Stress Biomarker Levels in Cardiac Tissues, Myocytes and Fibroblasts Cardiac tissue and neonatal rat cardiomyocytes and fibroblasts were harvested and lysed in lysis buffer (phosphate-buffered saline (PBS), $1\%$ NP-40, $0.5\%$ sodium deoxycholate, and $0.1\%$ sodium dodecyl sulfate (SDS)) containing 1 mM phenylmethylsulfonyl fluoride (PMSF) and incubated for 30 min on ice. After centrifugation at 4 °C and 12,000× g for 15 min, the supernatant was collected for protein quantification. MDA content and SOD activity were determined following the instructions of the lipid peroxidation assay kit and enzymatic activity assay kit. Cardiac myocytes and fibroblasts were collected, and ROS production was quantified using 2′,7′-dichlorodihydrofluorescein diacetate and analyzed by flow cytometry at an excitation wavelength of 495 nm and emission wavelength of 520 nm. ## 4.8. Determination of Mitochondrial Morphological Changes and the Mitochondrial Membrane Potential Cardiomyocytes were collected by centrifugation at 4 °C and 100× g for 5 min. Glutaraldehyde was added to the cells, followed by centrifugation at 4 °C and 200× g for 5 min and addition of $1\%$ osmium tetroxide for 90 min. PBS (0.1 M) was used to wash the cells after centrifugation, followed by incubation with 50, 70, and $90\%$ ethanol, $90\%$ acetone/$90\%$ ethanol (1:1), and $90\%$ acetone for 15 min. After centrifugation, the cells were washed three times with $100\%$ acetone for 10 min each and subsequently treated with $100\%$ acetone/embedding agent (1:1) overnight, embedding agent in a desiccator for 6 h, and 2,4,6-tris-(dimethylaminomethyl) phenol (DMP-30)-containing embedding agent in an oven at 40 °C for 3–5 days. The embedded samples were trimmed into a shape suitable for ultrathin slicing to observe the mitochondrial morphology using transmission electron microscopy (TEM). The mitochondrial membrane potential in cardiomyocytes and fibroblasts was assessed by 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolcarbocyanine iodide (JC-1) staining and a laser scanning confocal microscope as described previously [35]. JC-1 can accumulate and aggregate (red fluorescence) in normal mitochondria. Loss of the mitochondrial membrane potential prevents JC-1 entry into mitochondria, and monomeric JC-1 (green fluorescence) remains in the cytosol. Cardiomyocytes and fibroblasts from different treatment groups were washed, incubated with JC-1 at 37 °C for 15 min, washed again and mounted on a Leica fluorescence microscope for imaging. The ratio of the aggregated JC-1 fluorescence intensity to monomeric JC-1fluorescence intensity was used to quantify changes in the mitochondrial membrane potential. ## 4.9. Extraction of Protein from Cardiac Tissue and Cardiomyocytes Cardiomyocytes, cardiac tissues and fibroblasts were lysed in lysis buffer containing 1 mM PMSF by incubation for 30 min on ice and then centrifuged at 4 °C and 12,000× g for 15 min. The supernatant lysates were diluted in 2× or 5× SDS sample buffer and boiled for 8 min. Unless specified, whole-cell or tissue lysates were used for analysis. Mitochondrial and nuclear proteins were isolated using a mitochondria isolation kit (Thermo Fisher Scientific, Rockford, IL, USA) and NE-PER™ nuclear/cytoplasmic extraction reagent (Thermo Fisher Scientific, Rockford, IL, USA), respectively, according to the manufacturer’s instructions. ## 4.10. Western Blot Analysis The samples were resolved on an SDS-polyacrylamide gel and then transferred to polyvinylidene fluoride membranes (Millipore, Billerica, MA, USA) or nitrocellulose membranes (Millipore, Billerica, MA, USA). The membranes were blocked in tris-(hydroxymethyl) aminomethane (Tris)-buffered saline containing $5\%$ bovine serum albumin (20 mM Tris-HCl, 137 mM NaCl, and $0.1\%$ Tween 20) for 1 h at room temperature. The membranes were subsequently incubated overnight at 4 °C with primary antibodies. Following incubation with horseradish peroxidase-conjugated secondary antibodies, the immunoblots were exposed to film using enhanced chemiluminescence reagents (Thermo Fisher Scientific, Rockford, IL, USA). ImageJ software (v1.53), an open-source image-processing program, was used to quantify the band density. ## 4.11. Immunofluorescence Staining The cardiac fibroblasts were fixed with $4\%$ formalin solution and then treated with $0.1\%$ Triton X-100 in PBS. After 3 washes with PBS for 5 min each time, the cells were blocked in PBS with $1\%$ BSA, followed by incubation with a 1:200 dilution of antibodies against α-SMA and vimentin at 4 °C overnight. The cells were then incubated with the secondary antibodies Alexa Fluor 488-labeled goat antibody against mouse IgG and Alexa Fluor 647-labeled goat antibody against rabbit IgG (1:400 dilution) for 1 h. Finally, after washing three times, the cells were incubated with a 1:200 dilution of DAPI for 15 min and observed via fluorescence microscopy. ## 4.12. Statistical Analyses Data are expressed as the mean ± standard error of the mean (SEM). Statistical differences among groups were evaluated using one-way analysis of variance followed by Bonferroni post hoc analysis. Differences were considered statistically significant at $p \leq 0.05.$ Data were analyzed using SPSS software (Version 13.0, Chicago, IL, USA). ## 5. Conclusions In conclusion, the present study demonstrated for the first time that Ber prevents cardiac diastolic dysfunction and fibrosis in DOX-challenged rats by alleviating cardiomyocyte injury and CF differentiation. The molecular mechanism involves Ber-mediated activation of the Nrf-2 pathway, which attenuates DOX-induced oxidative stress and mitochondrial damage in both cardiomyocytes and cardiac fibroblasts. Considering that our previous study proved that Ber suppresses DOX-induced cardiomyocyte apoptosis, we suggest that Ber may be a promising therapeutic adjuvant that protects the heart from the serious side effects of DOX. ## References 1. Rivankar S.. An Overview of Doxorubicin Formulations in Cancer Therapy. J. Cancer Res. Ther. (2014) 10 853. DOI: 10.4103/0973-1482.139267 2. Duggan S.T., Keating G.M.. Pegylated Liposomal Doxorubicin: A Review of Its Use in Metastatic Breast Cancer, Ovarian Cancer, Multiple Myeloma and AIDS-Related Kaposi’s Sarcoma. Drugs (2011) 71 2531-2558. DOI: 10.2165/11207510-000000000-00000 3. Perez I.E., Taveras Alam S., Hernandez G.A., Sancassani R.. Cancer Therapy-Related Cardiac Dysfunction: An Overview for the Clinician. Clin. Med. Insights Cardiol. (2019) 13 1179546819866445. DOI: 10.1177/1179546819866445 4. Curigliano G., Cardinale D., Dent S., Criscitiello C., Aseyev O., Lenihan D., Cipolla C.M.. Cardiotoxicity of Anticancer Treatments: Epidemiology, Detection, and Management: Cardiotoxicity of Anticancer Treatments. CA Cancer J. Clin. (2016) 66 309-325. DOI: 10.3322/caac.21341 5. Fernández-Ruiz I.. Cardiotoxicity of Anticancer Therapy. Nat. Rev. Cardiol. (2016) 13 183. DOI: 10.1038/nrcardio.2016.39 6. Herrmann J.. Adverse Cardiac Effects of Cancer Therapies: Cardiotoxicity and Arrhythmia. Nat. Rev. Cardiol. (2020) 17 474-502. DOI: 10.1038/s41569-020-0348-1 7. Pugazhendhi A., Edison T.N.J.I., Velmurugan B.K., Jacob J.A., Karuppusamy I.. Toxicity of Doxorubicin (Dox) to Different Experimental Organ Systems. Life Sci. (2018) 200 26-30. DOI: 10.1016/j.lfs.2018.03.023 8. Monahan D.S., Flaherty E., Hameed A., Duffy G.P.. Resveratrol Significantly Improves Cell Survival in Comparison to Dexrazoxane and Carvedilol in a H9c2 Model of Doxorubicin Induced Cardiotoxicity. Biomed. Pharmacother. (2021) 140 111702. DOI: 10.1016/j.biopha.2021.111702 9. Bartlett J.J., Trivedi P.C., Pulinilkunnil T.. Autophagic Dysregulation in Doxorubicin Cardiomyopathy. J. Mol. Cell. Cardiol. (2017) 104 1-8. DOI: 10.1016/j.yjmcc.2017.01.007 10. Cappetta D., Rossi F., Piegari E., Quaini F., Berrino L., Urbanek K., De Angelis A.. Doxorubicin Targets Multiple Players: A New View of an Old Problem. Pharmacol. Res. (2018) 127 4-14. DOI: 10.1016/j.phrs.2017.03.016 11. Renu K., Abilash V.G., Tirupathi Pichiah P.B., Arunachalam S.. Molecular Mechanism of Doxorubicin-Induced Cardiomyopathy—An Update. Eur. J. Pharmacol. (2018) 818 241-253. DOI: 10.1016/j.ejphar.2017.10.043 12. Gonzalez Y., Pokrzywinski K.L., Rosen E.T., Mog S., Aryal B., Chehab L.M., Vijay V., Moland C.L., Desai V.G., Dickey J.S.. Reproductive Hormone Levels and Differential Mitochondria-Related Oxidative Gene Expression as Potential Mechanisms for Gender Differences in Cardiosensitivity to Doxorubicin in Tumor-Bearing Spontaneously Hypertensive Rats. Cancer Chemother. Pharmacol. (2015) 76 447-459. DOI: 10.1007/s00280-015-2786-8 13. Liao Z.Q., Jiang Y.N., Su Z.L., Bi H.L., Li J.T., Li C.L., Yang X.L., Zhang Y., Xie X.. Rutaecarpine Inhibits Doxorubicin-Induced Oxidative Stress and Apoptosis by Activating AKT Signaling Pathway. Front. Cardiovasc. Med. (2022) 8 809689. DOI: 10.3389/fcvm.2021.809689 14. Songbo M., Lang H., Xinyong C., Bin X., Ping Z., Liang S.. Oxidative Stress Injury in Doxorubicin-Induced Cardiotoxicity. Toxicol. Lett. (2019) 307 41-48. DOI: 10.1016/j.toxlet.2019.02.013 15. Tsutsui H., Kinugawa S., Matsushima S.. Oxidative Stress and Heart Failure. Am. J. Physiol.—Heart Circ. Physiol. (2011) 301 H2181-H2190. DOI: 10.1152/ajpheart.00554.2011 16. Wallace K.B., Sardão V.A., Oliveira P.J.. Mitochondrial Determinants of Doxorubicin-Induced Cardiomyopathy. Circ. Res. (2020) 126 926-941. DOI: 10.1161/CIRCRESAHA.119.314681 17. Tanaka R., Umemura M., Narikawa M., Hikichi M., Osaw K., Fujita T., Yokoyama U., Ishigami T., Tamura K., Ishikawa Y.. Reactive Fibrosis Precedes Doxorubicin-induced Heart Failure through Sterile Inflammation. ESC Heart Fail. (2020) 7 588-603. DOI: 10.1002/ehf2.12616 18. Purnomo Y., Piccart Y., Coenen T., Prihadi J., Lijnen P.. Oxidative Stress and Transforming Growth Factor-β1-Induced Cardiac Fibrosis. Cardiovasc. Haematol. Disord.-Drug Targets (2013) 13 165-172. DOI: 10.2174/1871529X11313020010 19. Cucoranu I., Clempus R., Dikalova A., Phelan P.J., Ariyan S., Dikalov S., Sorescu D.. NAD(P)H Oxidase 4 Mediates Transforming Growth Factor-Β1–Induced Differentiation of Cardiac Fibroblasts Into Myofibroblasts. Circ. Res. (2005) 97 900-907. DOI: 10.1161/01.RES.0000187457.24338.3D 20. do Nascimento T.C., Cazarin C.B.B., Maróstica M.R., Mercadante A.Z., Jacob-Lopes E., Zepka L.Q.. Microalgae Carotenoids Intake: Influence on Cholesterol Levels, Lipid Peroxidation and Antioxidant Enzymes. Food Res. Int. (2020) 128 108770. DOI: 10.1016/j.foodres.2019.108770 21. Bellezza I., Giambanco I., Minelli A., Donato R.. Nrf2-Keap1 Signaling in Oxidative and Reductive Stress. Biochim. Biophys. Acta (BBA)—Mol. Cell Res. (2018) 1865 721-733. DOI: 10.1016/j.bbamcr.2018.02.010 22. Mirzaei S., Zarrabi A., Hashemi F., Zabolian A., Saleki H., Azami N., Hamzehlou S., Farahani M.V., Hushmandi K., Ashrafizadeh M.. Nrf2 Signaling Pathway in Chemoprotection and Doxorubicin Resistance: Potential Application in Drug Discovery. Antioxidants (2021) 10. DOI: 10.3390/antiox10030349 23. Salerno L., Floresta G., Ciaffaglione V., Gentile D., Margani F., Turnaturi R., Rescifina A., Pittalà V.. Progress in the development of selective heme oxygenase-1 inhibitors and their potential therapeutic application. Eur. J. Med. Chem. (2019) 167 439-453. DOI: 10.1016/j.ejmech.2019.02.027 24. Barone E., Di Domenico F., Mancuso C., Butterfield D.A.. The Janus face of the heme oxygenase/biliverdin reductase system in Alzheimer disease: It’s time for reconciliation. Neurobiol. Dis. (2014) 62 144-159. DOI: 10.1016/j.nbd.2013.09.018 25. Howden R.. Oxidative Med. Cell. Longev. (2013) 2013 104308. DOI: 10.1155/2013/104308 26. Loboda A., Damulewicz M., Pyza E., Jozkowicz A., Dulak J.. Role of Nrf2/HO-1 System in Development, Oxidative Stress Response and Diseases: An Evolutionarily Conserved Mechanism. Cell. Mol. Life Sci. (2016) 73 3221-3247. DOI: 10.1007/s00018-016-2223-0 27. Azouz A.A., Abdel-Nassir Abdel-Razek E., Abo-Youssef A.M.. Amlodipine alleviates cisplatin-induced nephrotoxicity in rats through gamma-glutamyl transpeptidase (GGT) enzyme inhibition, associated with regulation of Nrf2/HO-1, MAPK/NF-kB, and Bax/Bcl-2 signaling. Saudi Pharm. J. (2020) 28 1317-1325. DOI: 10.1016/j.jsps.2020.08.022 28. Lu Y., Wu S., Xiang B., Li L., Lin Y.. Curcumin Attenuates Oxaliplatin-Induced Liver Injury and Oxidative Stress by Activating the Nrf2 Pathway. Drug Des. Dev. Ther. (2020) 14 73-85. DOI: 10.2147/DDDT.S224318 29. Fang Y., Ye J., Zhao B., Sun J., Gu N., Chen X., Ren L., Chen J., Cai X., Zhang W.. Formononetin ameliorates oxaliplatin-induced peripheral neuropathy via the KEAP1-NRF2-GSTP1 axis. Redox Biol. (2020) 36 101677. DOI: 10.1016/j.redox.2020.101677 30. Li Y., Feng Y.F., Liu X.T., Li Y.C., Zhu H.M., Sun M.R., Li P., Liu B., Yang H.. Songorine Promotes Cardiac Mitochondrial Biogenesis via Nrf2 Induction during Sepsis. Redox Biol. (2021) 38 101771. DOI: 10.1016/j.redox.2020.101771 31. Tabeshpour J., Imenshahidi M., Hosseinzadeh H.. A Review of the Effects of Berberis Vulgaris and Its Major Component, Berberine, in Metabolic Syndrome. Iran. J. Basic Med. Sci. (2017) 20 557-568. PMID: 28656091 32. Lau C.W., Yao X.Q., Chen Z.Y., Ko W.H., Huang Y.. Cardiovascular Actions of Berberine. Cardiovasc. Drug Rev. (2006) 19 234-244. DOI: 10.1111/j.1527-3466.2001.tb00068.x 33. Zhang P., Ma D., Wang Y., Zhang M., Qiang X., Liao M., Liu X., Wu H., Zhang Y.. Berberine Protects Liver from Ethanol-Induced Oxidative Stress and Steatosis in Mice. Food Chem. Toxicol. (2014) 74 225-232. DOI: 10.1016/j.fct.2014.10.005 34. Hsu Y.Y., Chen C.S., Wu S.N., Jong Y.J., Lo Y.C.. Berberine Activates Nrf2 Nuclear Translocation and Protects against Oxidative Damage via a Phosphatidylinositol 3-Kinase/Akt-Dependent Mechanism in NSC34 Motor Neuron-like Cells. Eur. J. Pharm. Sci. (2012) 46 415-425. DOI: 10.1016/j.ejps.2012.03.004 35. Lv X., Yu X., Wang Y., Wang F., Li H., Wang Y., Lu D., Qi R., Wang H.. Berberine Inhibits Doxorubicin-Triggered Cardiomyocyte Apoptosis via Attenuating Mitochondrial Dysfunction and Increasing Bcl-2 Expression. PLoS ONE (2012) 7. DOI: 10.1371/journal.pone.0047351 36. Chen J., Duan T., Fang W., Liu S., Wang C.. Analysis of Clinical Characteristics of Mesalazine-Induced Cardiotoxicity. Front. Pharmacol. (2022) 13 970597. DOI: 10.3389/fphar.2022.970597 37. Dorn G.W.. Mitochondrial Dynamics in Heart Disease. Biochim. Biophys. Acta (BBA)—Mol. Cell Res. (2013) 1833 233-241. DOI: 10.1016/j.bbamcr.2012.03.008 38. Russo M., Guida F., Paparo L., Trinchese G., Aitoro R., Avagliano C., Fiordelisi A., Napolitano F., Mercurio V., Sala V.. The novel butyrate derivative phenylalanine-butyramide protects from doxorubicin-induced cardiotoxicity. Eur. J. Heart Fail. (2019) 21 519-528. DOI: 10.1002/ejhf.1439 39. He H., Wang L., Qiao Y., Zhou Q., Li H., Chen S., Yin D., Huang Q., He M.. Doxorubicin Induces Endotheliotoxicity and Mitochondrial Dysfunction via ROS/eNOS/NO Pathway. Front. Pharmacol. (2020) 10 1531. DOI: 10.3389/fphar.2019.01531 40. Yao H., He Q., Huang C., Wei S., Gong Y., Li X., Liu W., Xu Z., Wu H., Zheng C.. Panaxatriol Saponin Ameliorates Myocardial Infarction-Induced Cardiac Fibrosis by Targeting Keap1/Nrf2 to Regulate Oxidative Stress and Inhibit Cardiac-Fibroblast Activation and Proliferation. Free Radic. Biol. Med. (2022) 190 264-275. DOI: 10.1016/j.freeradbiomed.2022.08.016 41. Sawicki K.T., Sala V., Prever L., Hirsch E., Ardehali H., Ghigo A.. Preventing and Treating Anthracycline Cardiotoxicity: New Insights. Annu. Rev. Pharmacol. Toxicol. (2021) 61 309-332. DOI: 10.1146/annurev-pharmtox-030620-104842 42. Bonaca M.P., Olenchock B.A., Salem J.E., Wiviott S.D., Ederhy S., Cohen A., Stewart G.C., Choueiri T.K., Di Carli M., Allenbach Y.. Myocarditis in the Setting of Cancer Therapeutics: Proposed Case Definitions for Emerging Clinical Syndromes in Cardio-Oncology. Circulation (2019) 140 80-91. DOI: 10.1161/CIRCULATIONAHA.118.034497 43. Awad H.H., El-Derany M.O., Mantawy E.M., Michel H.E., El-Naa M.M., Salah El-Din R.A., El-Brairy A.I., El-Demerdash E.. Comparative Study on Beneficial Effects of Vitamins B and D in Attenuating Doxorubicin Induced Cardiotoxicity in Rats: Emphasis on Calcium Homeostasis. Biomed. Pharmacother. (2021) 140 111679. DOI: 10.1016/j.biopha.2021.111679 44. Anjos M., Fontes-Oliveira M., Costa V.M., Santos M., Ferreira R.. An Update of the Molecular Mechanisms Underlying Doxorubicin plus Trastuzumab Induced Cardiotoxicity. Life Sci. (2021) 280 119760. DOI: 10.1016/j.lfs.2021.119760 45. Yang G., Song M., Hoang D.H., Tran Q.H., Choe W., Kang I., Kim S.S., Ha J.. Melatonin Prevents Doxorubicin-Induced Cardiotoxicity through Suppression of AMPKα2-Dependent Mitochondrial Damage. Exp. Mol. Med. (2020) 52 2055-2068. DOI: 10.1038/s12276-020-00541-3 46. Lee I.A., Hyun Y.J., Kim D.H.. Berberine Ameliorates TNBS-Induced Colitis by Inhibiting Lipid Peroxidation, Enterobacterial Growth and NF-ΚB Activation. Eur. J. Pharmacol. (2010) 648 162-170. DOI: 10.1016/j.ejphar.2010.08.046 47. Cappetta D., Esposito G., Coppini R., Piegari E., Russo R., Ciuffreda L.P., Rivellino A., Santini L., Rafaniello C., Scavone C.. Effects of Ranolazine in a Model of Doxorubicin-Induced Left Ventricle Diastolic Dysfunction: Ranolazine in Anthracycline Cardiotoxicity. Br. J. Pharmacol. (2017) 174 3696-3712. DOI: 10.1111/bph.13791 48. Quagliariello V., De Laurentiis M., Rea D., Barbieri A., Monti M.G., Carbone A., Paccone A., Altucci L., Conte M., Canale M.L.. The SGLT-2 Inhibitor Empagliflozin Improves Myocardial Strain, Reduces Cardiac Fibrosis and pro-Inflammatory Cytokines in Non-Diabetic Mice Treated with Doxorubicin. Cardiovasc. Diabetol. (2021) 20 150. DOI: 10.1186/s12933-021-01346-y 49. Olson R.D., Mushlin P.S.. Doxorubicin Cardiotoxicity: Analysis of Prevailing Hypotheses. FASEB J. (1990) 4 3076-3086. DOI: 10.1096/fasebj.4.13.2210154 50. Xu M.F., Tang P.L., Qian Z.M., Ashraf M.. Effects by Doxorubicin on the Myocardium Are Mediated by Oxygen Free Radicals. Life Sci. (2001) 68 889-901. DOI: 10.1016/S0024-3205(00)00990-5 51. Šimůnek T., Štěrba M., Popelová O., Adamcová M., Hrdina R., Geršl V.. Anthracycline-Induced Cardiotoxicity: Overview of Studies Examining the Roles of Oxidative Stress and Free Cellular Iron. Pharmacol. Rep. (2009) 61 154-171. DOI: 10.1016/S1734-1140(09)70018-0 52. Bashar T., Akhter N.. Study on Oxidative Stress and Antioxidant Level in Patients of Acute Myocardial Infarction before and after Regular Treatment. Bangladesh Med. Res. Counc. Bull. (2015) 40 79-84. DOI: 10.3329/bmrcb.v40i2.25226 53. De Geest B., Mishra M.. Role of Oxidative Stress in Heart Failure: Insights from Gene Transfer Studies. Biomedicines (2021) 9. DOI: 10.3390/biomedicines9111645 54. Wang Y., Branicky R., Noë A., Hekimi S.. Superoxide Dismutases: Dual Roles in Controlling ROS Damage and Regulating ROS Signaling. J. Cell Biol. (2018) 217 1915-1928. DOI: 10.1083/jcb.201708007 55. Goormaghtigh E., Chatelain P., Caspers J., Ruysschaert J.M.. Evidence of a Complex between Adriamycin Derivatives and Cardiolipin: Possible Role in Cardiotoxicity. Biochem. Pharmacol. (1980) 29 3003-3010. DOI: 10.1016/0006-2952(80)90050-7 56. Tallquist M.D., Molkentin J.D.. Redefining the Identity of Cardiac Fibroblasts. Nat. Rev. Cardiol. (2017) 14 484-491. DOI: 10.1038/nrcardio.2017.57 57. Kong P., Christia P., Frangogiannis N.G.. The Pathogenesis of Cardiac Fibrosis. Cell. Mol. Life Sci. (2014) 71 549-574. DOI: 10.1007/s00018-013-1349-6 58. Segura A.M., Frazier O.H., Buja L.M.. Fibrosis and Heart Failure. Heart Fail. Rev. (2014) 19 173-185. DOI: 10.1007/s10741-012-9365-4 59. Sies H., Berndt C., Jones D.P.. Oxidative Stress. Annu. Rev. Biochem. (2017) 86 715-748. DOI: 10.1146/annurev-biochem-061516-045037 60. Virbasius J.V., Scarpulla R.C.. Activation of the Human Mitochondrial Transcription Factor A Gene by Nuclear Respiratory Factors: A Potential Regulatory Link between Nuclear and Mitochondrial Gene Expression in Organelle Biogenesis. Proc. Natl. Acad. Sci. USA (1994) 91 1309-1313. DOI: 10.1073/pnas.91.4.1309 61. Kanki T., Ohgaki K., Gaspari M., Gustafsson C.M., Fukuoh A., Sasaki N., Hamasaki N., Kang D.. Architectural Role of Mitochondrial Transcription Factor A in Maintenance of Human Mitochondrial DNA. Mol. Cell. Biol. (2004) 24 9823-9834. DOI: 10.1128/MCB.24.22.9823-9834.2004 62. Bernard K., Logsdon N.J., Miguel V., Benavides G.A., Zhang J., Carter A.B., Darley-Usmar V.M., Thannickal V.J.. NADPH Oxidase 4 (Nox4) Suppresses Mitochondrial Biogenesis and Bioenergetics in Lung Fibroblasts via a Nuclear Factor Erythroid-Derived 2-like 2 (Nrf2)-Dependent Pathway. J. Biol. Chem. (2017) 292 3029-3038. DOI: 10.1074/jbc.M116.752261 63. Arunachalam S., Nagoor Meeran M.F., Azimullah S., Sharma C., Goyal S.N., Ojha S.. Nerolidol Attenuates Oxidative Stress, Inflammation, and Apoptosis by Modulating Nrf2/MAPK Signaling Pathways in Doxorubicin-Induced Acute Cardiotoxicity in Rats. Antioxidants (2021) 10. DOI: 10.3390/antiox10060984 64. Kumar H., Kumar R.M., Bhattacharjee D., Somanna P., Jain V.. Role of Nrf2 Signaling Cascade in Breast Cancer: Strategies and Treatment. Front. Pharm. (2022) 13 720076. DOI: 10.3389/fphar.2022.720076 65. Park Y.J., Yang H.J., Li W., Oh Y.-C., Go Y.. Menthae Herba Attenuates Neuroinflammation by Regulating CREB/Nrf2/HO-1 Pathway in BV2 Microglial Cells. Antioxidants (2022) 11. DOI: 10.3390/antiox11040649 66. Fang X., Wang H., Han D., Xie E., Yang X., Wei J., Gu S., Gao F., Zhu N., Yin X.. Ferroptosis as a Target for Protection against Cardiomyopathy. Proc. Natl. Acad. Sci. USA (2019) 116 2672-2680. DOI: 10.1073/pnas.1821022116 67. Koh J.S., Yi C., Heo R.W., Ahn J.-W., Park J.R., Lee J.E., Kim J.-H., Hwang J.-Y., Roh G.S.. Protective Effect of Cilostazol against Doxorubicin-Induced Cardiomyopathy in Mice. Free Radic. Biol. Med. (2015) 89 54-61. DOI: 10.1016/j.freeradbiomed.2015.07.016 68. Sunitha M.C., Dhanyakrishnan R., PrakashKumar B., Nevin K.G.. p-Coumaric acid mediated protection of H9c2 cells from Doxorubicin-induced cardiotoxicity: Involvement of augmented Nrf2 and autophagy. Biomed. Pharmacother. (2018) 102 823-832. DOI: 10.1016/j.biopha.2018.03.089 69. Ye M., Zhang L., Yan Y., Lin H.. Punicalagin protects H9c2 cardiomyocytes from doxorubicin-induced toxicity through activation of Nrf2/HO-1 signaling. Biosci. Rep. (2019) 39 BSR20190229. DOI: 10.1042/BSR20190229 70. El-Agamy D.S., El-Harbi K.M., Khoshhal S., Ahmed N., Elkablawy M.A., Shaaban A.A., Abo-Haded H.M.. Pristimerin protects against doxorubicin-induced cardiotoxicity and fibrosis through modulation of Nrf2 and MAPK/NF-kB signaling pathways. Cancer Manag. Res. (2019) 11 47-61. DOI: 10.2147/CMAR.S186696 71. Nazari Soltan Ahmad S., Sanajou D., Kalantary-Charvadeh A., Hosseini V., Roshangar L., Khojastehfard M., Haiaty S., Mesgari-Abbasi M.. b-LAPachone ameliorates doxorubicin-induced cardiotoxicity via regulating autophagy and Nrf2 signalling pathways in mice. Basic Clin. Pharmacol. Toxicol. (2020) 126 364-373. DOI: 10.1111/bcpt.13340 72. Wang Y., Wang Y., Yang D., Yu X., Li H., Lv X., Lu D., Wang H.. β1-Adrenoceptor Stimulation Promotes LPS-Induced Cardiomyocyte Apoptosis through Activating PKA and Enhancing CaMKII and IκBα Phosphorylation. Crit. Care (2015) 19 76. DOI: 10.1186/s13054-015-0820-1
--- title: Dyslipidemia Treatment and Lipid Control in US Adults with Diabetes by Sociodemographic and Cardiovascular Risk Groups in the NIH Precision Medicine Initiative All of Us Research Program authors: - Meleeka Akbarpour - Divya Devineni - Yufan Gong - Nathan D. Wong journal: Journal of Clinical Medicine year: 2023 pmcid: PMC9966763 doi: 10.3390/jcm12041668 license: CC BY 4.0 --- # Dyslipidemia Treatment and Lipid Control in US Adults with Diabetes by Sociodemographic and Cardiovascular Risk Groups in the NIH Precision Medicine Initiative All of Us Research Program ## Abstract Real-world data on lipid levels and treatment among adults with diabetes mellitus (DM) are relatively limited. We studied lipid levels and treatment status in patients with DM across cardiovascular disease (CVD) risk groups and sociodemographic factors. In the All of Us Research Program, we categorized DM as [1] moderate risk (≤1 CVD risk factor), [2] high risk (≥2 CVD risk factors), and [3] DM with atherosclerotic CVD (ASCVD). We examined the use of statin and non-statin therapy as well as LDL-C and triglyceride levels. We studied 81,332 participants with DM, which included $22.3\%$ non-Hispanic Black and $17.2\%$ Hispanic. A total of $31.1\%$ had ≤1 DM risk factor, $30.3\%$ had ≥2 DM risk factors, and $38.6\%$ of participants had DM with ASCVD. Only $18.2\%$ of those with DM and ASCVD were on high-intensity statins. Overall, $5.1\%$ were using ezetimibe and $0.6\%$ PCSK9 inhibitors. Among those with DM and ASCVD, only $21.1\%$ had LDL-C < 70 mg/dL. Overall, $1.9\%$ of participants with triglycerides ≥ 150 mg/dL were on icosapent ethyl. Those with DM and ASCVD were more likely to be on high-intensity statins, ezetimibe, and icosapent ethyl. Guideline-recommended use of high-intensity statins and non-statin therapy among our higher risk DM patients is lacking, with LDL-C inadequately controlled. ## 1. Introduction Atherosclerotic cardiovascular diseases (ASCVD) are major causes of morbidity and mortality in people with diabetes mellitus (DM) [1]. Dyslipidemia remains a significant ASCVD risk factor in those with DM. In the US Diabetes Collaborative Registry [2], among the 74,393 patients with DM, $48.6\%$ had controlled levels of low-density lipoprotein-cholesterol (LDL-C) but only $62\%$ were on a moderate- or high-intensity statin. Hypertriglyceridemia (HTG) also remains common in patients with DM. Among 1448 U.S. adults aged 20 years and over with diabetes in the US National Health and Nutrition Examination Survey, approximately $40\%$ had triglyceride levels of ≥ 150 mg/dL, regardless of statin use; and even among statin users with LDL-C < 70 mg/dL, one-third had borderline or elevated levels [3]. Moreover, clinical trials have shown that statin therapy, proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor use, and fish oil therapy using pure icosapent ethyl all reduce ASCVD risk, including among those with DM [4,5,6]. US and other international guidelines recommend statin therapy for all adults with DM, with high-intensity statins for those at higher risk and icosapent ethyl (pure EPA fish oil) for those at higher risk who have elevated triglycerides [5,7]. US and European guidelines for the management of dyslipidemias now include the use of PCSK9 inhibitors for very high-risk ASCVD patients (with or without DM) who are not adequately controlled for LDL-C on a maximum tolerated dose of statin and ezetimibe [8,9,10]. Data on the extent of dyslipidemia and lipid target attainment, as well as on the use of statin and newer non-statin therapies, are limited among recent real-world cohorts of diverse patient populations. The aim of our study was to examine disparities in lipid control and use of statin and newer lipid therapies according to sociodemographic and ASCVD risk groups in a large cohort of patients with DM representative of the diversity of the United States. Key objectives were to examine differences in [1] LDL-C and triglyceride control by sociodemographic and ASCVD risk groups, and [2] the use of statin, ezetimibe, PCSK9 inhibitor, and icosapent ethyl by sociodemographic and ASCVD risk groups. ## 2.1. All of Us Research Program The mission of the All of Us Research Program is to accelerate health research and medical breakthroughs, enabling individualized prevention, treatment, and care [11]. The All of Us Research Program is an ongoing program that aims to invite 1 million adults across the United States. There are currently over 541,000 participants that have been recruited from 590+ sites. Over $50\%$ of these participants represent racial and ethnic minorities, and over $80\%$ of them are underrepresented in biomedical research [11]. This work was performed on data collected using the All of Us Researcher Workbench, a cloud-based platform where approved researchers can access and analyze data [11]. The data currently includes surveys, electronic health records (EHR) data, and physical measurements (PM). Participants could choose not to answer specific questions. PM recorded at enrollment include systolic and diastolic blood pressure, height, weight, heart rate, waist and hip measurement, wheelchair use, and current pregnancy status. EHR data was linked for those participants who consented [11]. All participants provided informed consent to participate in the All of Us research program. The current analysis utilized de-identified data. ## 2.2. Study Sample On the researcher workbench, we created a cohort of 81,332 participants aged ≥ 18 years enrolled between 2018 and 2022 with DM based on ≥1 of the following from recorded personal or medical history: DM, DM without complications, type 2 DM, different diseases/conditions due to DM, secondary DM, on insulin treatment or DM medication, HbA1c ≥ $6.5\%$, fasting glucose ≥ 126 mg/dL, or non-fasting glucose ≥ 200 mg/dL. We excluded participants with Type 1 DM and variables with missing values in our analysis from participants. Ethnicity within our cohort included non-Hispanic White, non-Hispanic Black, Hispanic or Latino, Asian, and other. We categorized our ASCVD risk groups as moderate risk based on ≤1 CVD risk factor, high risk with ≥2 CVD risk factors, and DM with known ASCVD. Risk factors included were age ≥60 years, hypertension (blood pressure ≥ $\frac{130}{80}$ mmHg or being on antihypertensive therapy), low-density lipoprotein cholesterol (LDL-C) ≥ 160 mg/dL, cigarette smoking, and high-density lipoprotein cholesterol (HDL-C) < 40 mg/dL for males and <50 mg/dL for females (Table 1). We also analyzed these parameters across health insurance status, education, and income categories. ## 2.3. Definitions and Measurements We extracted information on each subject on demographics, survey data, cholesterol, LDL-C, and triglyceride levels, as well as use of statins and PCSK9 inhibitor use. ASCVD was defined based on all listed manifestations of coronary artery disease, cerebrovascular disease (excluding hemorrhagic stroke), and peripheral arterial disease. Statin use was defined as a documented prescription (generic or branded) of atorvastatin, cerivastatin, fluvastatin, lovastatin, pitavastatin, pravastatin, rosuvastatin, and/or simvastatin. Statin intensity was categorized into those at high and low/moderate intensities according to US guidelines [12]. Ezetimibe and icosapent ethyl use was also captured, and PCSK9 inhibitors included evolocumab and alirocumab. We additionally obtained survey data on health insurance status, types of health insurance, BMI, education level, cigarette smoking status, and income. ## 2.4. Statistical Analyses R programming was used for statistical analysis, utilizing the All of Us Research Program participants to project estimates to the US population. The Chi-squared test of proportions was used to compare icosapent ethyl and statin use according to risk group, sex, and ethnicity. We examined the percentage of people on low-, moderate-, and high-intensity statin therapy, and at LDL-C levels less than 70 mg/dL, 70–99 mg/dL, and 100 mg/dL or greater. The percentage of people on icosapent ethyl and with triglyceride levels less than 100 mg/dL, 100–149 mg/dL, 150 mg/dL to 199 mg/dL, and 200 mg/dL and greater were also analyzed using the Chi-squared test of proportions. We then performed logistic regressions that assessed the relation of demographic factors to high-statin, ezetimibe, PCSK9 inhibitor, and icosapent ethyl uses. Multiple logistic regressions were used to assess the relation of predetermined sociodemographic factors, risk groups, and individual risk factors, with odds ratios (ORs) and $95\%$ confidence intervals calculated. The p-values shown represent the significance levels across the strata of interest (e.g., sex, ethnicity, or DM risk group). ## 3. Results Our analysis includes 81,332 participants diagnosed with DM based on our inclusion criteria. Overall, $31.1\%$, $30.3\%$, and $38.6\%$ were at moderate risk, high risk, or with ASCVD, respectively. Our sample also comprised $22.3\%$ non-Hispanic Black, $17.2\%$ Hispanic or Latino, $52.3\%$ non-Hispanic White, and $1.8\%$ Asian participants, as well as $40.6\%$ males and $59.4\%$ females. Overall, $4.4\%$ did not have health insurance, and $34.1\%$ had a high school education or less (Table 1). Table 2 shows how the use of different therapies for dyslipidemia varied by risk and demographic groups. Within risk groups, sex, and ethnicity, there were significant differences in the use of statins. Approximately $33.5\%$ of people who have DM and ASCVD were not using any statins. High-intensity statin use also varied among groups, ranging from $5.9\%$ in those at lower risk to $18.2\%$ in those with DM and ASCVD ($p \leq 0.05$). Furthermore, across all risk groups, use of PCSK9 inhibitors and icosapent ethyl was universally low, being highest at $1.3\%$ and $1.7\%$, respectively, in those with both DM and ASCVD. Approximately $1.9\%$ of participants with TG levels greater than or equal to 150 mg/dL were on icosapent ethyl. A total of $5.1\%$ of participants were on ezetimibe ($p \leq 0.05$). Overall, $50.6\%$ of our participants had LDL-C levels < 100 mg/dL, although only $16.0\%$ were <70 mg/dL (Table 2). Figure 1 shows the proportion of participants with LDL-C < 70 mg/dL, 70–99 mg/dL, and ≥100 mg/dL according to sociodemographic and ASCVD risk groups. A total of $55.5\%$ of those with ≥ 2 risk factors had LDL-C ≥ 100 mg/dL, whereas $40.9\%$ of those with DM and ASCVD had LDL-C ≥ 100 mg/dL (with only $21.1\%$ having LDL-C < 70 mg/dL). A total of $56.3\%$ of females had LDL-C ≥ 100 compared to $39.3\%$ of males. Of the participants who had health insurance, $49.5\%$ had LDL-C ≥ 100 mg/dL, compared to $53.2\%$ of participants who had no insurance and had LDL-C ≥ 100 mg/dL. Overall, $64.4\%$ had triglyceride levels < 150 mg/dL, and only $31.6\%$ had levels < 100 mg/dL (Table 2). Figure 2 shows the proportion of participants with triglyceride levels of <100 mg/dL, 100–149 mg/dL, 150–199 mg/dL, and ≥200 mg/dL by sociodemographic and risk groups. A total of $21.3\%$ of participants with 2 or more risk factors had triglyceride levels ≥ 200 mg/dL, whereas $17.9\%$ of those with DM and ASCVD had triglyceride levels ≥ 200 mg/dL. A total of $15.6\%$ of females had triglyceride levels ≥ 200 mg/dL compared to $19.6\%$ of males. Additionally, $16.9\%$ of participants with health insurance had triglyceride levels ≥ 200 mg/dL, compared to $25.1\%$ of those without health insurance. Table 3 shows significant differences in the prevalence of high-intensity statin, ezetimibe, PCSK9 inhibitor, and icosapent ethyl use across health insurance, education, and income. For those with health insurance, $27.5\%$ were on high-intensity statins, compared to $23.7\%$ without health insurance. Ezetimibe use was greater in those with health insurance at $5.3\%$ compared to $1.4\%$ in those without health insurance, as was PCSK9 inhibitor use ($0.6\%$ and $0.1\%$, respectively). Moreover, $3.1\%$, $0.2\%$, and $0.5\%$ of participants with less than a high school degree were on ezetimibe, PCSK9 inhibitors, and icosapent ethyl, respectively. For those with a college or advanced degree, this was $6.4\%$, $0.8\%$, and $1.0\%$, respectively. Ezetimibe use was more common in those at higher versus lower income levels. Multiple logistic regression (Table 4) showed males to be significantly more likely to be on icosapent ethyl (OR = 2.98 [2.03, 4.48]) and high-intensity statins (OR = 1.73 [1.62, 1.85]) compared to females. Non-Hispanic Black participants were significantly less likely to be on icosapent ethyl (OR = 0.22 [0.12, 0.38]) and ezetimibe (OR = 0.62 [0.54, 0.72]) than non-Hispanic White participants, but were more likely to be on PCSK9 inhibitors and high-intensity statins. High-intensity statin use was significantly more likely in participants with hypertension (OR = 1.13 [1.07, 1.19]), and those with LDL-C ≥ 160 mg/dL (OR = 1.63 [1.43, 1.86). Ezetimibe use was significantly more likely in participants ≥60 years (OR = 1.27 [1.05, 1.54]) and among those with health insurance (OR = 1.52 [1.03, 2.35]). Hispanic or Latino participants were significantly less likely to be taking ezetimibe. Those with DM and ASCVD were significantly more likely to be on a high-intensity statin (OR = 3.66 [3.37, 3.97]) and ezetimibe (OR = 3.12 [2.66, 3.67]) as well as icosapent ethyl (OR = 2.21 [1.44, 3.47]). ## 4. Discussion We demonstrated continuing gaps in lipid treatment and inadequate control of LDL-C and triglycerides in an important current real-world cohort of US adults with DM. We analyzed these gaps across ASCVD risk groups and key underserved demographic groups of participants within the NIH Precision Medicine Initiative’s All of Us Study who have been underrepresented in health research. We found that LDL-C and triglyceride levels remain inadequately controlled, including among people with ASCVD, who despite having the strongest recommendations for treatment, remain suboptimally treated with high-intensity statins, ezetimibe, PCSK9i, and icosapent ethyl. Among participants with both DM and ASCVD, only $21.1\%$ had LDL-C < 70 mg/dL and $36.5\%$ had triglyceride levels ≥ 150 mg/dL, respectively. Additionally, ezetimibe, PCSK9i, and icosapent ethyl, while not widely used, were most prevalent among those with a college degree or higher, and PCSK9i was most used in those with health insurance. Furthermore, ezetimibe, PCSK9 inhibitor, and icosapent ethyl use were highest among non-Hispanic White populations compared to other minority racial/ethnic groups. These results are concerning because Hispanic or Latino populations and non-Hispanic Black populations had the highest proportions with LDL-C levels ≥ 100 mg/dL, and Hispanic or Latino populations and Asian populations had the highest proportion of uncontrolled triglyceride levels of 150 mg/dL or higher. Others have also shown minority groups are more likely to have high triglyceride levels and low HDL-C dyslipidemia [13]. In the US Diabetes Collaborative Registry [2], we recently showed Black persons to be less likely to be at LDL-C target ($42.7\%$) compared to White persons ($49.3\%$). Moreover, from analysis of electronic health record data from a large healthcare system [14], among those with diabetes, Black persons had a $36\%$ lower likelihood of being prescribed a statin compared to White persons in adjusted analysis. The Reasons for Geographic and Racial Differences in Stroke (REGARDS) study similarly showed underutilization of statins in non-Hispanic Black populations compared to non-Hispanic White populations [15]. Clinical trials have documented the efficacy of statin and ezetimibe therapy as well as PCSK9 inhibitors and icosapent ethyl, including among persons with DM. In 14 randomized statin trials, which included 18,686 people, researchers found that people with DM who were on statins for an average of 4.3 years had a $21\%$ decrease in major vascular events and a $9\%$ decrease in mortality compared to those who were not on statins [4]. Further reduction of LDL-C not satisfactorily achieved by high-intensity statins can be achieved by ezetimibe or PCSK9 inhibitors [4]. In the IMPROVE-IT trial comparing the addition of ezetimibe to statins alone in persons with a recent acute coronary syndrome, in subgroup analyses, those with DM (in addition to the recent acute coronary syndrome) compared to those without DM had a substantially greater reduction in risk of the primary composite cardiovascular endpoint [16]. In the Fourier trial of evolocumab in persons with prior ASCVD, pre-specified subgroup analyses showed that among the 11,031 ($40\%$) patients with DM, there was a similar $17\%$ risk reduction of the primary cardiovascular endpoint compared to the $13\%$ risk reduction in those without DM (interaction term not significant) [5]. Another study found that the rosuvastatin/ezetimibe combination is safe and effective in patients with hypercholesterolemia or dyslipidemia with or without DM and with or without cardiovascular disease [17,18]. The drug combination enabled higher proportions of patients to achieve recommended LDL-C goals than rosuvastatin monotherapy, without additional adverse events [17,18]. However, despite statin use, people with well-controlled LDL still have residual ASCVD risk associated in part with elevated triglycerides that may be lowered by omega-3 fatty acids, such as icosapent ethyl [5] or fibrate therapy. In the REDUCE-IT trial testing the efficacy of icosapent ethyl in persons with prior ASCVD or DM and multiple risk factors with triglycerides of 135–499 mg/dL on statin therapy, those with vs. without DM had a similar risk reduction in the primary endpoint ($23\%$ vs. $27\%$, with the interaction term not significant) [8]. The recently reported RESPECT-EPA trial [19], while of borderline significance for the primary endpoint, did achieve the secondary endpoint, with relative risk reductions due to icosapent ethyl therapy consistent with REDUCE-IT. However, the recently reported PROMINENT trial [20] involving pemafibrate failed to demonstrate any benefit from this therapy in reducing ASCVD risk in persons with DM who had elevated triglycerides and low HDL-C, and instead showed increased LDL-C levels in the treated group. Recent real-world evidence from population studies in those with DM shows use of lipid-lowering therapy is still limited, and acceptable LDL-C levels are often not achieved. While our recent report from the National Health and Nutrition Examination Survey 2013-2016 did show more than $80\%$ of those with DM were on lipid-lowering therapy, only $57\%$ (among those without ASCVD) had an LDL-C < 100 mg/L and only $26\%$ of those who had both DM and CVD had an LDL-C < 70 mg/dL [21]. Moreover, our recent report from the Diabetes Collaborative Registry showed that $49\%$ of those with DM were at LDL cholesterol targets < 100 mg/dL or < 70 mg/dL if with ASCVD, with two-thirds of these on moderate or high-intensity statins [2]. Our results from the All of Us cohort show lower levels of lipid treatment, as well as lower levels at appropriate LDL-C levels, likely due to the greater proportions of underrepresented and/or inadequately insured persons in our cohort. We have previously demonstrated in US adults with DM that despite statin therapy, triglycerides of ≥150 mg/dL are still present in $40\%$, and even if LDL-C < 100 mg/dL in those on statin therapy, more than a third of such persons still have triglycerides ≥ 150 mg/dL, warranting the consideration of additional triglyceride reducing therapies [3]. We found icosapent ethyl use to be only $1.9\%$ among participants with triglyceride levels greater than or equal to 150 mg/dL. This low use is consistent with other recent real-world data. A recent study by Derington et al. created cohorts using the National Health and Nutrition Examination Surveys (NHANES) 2009–2014 and the Optum Research Database (ORD) to see how many participants were eligible to receive icosapent ethyl [22]. They estimated 3.6 million US adults to be eligible and observed that the 5-year first event (composite of cardiovascular death, nonfatal myocardial infarction, nonfatal stroke, unstable angina requiring hospitalization, or coronary revascularization) rate without IPE was $19.0\%$ compared to $13.1\%$ with 5 years of IPE treatment, preventing 212,000 events. They also projected that the total 5-year event rate (first and recurrent) could be reduced from $42.5\%$ to $28.9\%$ with 5 years of IPE therapy, preventing around 490,000 events, which would amount to approximately USD 2.6 billion in net annual cost. In addition, because icosapent ethyl was approved for ASCVD risk reduction by the FDA recently in December of 2019, it is not surprising that uptake is low in the current study, especially given the wide range of demographic groups included in the All of Us research program. While our results show those with both DM and ASCVD were most likely to be on high-intensity statins, ezetimibe, and icosapent ethyl compared to people with DM who did not have ASCVD, their use was still suboptimal. High-intensity statins are recommended for those with DM and ASCVD [23,24], with further non-statin therapy indicated for further LDL-C lowering. Only $18.2\%$ of our patients with DM and ASCVD were on high-intensity statins, and only $9.1\%$, $1.3\%$, and $1.7\%$ were on ezetimibe, PCSK9 inhibitors, and icosapent ethyl, respectively. Our study has some strengths and limitations. The participants in this study reflect the diversity of the United States and the data are available in near-real time, which is valuable when trying to understand current lipid treatment and control patterns. While the data are extracted from an on-line platform for analysis, these data are from the NIH Precision Medicine Initiative All of Us Research Program that does have standardized methods for data collection regarding surveys and blood measurements. However, like with most research studies, participation is voluntary and thus the sample studied, while large, is not necessarily representative of the US population. Moreover, this is a cross-sectional study and we do not have multiple measures of medication use to assess adherence nor multiple laboratory measures to examine the effects of individual therapies, which would require a clinical trial design. There are also other limitations in using electronic health records (EHR) data, where there may be inconsistencies across study sites in capturing prescription and diagnostic data. Additionally, assuming the absence of a diagnostic code as an absence of disease may lead to information and/or selection bias. Further, it has been demonstrated that one key source of bias in EHRs is “informed presence” bias, where those with more medical encounters are more likely to be diagnosed with various conditions [25,26]. Lastly, as our study population is enriched in underserved and disadvantaged persons, results may differ compared to results from health claims data from insured persons. ## 5. Conclusions In summary, our cross-sectional analysis demonstrates important disparities in lipid control, as well as in the use of statins, ezetimibe, PCSK9 inhibitors, and icosapent ethyl in US adults with DM across sociodemographic and DM risk groups. Guideline-recommended use of high-intensity statins and ezetimibe among our higher risk DM patients is lacking, with many having inadequately controlled LDL-C levels. Moreover, icosapent ethyl use remains low, even among those with high TG levels. Continued provider and patient education needs to be prioritized—especially among those at highest risk. However, systematic approaches, including the use of EHR and other automated interventions, are needed to address both remaining clinical inertia and significant remaining gaps between evidence-based guidelines and actual care received. ## References 1. Leon B.M., Maddox T.M.. Diabetes and cardiovascular disease: Epidemiology, biological mechanisms, treatment recommendations and future research. World J. Diabetes (2015.0) 6 1246-1258. DOI: 10.4239/wjd.v6.i13.1246 2. Fan W., Song Y., Inzucchi S.E., Sperling L., Cannon C.P., Arnold S.V., Kosiborod M., Wong N.D.. Composite cardiovascular risk factor target achievement and its predictors in US adults with diabetes: The Diabetes Collaborative Registry. Diabetes Obes. Metab. (2019.0) 21 1121-1127. DOI: 10.1111/dom.13625 3. Fan W., Philip S., Granowitz C., Toth P.P., Wong N.D.. Residual Hypertriglyceridemia and Estimated Atherosclerotic Cardiovascular Disease Risk by Statin Use in U.S. Adults With Diabetes: National Health and Nutrition Examination Survey 2007-2014. Diabetes Care (2019.0) 42 2307-2314. DOI: 10.2337/dc19-0501 4. Kearney P.M., Blackwell L., Collins R., Keech A., Simes J., Peto R., Armitage J., Baigent C.. Efficacy of cholesterol-lowering therapy in 18,686 people with diabetes in 14 randomised trials of statins: A meta-analysis. Lancet (2008.0) 371 117-125. DOI: 10.1016/S0140-6736(08)60104-X 5. Bhatt D.L., Steg P.G., Miller M., Brinton E.A., Jacobson T.A., Ketchum S.B., Doyle R.T., Juliano R.A., Jiao L., Granowitz C.. Cardiovascular Risk Reduction with Icosapent Ethyl for Hypertriglyceridemia. N. Engl. J. Med. (2019.0) 380 11-22. DOI: 10.1056/NEJMoa1812792 6. Sabatine M.S., Leiter L.A., Wiviott S.D., Giugliano R.P., Deedwania P., De Ferrari G.M., Murphy S.A., Kuder J.F., Gouni-Berthold I., Lewis B.S.. Cardiovascular safety and efficacy of the PCSK9 inhibitor evolocumab in patients with and without diabetes and the effect of evolocumab on glycaemia and risk of new-onset diabetes: A prespecified analysis of the FOURIER randomised controlled trial. Lancet Diabetes Endocrinol. (2017.0) 5 941-950. DOI: 10.1016/S2213-8587(17)30313-3 7. Ganda O.P., Bhatt D.L., Mason R.P., Miller M., Boden W.E.. Unmet Need for Adjunctive Dyslipidemia Therapy in Hypertriglyceridemia Management. J. Am. Coll. Cardiol. (2018.0) 72 330-343. DOI: 10.1016/j.jacc.2018.04.061 8. Schwartz G.G., Steg P.G., Szarek M., Bhatt D.L., Bittner V.A., Diaz R., Edelberg J.M., Goodman S.G., Hanotin S., Har-rington R.A.. Alirocumab and Cardiovascular Outcomes after Acute Coronary Syndrome. N. Engl. J. Med. (2018.0) 379 2097-2107. DOI: 10.1056/NEJMoa1801174 9. Gallego-Colon E., Daum A., Yosefy C.. Statins and PCSK9 inhibitors: A new lipid-lowering therapy. Eur. J. Pharmacol. (2020.0) 878 173114. DOI: 10.1016/j.ejphar.2020.173114 10. Sabatine M.S., Giugliano R.P., Keech A.C., Honarpour N., Wiviott S.D., Murphy S.A., Kuder J.F., Wang H., Liu T., Wasserman S.M.. Evolocumab and Clinical Outcomes in Patients with Cardiovascular Disease. N. Engl. J. Med. (2017.0) 376 1713-1722. DOI: 10.1056/NEJMoa1615664 11. The All of Us Research Program. (2021.0) 12. Force U.P.S.T., Mangione C.M., Barry M.J., Nicholson W.K., Cabana M., Chelmow D., Coker T.R., Davis E.M., Donahue K.E., Jaén C.R.. Statin Use for the Primary Prevention of Cardiovascular Disease in Adults: US Preventive Services Task Force Recommendation Statement. JAMA (2022.0) 328 746-753. DOI: 10.1001/jama.2022.13044 13. Frank A.T., Zhao B., Jose P.O., Azar K.M., Fortmann S.P., Palaniappan L.P.. Racial/ethnic differences in dyslipidemia patterns. Circulation (2014.0) 129 570-579. DOI: 10.1161/CIRCULATIONAHA.113.005757 14. Dorsch M.P., Lester C.A., Ding Y., Joseph M., Brook R.D.. Effects of Race on Statin Prescribing for Primary Prevention With High Atherosclerotic Cardiovascular Disease Risk in a Large Healthcare System. J. Am. Heart Assoc. (2019.0) 8 e014709. DOI: 10.1161/JAHA.119.014709 15. Schroff P., Gamboa C.M., Durant R.W., Oikeh A., Richman J.S., Safford M.M.. Vulnerabilities to Health Disparities and Statin Use in the REGARDS (Reasons for Geographic and Racial Differences in Stroke) Study. J. Am. Heart Assoc. (2017.0) 6 e005449. DOI: 10.1161/JAHA.116.005449 16. Giugliano R.P., Cannon C.P., Blazing M.A., Nicolau J.C., Corbalán R., Špinar J., Park J.G., White J.A., Bohula E.A., Braunwald E.. Benefit of Adding Ezetimibe to Statin Therapy on Cardiovascular Outcomes and Safety in Patients With Versus Without Diabetes Mellitus: Results From IMPROVE-IT (Improved Reduction of Outcomes: Vytorin Efficacy International Trial). Circulation (2018.0) 137 1571-1582. DOI: 10.1161/CIRCULATIONAHA.117.030950 17. Boutari C., Karagiannis A., Athyros V.G.. Rosuvastatin and ezetimibe for the treatment of dyslipidemia and hypercholesterolemia. Expert Rev. Cardiovasc. Ther. (2021.0) 19 575-580. DOI: 10.1080/14779072.2021.1940959 18. Lazarte J., Hegele R.A.. Dyslipidemia Management in Adults With Diabetes. Can. J. Diabetes (2020.0) 44 53-60. DOI: 10.1016/j.jcjd.2019.07.003 19. 19.Available online: https://www.acc.org/Latest-in-Cardiology/Articles/2022/11/01/22/00/sun-7pm-respect-epa-aha-2022(accessed on 1 February 2023) 20. Das Pradhan A., Glynn R.J., Fruchart J.C., MacFadyen J.G., Zaharris E.S., Everett B.M., Campbell S.E., Oshima R., Amarenco P., Blom D.J.. Triglyceride Lowering with Pemafibrate to Reduce Cardiovascular Risk. N. Engl. J. Med. (2022.0) 387 1923-1934. DOI: 10.1056/NEJMoa2210645 21. Andary R., Fan W., Wong N.D.. Control of Cardiovascular Risk Factors Among US Adults With Type 2 Diabetes With and Without Cardiovascular Disease. Am. J. Cardiol. (2019.0) 124 522-527. DOI: 10.1016/j.amjcard.2019.05.035 22. Derington C.G., Bress A.P., Herrick J.S., Fan W., Wong N.D., Andrade K.E., Johnson J., Philip S., Abrahamson D., Jiao L.. The potential population health impact of treating REDUCE-IT eligible US adults with Icosapent Ethyl. Am J Prev Cardiol. (2022.0) 10 100345. DOI: 10.1016/j.ajpc.2022.100345 23. Grundy S.M., Stone N.J., Blumenthal R.S., Braun L.T., Heidenreich P.A., Lloyd-Jones D., Orringer C.E., Saseen J.J., Smith S.C., Sperling L.S.. High-Intensity Statins Benefit High-Risk Patients: Why and How to Do Better. Mayo Clin. Proc. (2021.0) 96 2660-2670. DOI: 10.1016/j.mayocp.2021.02.032 24. Draznin B., Aroda V.R., Bakris G., Benson G., Brown F.M., Freeman R., Green J., Huang E., Isaacs D.. 9. Pharmacologic Approaches to Glycemic Treatment: Standards of Medical Care in Diabetes-2022. Diabetes Care (2022.0) 45 S125-S143. DOI: 10.2337/dc22-S009 25. Gianfrancesco M.A., Goldstein N.D.. A narrative review on the validity of electronic health record-based research in epidemiology. BMC Med. Res. Methodol. (2021.0) 21. DOI: 10.1186/s12874-021-01416-5 26. Pareek M., Mason R.P., Bhatt D.L.. Icosapent ethyl: Safely reducing cardiovascular risk in adults with elevated triglycerides. Expert Opin. Drug Saf. (2022.0) 21 31-42. DOI: 10.1080/14740338.2021.1954158
--- title: 'Metformin-NSAIDs Molecular Salts: A Path towards Enhanced Oral Bioavailability and Stability' authors: - Francisco Javier Acebedo-Martínez - Alicia Domínguez-Martín - Carolina Alarcón-Payer - Carolina Garcés-Bastida - Cristóbal Verdugo-Escamilla - Jaime Gómez-Morales - Duane Choquesillo-Lazarte journal: Pharmaceutics year: 2023 pmcid: PMC9966766 doi: 10.3390/pharmaceutics15020449 license: CC BY 4.0 --- # Metformin-NSAIDs Molecular Salts: A Path towards Enhanced Oral Bioavailability and Stability ## Abstract According to the World Health Organization, more than 422 million people worldwide have diabetes. The most common oral treatment for type 2 diabetes is the drug metformin (MTF), which is usually formulated as a hydrochloride to achieve higher water solubility. However, this drug is also highly hygroscopic, thus showing stability problems. Another kind of worldwide prescribed drug is the non-steroidal anti-inflammatory drug (NSAID). These latter, on the contrary, show a low solubility profile; therefore, they must be administered at high doses, which increases the probability of secondary effects. In this work, novel drug-drug pharmaceutical solids combining MTF-NSAIDs have been synthesized in solution or by mechanochemical methods. The aim of this concomitant treatment is to improve the physicochemical properties of the parent active pharmaceutical ingredients. After a careful solid-state characterization along with solubility and stability studies, it can be concluded that the new molecular salt formulations enhance not only the stability of MTF but also the solubility of NSAIDs, thus giving promising results regarding the development of these novel pharmaceutical multicomponent solids. ## 1. Introduction According to the World Health Organization (WHO), diabetes is defined as a “chronic, metabolic disease characterized by elevated levels of blood glucose” [1]. WHO estimates as 422 million the number of worldwide adult patients diagnosed with diabetes, and this number rises to 537 million according to the International Diabetes Federation [2]. In 2045, the number of patients with diabetes is believed to reach 784 million, therefore being diabetes a real challenge for healthcare systems [1,2]. There are two types of diabetes. Type 1 diabetes, once so-called juvenile diabetes, occurs when the pancreas does not produce insulin, whereas type 2 diabetes occurs when the body becomes resistant to insulin, and it most often develops in adults. Nonetheless, the prevalence of overweight individuals and obesity in the younger age populations, along with insufficient physical activity, is directly related to the increasing occurrence of type 2 diabetes in younger populations, too [1,3]. Metformin (MTF) is a biguanide antihyperglycemic agent (Scheme 1) used as a first-line pharmacological treatment in the management of type 2 diabetes due to its efficacy, safety profile, and low cost for patients [4,5]. MTF exerts its antihyperglycemic action by increasing insulin sensitivity in peripheral tissues and reducing hepatic gluconeogenesis [6]. It is most commonly prescribed alone as metformin hydrochloride (MTF·HCl), i.e., chemically defined as a salt, in order to increase solubility and stability [7,8,9]. Although MTF·HCl solubility is outstanding [10], the low gastrointestinal permeability of MTF [11] eventually leads to gastrointestinal disorders in about 20–$30\%$ of chronic patients due to accumulation in the enterocytes within the small intestine [12,13]. Regarding the molecular stability of metformin, the instability profile and degradation products are detailed in its official Pharmacopeia monograph [14]. In addition, several stability assessments on aqueous media agree on metformin’s instability in alkaline solutions and its being less sensitive to oxidants [15,16]. Unfortunately, glycemic control with MTF in type 2 diabetes is still a challenge that concerns physicians due to its complexity [17]. Hence, its prescription, along with other antidiabetic drugs, is not rare in clinics. Non-steroidal anti-inflammatory drugs (NSAIDs) work to relieve pain, reduce inflammation, and bring down fever [18], thus being one of the most commonly prescribed medications worldwide. Despite their chemical diversity, they are generally poorly soluble [18], which is an important drawback of this kind of drug. This fact requires an increase in their dosage, increasing the probability of side effects and drug interactions. Therefore, enhancing the solubility of NSAIDs is also quite an active research area within the pharmaceutical industry. Diabetic neuropathy is a serious diabetes complication in which overall peripheral nerves are affected, whose symptoms include pain and numbness in extremities [19]. Thereby the concomitant administration of MTF and NSAIDs is very frequent [20], even if NSAIDs are not considered especially effective at treating neuropathic pain unless there is inflammation [19]. Despite the latter, their massive use is a reality in clinics, most likely due to: (i) the lack of non-opioid alternatives in the treatment of mild-moderate pain, (ii) the ‘over the counter’ (OTC) character of several NSAIDs, and (iii) their safety profile. Even if they are generally regarded as safe, a rare, although serious, condition known as lactic acidosis has been reported for long-term MTF treatments connected to acute renal failure [21,22]. On the other hand, NSAIDs are known for their nephrotoxicity [23]. Therefore, patients with concomitant and chronic treatments of MTF and NSAIDs are more susceptible to developing renal impairment [24]. In this work, crystal engineering tools have been used to yield drug–drug multicomponent pharmaceutical solids. This novel approach is rather interesting for the pharmaceutical industry since it allows for the improvement of the physicochemical properties of parent active pharmaceutical ingredients (APIs) without modifying their chemical structure, thus being a relatively inexpensive method compared to the development of a complete drug pipeline [25]. The proposed novel formulations involve the antidiabetic drug metformin (MTF) and five different NSAIDs, i.e., Niflumic acid (NIF), Diflunisal (DIF), Mefenamic acid (MEF), Tolfenamic acid (TLF), and Flurbiprofen (FLF) (Scheme 1). The aim of the study is to explore novel treatment alternatives that might increase the oral bioavailability of the aforementioned drugs via the formation of multidrug salts, as well as reduce the side effects associated with this combined therapy. To this purpose, five (MTF)(NSAIDs) salts have been synthesized. Structure-physicochemical properties relationships of the new multicomponent pharmaceutical solids have been assessed thanks to a comprehensive solid-state analysis of the crystallographic structures. Moreover, solubility and stability have been evaluated in the new formulations and compared to the isolated parent APIs, showing promising results. ## 2.1. Materials All APIs and solvents used in this work were purchased from commercial sources and used as received. Metfomin·HCl, tolfenamic acid, niflumic acid, and flurbiprofen were obtained from TCI Europe (Zwijndrecht, Belgium). Diflunisal and mefenamic acid were obtained from Sigma-Aldrich (St. Louis, CA, USA). Ethyl acetate HPLC grade and absolute ethanol were obtained from labkem (Barcelona, Spain). ## 2.2.1. Metformin Hydrochloride Neutralization To obtain the MTF base, MTF·HCl was neutralized by stirring 10 mmol of MTF·HCl (1.656 g) and 10 mmol of NaOH (0.4 g) in 60 mL of isopropanol at room temperature, using a magnetic stirrer and a sealed glass beaker to avoid evaporation. After 24 h, the solution was filtered using 0.22 µm syringe filters to remove the NaCl (insoluble in isopropanol). The solvent of the clear filtered obtained (containing the MTF base) was removed using a rotatory evaporator set at 50 °C and 30 rpm. The remaining powder was dried and characterized by powder X-ray diffraction (PXRD) to confirm the purity of the MTF base obtained. ## 2.2.2. Mechanochemical Synthesis The mechanochemical synthesis of MTF salts was conducted using Liquid-Assisted Grinding (LAG) in a Retsh MM2000 ball mill (Retsch, Haan, Germany), operating at a 25-Hz frequency, using stainless steel jars along with stainless steel balls 7 mm in diameter. MTF–MEF was obtained by LAG of a mixture of MTF (0.5 mmol, 64.58 mg) and MEF (0.5 mmol, 120.64 mg) in a 1:1 stoichiometric ratio with 100 µL of ultrapure water as a liquid additive. MTF–FLP and MTF–TLF were obtained using LAG of a mixture of MTF (0.5 mmol, 64.58 mg) and the respective coformer (130.85 mg of TLF, 122.13 mg of FLP) in a 1:1 stoichiometric ratio, along with 100 µL of ethanol as a liquid additive. MTF–DIF and MTF–NIF hydrates were obtained using LAG of a mixture of MTF (0.5 mmol, 64.58 mg) and the respective coformer (125.10 mg of DIF, 141.11 mg of NIF) in a 1:1 stoichiometric ratio, along with 100 µL of ultrapure water as a liquid additive. To obtain MTF–DIF and MTF–NIF as anhydrated salts, the product of the mechanochemical synthesis was heated at 100 °C for 2 h. All reaction syntheses lasted for 30 min and were repeated to ensure reproducibility. Bulk materials were further evaluated using PXRD to determine the salt formation. ## 2.2.3. Preparation of Single Crystals Single crystals of MTF–based salts were obtained by dissolving the product of the LAGs in ethanol (for MTF–NIF·2H2O, MTF–MEF, MTF–FLP, and MTF–TLF) and ethyl acetate (for MTF–NIF). MTF–DIF single crystals were obtained using a hydrothermal reaction by placing a mixture of 0.2 mmol of MTF (25.8 mg) and 0.2 mmol of DIF (50.4 mg) in a hydrothermal reactor along with 3 mL of distilled water. The reactor was sealed and heated at 110 °C for 24 h. After cooling down to room temperature, the reactor was opened, and single crystals were separated from the solution for further characterization. ## 2.3. X-ray Diffraction Analysis Single-crystal X-ray diffraction (SCXRD) data were acquired at room temperature on a Bruker D8 Venture diffractometer (Bruker-AXS, Karlsruhe, Germany) using CuKα radiation (λ = 1.54178 Å). The data were processed with the APEX4 suite [26]. The structures were solved with intrinsic phasing (SHELXT) [27] and refined with full-matrix least squares on F2 [28] using Olex2 as a graphical interface [29]. The non-hydrogen atoms were refined anisotropically. For all structures, hydrogen atoms were located in difference Fourier maps and included as fixed contributions riding on attached atoms with isotropic thermal displacement parameters 1.2 or 1.5 times those of the respective atom. Mercury [30], Platon [31], and Olex2 [29] were used for the analysis and visualization of the structures and also for graphic material preparation. All deposited CIF files are in the Cambridge Structural Database (CSD) under the CCDC numbers 2232928-2232933. Copies of the data can be obtained free of charge at https://www.ccdc.cam.ac.uk/structures/ (accessed on 29 December 2022). Powder X-ray diffraction (PXRD) analysis was performed at room temperature on a Bruker D8 Advance Vαrio diffractometer (Bruker-AXS, Karlsruhe, Germany) diffractometer equipped with a LYNXEYE detector and CuKα1 radiation (1.5406 Å). The diffractograms were collected over an angular range of 5–40° (2θ) with a step size of 0.02° (2θ) and a constant counting time of 5 s per step. ## 2.4. Differential Scanning Calorimetry A differential scanning calorimetry (DSC) study was performed with a Mettler-Toledo SC-822e calorimeter (Mettler Toledo, Columbus, OH, USA). Experimental conditions: aluminum crucibles of 40 µL volume, an atmosphere of dry nitrogen with 50 mL/min flow rate, and heating rates of 1 °C/min and 10 °C/min. The calorimeter was calibrated with indium of $99.99\%$ purity (m.p.: 156.4 °C; DH: 28.14 J/g). ## 2.5. Stability Studies Stability in an aqueous solution was evaluated through slurry experiments. An excess of powder samples of each phase was added to 1 mL of buffer phosphate (pH 6.8) and stirred for 24 h in sealed vials. The solids were collected, filtered, dried, and analyzed with PXRD. Stability at accelerated aging conditions was also studied: 200 mg of solid was placed in watch glasses and left at 40 °C in $75\%$ relative humidity using a Memmert HPP110 climate chamber (Memmert, Schwabach, Germany). Under the above-accelerated aging conditions, the stability of the solid forms was periodically monitored using PXRD for two months. ## 2.6. Solubility Studies Samples for the solubility studies were prepared following the shake-flask method [32]. Saturated solutions were obtained by stirring an excess amount of APIs in 10 mL of pH 6.8 phosphate buffer at 25 °C until the thermodynamic equilibrium was reached after 24 h. Thermodynamic equilibrium was tested using UV spectroscopy by following the global solubility of the molecular salt over time until the plateau region was reached (3 h) and sustained more than 24 h. The solutions were then centrifuged, filtered through 0.22 μm polyether sulfone (PES) filters, and directly measured using high-performance liquid chromatography (HPLC). Appropriate dilutions were made to obtain measurable absorbance values. The absorbance measurements were thereafter used to quantify the MTF dissolved in each sample. The remaining solids were analyzed using PXRD to identify the crystalline phases and, thus, to check the stability of the initial crystalline phase. HPLC experiments were performed with an Agilent 1200 HPLC system (Agilent Technologies, Santa Clara, CA, USA) equipped with a solvent degasser, pump, auto-sampler, and diode array detector. A Scharlau (Barcelona, Spain) 100 C18 chromatographic column (3 μm, 150 × 4.6 mm) was the thermostat at 25 °C and used for compound separation, using an isocratic elution method. The mobile phase was composed of a mixture of $10\%$ acetonitrile ($0.1\%$ Formic acid, v/v) and $90\%$ water ($0.1\%$ Formic acid, v/v). The flow rate was 1 mL/min, and the injection volume was 10 μL. The absorbance was measured at 233 nm, i.e., the maximum absorbance for metformin. Data acquisition and analysis were performed using the software ChemStation (Agilent Technologies, Santa Clara, CA, USA). The retention time for MTF was 1 min 54 s, and the concentration for the calibration curve was determined from the area under the MTF peak. The conditions are summarized in Supplementary Table S1. ## 3.1. Salt Synthesis Mechanochemistry, especially liquid-assisted grinding (LAG), has been widely used in the pharmaceutical industry for the synthesis of new multicomponent materials. This methodology is well known to be efficient, quick, and reproducible, requiring a minimum amount of organic solvents compared with other traditional techniques [33]. In this work, LAG reactions were performed using a 1:1 stoichiometric mixture of metformin base, previously neutralized, and the corresponding NSAIDs, along with ultrapure water or ethanol as an additive solvent. After 30 min of reaction, the powder obtained in each reaction was characterized using PXRD and compared with the corresponding parent APIs to evaluate the formation of the new salts. PXRD patterns demonstrated the formation of five new phases after LAG reactions and their purity (Figure 1 and Figure S1). To obtain more information about the crystalline structure of the new phases, the polycrystalline products of the LAGs were used for recrystallization by dissolving them in ethanol and ethyl acetate. After slow solvent evaporation at room temperature, suitable crystals for single-crystal X-ray diffraction (SCXRD) were obtained. From these experiments, MTF–TLF, MTF–MEF, MTF–FLP, and MTF–NIF structures were obtained. The structural information allowed the simulation of a calculated PXRD for each salt (Figure S1), which confirmed the phase purity of the bulk products obtained from LAG except for MTF–NIF, whose calculated powder pattern did not match the product obtained from the LAG reaction, thus indicating the formation of an intermediate phase or a hydrate phase. For that reason, the product of the LAG of MTF and NIF was heated at 100 °C for 2 h. After the thermal treatment, the PXRD pattern was in good agreement with the calculated powder pattern for MTF–NIF anhydrate salt. To confirm the formation of a hydrated phase during the LAG recrystallization, single crystals of MTF–NIF·2H2O were obtained and analyzed, which allowed the use of the calculated PXRD pattern to compare with the initial mechanochemical reaction. As expected, however, the PXRD patterns of MTF–NIF·2H2O and the product obtained from LAG synthesis did not match since they are different hydrated forms (Figure S2). Similar results were obtained for MTF–DIF. Although good-quality single crystals were obtained from hydrothermal reactions, the bulk material obtained from LAG did not match the PXRD pattern of the anhydride phase. Following the previous methodology, the product of the LAG was heated at 100 °C for 2 h. After the treatment, the PXRD patterns matched perfectly. For MTF–DIF, no crystalline structure of a hydrate form was found. ## 3.2. Salt Screening The Cambridge Structural Database (CSD version 5.43, update 4 from November 2022) was searched for MTF complexes resulting in 59 hits. After excluding entries corresponding to the MTF base molecule and its inorganic salts [4] as well as MTF metal complexes [29], the remaining dataset contained 26 molecular salts ($44\%$). From the remaining molecular salts, 13 entries corresponded to MTF–drug salts, mainly antidiabetic drugs [34,35,36,37,38]. Only three salt structures containing MTF and an NSAID as counterion have been reported: MTF–diclofenac [39], MTF–salicylate [40], and MTF–aspirin [41] salts. The high number of observed molecular salts agrees with the strong basic nature of MTF. Table 1 evidences a significant difference in pKa values between the ionizable groups of MTF and the selected NSAIDs. Thereby, proton transfer is expected from the carboxylic group present in NSAIDs to an amine moiety in MTF, according to the well-known pKa rule widely used in the pharmaceutical industry [42,43,44]. Indeed, crystal structures of the multi-component drug–drug materials reported in this work confirm the molecular salt nature of these solids, which is consistent with other multicomponent pharmaceutical solids involving MTF published very recently [41,45,46]. ## 3.3. Crystal Structure Analysis of Molecular Salts Single crystals of molecular salts suitable for structural purposes were obtained, and their structures were determined using SCXRD. Crystallographic data for these salts are summarized in Table 2. Asymmetric units of the salts are represented in Figure S3, and hydrogen bond information is presented in Table S2. SCXRD data confirmed the proton transfer from the acid groups of NSAIDs to the basic nitrogen site of metformin. This finding was also evidenced in the experimental electron density map and confirmed by the analysis of the C–O bond distances of the carboxylate group of NSAIDs, with ΔDC–O values being similar to those observed in salts, in the range 0.008–0.024 Å [53]. Due to the protonation of MTF, the resulting MTF+ cation is able to participate in hetero-synthons with the NSAID anion through guanidinium···carboxylate synthons, engaged through the R228 ring motif [54,55]. ## 3.3.1. MTF–MEF and MTF–TLF Salts Molecular salts of MTF and the fenamic acids reported in this study (MEF and TLF) are isostructural (have the same crystal structure) and, in addition, isomorphous (have the same unit-cell dimensions and spacegroup [56]). The unit-cell parameters of two crystal structures were used to calculate the unit-cell similarity index Π (Equation [1]) [57]. A Π value of 0.003 confirms the isomorphous nature of the MTF–fenamate pairs. Moreover, PXRD similarity index scores for each pair and the RMSDs (root-mean-square deviations) were calculated from the packing similarity tool in Mercury [30] (overlay with 20 molecules and allowing molecular differences). The results obtained from these calculations showed that 20 out of 20 molecules were matched in the pairs of fenamate salts (PXRD similarity: 0.986; RMSD (Å): 0.131), suggesting that these molecular salts have identical intermolecular interactions and therefore affording the same crystal packing. [ 1]Π=a+b+ca′+b′+c′−1 These molecular salts crystallize in the monoclinic P21/c space group, with one monoprotonated MTF+ cation and one fenamate (MEF or TLF) anion in the asymmetric unit as ionic pair. In the molecular salt, the ions are associated through the R228 ring motif built by the COO− group and the terminal amines moiety of the MTF+ cation (Figure 2). Ionic pairs are aligned in a 1D-chain along [010] direction by the N–H⋯O H-bond (N2–H2D⋯O1) formed between the NH2 group in MTF+ cations and the COO− group of fenamate anions. Along [001] direction, adjacent chains are held together through the R428 homo dimeric motif between two MTF+ fenamate pairs related by an inversion center, generating a ribbon structure. A layered 2D structure is further generated with π,π-stacking interactions between the substituted rings of fenamate ions. The supramolecular structure is then obtained by stacking these layers with H-bonds involving amine groups of MTF+ and carboxylate groups of fenamates (Figure 2c). ## 3.3.2. MTF–NIF Salts and MTF–NIF·2H2O Hydrate Salt MTF–NIF salt crystallizes in the monoclinic system spacegroup P21/n. The asymmetric unit of this crystal phase contains one MTF+ cation and a nifluminate anion. The ion pair is associated with a charge-assisted hydrogen bond involving the guanidium moiety of MTF and the carboxylate group of NIF. As in the previously described salt structures containing MEF and TLF, adjacent ion pairs are H-bonded to build a 1D chain (along the a-axis). π,π-stacking interactions between each one of the rings of NIF reinforce the chains. A 2D layered structure is then generated through H-bonds between amine groups of MTF cations and carboxylate groups of NIF anions. Finally, the supramolecular structure is obtained by stacking these layers (along the b-axis) facing –CF3 groups of NIF (Figure 3). MTF–NIF·2H2O crystallized in the monoclinic system spacegroup P21/c. The asymmetric unit of this crystal phase contains one MTF+ cation, a niflumic anion, and two water molecules, resulting in a salt hydrate. In the asymmetric unit, the ions are connected by charge-assisted hydrogen bonds between the guanidinium moiety of MTF and a carboxylate group of NIF. Unlike the previously described structures, this interaction involves only one oxygen atom from the carboxylate group of NIF (R216 graph set). Water molecules and ion pars are connected by H-bonding interactions to build a ribbon structure that extends along the b-axis, locating the –CF3 groups in the periphery. Additional H-bonding interactions involving the terminal amino group of the MTF+ cation and water molecules connect ribbons to build a 2D layered structure that extends parallel to the bc- plane of the crystal. Finally, the supramolecular 3D structure is generated by stacking these layers along the a-axis facing the -CF3 containing rings of NIF anions (Figure 4). A possible explanation for the transformation of the anhydrate salt into the hydrate salt form can be obtained from the study of the predicted crystal morphology of MTF–NIF (Figure 5). BFDH morphology of the anhydrate MTF–NIF salt was calculated by using the Bravais–Friedel–Donnay–Harker (BFDH) method included in the latest release of the visualization software package Mercury [30]. In the case of MTF–NIF, the faces {−110}, {−1–10}, {1 1 0} and {1 −1 0} (corresponding to 14.4 % of the total surface) and the faces {1 −1 −1}, {1 1 −1}, {−1 −1 1} and {−1 1 1} (corresponding to 32 % of the total surface) expose amino groups of MTF and carboxylate groups of NIF to the surface and coincide with the crystal structure region where the ion pairs form the tape-like structures. Therefore, it seems reasonable that water molecules can access and form additional H-bonds with the groups involved resulting in the hydrated structure as reported herein. ## 3.3.3. MTF–DIF Salt MTF–DIF crystallized in the monoclinic system spacegroup C2/c, with one MTF+ cation and one DIF anion in the asymmetric unit. A trimer was formed by MTF and two DIF ions through guanidinium···carboxylate, N4–H4E⋯O2 (3.01 Å) and N5–H5B⋯O1 (2.96 Å) (R228 graph set), and N2–H2A···O2 (2.92 Å) hydrogen bonds (Figure 6). The trimers are then linked to build a ribbon structure through N5–H5A⋯O2 (3.05 Å) hydrogen bonds along the b-axis (Figure 6b). Additional H-bond involving amine groups of MTF+ and carboxylate groups of DIF generate a 2D layered structure running parallel to the bc-plane of the crystal. C-H···F contacts participate in the cohesion of these structures. The supramolecular structure is finally built with stacks of layers facing 2,4-difluorophenyl rings of DIF (Figure 6). ## 3.3.4. MTF–FLP Salt This molecular salt crystallizes in the triclinic system space group P-1, with one monoprotonated MTF+ cation and one FLP anion in the asymmetric unit as ionic pair. Both ions are associated with a charge-assisted H-bond through the guanidinium moiety of MTF+ and the carboxylate group of FLP anion (R228 graph set). Adjacent pairs are further connected by H-bonds (R428 graph set) to build a ribbon structure that extends along the c-axis, locating the FLP ions outside. 2D-layered structures are then generated by connecting ribbons involving non-terminal amine groups of MTF+ and carboxylate groups. The 3D supramolecular architecture is built through π,π-stacking interactions between the non-substituted aromatic ring of FLP anions that connect adjacent layers (Figure 7). ## 3.4. Thermal Analysis Differential scanning calorimetry (DSC) was used to evaluate the thermal stability and determine the melting point of the new phases. Figure 8 shows the DSC traces and the endothermic events occurring during the experiments, which correspond to the melting point of the salts. The existence of one single and well-defined endothermic event confirms the purity of the phases, which is in good agreement with the results obtained using PXRD. In addition, the presence of only one endothermic event demonstrates the stability of the phase under the melting point. Over the transition state, other endothermic events are observed, ascribed to the degradation of the samples. With the exception of MTF–FLP, all the molecular salts show a melting point in a range between the melting point of the MTF base (117 °C) and the corresponding NSAID coformer (MEF 230 °C, DIF 210–211 °C, TLF 207 °C, NIF 204 °C), increasing the thermal stability of the MTF base in all cases, although still under the melting point of the commercially available MTF·HCl (231.5 °C) [46]. This behavior has already been described by other researchers, demonstrating the ability of multicomponent materials to modulate the thermal behavior of APIs [58]. Interestingly, the melting point of MTF–FLP is 200 °C, while the melting point of the components goes from 110 to 117 °C, increasing the thermal stability and the melting point of both APIs by more than 80 °C. ## 3.5. Stability Studies The thermodynamic stability of molecular salts was studied by conducting aqueous slurry experiments at 25 °C. After 24 h, the suspensions were filtered, air-dried at room temperature, and characterized using PXRD (Figure S4). No observable changes in color or texture were observed for the samples. Despite MTF–NIF and MTF–DIF displaying a phase transition indicating low stability in solution, MTF–TLF, MTF–MEF, and MTF–FLP remained stable upon slurrying, as the crystallinity and the initial crystal structures remained intact. SCXRD and further PXRD confirmed the formation of the dihydrate salt of MTF–NIF, but it was impossible to determine the final phase of MTF–DIF. Interestingly, this unknown phase agrees with the phase obtained directly from the LAG reaction using water as a liquid additive (Figure S5), suggesting the formation of a hydrate form of MTF–DIF salt. The molecular salts were also stored under accelerated aging conditions (40 °C and $75\%$ RH). Under these conditions, all the samples remained stable after four months, with the exception of MTF–NIF (Figure S6), which presented a partial phase transition on the third day, with a final conversion at one week. This new phase matched neither the already determined dihydrate salt form nor the product of the LAG in water, suggesting the formation of different grades of hydrate salt for the MTF–NIF system. ## 3.6. Solubility Studies Solubility and oral bioavailability are closely related. Therefore, solubility enhancement is one of the most common approaches to increasing the bioavailability of poorly soluble drugs [59,60,61]. In this work, HPLC was used to determine the solubility of the proposed MTF molecular salts. The solubility improvement of NSAID coformers was also assessed through data normalization, using the corresponding stoichiometry and molecular weight of MTF-NSAID salts and isolated components. It should be noted that this procedure can only be applied to stable phases in which the stoichiometry MTF:NSAID is maintained. Low stability was observed for MTF–NIF, and MTF–DIF in the stability test performed in aqueous media. A phase transformation for MTF–NIF was confirmed through the obtention of MTF–NIF·2H2O single crystals. For this reason, the solubility of MTF–NIF could not be determined at the reported experimental conditions. Instead, the solubility corresponding to the MTF–NIF salt hydrate was obtained. On the other hand, the phase transformation for MTF–DIF could not be determined, thus, making MTF–DIF unsuitable for solubility studies. Table 3 shows a notorious increase in the solubility of NSAIDs when compared with the reported isolated APIs: MTF–TLF increased solubility 111 times, MTF–MEF 208 times, MTF–NIF·2H2O 574 times, and finally MTF–FLP 1110 times, while exhibiting good stability. Furthermore, an interesting modulation of MTF solubility is also observed in the novel multicomponent molecular salts when compared with the solubility of MTF·HCl (ranging from 250 mg/mL [46,62,63]). In all reported MTF–NSAID salts, the solubility is reduced more than 50 times, with MTF–TLF decreasing the solubility of MTF by 90 times. Bearing in mind that clinical side effects of metformin were mainly caused by the gastrointestinal accumulation of MTF·HCl due to its extreme solubility [12], the obtention of a less soluble phase might reduce such side effects. The remaining solids were finally analyzed using PXRD to identify the crystalline phases, thus confirming the stability of the initial phases and the congruent solubility of these new salts. The remarkable solubility values obtained for the novel MTF–NSAID salts are consistent with the formation of molecular salts [7], as already expected from their pKa values. Furthermore, it is well known that the physicochemical properties of solids, such as solubility, are also strongly dependent on their intimate crystal structure. Herein, the layered arrangement exhibited by the APIs and the presence of charge-assisted hydrogen bonds in the crystal structure [64,65] are eventually responsible for the new solubility properties. Table 3 \| Compound \| MTFSolubility in the Salts [mg/mL] \| NSAIDSolubility in the Salts [mg/mL] \| SolubilityEnhancement over MTF·HCl/NSAID \| NSAIDSolubility[mg/mL] \| Ref. \| \| --- \| --- \| --- \| --- \| --- \| --- \| \| MTF–NIF·2H2O \| 5.29 \| 11.493 \| 0.0212x/574x \| <0.02 \| [66] \| \| MTF–FLP \| 4.775 \| 5.552 \| 0.0191x/1110x \| <0.005 \| [67] \| \| MTF–MEF \| 3.185 \| 5.95 \| 0.0127x/119x \| <0.05 \| [68,69] \| \| MTF–TLF \| 2.74 \| 10.397 \| 0.0110x/208x \| <0.05 \| [68] \| ## 4. Conclusions The novel multicomponent MTF–NSAID solids succeeded in overcoming two of the main problems that these APIs have when administered separately. On the one hand, MTF stability has been improved while achieving an outstanding solubility profile. For instance, the reported MTF–NSAID solids are more soluble than MTF base but less soluble than MTF·HCl salt, thus potentially reducing improper intestinal accumulation, which is one of the most common side effects associated with the current MTF·HCl chronic treatment. On the other hand, MTF–NSAID molecular salts, thanks to the salification strategy, are able to significantly enhance the solubility of NSAIDs, thus reducing the dosage and the undesired gastrointestinal dose-dependent side effects of these drugs. Interestingly, structure-properties relationships could be gathered from the thorough study of the intimate crystal structure of the new multicomponent MTF–NSAIDs molecular salts. The disruption of the robust NSAID–NSAID dimers present in the crystal structure of isolated NSAIDs is key for understanding their enhancement in solubility. Moreover, the new layered NSAID–MTF–NSAID structure in the novel multicomponent materials protects MTF from water, explaining the higher stability and the new solubility profile. Considering all the above, the novel drug–drug molecular salts are worthy of further investigation. Our results confirm improved physicochemical properties thanks to the novel formulation, among which a better solubility profile should be remarked. Improving solubility should not be underestimated since a poor solubility profile is currently the most important limitation of oral biopharmaceutics in the pharmaceutical industry. Likewise, optimized solubility opens the door to dosage reduction and consequently might reduce those side effects associated with MTF–NSAID treatments. ## Figures, Scheme and Tables Scheme 1: *Chemical formula of metformin (MTF), mefenamic acid (MEF), tolfenamic acid (TLF), niflumic acid (NIF), diflunisal (DIF), and flurbiprofen (FLP).* Figure 1: PXRD patterns of MTF base and the novel salts obtained by mechanochemical synthesis. Figure 2: (a) Fragment of 1D-chain structure in the crystal structure of MTF–MEF. (b) Detailed view of the ribbon structure in the molecular salt MTF–MEF. MTF+ cations are represented as balls and sticks. (c) Detailed view of the packing arrangement of MTF cations (blue) and MEF anions (green) in the crystal structure of MTF–MEF. MTF+ cations are represented as balls and sticks. (d) A pair of stacked MEF–MEF anions. Figure 3: (a) Fragment of 1D-chain structure in the crystal structure of MTF–NIF. (b) Detailed view of multi-stacking π,π-interaction between NIF anions. (c) Detailed view of the ribbon structure in the molecular salt MTF–NIF. MTF+ cations are represented as balls and sticks. (d) Fragment of the 2D-layered structure in the molecular salt MTF–NIF. (e) Packing arrangement of MTF cations (blue) and NIF anions (green) in the crystal structure of MTF–NIF. Figure 4: (a) Fragment of ribbon structure in the crystal structure of MTF–NIF·2H2O. (b) Detailed view voids where water molecules are located in MTF–NIF·2H2O. (c) Packing arrangement of MTF cations (blue), water molecules (red), and NIF anions (green) in the crystal structure of MTF–NIF·2H2O. MTF cations are represented as balls and sticks. Figure 5: BFDH−predicted morphology of MTF–NIF salt. Figure 6: (a) Trimer structure generated by H-bonding interactions in the MTF–DIF salt. (b)Fragment of the ribbon structure generated using connecting trimers with H-bonding interactions. (c) Detailed view of the packing arrangement of MTF cations (blue) and DIF anions (green) in the crystal structure of MTF–DIF. MTF cations are represented as balls and sticks. Figure 7: (a) Fragment of ribbon structure in the crystal structure of MTF–FLP. (b) Fragment of the 2D-layered structure in the molecular salt MTF–FLP. (c) Packing arrangement of MTF cations (blue) and NIF anions (green) in the crystal structure of MTF–DIF. (d) A pair of stacked FLP–FLP anions. Figure 8: DSC traces of the pure molecular salts and MTF base. Blue-dotted line corresponds to the melting point of MTF·HCl. TABLE_PLACEHOLDER:Table 1 TABLE_PLACEHOLDER:Table 2 ## References 1. World Health Organization (WHO) Diabetes 2. International Diabetes Federation Diabetes 3. Type 2 Diabetes 4. 4. U.S. Food and Drug Administration (FDA) Approved Drug Products: MetforminFDASilver Spring, MD, USA2018. (FDA) Approved Drug Products: Metformin (2018.0) 5. Viollet B., Guigas B., Sanz Garcia N., Leclerc J., Foretz M., Andreelli F.. Cellular and Molecular Mechanisms of Metformin: An Overview. Clin. Sci. (2012.0) 122 253. DOI: 10.1042/CS20110386 6. Rena G., Hardie D.G., Pearson E.R.. The Mechanisms of Action of Metformin. Diabetologia (2017.0) 60 1577. DOI: 10.1007/s00125-017-4342-z 7. Serajuddin A.T.M.. Salt Formation to Improve Drug Solubility. Adv. Drug Deliv. Rev. (2007.0) 59 603-616. DOI: 10.1016/j.addr.2007.05.010 8. Sun X., Du S., Sun Y., Li H., Yu C.P., Guo J., Wang Y., Yu S., Cheng Y., Xue F.. Solubility Measurement and Data Correlation of Metformin Hydrochloride in Four Aqueous Binary Solvents and Three Pure Solvents from 283.15 to 323.15 K. J. Chem. Eng. Data (2021.0) 66 3282-3292. DOI: 10.1021/acs.jced.1c00338 9. Klepser T.B., Kelly M.W.. Metformin Hydrochloride: An Antihyperglycemic Agent. Am. J. Health Syst. Pharm. (1997.0) 54 893-903. DOI: 10.1093/ajhp/54.8.893 10. Bretnall A.A., Clarke G.S., Brittain H.. Analytical Profiles of Drug Substances and Excipients (1998.0) Volume 25 11. Nicklin P., Keates A.C., Page T., Bailey C.J.. Transfer of Metformin across Monolayers of Human Intestinal Caco-2 Cells and across Rat Intestine. Int. J. Pharm. (1996.0) 128 155-162. DOI: 10.1016/0378-5173(95)04259-8 12. McCreight L.J., Bailey C.J., Pearson E.R.. Metformin and the Gastrointestinal Tract. Diabetologia (2016.0) 59 426-435. DOI: 10.1007/s00125-015-3844-9 13. Wilcock C., Bailey C.J.. Accumulation of Metformin by Tissues of the Normal and Diabetic Mouse. Xenobiotica (1994.0) 24 49-57. DOI: 10.3109/00498259409043220 14. 14. Council of Europe European Pharmacopeia9th ed.Council of EuropeStrasbourg, France2016. European Pharmacopeia (2016.0) 15. Hamdan I.I., Jaber A.K.B., Abushoffa A.M.. Development and Validation of a Stability Indicating Capillary Electrophoresis Method for the Determination of Metformin Hydrochloride in Tablets. J. Pharm. Biomed. Anal. (2010.0) 53 1254-1257. DOI: 10.1016/j.jpba.2010.03.017 16. Gumieniczek A., Berecka-Rycerz A., Mroczek T., Wojtanowski K.. Determination of Chemical Stability of Two Oral Antidiabetics, Metformin and Repaglinide in the Solid State and Solutions Using LC-UV, LC-MS, and FT-IR Methods. Molecules (2019.0) 24. DOI: 10.3390/molecules24244430 17. Inzucchi S.E., Bergenstal R.M., Buse J.B., Diamant M., Ferrannini E., Nauck M., Peters A.L., Tsapas A., Wender R., Matthews D.R.. Management of Hyperglycaemia in Type 2 Diabetes: A Patient-Centered Approach. Position Statement of the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia (2012.0) 55 1577-1596. DOI: 10.1007/s00125-012-2534-0 18. NHS Non-Steroidal Anti-Inflammmatory Drugs 19. Schreiber A.K., Nones C.F., Reis R.C., Chichorro J.G., Cunha J.M.. Diabetic Neuropathic Pain: Physiopathology and Treatment. World J. Diabetes (2015.0) 6 432. DOI: 10.4239/wjd.v6.i3.432 20. Moore R.A., Chi C.C., Wiffen P.J., Derry S., Rice A.S.C.. Oral Nonsteroidal Anti-Inflammatory Drugs for Neuropathic Pain. Cochrane Database Syst. Rev. (2015.0) 10 CD010902. DOI: 10.1002/14651858.CD010902.pub2 21. Assan R., Heuclin C., Ganeval D., Bismuth C., George J., Girard J.R.. Metformin-Induced Lactic Acidosis in the Presence of Acute Renal Failure. Diabetologia (1977.0) 13 211-217. DOI: 10.1007/BF01219702 22. Vecchio S., Giampreti A., Petrolini V.M., Lonati D., Protti A., Papa P., Rognoni C., Valli A., Rocchi L., Rolandi L.. Metformin Accumulation: Lactic Acidosis and High Plasmatic Metformin Levels in a Retrospective Case Series of 66 Patients on Chronic Therapy. Clin. Toxicol. (2014.0) 52 129-135. DOI: 10.3109/15563650.2013.860985 23. Whelton A.. Nephrotoxicity of Nonsteroidal Anti-Inflammatory Drugs: Physiologic Foundations and Clinical Implications. Am. J. Med. (1999.0) 106 13S-24S. DOI: 10.1016/S0002-9343(99)00113-8 24. Chan N.N., Fauvel N.J., Feher M.D.. Non-Steroidal Anti-Inflammatory Drugs and Metformin: A Cause for Concern?. Lancet (1998.0) 352 201. DOI: 10.1016/S0140-6736(05)77806-5 25. Berry D.J., Steed J.W.. Pharmaceutical Cocrystals, Salts and Multicomponent Systems; Intermolecular Interactions and Property Based Design. Adv. Drug Deliv. Rev. (2017.0) 117 3-24. DOI: 10.1016/j.addr.2017.03.003 26. Bruker APEX4. APEX4 V2022.1 (2022.0) 27. Sheldrick G.M.. SHELXT—Integrated Space-Group and Crystal-Structure Determination. Acta Cryst. A (2015.0) 71 3-8. DOI: 10.1107/S2053273314026370 28. Sheldrick G.M.. Crystal Structure Refinement with SHELXL. Acta Cryst. C Struct. Chem. (2015.0) 71 3-8. DOI: 10.1107/S2053229614024218 29. Dolomanov O.V., Bourhis L.J., Gildea R.J., Howard J.A.K., Puschmann H.. OLEX2: A Complete Structure Solution, Refinement and Analysis Program. J. Appl. Cryst. (2009.0) 42 339-341. DOI: 10.1107/S0021889808042726 30. MacRae C.F., Sovago I., Cottrell S.J., Galek P.T.A., McCabe P., Pidcock E., Platings M., Shields G.P., Stevens J.S., Towler M.. Mercury 4.0: From Visualization to Analysis, Design and Prediction. J. Appl. Cryst. (2020.0) 53 226-235. DOI: 10.1107/S1600576719014092 31. Spek A.L.. Structure Validation in Chemical Crystallography. Acta Cryst. D Biol. Cryst. (2009.0) 65 148-155. DOI: 10.1107/S090744490804362X 32. Glomme A., März J., Dressman J.B.. Comparison of a Miniaturized Shake-Flask Solubility Method with Automated Potentiometric Acid/Base Titrations and Calculated Solubilities. J. Pharm. Sci. (2005.0) 94 1-16. DOI: 10.1002/jps.20212 33. Braga D., Maini L., Grepioni F.. Mechanochemical Preparation of Co-Crystals. Chem. Soc. Rev. (2013.0) 42 7638-7648. DOI: 10.1039/c3cs60014a 34. Jia L., Wu S., Gong J.. A Tolbutamide Metformin Salt Based on Antidiabetic Drug Combinations: Synthesis, Crystal Structure Analysis and Pharmaceutical Properties. Acta Cryst. C Struct. Chem (2019.0) 75 1250-1258. DOI: 10.1107/S2053229619010647 35. Bian X., Jiang L., Gan Z., Guan X., Zhang L., Cai L., Hu X.. A Glimepiride-Metformin Multidrug Crystal: Synthesis, Crystal Structure Analysis, and Physicochemical Properties. Molecules (2019.0) 24. DOI: 10.3390/molecules24203786 36. Bian X., Jiang L., Zhou J., Guan X., Wang J., Xiang P., Pan J., Hu X.. Improving Dissolution and Cytotoxicity by Forming Multidrug Crystals. Molecules (2020.0) 25. DOI: 10.3390/molecules25061343 37. Putra O.D., Furuishi T., Yonemochi E., Terada K., Uekusa H.. Drug-Drug Multicomponent Crystals as an Effective Technique to Overcome Weaknesses in Parent Drugs. Cryst. Growth Des. (2016.0) 16 3577-3581. DOI: 10.1021/acs.cgd.6b00639 38. Sun J., Jia L., Wang M., Liu Y., Li M., Han D., Gong J.. Novel Drug-Drug Multicomponent Crystals of Epalrestat-Metformin: Improved Solubility and Photostability of Epalrestat and Reduced Hygroscopicity of Metformin. Cryst. Growth Des. (2022.0) 22 1005-1016. DOI: 10.1021/acs.cgd.1c00810 39. Feng W.Q., Wang L.Y., Gao J., Zhao M.Y., Li Y.T., Wu Z.Y., Yan C.W.. Solid State and Solubility Study of a Potential Anticancer Drug-Drug Molecular Salt of Diclofenac and Metformin. J. Mol. Struct. (2021.0) 1234 130166. DOI: 10.1016/j.molstruc.2021.130166 40. Pérez-Fernández R., Fresno N., Goya P., Elguero J., Menéndez-Taboada L., García-Granda S., Marco C.. Structure and Thermodynamical Properties of Metformin Salicylate. Cryst. Growth Des. (2013.0) 13 1780-1785. DOI: 10.1021/cg400150f 41. Zhou W.-X., Zhao H.-W., Chen H.-H., Zhang Z.-Y., Chen D.-Y.. Characterization of Drug–Drug Salt Forms of Metformin and Aspirin with Improved Physicochemical Properties. CrystEngComm (2019.0) 21 3770-3773. DOI: 10.1039/C9CE00377K 42. Johnson S.L., Rumon K.A.. Infrared Spectra of Solid 1:1 Pyridine-Benzoic Acid Complexes; the Nature of the Hydrogen Bond as a Function of the Acid-Base Levels in the Complex. J. Phys. Chem. (1965.0) 69 74-86. DOI: 10.1021/j100885a013 43. Tong W.Q., Whitesell G.. In Situ Salt Screening-A Useful Technique for Discovery Support and Preformulation Studies. Pharm. Dev. Technol. (1998.0) 3 215-223. DOI: 10.3109/10837459809028498 44. Cruz-Cabeza A.J.. Acid-Base Crystalline Complexes and the PKa Rule. CrystEngComm (2012.0) 14 6362-6365. DOI: 10.1039/c2ce26055g 45. Verdugo-Escamilla C., Alarcón-Payer C., Acebedo-Martínez F.J., Domínguez-Martín A., Choquesillo-Lazarte D.. New Metformin-Citric Acid Pharmaceutical Molecular Salt: Improving Metformin Physicochemical Properties. Crystals (2022.0) 12. DOI: 10.3390/cryst12121748 46. Diniz L.F., Carvalho P.S., Gonçalves J.E., Diniz R., Fernandes C.. Solid-State Landscape and Biopharmaceutical Implications of Novel Metformin-Based Salts. New J. Chem. (2022.0) 46 13725-13737. DOI: 10.1039/D2NJ00453D 47. Jones O.A.H., Voulvoulis N., Lester J.N.. Aquatic Environmental Assessment of the Top 25 English Prescription Pharmaceuticals. Water Res. (2002.0) 36 5013-5022. DOI: 10.1016/S0043-1354(02)00227-0 48. Avdeef A.. Solubility of Sparingly-Soluble Ionizable Drugs. Adv. Drug Deliv. Rev. (2007.0) 59 568-590. DOI: 10.1016/j.addr.2007.05.008 49. Pentikäinen P.J., Tokola O., Alhava E., Penttilä A.. Pharmacokinetics of Tolfenamic Acid: Disposition in Bile, Blood and Urine after Intravenous Administration to Man. Eur. J. Clin. Pharmacol. (1984.0) 27 349-354. DOI: 10.1007/BF00542174 50. Wishart D.S., Feunang Y.D., Guo A.C., Lo E.J., Marcu A., Grant J.R., Sajed T., Johnson D., Li C., Sayeeda Z.. DrugBank 5.0: A Major Update to the DrugBank Database for 2018. Nucleic Acids Res. (2018.0) 46 D1074-D1082. DOI: 10.1093/nar/gkx1037 51. Somasundaram S., Rafi S., Hayllar J., Sigthorsson G., Jacob M., Price A.B., Macpherson A., Mahmod T., Scott D., Wrigglesworth J.M.. Mitochondrial Damage: A Possible Mechanism of the “topical” Phase of NSAID Induced Injury to the Rat Intestine. Gut (1997.0) 41 344-353. DOI: 10.1136/gut.41.3.344 52. ChemAxon MarvinSketch 2020 53. Childs S.L., Stahly G.P., Park A.. The Salt-Cocrystal Continuum: The Influence of Crystal Structure on Ionization State. Mol. Pharm. (2007.0) 4 323-338. DOI: 10.1021/mp0601345 54. Etter M.C., MacDonald J.C., Bernstein J.. Graph-Set Analysis of Hydrogen-Bond Patterns in Organic Crystals. Acta Cryst. B (1990.0) 46 256-262. DOI: 10.1107/S0108768189012929 55. Etter M.C.. Encoding and Decoding Hydrogen-Bond Patterns of Organic Compounds. Acc. Chem. Res. (1990.0) 23 120-126. DOI: 10.1021/ar00172a005 56. Kálmán A., Párkányi L., Argay G.. Classification of the Isostructurality of Organic Molecules in the Crystalline State. Acta Cryst. Sect. B (1993.0) 49 1039-1049. DOI: 10.1107/S010876819300610X 57. Oliveira M.A., Peterson M.L., Klein D.. Continuously Substituted Solid Solutions of Organic Co-Crystals. Cryst. Growth Des. (2008.0) 8 4487-4493. DOI: 10.1021/cg800454r 58. Perlovich G.. Melting Points of One- and Two-Component Molecular Crystals as Effective Characteristics for Rational Design of Pharmaceutical Systems. Acta Cryst. B Struct. Sci. Cryst. Eng. Mater. (2020.0) 76 696-706. DOI: 10.1107/S2052520620007362 59. Sathisaran I., Dalvi S.V.. Engineering Cocrystals of Poorlywater-Soluble Drugs to Enhance Dissolution in Aqueous Medium. Pharmaceutics (2018.0) 10. DOI: 10.3390/pharmaceutics10030108 60. Varshosaz J., Ghassami E., Ahmadipour S.. Crystal Engineering for Enhanced Solubility and Bioavailability of Poorly Soluble Drugs. Curr. Pharm. Des. (2018.0) 24 2473-2496. DOI: 10.2174/1381612824666180712104447 61. Patel J.N., Rathod D.M., Patel N.A., Modasiya M.K.. Techniques to Improve the Solubility of Poorly Soluble Drugs. Int. J. Pharm. Life Sci. (2012.0) 3 1459-1469 62. Cheng C.L., Yu L.X., Lee H.L., Yang C.Y., Lue C.S., Chou C.H.. Biowaiver Extension Potential to BCS Class III High Solubility-Low Permeability Drugs: Bridging Evidence for Metformin Immediate-Release Tablet. Eur. J. Pharm. Sci. (2004.0) 22 297-304. DOI: 10.1016/j.ejps.2004.03.016 63. Kim D.W., Park J.B.. Development and Pharmaceutical Approach for Sustained-Released Metformin Succinate Tablets. J. Drug Deliv. Sci. Technol. (2015.0) 30 90-99. DOI: 10.1016/j.jddst.2015.09.019 64. Gould P.L.. Salt Selection for Basic Drugs. Int. J. Pharm. (1986.0) 33 201-217. DOI: 10.1016/0378-5173(86)90055-4 65. Putra O.D., Umeda D., Fujita E., Haraguchi T., Uchida T., Yonemochi E., Uekusa H.. Solubility Improvement of Benexate through Salt Formation Using Artificial Sweetener. Pharmaceutics (2018.0) 10. DOI: 10.3390/pharmaceutics10020064 66. Kim S., Chen J., Cheng T., Gindulyte A., He J., He S., Li Q., Shoemaker B.A., Thiessen P.A., Yu B.. PubChem in 2021: New Data Content and Improved Web Interfaces. Nucleic Acids Res. (2021.0) 49 D1388-D1395. DOI: 10.1093/nar/gkaa971 67. Oh D.H., Park Y.J., Kang J.H., Yong C.S., Choi H.G.. Physicochemical Characterization and in Vivo Evaluation of Flurbiprofen-Loaded Solid Dispersion without Crystalline Change. Drug Deliv. (2011.0) 18 46-53. DOI: 10.3109/10717544.2010.509365 68. Avdeef A., Bendels S., Tsinman O., Tsinman K., Kansy M.. Solubility-Excipient Classification Gradient Maps. Pharm. Res. (2007.0) 24 530-545. DOI: 10.1007/s11095-006-9169-0 69. Gao L., Zheng W.Y., Yang W.L., Zhang X.R.. Drug-Drug Salt Forms of Vortioxetine with Mefenamic Acid and Tolfenamic Acid. J. Mol. Struct. (2022.0) 1268 133725. DOI: 10.1016/j.molstruc.2022.133725
--- title: Association of Red Blood Cell Distribution Width and Neutrophil-to-Lymphocyte Ratio with Calcification and Cardiovascular Markers in Chronic Kidney Disease authors: - Stefanos Roumeliotis - Ioannis E. Neofytou - Cecile Maassen - Petra Lux - Konstantia Kantartzi - Evangelos Papachristou - Leon J. Schurgers - Vassilios Liakopoulos journal: Metabolites year: 2023 pmcid: PMC9966770 doi: 10.3390/metabo13020303 license: CC BY 4.0 --- # Association of Red Blood Cell Distribution Width and Neutrophil-to-Lymphocyte Ratio with Calcification and Cardiovascular Markers in Chronic Kidney Disease ## Abstract We aimed to investigate the association between Red Blood Cell Distribution Width (RDW) and Neutrophil-to-Lymphocyte Ratio (NLR), simple, rapidly assessed markers from the complete blood count with vascular calcification (VC)/stiffness and cardiovascular disease (CVD) in chronic kidney disease (CKD). Dephosphorylated, uncarboxylated matrix Gla-protein (dp-ucMGP), and central/peripheral hemodynamics’ parameters were measured in 158 CKD patients, including Hemodialysis and Peritoneal Dialysis. Spearman’s rho analysis showed that RDW correlated with C-reactive protein (CRP) ($r = 0.29$, $p \leq 0.001$), dp-ucMGP ($r = 0.43$, p = < 0.0001), central diastolic blood pressure (DBP) (r = −0.19, $$p \leq 0.02$$), and albuminuria (r = −0.17, $$p \leq 0.03$$). NLR correlated with the duration of CVD ($r = 0.32$, $p \leq 0.001$), CRP ($r = 0.27$, $$p \leq 0.01$$), dp-ucMGP ($r = 0.43$, $p \leq 0.0001$), central DBP (r = −0.32, $p \leq 0.0001$) and eGFR (r = −0.25, $$p \leq 0.04$$). In multiple regression models, circulating dp-ucMGP was an independent predictor of RDW (β = 0.001, $$p \leq 0.001$$) and NLR (β = 0.002, $$p \leq 0.002$$). In CKD patients, RDW and NLR are associated with traditional and novel markers of VC and CVD. ## 1. Introduction Cardiovascular disease (CVD) is highly prevalent, in particular in the early stages of chronic kidney disease (CKD), it progresses in parallel with the deterioration of kidney function and accounts for more than $50\%$ of all deaths in patients with end-stage kidney disease (ESKD) [1]. CVD is the leading cause of death not only in ESKD but also in CKD patients; patients at the early stages of CKD are more likely to die due to CVD or heart failure (HF) than progress to EKSD and require dialysis [2]. This CV burden in uremic patients might be attributed to the fact that arterial calcification and stiffness—which predispose to atherosclerosis and CVD—are gradually increased with the progression of CKD to ESKD and are further exacerbated in dialysis [3]. In CKD, accumulating data suggest that dephosphorylated uncarboxylated matrix Gla-protein (dp-ucMGP), the inactive form of MGP, is a strong and reliable marker of vascular calcification (VC) and predicts CVD morbidity and mortality [4,5,6,7], whereas pulse wave velocity (PWV) has been shown to be a surrogate marker for arterial stiffness in these patients [8]. In order to prevent and manage CV morbidity and mortality, the early and precise identification of patients at risk for arterial calcification or stiffness is crucial in CKD. In this direction, various markers have been proposed; however, the major limitations for the broad clinical use of these markers include high cost and complex, time-consuming methods of measurement. On the other hand, there is a growing interest in the identification of novel, simple, low-cost and easily accessible tools that can be used in everyday clinical practice to classify CKD patients at high risk for VC, vascular stiffness (VS) and CVD. Red Blood Cell Distribution Width (RDW) is a measure of difference in volume and size of erythrocytes calculated by automated hematology analyzers and routinely reported in the total blood count lab test. RDW is used as a marker of anisocytosis and traditionally has been used for the differential diagnosis of different types of anemias [9]. Increased RDW levels reflect short red blood cell (RBC) life span due to limited erythropoiesis and accelerated RBC destruction. However, large cohort studies during the past decade coherently showed that RDW is steadily increased in CVD [10], heart failure (HF) [11,12,13,14] and CKD [15]. Moreover, it has been reported to be a strong and independent predictor of CV mortality and VC in CKD populations [16,17,18,19]. Although the exact pathophysiology linking increased RDW with CVD has not yet been fully elucidated, VC and VS, chronic inflammation and nutritional disorders have been proposed as potential underlying mechanisms. The Neutrophil-to-Lymphocyte Ratio (NLR) is a marker also derived from the full blood count lab test, calculated from deriving the absolute number of neutrophils to the absolute count of lymphocytes. Similarly to RDW, NLR has been demonstrated by large studies to be an indicator of CVD, VC and inflammation [20], especially in CKD and ESKD patients [21]. However, until now, no study has assessed the possible clinical value of these two markers combined in CKD cohorts. The aim of this study is to investigate the association of simple, quick and low-cost markers derived from the total blood count, such as RDW and NLR with traditional or non-traditional risk factors of CVD and VC/VS assessed by dp-ucMGP and PWV. ## 2.1. Patients In this cross-sectional, single-center study, we recruited 158 patients at different CKD stages. Eligible patients were those with a documented diagnosis of CKD. A total of 158 patients were included in the study, divided into 73 pre-dialysis outpatients and 84 dialysis patients. All pre-dialysis CKD subjects were outpatients followed in the Division of Nephrology and Hypertension of the University General AHEPA Hospital of Thessaloniki (Greece), whereas dialysis patients were undergoing maintenance hemodialysis in the Hemodialysis Unit or Peritoneal Dialysis, followed in the Peritoneal Dialysis Unit of the University General AHEPA Hospital of Thessaloniki (Greece). All patients were enrolled between 1 November 2021 and 1 March 2022. Initially, 173 patients were screened to participate in the study. Three declined participation and 12 met the exclusion criteria, as described in the patients’ enrolment flow chart (Figure 1). Among pre-dialysis patients, 11 were at stages 1–2, 2 at stage 3, 33 at stage 4 and 3 at stage 5 (end-stage kidney disease-ESKD), whereas 25 patients were under chronic peritoneal dialysis (PD) and 59 under maintenance hemodialysis (HD) treatment. The estimated glomerular filtration rate was calculated using the CKD-EPI calculation and a diagnosis/classification of CKD was based on the National Kidney Foundation Kidney Disease Outcomes Quality Initiative criteria [22]. Patients with acute kidney injury, hospitalized or acute illness were excluded from the study. The definition of AKI was in accordance with the Acute Kidney Injury Working Group of KDIGO (Kidney Disease: Improving Global Outcomes) [23]. Acute illness was defined as any acute condition that would limit the ability of the patients to participate in the study and might affect the markers measured, such as infection, newly diagnosed tumor and current hospitalization (none of the patients assessed was recently hospitalized, within 3 months from enrollment). At recruitment, we documented demographic, anthropometric, clinical and laboratory data, including documented history of hypertension (HT), type 2 diabetes mellitus (T2DM), heart failure (HF), CVD and measured pulse-wave velocity (PWV), and indices of central blood pressure (BP). CVD included coronary heart disease, heart failure, angina, stroke or peripheral arterial disease. We obtained serum, plasma and whole blood from all patients and urine albumin to creatin ratio (UACR) was measured in a spot urine sample. We measured dp-ucMGP from plasma in 116 patients and obtained RDW and NLR values from all patients from the complete blood count. All patients provided informed consent at enrollment. The study was conducted in accordance with the Helsinki Declaration of Human Rights and was approved by the Ethics Committee/Scientific Council of the Medical School of Aristotle University of Thessaloniki ($\frac{235}{14}$ May2021). ## 2.2. Laboratory Analyses We drew fasting blood from all patients into tubes containing EDTA and tubes without anticoagulant and obtained plasma, serum and whole blood. Samples for creatinine, calcium, phosphorus, C-reactive protein (CRP), parathormone, glycated hemoglobin (HbA1c), triglycerides, total, low density lipoprotein (LDL) and high-density lipoprotein (HDL) cholesterol and serum albumin and total blood count for white blood cells, hemoglobin and RDW were transferred to the laboratory and immediately assessed, whereas for dp-ucMGP, we centrifuged the samples and stored plasma at −20 °C, until analysis. We collected blood samples from HD patients after eight hours of fasting overnight and before the start of a mid-week dialysis session. The presence of albuminuria (UACR) was defined in 2 out of 3 consecutive measurements in morning spot urine samples, during a 3-month period. Plasma dp-ucMGP levels were measured in a single run by the Laboratory of Coagulation Profile (Maastricht, the Netherlands) using the commercially available IVD CE-marked chemiluminescent InaKtif MGP assay on the IDS-iSYS system (IDS, Boldon, UK), which has been described elsewhere [24]. In brief, plasma samples and internal calibrators were incubated using magnetic particles that were coated with murine monoclonal antibodies against dp-MGP, acridinium-labelled murine monoclonal antibodies against ucMGP and an assay buffer. The magnetic particles were captured using a magnet and washed to remove any unbound analyte. Trigger reagents were added, and the resulting light emitted by the acridinium label was directly proportional to the level of dp-ucMGP in the sample. The assay-measuring range was between 300 and 12,000 pmol/L and was linear up to 11,651 pmol/L. The within-run and total variations of this assay were 0.8–$6.2\%$ and 3.0–$8.2\%$, respectively. RDW and neutrophils/lymphocytes were measured as part of the routine total blood cell count, using the automatic hematology analyzer Sysmex XE-5000 (Sysmex Corporation, Kobe, Japan). According to our laboratory, the reference range for RDW was 12.0–$14.0\%$. To assess the potential clinical utility of RDW and NLR, all patients were divided into 2 groups according to the presence of CVD and/or HF and into quartiles according to median RDW and NLR. ## 2.3. PWV, Peripheral and Central Hemodynamic Parameters Measurement Peripheral (brachial) and central aortic BP, diastolic (DBP), systolic (SBP), pulse wave velocity (PWV), cardiac rhythm, pulse pressure and augmentation index (AI) and other peripheral and central hemodynamic parameters were determined with the Mobil-O-Graph device (IEM, Stolberg, Germany), as described before [25]. The BP-detection unit was validated according to the ESH/ESC criteria, as described elsewhere [26]. ## 2.4. Statistics We performed all statistical analyses using the IBM Statistical Package for Social Sciences (SPSS 18.0 for Windows, IBM, Chicago, IL, USA). The Kolmogorov–Smirnov test was used to test data for normality. Normally distributed continuous variables are presented as mean ± standard deviation and non-normally distributed continuous variables are presented as median (range, minimum to maximum values). Patient characteristics were compared among groups of HF/CVD diagnosis and RDW/NLR quartiles using the chi-square test for categorical variables and the Mann–Whitney test for continuous variables. To compare the values of NLR among groups with different cause of nephropathy, the Kruskal–Wallis test was used. Bivariate associations between RDW, NLR and other variables were examined with Spearman’s correlation co-efficient. We divided our cohort into quartiles according to median RDW ($14.4\%$) and median NLR (3.3)—quartile 1: below median RDW below median NLR; quartile 2: below median NLR, above median RDW; quartile 3: above median NLR, below median RDW and quartile 4: above median RDW, above median NLR. The Kruskal–Wallis test or chi-square test was used to compare for differences of variables among groups, accordingly. To investigate the possible predictors of RDW and NLR, we conducted a multiple regression analysis (forward, stepwise) with RDW and NLR as the dependent variables in the absence and presence of possible predictors, as determined in the correlation bivariate analyses. The multiple regression models where RDW was the dependent variable were adjusted for serum albumin, C-reactive protein, urinary albumin–creatinine ratio, calcium, phosphorus, cardiac rhythm, central diastolic blood pressure, warfarin and erythropoietin-stimulating agent, whereas the models with NLR as the dependent variable were adjusted for age, presence of cardiovascular disease, duration of cardiovascular disease, serum albumin, C-reactive protein, estimated glomerular filtration rate, calcium, parathormone, central diastolic blood pressure, cause of nephropathy and erythropoietin-stimulating agent. Non-normally distributed variables were log-transformed before entering the regression analyses. Significance was set at $p \leq 0.05.$ ## 3. Results In the study cohort, 41 patients had HF ($25.9\%$), 61 CVD ($38.6\%$) and 81 ($51.3\%$) had HF and/or CVD. Clinical, anthropometric, hematologic and biochemical characteristics of 158 CKD patients according to the presence of HF/CVD are shown in Table 1. Patients with a documented history of HF or CVD were mostly male, had significantly higher RDW, NLR and dp-ucMGP, were older and had significantly lower eGFR and HDL cholesterol. They also had a $28.3\%$ increased prevalence of T2DM diagnosis, longer period of T2DM diagnosis and significantly higher HbA1c values. Compared to patients with no HF or CVD, those with HF and/or CVD presented significantly lower diastolic, central diastolic and mean BP. There was no difference among groups regarding BMI, SBP, PWV, AI, calcium, phosphorus, parathormone, CRP, UACR or treatment with warfarin or erythropoietin. In the subgroup of 84 ESKD patients undergoing maintenance dialysis (25 PD and 59 HD), time on dialysis, Kt/V (indicator of dialysis efficiency) and peritoneal equilibration test and type of PD transporter did not differ significantly among groups of HF or CVD (results not shown). Moreover, the prevalence of HF/CVD did not differ significantly among PD and HD ($56\%$ in PD and $52.5\%$ in HD). The correlation matrix analysis of RDW and NLR with various traditional and non-traditional risk factors (categorized as anthropometric, hematologic, nutrition, inflammation, CKD and calcification/stiffness markers) for HF and CVD in CKD is shown in Table 2. RDW was positively correlated with NLR ($r = 0.23$, $$p \leq 0.004$$), total WBC ($r = 0.17$, $$p \leq 0.04$$), CRP ($r = 0.29$, $p \leq 0.001$), phosphorus ($r = 0.27$, $$p \leq 0.001$$) and dp-ucMGP ($r = 0.43$, p = < 0.0001). There was a significant inverse correlation between RDW and hemoglobin (r = −0.36, $p \leq 0.0001$), DBP (r = −0.19, $$p \leq 0.02$$), central DBP (r = −0.19, $$p \leq 0.02$$), cardiac rhythm (r = −0.16, $$p \leq 0.05$$), serum albumin (r = −0.24, $$p \leq 0.002$$), UACR (r = −0.17, $$p \leq 0.03$$) and calcium (r = −0.23, $$p \leq 0.004$$). NLR was positively correlated with total WBC ($r = 0.34$, $p \leq 0.001$), age ($r = 0.19$, $$p \leq 0.02$$), duration of CVD ($r = 0.32$, $p \leq 0.001$), CRP ($r = 0.27$, $$p \leq 0.01$$), dp-ucMGP ($r = 0.43$, $p \leq 0.0001$) and parathormone ($r = 0.18$, $$p \leq 0.024$$) and negatively with DBP (r = −0.30, $p \leq 0.0001$), central DBP (r = −0.32, $p \leq 0.0001$), mean BP (r = −0.24, $$p \leq 0.003$$), hemoglobin (r = −0.37, $p \leq 0.0001$), serum albumin (r = −0.29, $p \leq 0.0001$), calcium (r = −0.23, $$p \leq 0.005$$) and eGFR (r = −0.25, $$p \leq 0.04$$). Moreover, the Kruskal–Wallis analysis showed that NLR was significantly different among groups with a different cause of nephropathy ($$p \leq 0.017$$), whereas men presented significantly higher values of NLR compared to women (3.56 vs. 2.78, $$p \leq 0.008$$). Patients receiving an erythropoietin-stimulating agent and those with documented CVD had significantly higher values of NLR compared to the others ($p \leq 0.0001$ and $p \leq 0.0001$, respectively). In the group analysis, RDW was associated with use of warfarin and use of erythropoietin; patients treated with warfarin had significantly higher values of RDW compared to those who did not receive warfarin (17.1 vs. 14.4, $$p \leq 0.006$$). Similarly, erythropoietin was associated with an increase of $1.3\%$ in RDW ($p \leq 0.0001$). NLR values according to different causes of nephropathy are shown in Figure 2. Patients with obstructive nephropathy had lower NLR values and patients with cardiorenal syndrome, due to heart failure with reduced ejection fraction, had the highest (2 and 5.2, respectively). Compared to patients with diabetic, hypertensive and polycystic nephropathy, those with glomerulopathy had significantly higher NLR values (3, 3.3, 3.5 and 4.1, respectively). Therefore, NLR was significantly higher in patients with CVD-derived CKD than those with glomerulopathy, where inflammation is a major pathogenetic mechanism. We divided our cohort to quartiles according to median RDW ($14.4\%$) and median NLR (3.3)—quartile 1: below median RDW below median NLR; quartile 2: below median NLR, above median RDW; quartile 3: above median NLR, below median RDW and quartile 4: above median RDW, above median NLR (Table 3). Dp-ucMGP and the prevalence of CVD were progressively increased across quartiles. Compared to the first quartile, patients at the fourth quartile had significantly lower Hb, DBP, mean BP, central DBP, calcium and serum albumin values and higher duration of CVD, phosphorus and CRP levels. Table 4 shows the multiple regression analysis in the study cohort with RDW as the dependent variable. In both unadjusted and adjusted for several well-established confounders affecting RDW values (serum albumin, CRP, UACR, calcium, phosphorus, cardiac rhythm, central DBP, warfarin, ESA), only circulating dp-ucMGP was an independent predictor of RDW (β = 0.001, $$p \leq 0.001$$). Table 5 shows the multiple regression analysis in the study cohort with NLR as the dependent variable. Even after adjustment for all variables correlated with NLR in the previous analyses (age, CVD, duration of CVD, serum albumin, CRP, eGFR, calcium, parathormone, central DBP, cause of nephropathy, ESA), dp-ucMGP, mean BP and gender were strong, independent predictors of NLR. ## 4. Discussion CVD still remains the leading cause of morbidity and mortality in CKD patients, accounting for more than $50\%$ of all deaths in ESKD. VC is highly prevalent in CKD patients and might explain the heavy CV burden that these patients carry. Therefore, the investigation of early, accurate biomarkers that might identify and classify CKD patients at risk for VC/VS is mandatory. However, although several markers have shown promising results, their use in everyday clinical practice is discouraged due to several limitations including high cost and/or complex or time-consuming measurement. We found that CVD was highly prevalent in our study cohort ($38.6\%$) and associated with RDW, NLR and dp-ucMGP. Further, when we divided our cohort into quartiles according to median values of RDW and NLR, we found that compared to the lowest quartile, patients with values above median of both RWD and NLR had significantly higher dp-ucMGP, prevalence of CVD, longer duration of T2DM, lower diastolic/mean and central diastolic BP, lower calcium and albumin and higher CRP and phosphorus, known traditional and non-traditional CV risk factors. Moreover, RDW was positively correlated with NLR, CRP, dp-ucMGP and phosphorus and negatively with peripheral and central DBP, UACR, calcium and albumin, whereas NLR was positively correlated with age, duration of CVD, CRP, PTH, dp-ucMGP, and inversely with peripheral and central DBP, calcium and albumin. RDW is a marker measuring the difference/variation in the size and volume of erythrocytes routinely reported in the total blood count. It is calculated from the division of the standard deviation of the mean corpuscular volume (MCV) by the MCV multiplied by 100 to yield a percentage value. The normal range of RDW is roughly between 11.5 and $15.4\%$. RDW was long used for the differential diagnosis of anemias; however, in recent years, it became evident that it might serve as a novel predictive marker of CVD and mortality in various settings, including the general population [27], HF [14,28,29] and CVD [30,31,32]. In line with these, a meta-analysis including 28 studies and approximately 103,000 patients with CVD showed that RDW was an independent predictor of adverse events including major CV events [33]. However, in CKD populations, the data regarding the association of RDW with CVD and mortality are mainly focused in ESKD populations, either HD [34,35] or PD [36,37,38]. In these patients, even a modest $1\%$ increase in RDW levels increases the risk for all-cause mortality by $47\%$ [39]. In pre-dialysis CKD, the data remain limited and are mainly derived from three retrospective studies [16,17,18,40]. In agreement with our results, we have previously showed in a cohort of 142 diabetic CKD patients in various CKD stages that RDW was an independent predictor of CV disease and mortality [17]. This association might be explained by the fact that VC (assessed by carotid intima media thickness) was the strongest independent factor predicting RDW values. In agreement with these findings, we found that dp-ucMGP (a well-established marker of VC) was independently associated with RDW values. Although VC was long considered a passive, degenerative process of the chronic accumulation of calcium in the arterial wall, this perspective recently changed and it is now well accepted that the calcification process is an active process that is regulated by molecules and proteins that act either as inhibitors or promoters [5,41]. Among these, MGP, a vitamin K-dependent protein, is the strongest natural inhibitor of VC that might abbrogate or even reverse the calcification process. To become active, MGP must undergo carboxylation and phosphorylation. MGP carboxylation requires vitamin K to be a cofactor to drive this reaction [42]. The fully inactive form of MGP, dp-ucMGP, reflects vitamin K deficiency and has been shown to be a reliable marker of arterial calcification and stiffness in various settings, including uremia. In CKD, dp-ucMGP is progressively increased in parallel with CKD stages [4,7], is tightly correlated with various surrogate markers of VC or VS (including cIMT, PWV, coronary calcification, Agatston score, etc.) and has been repeatedly shown to be a strong independent predictor of CVD [6,43]. Therefore, since vitamin K deficiency (assessed by increased dp-ucMGP) has been coherently shown to be a novel, non-traditional risk factor for CVD in CKD, several ongoing randomized controlled studies are currently investigating whether exogenous supplementation with vitamin K might ameliorate VC and protect from CVD by reducing circulating dp-ucMGP in CKD, HD and PD patients [42,44,45]. In our study, we found that both RDW and NLR were strongly associated with circulating dp-ucMGP; moreover, dp-ucMGP was the strongest independent factor predicting the values of these two markers, thus indicating that both RDW and NLR are associated with accelerated VC and an increased risk for CVD. According to our knowledge, this is the first study to report an association between RDW, NLR and dp-ucMGP. However, RDW has been repeatedly associated with other surrogate calcification markers, such as IMT [17,19,46,47,48], coronary calcium scores [49] and PWV [50,51]. In disagreement with Fornal et al. [ 52], we failed to show any association between RDW and PWV. However, this study included only hypertensive patients. Similar results were reported by two other small cohort studies in kidney transplant recipients [50,51]. One possible explanation might be the different population we enrolled and the fact that we used the optimal oscillometric method for measuring PWV, whereas the tonometric method is more sensitive and accurate [53]. Nonetheless, some indices of PWV (central diastolic blood pressure, diastolic blood pressure and heart rate) were found to be significantly correlated with RDW and NLR. Moreover, we found that RDW and NLR were associated with calcium, phosphate and PTH well-established traditional factors favoring VC in HD patients [54,55,56]. The associations of RDW with markers of VC and CVD might be attributed to shortened RBC life span, inhibited erythropoietin response, anemia and impaired iron metabolism. However, the correlation of RDW with VC was independent of Hb levels and the use of ESAs. Moreover, we found that patients receiving warfarin (a vitamin K antagonist) had significantly higher RDW levels, thus indicating that RDW is directly associated with vitamin K deficiency. Another explanation for our findings might be that anisocytosis might reflect a high inflammatory state and progressive CKD (as assessed by high UACR). The association between RDW, albuminuria [17,57,58] and inflammation in CKD settings [17,19,35,59] has been repeatedly reported before. We also found that besides RDW, NLR was associated with CRP, dp-ucMGP, CVD duration, indices of peripheral and central hemodynamics and eGFR. Although several biomarkers have been proposed as markers of VC and CVD in CKD, most of them necessitate special laboratory equipment or assays, need time-consuming methods or are expensive. In contrast, both RDW and NLR are simple, low in cost and quickly extracted by the complete blood count. NLR in peripheral blood is calculated by dividing the absolute number of neutrophils by the absolute number of lymphocytes, with data obtained by the total blood count. In healthy subjects, the normal NLR values are approximately between 0.8 and 3, whereas higher values indicate stress; in critically ill patients, NLR values might even reach 30 times higher than normal range values. NLR is a surrogate and reliable marker of systemic inflammation and has emerged during the past decade as a predictive biomarker for mortality and CVD in various populations, especially in CKD patients [21,60,61]. In our study, we found a tight correlation between NLR and CVD duration and traditional/non-traditional markers of CVD/VC. In agreement with our results, a very recent meta-analysis including 13 studies and 116,709 CKD patients showed that high NLR was a significant, strong predictor of both all-cause and CVD mortality (HR 1.93, $95\%$ CI 1.87–2.00; $p \leq 0.00001$ and HR 1.45, $95\%$ CI 1.18–1.79, $p \leq 0.001$, respectively) [62]. The association between NLR and VC is scarcely documented in CKD patients. Wang et al. reported that NLR was associated with coronary artery calcification in CKD stage 3–5 patients [63], whereas Li et al. showed that NLR was an independent risk factor of cardiac valve calcification in CKD patients [64] and another three small studies showed that VC was correlated with NLR in dialysis populations [65,66,67]. In our study, we found a strong association between NLR and the disturbed homeostasis of calcium–phosphate–PTH, which is a well-known molecular pathway underlying VC in uremic patients [68]. Moreover, we found that high NLR was correlated with peripheral and central hemodynamic parameters obtained by PWV measurements, which is in agreement with a previous study showing a close association between NLR and arterial stiffness in PD patients [69]. However, the main finding of our study was that dp-ucMGP, a reliable and novel marker of VC and vitamin K deficiency, was the strongest independent predictor of NLR values. According to our knowledge, this is the first study to ever report this association. We found that NLR was associated with eGFR and the presence of cardiorenal syndrome and/or glomerulopathy in our cohort. This finding could be interpreted by the fact that NLR is an independent risk factor for both CKD and HF/CVD [70,71]. Glomerulonephritis, on the other hand, is primarily an inflammatory disease which results in a congregation of inflammatory molecules that damages the glomeruli. Likewise, several investigators have previously reported that NLR was not only associated with renal function [72,73,74] but also predicted the deterioration of kidney function and CVD [21,75,76,77] in various types of glomerulopathies including lupus, vasculitis and IgA nephropathy. Our findings that both RDW and NLR were associated with VC, CVD and kidney function in a cohort of CKD patients might be explained by the fact that these two markers reflect systemic inflammation [78,79,80]. Since we found associations with both VC and VS, it could be hypothesized that both markers might reflect endothelial dysfunction in these patients. In this study, we found that parameters from the total blood count (RDW and NLR) are associated with inflammation, markers of CKD and traditional/non-traditional markers of endothelial dysfunction in a cohort of 158 CKD patients. Moreover, we found that circulating dp-ucMGP was the strongest predictor of both NLR and RDW. According to our knowledge, this is the first study to investigate the association between RDW, NLR and dp-ucMGP and other markers of VC or stiffness in a cohort of CKD patients in different stages. The strengths of our study include the large sample size, comprehensive data and the inclusion of several parameters to improve the statistical validity of our results. However, there are various limitations that should be acknowledged. Firstly, due to the cross-sectional, observational design, no causality can be established, and our results should be interpreted with caution. However, based on the results we obtained, we started following this cohort prospectively to investigate the possible predictive effect of the markers tested for CV events, mortality and CV mortality. Secondly, we cannot draw any definite conclusions regarding the mechanisms underlying the associations that were found, and thirdly, the single center design and the fact that other important parameters-markers of VC/VS were not included in our study. However, our study showed that information derived from a very simple, cheap and quick exam, the full blood count, might give a rough idea regarding the status of VC and/or CVD in uremic patients that carry a heavy CV burden. Therefore, our conclusions are speculative and require further, larger, prospective cohort studies in order to draw definite conclusions regarding the possible clinical utility of RDW and NLR. ## References 1. Parfrey P.S., Foley R.N.. The clinical epidemiology of cardiac disease in chronic renal failure. J. Am. Soc. Nephrol. (1999) 10 1606-1615. DOI: 10.1681/ASN.V1071606 2. Foley R.N., Murray A.M., Li S., Herzog C.A., McBean A.M., Eggers P.W., Collins A.J.. Chronic kidney disease and the risk for cardiovascular disease, renal replacement, and death in the United States Medicare population, 1998 to 1999. J. Am. Soc. Nephrol. (2005) 16 489-495. DOI: 10.1681/ASN.2004030203 3. Russo D., Palmiero G., De Blasio A.P., Balletta M.M., Andreucci V.E.. Coronary artery calcification in patients with CRF not undergoing dialysis. Am. J. Kidney Dis. (2004) 44 1024-1030. DOI: 10.1053/j.ajkd.2004.07.022 4. Roumeliotis S., Roumeliotis A., Stamou A., Leivaditis K., Kantartzi K., Panagoutsos S., Liakopoulos V.. The association of dp-ucMGP with cardiovascular morbidity and decreased renal function in diabetic chronic kidney disease. Int. J. Mol. Sci. (2020) 21. DOI: 10.3390/ijms21176035 5. Roumeliotis S., Roumeliotis A., Dounousi E., Eleftheriadis T., Liakopoulos V.. Biomarkers of vascular calcification in serum. Adv. Clin. Chem. (2020) 98 91-147. PMID: 32564789 6. Roumeliotis S., Dounousi E., Eleftheriadis T., Liakopoulos V.. Association of the inactive circulating matrix Gla protein with vitamin K intake, calcification, mortality, and cardiovascular disease: A review. Int. J. Mol. Sci. (2019) 20. DOI: 10.3390/ijms20030628 7. Schurgers L.J., Barreto D.V., Barreto F.C., Liabeuf S., Renard C., Magdeleyns E.J., Vermeer C., Choukroun G., Massy Z.A.. The circulating inactive form of matrix gla protein is a surrogate marker for vascular calcification in chronic kidney disease: A preliminary report. Clin. J. Am. Soc. Nephrol. (2010) 5 568-575. DOI: 10.2215/CJN.07081009 8. Townsend R., Anderson A., Chirinos J., Feldman H., Grunwald J., Nessel L., Roy J., Weir M., Wright Jr J., Bansal N.. CRIC Study Investigators: Association of pulse wave velocity with chronic kidney disease progression and mortality: Findings from the CRIC study (Chronic Renal Insufficiency Cohort). Hypertension (2018) 71 1101-1107. DOI: 10.1161/HYPERTENSIONAHA.117.10648 9. Buttarello M.. Laboratory diagnosis of anemia: Are the old and new red cell parameters useful in classification and treatment, how?. Int. J. Lab. Hematol. (2016) 38 123-132. DOI: 10.1111/ijlh.12500 10. Zalawadiya S.K., Veeranna V., Niraj A., Pradhan J., Afonso L.. Red cell distribution width and risk of coronary heart disease events. Am. J. Cardiol. (2010) 106 988-993. DOI: 10.1016/j.amjcard.2010.06.006 11. van Kimmenade R.R., Mohammed A.A., Uthamalingam S., van der Meer P., Felker G.M., Januzzi J.L.. Red blood cell distribution width and 1-year mortality in acute heart failure. Eur. J. Heart Fail. (2010) 12 129-136. DOI: 10.1093/eurjhf/hfp179 12. Lappegård J., Ellingsen T.S., Hindberg K., Mathiesen E.B., Njølstad I., Wilsgaard T., Løchen M.-L., Brækkan S.K., Hansen J.-B.. Impact of chronic inflammation, assessed by hs-CRP, on the association between red cell distribution width and arterial cardiovascular disease: The Tromsø Study. TH Open (2018) 2 e182-e189. DOI: 10.1055/s-0038-1651523 13. Sotiropoulos K., Yerly P., Monney P., Garnier A., Regamey J., Hugli O., Martin D., Metrich M., Antonietti J.P., Hullin R.. Red cell distribution width and mortality in acute heart failure patients with preserved and reduced ejection fraction. ESC Heart Fail. (2016) 3 198-204. DOI: 10.1002/ehf2.12091 14. Felker G.M., Allen L.A., Pocock S.J., Shaw L.K., McMurray J.J., Pfeffer M.A., Swedberg K., Wang D., Yusuf S., Michelson E.L.. Red cell distribution width as a novel prognostic marker in heart failure: Data from the CHARM Program and the Duke Databank. J. Am. Coll. Cardiol. (2007) 50 40-47. DOI: 10.1016/j.jacc.2007.02.067 15. Deng X., Gao B., Wang F., Zhao M.-h., Wang J., Zhang L.. Red blood cell distribution width is associated with adverse kidney outcomes in patients with chronic kidney disease. Front. Med. (2022) 9. DOI: 10.3389/fmed.2022.877220 16. Hsieh Y.-P., Chang C.-C., Kor C.-T., Yang Y., Wen Y.-K., Chiu P.-F.. The predictive role of red cell distribution width in mortality among chronic kidney disease patients. PLoS ONE (2016) 11. DOI: 10.1371/journal.pone.0162025 17. Roumeliotis S., Stamou A., Roumeliotis A., Theodoridis M., Leivaditis K., Panagoutsos S., Liakopoulos V.. Red blood cell distribution width is associated with deterioration of renal function and cardiovascular morbidity and mortality in patients with diabetic kidney disease. Life (2020) 10. DOI: 10.3390/life10110301 18. Kor C.-T., Hsieh Y.-P., Chang C.-C., Chiu P.-F.. The prognostic value of interaction between mean corpuscular volume and red cell distribution width in mortality in chronic kidney disease. Sci. Rep. (2018) 8 11870. DOI: 10.1038/s41598-018-19881-2 19. Solak Y., Gaipov A., Turk S., Kayrak M., Yilmaz M.I., Caglar K., Verim S., Unal H.U., Gok M., Demirkaya E.. Red cell distribution width is independently related to endothelial dysfunction in patients with chronic kidney disease. Am. J. Med. Sci. (2014) 347 118-124. DOI: 10.1097/MAJ.0b013e3182996a96 20. Forget P., Khalifa C., Defour J.-P., Latinne D., Van Pel M.-C., De Kock M.. What is the normal value of the neutrophil-to-lymphocyte ratio?. BMC Res. Notes (2017) 10. DOI: 10.1186/s13104-016-2335-5 21. Zhao W.-M., Tao S.-M., Liu G.-L.. Neutrophil-to-lymphocyte ratio in relation to the risk of all-cause mortality and cardiovascular events in patients with chronic kidney disease: A systematic review and meta-analysis. Ren. Fail. (2020) 42 1059-1066. DOI: 10.1080/0886022X.2020.1832521 22. Levey A.S., Coresh J., Balk E., Kausz A.T., Levin A., Steffes M.W., Hogg R.J., Perrone R.D., Lau J., Eknoyan G.. National Kidney Foundation practice guidelines for chronic kidney disease: Evaluation, classification, and stratification. Ann. Intern. Med. (2003) 139 137-147. DOI: 10.7326/0003-4819-139-2-200307150-00013 23. Workgroup K.. KDIGO clinical practice guideline for acute kidney injury. Nephron Clin. Pract. (2012) 2 179-184 24. Jaminon A.M., Dai L., Qureshi A.R., Evenepoel P., Ripsweden J., Söderberg M., Witasp A., Olauson H., Schurgers L.J., Stenvinkel P.. Matrix Gla protein is an independent predictor of both intimal and medial vascular calcification in chronic kidney disease. Sci. Rep. (2020) 10 1-9. DOI: 10.1038/s41598-020-63013-8 25. Vaios V., Georgianos P.I., Vareta G., Dounousi E., Eleftheriadis T., Papagianni A., Zebekakis P.E., Liakopoulos V.. A comparative analysis of ambulatory BP profile and arterial stiffness between CAPD and APD. Hum. Hypertens. (2022) 36 254-262. DOI: 10.1038/s41371-021-00516-5 26. Franssen P.M., Imholz B.P.. Evaluation of the Mobil-O-Graph new generation ABPM device using the ESH criteria. Blood Press. Monit. (2010) 15 229-231. DOI: 10.1097/MBP.0b013e328339be38 27. Perlstein T.S., Weuve J., Pfeffer M.A., Beckman J.A.. Red blood cell distribution width and mortality risk in a community-based prospective cohort. Arch. Intern. Med. (2009) 169 588-594. DOI: 10.1001/archinternmed.2009.55 28. Allen L.A., Felker G.M., Mehra M.R., Chiong J.R., Dunlap S.H., Ghali J.K., Lenihan D.J., Oren R.M., Wagoner L.E., Schwartz T.A.. Validation and potential mechanisms of red cell distribution width as a prognostic marker in heart failure. J. Card. Fail. (2010) 16 230-238. DOI: 10.1016/j.cardfail.2009.11.003 29. Al-Najjar Y., Goode K.M., Zhang J., Cleland J.G., Clark A.L.. Red cell distribution width: An inexpensive and powerful prognostic marker in heart failure. Eur. J. Heart Fail. (2009) 11 1155-1162. DOI: 10.1093/eurjhf/hfp147 30. Ye Z., Smith C., Kullo I.J.. Usefulness of red cell distribution width to predict mortality in patients with peripheral artery disease. Am. J. Cardiol. (2011) 107 1241-1245. DOI: 10.1016/j.amjcard.2010.12.023 31. Azab B., Torbey E., Hatoum H., Singh J., Khoueiry G., Bachir R., McGinn J.T., McCord D., Lafferty J.. Usefulness of red cell distribution width in predicting all-cause long-term mortality after non-ST-elevation myocardial infarction. Cardiology (2011) 119 72-80. DOI: 10.1159/000329920 32. Dabbah S., Hammerman H., Markiewicz W., Aronson D.. Relation between red cell distribution width and clinical outcomes after acute myocardial infarction. Am. J. Cardiol. (2010) 105 312-317. DOI: 10.1016/j.amjcard.2009.09.027 33. Hou H., Sun T., Li C., Li Y., Guo Z., Wang W., Li D.. An overall and dose-response meta-analysis of red blood cell distribution width and CVD outcomes. Sci. Rep. (2017) 7 43420. DOI: 10.1038/srep43420 34. Fukasawa H., Ishibuchi K., Kaneko M., Niwa H., Yasuda H., Kumagai H., Furuya R.. Red blood cell distribution width is associated with all-cause and cardiovascular mortality in hemodialysis patients. Ther. Apher. Dial. (2017) 21 565-571. DOI: 10.1111/1744-9987.12579 35. Vashistha T., Streja E., Molnar M.Z., Rhee C.M., Moradi H., Soohoo M., Kovesdy C.P., Kalantar-Zadeh K.. Red cell distribution width and mortality in hemodialysis patients. Am. J. Kidney Dis. (2016) 68 110-121. DOI: 10.1053/j.ajkd.2015.11.020 36. Sun I.O., Chung B.H., Yoon H.J., Kim J.H., Choi B.S., Park C.W., Kim Y.S., Yang C.W., Lee K.Y.. Clinical significance of red blood cell distribution width in the prediction of mortality in patients on peritoneal dialysis. Kidney Res. Clin. Pract. (2016) 35 114-118. DOI: 10.1016/j.krcp.2016.03.003 37. Soohoo M., Molnar M.Z., Ujszaszi A., Obi Y., Kovesdy C.P., Kalantar-Zadeh K., Streja E.. Red blood cell distribution width and mortality and hospitalizations in peritoneal dialysis patients. Ephrol. Dial. Transplant. (2019) 34 2111-2118. DOI: 10.1093/ndt/gfy196 38. Cao H.-X., Zhao X.-d., Yan L., Fan X.-G., Shao F.-M.. Correlation between red blood cell distribution width and cardiovascular events in the patients receiving peritoneal dialysis: A Strobe-compliant article. Medicine (2019) 98 e14376. DOI: 10.1097/MD.0000000000014376 39. Zhang T., Li J., Lin Y., Yang H., Cao S.. Association between red blood cell distribution width and all-cause mortality in chronic kidney disease patients: A systematic review and meta-analysis. Arch. Med. Res. (2017) 48 378-385. DOI: 10.1016/j.arcmed.2017.06.009 40. den Harder A.M., de Jong P.A., de Groot M.C., Wolterink J.M., Budde R.P., Iŝgum I., van Solinge W.W., Maarten J., Lutgens E., Veldhuis W.B.. Commonly available hematological biomarkers are associated with the extent of coronary calcifications. Atherosclerosis (2018) 275 166-173. DOI: 10.1016/j.atherosclerosis.2018.06.017 41. Roumeliotis S., Dounousi E., Salmas M., Eleftheriadis T., Liakopoulos V.. Vascular calcification in chronic kidney disease: The role of vitamin K-dependent matrix Gla protein. Front. Med. (2020) 7 154. DOI: 10.3389/fmed.2020.00154 42. Roumeliotis S., Duni A., Vaios V., Kitsos A., Liakopoulos V., Dounousi E.. Vitamin K Supplementation for Prevention of Vascular Calcification in Chronic Kidney Disease Patients: Are We There Yet?. Nutrients (2022) 14. DOI: 10.3390/nu14050925 43. Roumeliotis S., Roumeliotis A., Panagoutsos S., Giannakopoulou E., Papanas N., Manolopoulos V.G., Passadakis P., Tavridou A.. Matrix Gla protein T-138C polymorphism is associated with carotid intima media thickness and predicts mortality in patients with diabetic nephropathy. J. Diabetes Its Complicat. (2017) 31 1527-1532. DOI: 10.1016/j.jdiacomp.2017.06.012 44. Roumeliotis S., Roumeliotis A., Dounousi E., Eleftheriadis T., Liakopoulos V.. Vitamin K for the treatment of cardiovascular disease in End-Stage Renal Disease patients: Is there hope?. Curr. Vasc. Pharmacol. (2021) 19 77-90. DOI: 10.2174/18756212MTA1sNDEzz 45. Roumeliotis S., Roumeliotis A., Georgianos P.I., Thodis E., Schurgers L.J., Maresz K., Eleftheriadis T., Dounousi E., Tripepi G., Mallamaci F.. VItamin K In PEritonial DIAlysis (VIKIPEDIA): Rationale and study protocol for a randomized controlled trial. PLoS ONE (2022) 17. DOI: 10.1371/journal.pone.0273102 46. Lappegård J., Ellingsen T.S., Vik A., Skjelbakken T., Brox J., Mathiesen E.B., Johnsen S.H., Brækkan S.K., Hansen J.-B.. Red cell distribution width and carotid atherosclerosis progression The Tromsø Study. Thromb. Haemost. (2015) 113 649-654. PMID: 25631329 47. Söderholm M., Borné Y., Hedblad B., Persson M., Engström G.. Red cell distribution width in relation to incidence of stroke and carotid atherosclerosis: A population-based cohort study. PLoS ONE (2015) 10. DOI: 10.1371/journal.pone.0124957 48. Sicaja M., Pehar M., Đerek L., Starčević B., Vuletić V., Romić Ž., Božikov V.. Red blood cell distribution width as a prognostic marker of mortality in patients on chronic dialysis: A single center, prospective longitudinal study. Croat. Med. J. (2013) 54 25-32. DOI: 10.3325/cmj.2013.54.25 49. Khalil A., Shehata M., Abdeltawab A., Onsy A.. Red blood cell distribution width and coronary artery disease severity in diabetic patients. Future Cardiol. (2019) 15 355-366. PMID: 31496273 50. Czyzewski L., Wyzgal J., Czyzewska E., Kurowski A., Sierdzinski J., Truszewski Z., Szarpak L.. Assessment of arterial stiffness, volume, and nutritional status in stable renal transplant recipients. Medicine (2016) 95 e2819. DOI: 10.1097/MD.0000000000002819 51. Pacek A., Saran M., Wyzgał J.. 24-hour Arterial stiffness monitoring in kidney transplant recipients in the early postoperative period. Trans. Proc. (2018) 50 1824-1828. DOI: 10.1016/j.transproceed.2018.04.004 52. Fornal M., Wizner B., Cwynar M., Królczyk J., Kwater A., Korbut R.A., Grodzicki T.. Association of red blood cell distribution width, inflammation markers and morphological as well as rheological erythrocyte parameters with target organ damage in hypertension. Clin. Hemorheol. Microcirc. (2014) 56 325-335. DOI: 10.3233/CH-131745 53. Del Giorno R., Troiani C., Gabutti S., Stefanelli K., Gabutti L.. Comparing oscillometric and tonometric methods to assess pulse wave velocity: A population-based study. Ann. Med. (2021) 53 1-16. DOI: 10.1080/07853890.2020.1794538 54. Giachelli C.M., Speer M.Y., Li X., Rajachar R.M., Yang H.. Regulation of vascular calcification: Roles of phosphate and osteopontin. Circ. Res. (2005) 96 717-722. DOI: 10.1161/01.RES.0000161997.24797.c0 55. Giachelli C.M.. The emerging role of phosphate in vascular calcification. Kidney Int. (2009) 75 890-897. DOI: 10.1038/ki.2008.644 56. Shroff R., Long D.A., Shanahan C.. Mechanistic insights into vascular calcification in CKD. Am. Soc. Nephrol. (2013) 24 179-189. DOI: 10.1681/ASN.2011121191 57. Assulyn T., Khamisy-Farah R., Nseir W., Bashkin A., Farah R.. Neutrophil-to-lymphocyte ratio and red blood cell distribution width as predictors of microalbuminuria in type 2 diabetes. J. Clin. Lab. Anal. (2020) 34 e23259. DOI: 10.1002/jcla.23259 58. Uslu A.U., Yonem O., Aydin B., Uncu T., Seven D., Balta S., Cicekli E.. Red cell distribution width is associated with albuminuria in adults with familial Mediterranean fever. Kaohsiung J. Med. Sci. (2016) 32 216-220. DOI: 10.1016/j.kjms.2016.03.009 59. Peng F., Li Z., Zhong Z., Luo Q., Guo Q., Huang F., Yu X., Yang X.. An increasing of red blood cell distribution width was associated with cardiovascular mortality in patients on peritoneal dialysis. Int. J. Cardiol. (2014) 176 1379-1381. DOI: 10.1016/j.ijcard.2014.08.030 60. Li W.-J., Chen X.-M., Nie X.-Y., Zhang J., Cheng Y.-J., Lin X.-X., Wu S.-H.. Cardiac troponin and C-reactive protein for predicting all-cause and cardiovascular mortality in patients with chronic kidney disease: A meta-analysis. Clinics (2015) 70 301-311. DOI: 10.6061/clinics/2015(04)14 61. Goicoechea M., Quiroga B., García de Vinuesa S., Verdalles Ú., Reque J., Panizo N., Arroyo D., Santos A., Macías N., Luño J.. Intraindividual interleukin-6 variations on the cardiovascular prognosis of patients with chronic renal disease. Ren. Fail. (2012) 34 1002-1009. DOI: 10.3109/0886022X.2012.696469 62. Ao G., Wang Y., Qi X., Wang F., Wen H.. Association of neutrophil-to-lymphocyte ratio and risk of cardiovascular or all-cause mortality in chronic kidney disease: A meta-analysis. Clin. Exp. Nephrol. (2021) 25 157-165. DOI: 10.1007/s10157-020-01975-9 63. Wang X.-r., Yuan L.-., Shi R., Li H., Wang D.-G., Wu Y.-G.. Predictors of coronary artery calcification and its association with cardiovascular events in patients with chronic kidney disease. Ren. Fail. (2021) 43 1172-1179. DOI: 10.1080/0886022X.2021.1953529 64. Li B., Hu X., Pan H., Xiao Y., Dong J., Bao Y., Fang M.. Cardiac valve calcification prevalence and association with neutrophil-to-lymphocyte ratio in newly diagnosed patients with non-dialysis chronic kidney disease stage 3-5. Bratisl. Med. J. (2022) 123 523-527. DOI: 10.4149/BLL_2022_084 65. Turkmen K., Guney I., Yerlikaya F.H., Tonbul H.Z.. The relationship between neutrophil-to-lymphocyte ratio and inflammation in end-stage renal disease patients. Ren. Fail. (2012) 34 155-159. DOI: 10.3109/0886022X.2011.641514 66. Yaprak M., Turan M.N., Dayanan R., Akın S., Değirmen E., Yıldırım M., Turgut F.. Platelet-to-lymphocyte ratio predicts mortality better than neutrophil-to-lymphocyte ratio in hemodialysis patients. Int. Urol. Nephrol. (2016) 48 1343-1348. DOI: 10.1007/s11255-016-1301-4 67. Chandra A., Raj G., Awasthi N.P., Rao N., Srivastava D.. Evaluation of the relationship between blood cell parameters and vascular calcification in dialysis-dependent end-stage renal disease patients. Saudi J. Kidney Dis. Transplant. (2020) 31 136 68. Shanahan C.M., Crouthamel M.H., Kapustin A., Giachelli C.M.. Arterial calcification in chronic kidney disease: Key roles for calcium and phosphate. Circ. Res. (2011) 109 697-711. DOI: 10.1161/CIRCRESAHA.110.234914 69. Cai K., Luo Q., Zhu B., Han L., Wu D., Dai Z., Wang K.. Neutrophil-lymphocyte ratio is associated with arterial stiffness in patients with peritoneal dialysis. BMC Nephrol. (2016) 17. DOI: 10.1186/s12882-016-0394-4 70. Delcea C., Buzea C.A., Dan G.A.. The neutrophil to lymphocyte ratio in heart failure: A comprehensive review. Rom. J. Intern. Med. (2019) 57 296-314. DOI: 10.2478/rjim-2019-0018 71. Durmus E., Kivrak T., Gerin F., Sunbul M., Sari I., Erdogan O.. Neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio are predictors of heart failure. Arq. Bras. Cardiol. (2015) 105 606-613. DOI: 10.5935/abc.20150126 72. Li Q., Shi S., Liu L., Lv J., Zhu L., Zhang H.. Neutrophil-to-lymphocyte ratio as an independent inflammatory indicator for poor renal prognosis in adult IgA vasculitis with nephritis. Int. Immunopharmacol. (2022) 111 109178. DOI: 10.1016/j.intimp.2022.109178 73. Tang D., Tang Q., Zhang L., Wang H.. High Neutrophil-lymphocyte Ratio Predicts Serious Renal Insufficiency in Patients with Lupus Nephritis. Iran. J. Immunol. (2022) 19 49-57 74. Fonseca J.A., Gameiro J., Duarte I., Jorge S., Lopes J.A.. The neutrophil-to-lymphocyte ratio as a marker of vasculitis activity, severe infection and mortality in anca-associated vasculitis: A retrospective study. Nefrologia (2021) 41 321-328. DOI: 10.1016/j.nefro.2020.07.013 75. Lu X., Wang S., Zhang G., Xiong R., Li H.. High neutrophil-to-lymphocyte ratio is a significant predictor of cardiovascular and all-cause mortality in patients undergoing peritoneal dialysis. Kidney Blood Press. Res. (2018) 43 490-499. DOI: 10.1159/000488696 76. Yuan Q., Wang J., Peng Z., Zhou Q., Xiao X., Xie Y., Wang W., Huang L., Tang W., Sun D.. Neutrophil-to-lymphocyte ratio and incident end-stage renal disease in Chinese patients with chronic kidney disease: Results from the Chinese Cohort Study of Chronic Kidney Disease (C-STRIDE). J. Transl. Med. (2019) 17 86. DOI: 10.1186/s12967-019-1808-4 77. Liao J., Wei D., Sun C., Yang Y., Wei Y., Liu X.. Prognostic value of the combination of neutrophil-to-lymphocyte ratio, monocyte-to-lymphocyte ratio and platelet-to-lymphocyte ratio on mortality in patients on maintenance hemodialysis. MC Nephrol. (2022) 23 393. DOI: 10.1186/s12882-022-03020-1 78. Zahorec R.. Neutrophil-to-lymphocyte ratio, past, present and future perspectives. Bratisl. Lek. Listy (2021) 122 474-488. DOI: 10.4149/BLL_2021_078 79. Okyay G.U., İnal S., Öneç K., Er R.E., Paşaoğlu Ö., Paşaoğlu H., Derici Ü., Erten Y.. Neutrophil to lymphocyte ratio in evaluation of inflammation in patients with chronic kidney disease. Ren. Fail. (2013) 35 29-36. DOI: 10.3109/0886022X.2012.734429 80. Lippi G., Targher G., Montagnana M., Salvagno G.L., Zoppini G., Guidi G.C.. Relation between red blood cell distribution width and inflammatory biomarkers in a large cohort of unselected outpatients. Arch. Pathol. Lab. Med. (2009) 133 628-632. DOI: 10.5858/133.4.628
--- title: Design and Evaluation of Pegylated Large 3D Pore Ferrisilicate as a Potential Insulin Protein Therapy to Treat Diabetic Mellitus authors: - B. Rabindran Jermy - Mohammed Salahuddin - Gazali Tanimu - Hatim Dafalla - Sarah Almofty - Vijaya Ravinayagam journal: Pharmaceutics year: 2023 pmcid: PMC9966771 doi: 10.3390/pharmaceutics15020593 license: CC BY 4.0 --- # Design and Evaluation of Pegylated Large 3D Pore Ferrisilicate as a Potential Insulin Protein Therapy to Treat Diabetic Mellitus ## Abstract An iron-based SBA-16 mesoporous silica (ferrisilicate) with a large surface area and three-dimensional (3D) pores is explored as a potential insulin delivery vehicle with improved encapsulation and loading efficiency. Fe was incorporated into a framework of ferrisilicate using the isomorphous substitution technique for direct synthesis. Fe3+ species were identified using diffuse reflectance spectroscopy. The large surface area (804 m2/g), cubic pores (3.2 nm) and insulin loading were characterized using XRD, BET surface area, FTIR and TEM analyses. For pH sensitivity, the ferrisilicate was wrapped with polyethylene glycol (MW = 400 Daltons) (PEG). For comparison, Fe (10 wt%) was impregnated on a Korea Advanced Institute of Science and Technology Number 6 (KIT-6) sieve and Mesocellular Silica Foam (MSU-F). Insulin loading was optimized, and its release mechanism was studied using the dialysis membrane technique (MWCO = 14,000 Da) at physiological pH = 7.4, 6.8 and 1.2. The kinetics of the drug’s release was studied using different structured/insulin nanoformulations, including Santa Barbara Amorphous materials (SBA-15, SBA-16), MSU-F, ultra-large-pore FDU-12 (ULPFDU-12) and ferrisilicates. A different insulin adsorption times (0.08–1 h), insulin/ferrisilicate ratios (0.125–1.0) and drug release rates at different pH were examined using the Korsmeyer–Peppas model. The rate of drug release and the diffusion mechanisms were obtained based on the release constant (k) and release exponent (n). The cytotoxicity of the nanoformulation was evaluated by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay using human foreskin fibroblast (HFF-1) cells. A low cytotoxicity was observed for this nanoformulation starting at the highest concentrations used, namely, 400 and 800 μg. The hypoglycemic activity of insulin/ferrisilicate/PEG on acute administration in Wistar rats was studied using doses of 2, 5 and 10 mg/kg body weight. The developed facile ferrisilicate/PEG nanoformulation showed a high insulin encapsulation and loading capacity with pH-sensitive insulin release for potential delivery through the oral route. ## 1. Introduction Diabetic mellitus (DM) is a metabolic disorder categorized by hyperglycemia due to inopportune insulin secretion. DM is differentiated as type I (insulin-reliant) and type II diabetes (insulin resistance) [1,2]. Diabetic mellitus, a disease termed as a life style disease, is quickly turning into a global epidemic. The prime reason for this is attributed due to change in life style, unhealthy diets and lack of awareness. In 2019, diabetic severity resulted in 1.5 million mortalities, and notably, $48\%$ of these deaths occurred before 70 years of age. Among the two types of diabetes, type II diabetes is dominant, accounting for $95\%$ (WHO). Recent data show that worldwide, about 537 million people have diabetes, and this number is expected to reach 783 million by 2045 [3]. Such a high percentage and increasing rate of DM is primarily attributed to obesity and changes in life style. The healthcare expenditure spent on diabetic treatment was estimated to be USD 966 billion in 2021. Insulin was discovered by Frederick Banting in 1921, while Charles Best developed the clinical use of insulin in 1922 [4]. Insulin is administered to control the blood glucose level. Insulin helps to uptake glucose by binding with the insulin receptor and intitiating several protein activations cascades (e.g., Glut-4 transporter to plasma member, influx of glucose, synthesis of glycogen, glycolysis and triglyceride production). Type 1 DM (due to defective pancreatic β cells) depends on the lifelong supply of insulin. In the case of type 2 DM patients, the peripheral cells resist the administration of insulin, while some patients at the lateral life stage also require insulin. In order to treat type 1 DM, a common mode of insulin administration is through the subcutaneous route with about four injections per day. The treatment affects patient compliance and induces several side effects (lipoatrophy). Still, the subcutaneous administration route is preferred due to insulin’s low bioavailability and challenges in developing an effective drug delivery system due to low membrane permeability and molecular size constraints. In order to limit the number of injection cycles, a controlled insulin release strategy was followed using zinc and protamine. The used formulation showed poor reproducible kinetic parameters and the effect between meal periods was minimal and led to hypoglycemic events [5]. Accordingly, similar to insulin, insulin detemir and insulin glargine were shown to subvert the hyproglycemic action, but the required dose level was almost double compared to the normal dose of insulin for one day [6]. It has been reported that diabetic mediators similar to insulin could also alter the mitogenic pathway and can be a potential carcinogen in the long run [7]. The construction of a stimuli-responsive smart drug release system is the most recent attractive research direction involving interdisciplinary research between material science and medical science. An appropriately constructed nanovehicle with controlled insulin delivery using biocompatible nanosilica is proposed to overcome the deficiencies in subcutaneous therapy, improve therapeutic efficiency, enhance the stability of drug release and make ease diet control and exercise regiments. Several glucose-sensitive smart drug delivery system based on phenylboronic acid (PBA) and proteins such as concanavalin and glucose oxidase have been reported [8]. However, such a glucose-sensitive, sensor-based nanovehicle requires a multi-step synthesis procedure, the use of solvents and an advanced chemical set-up. Recently, a biocompatible drug delivery system based on structured silica/polymeric nanocomposites are shown to be a promising nanovehicle to carry insulin [9]. A microneedle design based on mesoporous silica capped with zinc oxide in the form of an insulin reservoir has been reported to effectively control insulin delivery for prolonged periods of time [10]. Several studies are ongoing to improve the efficacy of protein delivery using mesoporous silica/chitosan and poly(lactic-co-glycolic) acid nanoformulations and improve their permeability [11]. The isomorphous substitution of biocompatible metals such as Fe, Zn, Ti, etc., into the silica framework is gaining importance in biomedical applications [12,13]. The use of Fe cations (Fe3+ and Fe2+) with particle sizes ranging between 3 and 15 nm is gaining attention in multifunctional therapeutics as contrasting agents for magnetic resonance imaging, in hyperthermia treatments and as drug delivery agents [14]. The presence of iron oxide nanoparticles (FeNPs) favors biocompatibility [15], and as such, they are applied in hyperthermia for their anticancer [16,17] and antibacterial activity [18], as well as in tissue engineering [19]. Previously, we reported a direct hydrothermal synthesis of Iron-incorporated Santa Barbara Amorphous 16 (FeSBA-16) [20]. The presence of large 3D cubic pores of ferrisilicate could be exploited for insulin entrapment/loading capacity and insulin release. The wrapping of the nanocarrier with polyethylene glycol is reported to improve the bioavailability and drug stabilization and facilitate the transport of protein across human gastrointestinal fluid [21]. In this study, we investigated the effect of a pegylated, large 3D porous ferrisilicate/insulin nanoformulation for diabetes management. The textural characteristics are investigated using different physico-chemical characterization techniques. The insulin encapsulation/loading capacity and the pH-based, smart kinetic release behavior in response to stimuli were studied for insulin release. Furthermore, the nanoformulation toxicity in vitro and hypoglycemic effect in vivo were assessed. ## 2. Material and Methods The silica source tetraethylorthosilicate (reagent grade, $98\%$, Sigma Aldrich, Darmstadt, Germany) and non-ionic template Pluronic F127 (BioReagent, suitable for cell culture, BASF, Wyandotte, MI, USA), iron(III) nitrate nonahydrate (≥$99.95\%$, BioReagent, suitable for cell culture, Sigma Aldrich, Saint Louis, MO, USA), n-butanol (≥$99\%$, anhydrous, Sigma Aldrich, Saint Louis, MO, USA), human recombinant insulin (rHu, dry powder, Sigma-Aldrich Chemie Holding GmbH, Taufkirchen, Germany) and polyethylene glycol (BioUltra, MW = 400 Daltons, Sigma-Aldrich Chemie Holding GmbH, Taufkirchen, Germany) were obtained from Sigma Aldrich. All chemicals were used as received without any further purification. ## 2.1.1. Synthesis of Ferrisilicate Using Hydrothermal Technique Fe-SBA-16, termed as ferrisilicate, was prepared using sol–gel technique. The ferrisilicate containing Fe species can be tuned between SiO2/Fe2O3 ratios of 50 and 250. In the present study, the Fe content can reach a SiO2/Fe2O3 ratio of 50. In brief, 5 g of F127 was dissolved in acidic HCl solution (2 M) and stirred for 1 h. Then, 16 g of n-butanol (co-solvent) was added along with 24 g of tetraethylorthosilicate and the iron source (0.186 g of iron nitrate nonahydrate (Si/Fe ratio 250)) and stirred for 24 h. The mixture was stored in a polypropylene bottle (Nalgene, Thermo Fisher Scientific, NY, USA) and transferred to an oven to be hydrothermally aged at 100 °C for 24 h. The precipitate was filtered, washed several times with excess water and dried at 100 °C for 12 h. The as-synthesized sample was finally calcined at 550 °C for 6 h. ## 2.1.2. Synthesis of Iron-Impregnated Structured Silica (10 wt% Fe/KIT-6 and 10 wt% Fe/MSU-F) Using Impregnation Technique Firstly, 0.7235 g of iron nitrate nonahydrate was dissolved in 80 ml of distilled water. Then, 1.0 g of KIT-6, mesosilicalite, or Mesocellular Silica Foam (MSU-F) was added and stirred for 24 h at room temperature (RT). The solution was dried at 120 °C for 3 h and calcined at 500 °C for 2 h. ## 2.1.3. Insulin/Ferrisilicate For insulin loading, 80 mg of insulin was added to 8 ml of 0.01 M HCl solution and stirred for 20 min. Then, 160 mg of ferrisilicate was added and stirred at 300 rpm overnight in an ice-cold environment. After that, the mixture was filtered, washed with 5 ml of distilled water and dried at RT (5 h) and stored at 4 °C. ## 2.1.4. Insulin/Ferrisilicate/PEGylation For PEGylation, 14 μl of PEG (Molecular weight = 400) was added in 3 ml of deionized water, stirred for 20 min under argon atmosphere and then 150 mg of Insulin/Ferrisilicate was added and stirred under an ice-cold environment for 24 h. Then, the mixture was freeze-dried using the lyophilization technique. ## 2.2. Characterization Techniques The phase of insulin, insulin/ferrisilicate/PEG, was identified using benchtop XRD (Miniflex 600, Rigaku, Tokyo, Japan). The textural features, including BET surface area, pore size and pore volume, were measured using the nitrogen adsorption technique (ASAP-2020 plus, Micromeritics, Norcross, GA, USA). The ferrite nanoparticles’ chemical coordination was analyzed using DRS-UV-visible spectroscopy analysis (V-750, JASCO, Tokyo, Japan). The insulin functional groups of our nanoformulation were determined using FT-IR spectroscopy (L160000A, Perkin Elmer, Waltham, MA, USA). The morphological variations of insulin/ferrisilicate/PEG were investigated using transmission electron microscopy (TEM, JEM2100F, JEOL, Tokyo, Japan). ## 2.3. Insulin Entrapment Efficiency and Loading Capacity Insulin entrapment efficiency (EE %) and loading capacity (LC %) over ferrisilicate were estimated using UV-visible spectroscopy at the specific wavelength of 275 nm. EE %=Amount of insulin in ferrisilicate¯Initial amount of insulin×100LC%=Initial amount of insulin −Insulin in supernatant¯Amount of ferrisilicate +insulin×100 ## 2.4. Insulin Release Study The release trends of different nanoformulations were studied using the dialysis membrane technique. First, 30 mg of the nanoformulation was placed inside 3 mL of PBS solution inside the dialysis membrane. The release of insulin was monitored under different pH solutions (7.4, 6.8 and 1.2) at 37 °C. At regular time intervals, 10 mL of solution was withdrawn and replaced with an equal volume of fresh solution. The amount of released insulin was identified at a specific wavelength of 275 nm. In order to measure the insulin release, a standard curve for insulin was established through calibration. An initial stock solution of 1000 μg/mL of insulin using phosphate buffer solution was prepared. We prepared 10 mL of 6 different concentrations of 5, 10, 15, 20, 25 and 30 μg/mL from stock solution using the working PBS solution (pH = 1.2 or 6.8 or 7.4), and then a calibration plot was established at the maximum absorption wavelength (λmax = 275 nm). The linear regression was found to be $y = 0.0069$x + 0.0071, where y corresponds to absorbance and x to the concentration of the released drug (μg/mL). A linear calibration plot with a correlation coefficient of 0.993 was used to quantify insulin release from the nanoformulations. Each experiment was performed in triplicates. ## 2.5. Cytotoxicity of Insulin/Ferrisilicate Nanoformulation against HFF-1 Cells Human foreskin fibroblast (HFF-1) cells were obtained as (SCRC-1041TM, ATCC, Manassas, VA, USA) and maintained in DMEM supplied with $10\%$ fetal bovine serum, $1\%$ L-glutamine and $1\%$ penicillin-streptomycin (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) in a humidified $5\%$ CO2 incubator (Galaxy® 170S, Eppendorf, Stevenage, UK) at 37 °C. The cells were seeded in a 96-well plate at (104 cells/well) and treated with group (a), insulin/ferrisilicate/PEG, insulin/10 wt% Fe/KIT-6/PEG, ferrisilicate/PEG and 10 wt% Fe/KIT-6/PEG using 25, 50, 100, 200, 400 and 800 μg/mL, and group (b), free insulin at 12.5, 25, 50, 100, 200 and 400 μg/mL, for 24, 48 and 72 h, including (c) untreated cells as control cells in each experiment. Both groups of concentrations are labeled as 1, 2, 3, 4 and 5 and were applied corresponding to the functionalized encapsulated insulin. MTT assay was performed to determine the cytotoxicity of the ferrisilicate nanoformulations by using a colorimetric-based reaction using MTT reagent 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide to assess the metabolic activity of the cells. For the assay, 10 μl of MTT reagent ($98\%$, 2128-1G, Sigma-Aldrich Chemie Holding GmbH, Taufkirchen, Germany) was added to obtain a final dilution of 1:10 and incubated for 4 h in a CO2 incubator at 37 °C. The formed formazan blue dye was solubilized by adding 100 μl of DMSO (Dimethyl sulfoxide) and read at 570 nm wavelength by a SYNERGY Neo2 multi-mode microplate reader (BioTek Instruments, Winooski, VT, USA). Cell viability was calculated as: Cell viability %=Absorbance of Sample¯Absorbance of Control×100 ## 2.6. Statistics The cytotoxicity assay was performed in three independent experiments. Statistical analysis was performed using Prism 9 software (GraphPad, La Jolla, CA, USA). The analysis was performed using two-way ANOVA with Dunnett’s multiple comparison test. Statistically non-significant p values were indicated as (ns). The data analysis of drug delivery was conducted using Prism 8 software and SPSS software version 20.0 (IBM Corp., Armonk, NY, USA). ## 2.7. In Vivo Study Wistar rats of either sex weighing 180–200 g were used for the experiments. The animals are obtained from the IRMC animal house and were maintained and treated according to the IACUC policy. The study was approved by Imam Abdulrahman Bin Faisal University IRB through IRB number IRB-2020-13-278 with approval date of 30 September 2020. A single-dose study was performed to learn the effect of testing samples in diabetic rats using acute administration. The diabetic rats were divided into three groups, as follows: Experimental Grouping: Group 1: Diabetic rats orally administrated with normal saline (marketed dehydration fluid and electrolyte replenishment of sodium chloride) without the nanoformulation; Group 2, 3 and 4: Diabetic rats treated with nanoformulations (orally administered) at three gradient doses of 2, 5 and 10 IU/kg body weight, respectively. Blood glucose levels were measured at the start of the study (will be considered as initial blood glucose level) for all the animals included in the study. The glucose solution (2 g/kg) was administered orally 30 min after the administration of the testing samples, and blood samples were collected at 1, 2, 3, 4, 5 and 6 h after glucose administration to estimate the blood glucose levels. ## 3.1. Characterization of Ferrisilicate/Insulin/PEG Figure 1 depicts the XRD patterns of (A) insulin, (B) ferrisilicate and the (C) insulin/ferrisilicate/PEG nanoformulation. The recombinant insulin powder exhibited a crystalline peaks (2θ range 5–15°), characteristic of macrostructured protein. The ferrisilicate exhibited a broad amorphous silica peak between 10 and 40°. In the insulin/ferrisilicate/PEG nanoformulation, the crystalline peak of insulin is absent, which clearly indicates the molecular dispersion or amorphous transformation of ferrisilicate. A nitrogen adsorption isotherm was used to characterize the surface texture and pore diameter of mesoporous materials in the range of 2–50 nm. In our case, the changes in the ferrisilicate’s surface texture and three-dimensional cubic pores before and after insulin and PEG modifications were analyzed (Figure 2A,B). Parent ferrisilicate with an Im3m space group in the calcined form exhibited a type IV isotherm pattern with surface area of 804 m2/g, pore volume of 0.65 cm3/g and pore size centered at about 3.2 nm. After insulin loading and PEG wrapping, the quantity of nitrogen adsorbed reflected in the peak height that steeply decreased with capillary condensation (Figure 2A(a,b)). The hysteresis loop slightly decreases, indicating the occupation of pores by insulin. In parallel, the surface area (335 m2/g) and pore volume (0.28 cm3/g) decreases while the pore size stays at about 3.3 nm (Figure 2B(c,d)). The textural changes in the samples before and after insulin/PEG wrapping were analyzed using nitrogen adsorption isotherm, and the results are presented in Table 1. A systematic change in the surface and pore volume indicates a successful insulin inclusion into the cubic pores of the ferrisilicate, while PEG wrapping occupies about $42\%$ of the area around the 3D ferrisilicate. FeKIT-6 and mesosilicalite exhibited a similar high surface area and porous architecture. Overall, the textural modification indicates the insulin deposition in cubic cage pores of ferrisilicate (Figure 2A,B). The diffuse reflectance UV-visible spectra of ferrisilicate, insulin, and insulin/ferrisilicate/PEG are shown in Figure 2C(e–g). Ferrisilicate indicates a broad adsorption band in the 200–300 nm range, with peak maxima at 216 and 245 nm. The presence of a high-energy band is ascribed to the ligand-to-metal charge transfer due to tetrahedral Fe3+ species [22]. The band at 216 nm and 245 nm is ascribed to th(e electronic transition of O2− to t2g and eg orbitals of Fe3+ in the iron oxide cluster. Unlike iron oxide-loaded mesoporous silica, ferrisilicate also shows a broad unresolved absorption band expanding between 400 and 500 nm. The absorption at 400 and 500 nm are ascribed to the quantum size effect of alpha-Fe2O3 species. Ferrisilicate shows the presence of Fe3+ species and the presence of a few hexacoordinated alpha-Fe2O3 species (Figure 2C(e)). Such iron oxide species occur in aggregated form, with octahedral or distorted octahedral coordination [23], inside the cubic pore channels of ferrisilicate. The insulin binding ability of ferrisilicate was measured using DRS-UV-Visible spectra. Insulin showed a strong absorption band at about 285 nm. The administering of ferrisilicate loaded with insulin (50 wt/wt%) followed by washing showed a similar peak to that of insulin at about 281 nm. The PEG wrapping shows effective conjugation with insulin through hydrogen bonding (Figure 2C(f,g)). Hinds et al. [ 2002] [24] stated that the PEG hydroxyl group interact with the amine functional moiety of the aminoacids of insulin. The FTIR spectra of ferrisilicate, insulin/ferrisilicate/PEG and insulin are shown in Figure 2D(h–j). Ferrisilicate exhibited characteristic peaks at the hydroxyl region between 3730 cm−1 and 3610 cm−1. The presence of a broad band at about 3610 cm−1 is attributed to the hydroxyl group bridging the signalling between the Bronsted sites related to the isomorphous substitution of Fe for Si in the ferrisilicate framework [25]. Insulin showed characteristic carbonyl (C=O) stretching vibration bands due to the presence of amide at 1646 cm−1 and 1533 cm−1. In case of insulin/ferrisilicate/PEG, the characteristic insulin vibration bands appear at about 1639 cm−1 and 1518 cm−1. An increase in such vibrational peaks clearly shows the coupling of insulin with ferrisilicate. Moreover, there is a shift in the main vibration bands of insulin occurring from 1646 cm−1 to 1639 cm−1. Such a shift in bands after the functionalization of insulin indicates the effective interaction between the insulin and ferrisilicate [26]. The loading capacity and entrapment of insulin inside the pore channels play a critical role in its release and anti-diabetic activity. Elongation in the broad hydroxyl (3000 cm−1) and amine stretching (3291 cm−1) bands compared to ferrisilicate indicates the effective functionalization of insulin in the insulin/ferrisilicate/PEG nanoformulation. Figure 3A–D shows the morphological analyses of ferrisilicate and insulin-loaded, PEG-coated ferrisilicate. Ferrisilicate exhibits clear pore channels running on the parallel axis (Figure 3A). The analysis of the insulin-loaded, PEG-wrapped sample clearly shows the occurrence of structural transformation and enveloping by the polymeric layers. This indicates successful insulin loading and wrapping by PEG. In addition, magnification to 10 nm shows the presence of pore channels without any major changes inside the ferrisilicate (Figure 3C,D). The effect of different porous-structured nanoformulations, entrapment efficiencies, loading capacities and insulin adsorption times ($t = 0.08$ h, 0.5 h, 0.75 h and 1.0 h) on insulin release are shown in Figure 4A–D. The insulin entrapment efficiency of hexagon-shaped SBA-15, cubic-shaped SBA-16, mesocellular forms with large pore windows, ultralarge-pore FDU-12 and ferrisilicate were studied. MSU-F and ULPFDU-12 contain large pores of 22 nm and 25 nm, respectively. SBA-15, SBA-16 and ferrisilicate have pore sizes of 8 nm, 5 nm and 3.2 nm, respectively. The entrapment efficiency was in the range of 82–$93\%$, while the loading capacity was in the range of 38–$46\%$. Interestingly, ferrisilicate showed high entrapment ($92\%$) efficiency compared to large-pore nanomaterials MSU-F ($91\%$) and ULPFDU-12 ($93\%$). SBA-15 showed a slightly lower entrapment efficiency of $82\%$ (Figure 4A). The insulin release profiles of encapsulated samples were studied using the dialysis membrane technique (Figure 4B). As expected, in the absence of polymer wrapping, the samples showed a high percentage cumulative insulin release. SBA-16, in the absence of iron species, showed a lower release of $68\%$ at 72 h. The introduction of iron into SBA-16 silica (ferrisilicate) improves the insulin release by $95\%$ compared to its silica counterpart, SiSBA-16. The presence of large pores favors the high release profile for MSU-F ($91\%$), ULPFDU-12 ($98\%$) and SBA-15 ($91\%$). The effect of insulin adsorption over different adsorption times ($t = 0.08$ h, 0.5 h, 0.75 h and 1.0 h) on insulin release is shown in Figure 4C,D. Ferrisilicate shows an increase in insulin adsorption with time ($15.7\%$ in 0.08 h, $18.6\%$ in 0.5 h, $20.5\%$ in 0.75 h and $22\%$ within 1 h). This result demonstrates that a higher adsorption time reduces insulin release. It indicates an effective entrapment with higher adsorption time. The entrapment efficiency and loading capacity reveals the efficiency of the drug-loaded nanoformulation [27]. A glucose-responsive system has been reported from PBA containing structured silica coated with diol-based copolymers (N-acryloyl glucosamine and N-isopropyl acrylamide). The formulation showed a high loading capacity ($14.7\%$) and encapsulation efficiency ($85.9\%$) with glucose responsive release at pH = 7.4 [28]. It has been reported that PBA- and diol-based block copolymers along with post-modification improved the glucose-sensitive release of insulin [29]. Replacing PBA with fluorophenylboronic acid has been reported to improve insulin loading, resulting in a high encapsulation efficiency and better glucose responses in physiological conditions [30]. Glucose-responsive sulfonamide–PBA, with its temperature-responsive properties, was reported to improve both loading capacity and encapsulation efficiency. These nanoparticles are reported to be safe and effective for the subcutaneous injection of insulin [31]. In our case, the entrapment efficiency was $92\%$, while the drug loading capacity reached up to $46\%$. The insulin release over pegylated ferrisilicate, 10 wt% Fe/mesosilicalite and Fe/KIT-6 at different pH conditions (7.4, 6.8 and 1.2) are studied for 530 h (Figure 5). Ferrisilicate displays a high percentage of cumulative release of insulin of about 40–$50\%$. The 10 wt% Fe/mesosilicalite and Fe/KIT-6, which contain 3D pore architectures, showed a release of about 20–$30\%$ over 530 h. This indicates the ink-shaped pores of SBA-16 (about 3.3 nm) are slightly restricted with Fe impregnation, showing a sustained release behavior with respect to insulin. The presence of 3D cage-type pores with Ia3d structures in KIT-6 was found to favor the slow release of insulin ($20\%$ for 530 h), while mesosilicalite with hexagonal pores of MCM-41 showed a release of about $30\%$ over 530 h. This suggests that insulin tends to functionalize on the external micropores of mesosilicalite, while cage-type pores are able to accommodate the insulin inside the mesopores. Polysaccharide pullulan hydrogel in the form of carboxylation was reacted with concanavalin A (Con A) using amidization reaction. The nanoformulation was shown to control the release of insulin due to the specific bonding occurring between protein andglucose binding and to the crosslinked structure with uniform pores that accommodates insulin [32]. A detailed study on the glucose-responsive insulin release behavior of hydrogel/microgels-Con A nanoformulations shows that the release trend is controlled by the bolus and basal insulin release and network composition [33]. Konjac Mannan (heteropolysaccharide) fabricated with Con A through the crosslinking technique exhibited a targeted insulin release. The nanoformulation with a particle size of about 500 nm facilitated a glucose-responsive insulin release with a reversible pattern with different levels of glucose. Furthermore, the in vivo study reveals that the nanoparticles are non-toxic and able to control the blood sugar level for 6 h [34]. An immobilization of glucose oxidase on a biocompatible linear polysaccharide alginate/phenylboronic acid-derived composite exhibited an improved glucose sensitivity with a potential option for subcutaneous insulin delivery [35]. However, the protein stability, antigenicity and synthesis cost limit the design of such protein-based nanoformulations for clinical translation [36,37]. In case of phenylboronic acid functionalization, the cytotoxicity and lower solubility due to higher pKa (~9.0) are some limitations [38]. In the present study, the facile and simple nanocomposite formation between ferrisilicate and PEG can perform similar pH-sensitive insulin release. ## 3.2. Kinetics of Different Ferrisilicate/Insulin Nanoformulation Drug Release Using the Korsmeyer–Peppas Model The different ferrisilicate/insulin nanoformulation drug release profiles at different pHs were examined using the Korsmeyer–Peppas model, expressed using the equation:R %=ktn where R% is the different ferrisilicalite/insulin drug percentage release at time (t) and k and n are the kinetic release rate constant and the release exponent, respectively. The kinetic parameters with their $95\%$ confidence intervals are presented in Table 2. For all the base materials; SiSBA-15, SiSBA-16, SiMSU-F and SiULPFDU-12, the rate of drug release is enhanced as observed from the higher release constants which are all higher than that of the modified FeSBA-16 base material. Similarly, the release mechanisms of the drugs nanoformulation obtained from the base materials followed the fickian diffusion mechanism (n < 0.45), while the modified FeSBA-16 drug release followed the non-fickian (0.45 < 0.488 < 0.89) diffusion mechanism. These indicate that modifying the base materials affect both the rate of release and the release diffusion mechanism. The effect of four different insulin adsorption times (5, 30, 45 and 60 min) on the rate of release and diffusion mechanism of ferrisilicate nanoformulation drug shown that, the drug release mechanism at all the insulin adsorption times followed the fickian (n < 0.45) diffusion. However, the rate of drug release as signified by the released constant revealed that, the rate of drug release is highest after 30 min of insulin adsorption. Also, there is no direct correlation between the adsorption times studied and the rate of drug release, because the rate increased up to a maximum at 30 min adsorption, then continued to decline at higher insulin adsorption. This indicates that, there is a saturation limit of insulin adsorption, above which the release rate of the drug is affected. Different insulin/ferrisilicate ratios of 0.125, 0.25, 0.75 and 1.0 were utilized in the nanoformulation of drugs. The rate of drug release is highest at higher ratio of 1.0, indicating that the higher the insulin/ferrisilicate ratio, the greater the rate of drug release, except at the ratio of 0.75. The nanoformulation at all ratios followed the fickian diffusion mechanism. This confirmed that, the kinetic release profile of the nanoformed insulin/ferrisilicate drug is affected by the ratio of the insulin to the ferrisilicate base material. The effect of different pH of ferrisilicate on the diffusion mechanisms and rate of ferrisilicate/insulin/PEG drug release revealed that, at close to neutral pH of 6.8, the rate of drug release is highest relative to highly acidic (pH = 1.2) and slightly basic (pH = 7.4) conditions. However, at both pH of 1.2 and 6.8, the release exponent signified fickian diffusion mechanism, while at pH of 7.4, the release mechanism changed to non-fickian diffusion. For these nanoformulation, it can be inferred that the pH of ferrisilicate affects both the rate of drug release and the diffusion mechanism. Four drug nanoformulation with different base materials modification of ferrisilicate but having the same PEG coating at the same pH revealed that, the rate of drug release is enhanced with the modification, because of their higher release constant compared to the unmodified ferrisilicate. Also, the diffusion mechanism is affected by the modification as follows; 10 wt% Fe/KIT-6 and 10 wt% Fe/Mesosilicalite based nanoformulation drugs release followed the fickian (n < 0.45) diffusion mechanism while unmodified ferrisilicatefollowed the non-fickian (0.45 < n < 0.89) diffusion mechanism. For the null hypothesis: there is no significant differences between the various release exponent and 0.45, one-paired t-test was carried out and the calculated p-value (0.00003736) which is <<0.05 showed that the null hypothesis is invalid, rather the alternative hypothesis: that there are significant differences between the release exponents (n) and 0.45 is valid. ## 3.3. In Vitro Study The cytotoxicity of ferrisilicate nanoformulations were studied on Human foreskin fibroblast (HFF-1) cells (Figure 6). Insulin/Ferrisilicate/PEG and insulin/10 wt% Fe/KIT-6/PEG were obliviously showing no cytotoxicity at 25, 50, 100, 200 μg/mL after each timepoint even after 72 h, and started to show cytotoxic effect by less than $50\%$ cell viability at the highest concentration 800 μg/mL. Ferrisilicate, insulin/10 wt% Fe/KIT-6 and insulin were used as control groups to assess the cytotoxicity of empty vectors or NP material, as shown ferrisilicate and 10 wt% Fe/KIT-6 were not toxic even at the highest concentration 800 μg/mL post 72 h of treatment. Expectedly, insulin has stimulated cell growth to reach to $130\%$ in comparison to DMEM- treated cells which have a $100\%$ cell viability and used as control values in this experiment. ## 3.4. In Vivo Study To find out the hypoglycemic effect of the prepared insulin nanoformulation, it was administered orally to the diabetic animals at three gradient doses 2, 5 and 10 (Insulin Unit) IU/kg body weight (Figure 7). The blood glucose level was measured at different intervals up to 6 h. The results of the study indicate the hypoglycemic effect of insulin nanoformulation at a dose of 5 and 10 IU/kg, and the formulation was found to be effective after 2 h of administration at a dose of 5 mg/kg. Furthermore, it reduced the blood glucose level significantly ($p \leq 0.0001$) until 3 h, after which it starts to move towards the initial blood glucose values. At 5 IU/kg, the insulin/ferrisilicate/PEG nanoformulation decreases the blood glucose level from 501 to 375 mg/dL (reduction of $25\%$). Moreover, the formulation at 10 IU/kg body weight showed a significant reduction ($p \leq 0.0001$) in blood glucose level after 1 h of administration and was found to be significant until 3 h of administration. It reduced the blood glucose level from 417 to 268 mg/dL ($35\%$ reduction). These effects of the nanoformulation at a dose of 5 and 10 IU/kg body weight were found to be significant ($p \leq 0.0001$), as compared to diabetic control. No significant effect of insulin formulation was found at 2 IU/kg body weight. ## 4. Conclusions In the present investigation, pegylated 3D cubic porous ferrisilicate was explored for diabetic management. The physico-chemical characterization reveals the incorporation of Fe3+ species into the framework (via isomorphous substitution) of SBA-16. A large surface area and pore size were found to be sufficient for a high encapsulation of insulin ($46\%$) at 1.5 h. The adsorption increases with time to about $57\%$ at 1.5 h. Ferrisilicate exhibited a high percentage of cumulative insulin release of about $32\%$ in 530 h at pH 7.4. Insulin release was notably improved by PEG wrapping (0.08 μL/mg) to about $41\%$ in 530 h at pH 6.8. As a comparison, Fe/Mesocellular foam, Fe/Mesosilicalite and Fe/KIT-6 showed about $40\%$, $28\%$ and $19.2\%$ insulin release in 530 h. The kinetic studies using the Korsmeyer–Peppas model revealed that the nature of the base materials, different insulin adsorption times, varying insulin/ferrisilicate ratios, different pHs and different base material modifications all affect the rate of insulin release and its diffusion mechanism. The nanoformulation showed very low toxicity in in vitro study and a hypoglycemic effect in in vivo study. Doses of 5 and 10 mg/kg body weight of pegylated ferrisilicate were found to be significant ($p \leq 0.0001$) compared to the diabetic control. Overall, the developed 3D porous ferrisilicate wrapped with polyethylene glycol mimics smart stimuli-responsive behavior with a pH-sensitive insulin release with high cell viability for potential application in diabetic management. ## References 1. Von Scholten B.J., Kreiner F.F., Gough S.C.L., von Herrath M.. Current and future therapies for type 1 diabetes. Diabetologia (2021.0) 64 1037-1048. DOI: 10.1007/s00125-021-05398-3 2. Reed J., Bain S., Kanamarlapudi V.. A review of current trends with Type 2 diabetes epidemiology, aetiology, pathogenesis, treatments and future perspectives. Diabetes Metab. Syndr. Obes. (2021.0) 14 3567. DOI: 10.2147/DMSO.S319895 3. IDF Diabetes Atlas 4. Vecchio I., Tornali C., Bragazzi N.L., Martini M.. The discovery of insulin: An important milestone in the history of medicine. Front. Endocrinol. (2018.0) 9 613. DOI: 10.3389/fendo.2018.00613 5. Holleman F., Gale E.A.. Nice insulins, pity about the evidence. Diabetologia (2007.0) 50 1783-1790. DOI: 10.1007/s00125-007-0763-4 6. Girish C., Manikandan S., Jayanthi M.. Practitioners section-Newer insulin analogues and inhaled insulin. Indian J. Med. Sci. (2006.0) 60 117-123. DOI: 10.4103/0019-5359.22765 7. Jonasson J., Ljung R., Talbäck M., Haglund B., Gudbjörnsdòttir S., Steineck G.. Insulin glargine use and short-term incidence of malignancies—A population-based follow-up study in Sweden. Diabetologia (2009.0) 52 1745-1754. DOI: 10.1007/s00125-009-1444-2 8. Najmeddine A.A., Saeed M., Beadham I.G., ElShaer A.. Efficacy and safety of glucose sensors for delivery of insulin: A Systematic Review. PharmaNutrition (2021.0) 18 100280. DOI: 10.1016/j.phanu.2021.100280 9. Kuang Y., Zhai J., Xiao Q., Zhao S., Li C.. Polysaccharide/mesoporous silica nanoparticle-based drug delivery systems: A review. Int. J. Biol. Macromol. (2021.0) 193 457-473. DOI: 10.1016/j.ijbiomac.2021.10.142 10. Fu Y., Liu P., Chen M., Jin T., Wu H., Hei M., Wang C., Xu Y., Qian X., Zhu W.. On-demand transdermal insulin delivery system for type 1 diabetes therapy with no hypoglycemia risks. J. Colloid Interface Sci. (2022.0) 605 582-591. DOI: 10.1016/j.jcis.2021.07.126 11. Wang M., Wang C., Ren S., Pan J., Wang Y., Shen Y., Zeng Z., Cui H., Zhao X.. Versatile Oral Insulin Delivery Nanosystems: From Materials to Nanostructures. Int. J. Mol. Sci. (2022.0) 23. DOI: 10.3390/ijms23063362 12. Tong X., Li Z., Chen W., Wang J., Li X., Mu J., Tang Y., Li L.. Efficient catalytic ozonation of diclofenac by three-dimensional iron (Fe)-doped SBA-16 mesoporous structures. J. Colloid Interface Sci. (2020.0) 578 461-470. DOI: 10.1016/j.jcis.2020.06.003 13. Kunde G.B., Yadav G.D.. Green strategy for the synthesis of mesoporous, free-standing MAl2O4 (M = Fe, Co, Ni, Cu) spinel films by sol–gel method. Mater. Sci. Eng. B (2021.0) 271 115244. DOI: 10.1016/j.mseb.2021.115244 14. Samrot A.V., Sahithya C.S., Selvarani J., Purayil S.K., Ponnaiah P.. A review on synthesis, characterization and potential biological applications of superparamagnetic iron oxide nanoparticles. Curr. Res. Green Sustain. Chem. (2021.0) 4 100042. DOI: 10.1016/j.crgsc.2020.100042 15. Materón E.M., Miyazaki C.M., Carr O., Joshi N., Picciani P.H., Dalmaschio C.J., Davis F., Shimizu F.M.. Magnetic nanoparticles in biomedical applications: A review. Appl. Surf. Sci. Adv. (2021.0) 6 100163. DOI: 10.1016/j.apsadv.2021.100163 16. Jermy B.R., Alomari M., Ravinayagam V., Almofty S.A., Akhtar S., Borgio J.F., AbdulAzeez S.. SPIONs/3D SiSBA-16 based Multifunctional Nanoformulation for target specific cisplatin release in colon and cervical cancer cell lines. Sci. Rep. (2019.0) 9 14523. DOI: 10.1038/s41598-019-51051-w 17. Balasamy R.J., Ravinayagam V., Alomari M., Ansari M.A., Almofty S.A., Rehman S., Dafalla H., Marimuthu P.R., Akhtar S., Al Hamad M.. Cisplatin delivery, anticancer and antibacterial properties of Fe/SBA-16/ZIF-8 nanocomposite. RSC Adv. (2019.0) 9 42395-42408. DOI: 10.1039/C9RA07461A 18. Jose S., Senthilkumar M., Elayaraja K., Haris M., George A., Raj A.D., Sundaram S.J., Bashir A.K., Maaza M., Kaviyarasu K.. Preparation and characterization of Fe doped n-hydroxyapatite for biomedical application. Surf. Interfaces (2021.0) 25 101185. DOI: 10.1016/j.surfin.2021.101185 19. Janarthanan G., Noh I.. Recent trends in metal ion based hydrogel biomaterials for tissue engineering and other biomedical applications. J. Mater. Sci. Technol. (2021.0) 63 35-53. DOI: 10.1016/j.jmst.2020.02.052 20. Jermy B.R., Kim S.Y., Bineesh K.V., Selvaraj M., Park D.W.. Easy route for the synthesis of Fe-SBA-16 at weak acidity and its catalytic activity in the oxidation of cyclohexene. Microporous Mesoporous Mater. (2009.0) 121 103-113. DOI: 10.1016/j.micromeso.2009.01.010 21. Andreani T., de Souza A.L.R., Kiill C.P., Lorenzon E.N., Fangueiro J.F., Calpena A.C., Chaud M.V., Garcia M.L., Gremiao M.P.D., Silva A.M.. Preparation and characterization of PEG-coated silica nanoparticles for oral insulin delivery. Int. J. Pharm. (2014.0) 473 627-635. DOI: 10.1016/j.ijpharm.2014.07.049 22. Nakai M., Miyake K., Inoue R., Ono K., Al Jabri H., Hirota Y., Uchida Y., Miyamoto M., Nishiyama N.. *Synthesis of high silica BEA type ferrisilicate (Fe-Beta) by dry gel conversion method using dealuminated zeolites and its catalytic performance on acetone to olefins (ATO) reaction*. Microporous Mesoporous Mater.* (2019.0) 273 189-195. DOI: 10.1016/j.micromeso.2018.06.008 23. Bhaumik A., Samanta S., Mal N.K.. Iron oxide nanoparticles stabilized inside highly ordered mesoporous silica. Pramana (2005.0) 65 855-862. DOI: 10.1007/BF02704085 24. Hinds K.D., Kim S.W.. Effects of PEG conjugation on Insulin properties. Adv. Drug Deliv. Rev. (2002.0) 54 505-530. DOI: 10.1016/S0169-409X(02)00025-X 25. Yu Q., Feng Y., Tang X., Yi H., Zhao S., Gao F., Zhou Y., Zhang Y., Zhuang R.. A novel ferrisilicate MEL zeolite with bi-functional adsorption/catalytic oxidation properties for non-methane hydrocarbon removal from cooking oil fumes. Microporous Mesoporous Mater. (2020.0) 309 110509. DOI: 10.1016/j.micromeso.2020.110509 26. Massaro M., Cavallaro G., Colletti C.G., D’Azzo G., Guernelli S., Lazzara G., Pieraccini S., Riela S.. Halloysite nanotubes for efficient loading, stabilization and controlled release of insulin. J. Colloid Interface Sci. (2018.0) 524 156-164. DOI: 10.1016/j.jcis.2018.04.025 27. Kim M.K., Ki D.H., Na Y.G., Lee H.S., Baek J.S., Lee J.Y., Lee H.K., Cho C.W.. Optimization of Mesoporous Silica Nanoparticles through Statistical Design of Experiment and the Application for the Anticancer Drug. Pharmaceutics (2021.0) 13. DOI: 10.3390/pharmaceutics13020184 28. Huang Q., Yu H., Wang L., Shen D., Chen X., Wang N.. Synthesis and testing of polymer grafted mesoporous silica as glucose-responsive insulin release drug delivery systems. Eur. Polym. J. (2021.0) 157 110651. DOI: 10.1016/j.eurpolymj.2021.110651 29. Gaballa H., Theato P.. Glucose-responsive polymeric micelles via boronic acid–diol complexation for insulin delivery at neutral pH. Biomacromolecules (2019.0) 20 871-881. DOI: 10.1021/acs.biomac.8b01508 30. Luo L., Song R., Chen J., Zhou B., Mao X., Tang S.. Fluorophenylboronic acid substituted chitosan for insulin loading and release. React. Funct. Polym. (2020.0) 146 104435. DOI: 10.1016/j.reactfunctpolym.2019.104435 31. Zhang P., Ma Q., He D., Liu G., Tang D., Liu L., Wu J.. A new Glucose-Responsive delivery system based on Sulfonamide-phenylboronic acid for subcutaneous insulin injection. Eur. Polym. J. (2021.0) 157 110648. DOI: 10.1016/j.eurpolymj.2021.110648 32. Lin K., Yi J., Mao X., Wu H., Zhang L.M., Yang L.. Glucose-sensitive hydrogels from covalently modified carboxylated pullulan and concanavalin A for smart controlled release of insulin. React. Funct. Polym. (2019.0) 139 112-119. DOI: 10.1016/j.reactfunctpolym.2019.01.016 33. Yin R., Bai M., He J., Nie J., Zhang W.. Concanavalin A-sugar affinity based system: Binding interactions, principle of glucose-responsiveness, and modulated insulin release for diabetes care. Int. J. Biol. Macromol. (2019.0) 124 724-732. DOI: 10.1016/j.ijbiomac.2018.11.261 34. Xu M., Huang J., Jiang S., He J., Wang Z., Qin H., Guan Y.Q.. Glucose sensitive konjac glucomannan/concanavalin A nanoparticles as oral insulin delivery system. Int. J. Biol. Macromol. (2022.0) 202 296-308. DOI: 10.1016/j.ijbiomac.2022.01.048 35. Chai Z., Dong H., Sun X., Fan Y., Wang Y., Huang F.. Development of glucose oxidase-immobilized alginate nanoparticles for enhanced glucose-triggered insulin delivery in diabetic mice. Int. J. Biol. Macromol. (2020.0) 159 640-647. DOI: 10.1016/j.ijbiomac.2020.05.097 36. Wu J.Z., Williams G.R., Li H.Y., Wang D., Wu H., Li S.D., Zhu L.M.. Glucose-and temperature-sensitive nanoparticles for insulin delivery. Int. J. Nanomed. (2017.0) 12 4037. DOI: 10.2147/IJN.S132984 37. Jana B.A., Shinde U., Wadhwani A.. Synthetic enzyme-based nanoparticles act as smart catalyst for glucose responsive release of insulin. J. Biotechnol. (2020.0) 324 1-6. DOI: 10.1016/j.jbiotec.2020.09.023 38. Ma R., Yang H., Li Z., Liu G., Sun X., Liu X., An Y., Shi L.. Phenylboronic acid-based complex micelles with enhanced glucose-responsiveness at physiological pH by complexation with glycopolymer. Biomacromolecules (2012.0) 13 3409-3417. DOI: 10.1021/bm3012715
--- title: Integrated Computational Approaches for Inhibiting Sex Hormone-Binding Globulin in Male Infertility by Screening Potent Phytochemicals authors: - Suvro Biswas - Mohasana Akter Mita - Shamima Afrose - Md. Robiul Hasan - Md. Tarikul Islam - Md. Ashiqur Rahman - Mst. Jasmin Ara - Md. Bakhtiar Abid Chowdhury - Habibatun Naher Meem - Md. Mamunuzzaman - Tanvir Ahammad - Istiaq Uddin Ashik - Munjed M. Ibrahim - Mohammad Tarique Imam - Mohammad Akbar Hossain - Md. Abu Saleh journal: Life year: 2023 pmcid: PMC9966787 doi: 10.3390/life13020476 license: CC BY 4.0 --- # Integrated Computational Approaches for Inhibiting Sex Hormone-Binding Globulin in Male Infertility by Screening Potent Phytochemicals ## Abstract Male infertility is significantly influenced by the plasma-protein sex hormone-binding globulin (SHBG). Male infertility, erectile dysfunction, prostate cancer, and several other male reproductive system diseases are all caused by reduced testosterone bioavailability due to its binding to SHBG. In this study, we have identified 345 phytochemicals from 200 literature reviews that potentially inhibit severe acute respiratory syndrome coronavirus 2. Only a few studies have been done using the SARS-CoV-2 inhibitors to identify the SHBG inhibitor, which is thought to be the main protein responsible for male infertility. In virtual-screening and molecular-docking experiments, cryptomisrine, dorsilurin E, and isoiguesterin were identified as potential SHBG inhibitors with binding affinities of −9.2, −9.0, and −8.8 kcal/mol, respectively. They were also found to have higher binding affinities than the control drug anastrozole (−7.0 kcal/mol). In addition to favorable pharmacological properties, these top three phytochemicals showed no adverse effects in pharmacokinetic evaluations. Several molecular dynamics simulation profiles’ root-mean-square deviation, radius of gyration, root-mean-square fluctuation, hydrogen bonds, and solvent-accessible surface area supported the top three protein–ligand complexes’ better firmness and stability than the control drug throughout the 100 ns simulation period. These combinatorial drug-design approaches indicate that these three phytochemicals could be developed as potential drugs to treat male infertility. ## 1. Introduction Infertility is a reproductive system disease resulting in an inability to achieve a clinical pregnancy despite regular unprotected sexual intercourse for ≥12 months, impacting approximately 72.4 million couples globally [1,2,3]. Among the estimated 8–$12\%$ of reproductive-aged couples affected worldwide, 20–$30\%$ of infertility cases are exclusively due to male infertility, contributing to $50\%$ of overall cases [4,5,6]. Male infertility impedes spermatogenesis, diminishing the quality and quantity of sperm, and is often observed as altered sperm concentration, motility, and morphology in nearly $7\%$ of all males [1,7,8,9]. Male infertility can be categorized into defective spermatogenesis, defective transport, and ineffective delivery. According to the US Centers for Disease Control and Prevention, about 40–$50\%$ of cases are due to male-factor infertility, and $2\%$ are due to suboptimal sperm parameters [1,10]. The reasons underlying male infertility include chronic liver diseases, diabetes mellitus, chronic smoking, insufficient vitamins, coronary heart diseases, and a few genetic factors that adversely affect spermatogenesis [9,11]. An in-depth literature review indicated that infertility could arise due to decreased androgen levels, which play a major role in normal spermatogenesis maintenance. Infertility due to reduced testicular function is also common, the symptoms reflecting reduced testosterone production and serum and intratesticular levels due to reduced gonadotropin (e.g., follicle-stimulating hormone and luteinizing hormone) production regulated at the pituitary level by estrogen [12,13]. Therefore, estrogen has a direct deleterious effect on spermatogenesis since reduced testosterone–estrogen ratios are observed in infertility cases [12,14,15]. Notably, a balance between serum androgens and estrogens is required for normal semen parameters, suggesting a disrupted endocrine mechanism through binding to nuclear receptors, including the estrogen and androgen receptors, because their entry into target cells is essentially regulated by a few serum proteins. Therefore, a crucial transport protein in the serum capable of influencing sex hormone activity, is sex hormone binding globulin (SHBG), which has varying concentrations among individuals. SHBG has been studied extensively. It can bind estrogens and androgens, altering their bioavailability for entry into target cells and tissues since it binds estradiol and testosterone with high affinity, resulting in selective sex-hormone transport in plasma [16,17,18,19]. SHBG is a plasma glycoprotein secreted by the liver that exists as a homodimer comprising two identical monomers, encoded by the 4 kb SHBG gene located on the short arm of chromosome 17 (p12–p13 bands) that comprises seven introns and eight exons [20,21,22]. The SHBG gene is translated into a 402 amino-acid protein cleaved to release its 29-amino-acid N-terminal sorting peptide. SHBG homodimers have a sex hormone-binding site created by the two monomers, which form a sandwich-like structure capable of binding a single sex hormone, indicating the requirement of SHBG monomer polymerization for the sex hormone-binding site [20,21,22,23]. In the hepatocytes, the SHBG gene’s transcription unit is expressed under the control of a promoter region that is around 800-bp long. The mature SHBG monomer is made up of two laminin G-like (LG) domains and the signal polypeptide that is necessary for secretion, which is encoded by the exons. The translation initiation site for the SHBG-prototype polypeptide sequence, which consists of the signal polypeptide sequence that is terminated during the secretion of the mature polypeptide and the three amino-terminal residues pertaining to the mature SHBG protein, is included in exon 1, which also contains a 60-bp untranslated region. The highly conserved steroid-binding location for vertebrate species is found in the amino-terminal LG domain, which is encoded by exons 2 to 5. A serine residue deep within the binding pocket, like Ser42 in human SHBG, seems to be essential for steroid binding [24,25,26]. The SHBG dimer shows a binding affinity towards sex steroids such as testosterone, dihydrotestosterone, and estradiol, to a lesser extent. It transports sex hormones, regulating their plasma levels and bioavailability for responsive tissues and, overall, demonstrating SHBG’s ability to orchestrate reproductive function and sexual features in males and females [20,21,25,27]. SHBG reduces free testosterone levels, inhibiting the biological induction in reproductive organs by sex hormones and impacting normal reproductive system activity. Testosterone binding to SHBG reduces its bioavailability, preventing it from completing its physiological functions. This disruption causes male infertility, gonadal and erectile dysfunction, prostate cancer, and several male reproductive system diseases with testosterone-dominated male sex-hormone symptoms [21,28]. However, increased bioavailable testosterone levels result in metabolic and reproductive phenotypes. They arise when SHBG’s plasma concentration is decreased, reflecting its role in human metabolism since SHBG concentrations vary in cancer, type 2 diabetes, and dyslipidemia [29,30,31,32]. Additionally, mutations, including single nucleotide polymorphisms, which alter SHBG expression and functions that regulate sperm count and semen quality, are associated with human male infertility [21,33]. Therefore, it can be concluded that the high molecular-weight SHBG plasma protein has a central role in maintaining the balance between bound and unbound sex steroids by attaching to androgens and estrogens with high ligand-binding affinity. SHBG likely alters their access and distribution to their objective tissues through changes in its concentration, identifying SHBG as a therapeutic target for preventing male sterility [25,34]. Besides natural steroid hormones, including testosterone, dihydrotestosterone, and estradiol, SHBG binds numerous endocrine-disrupting chemicals such as phthalates esters, a potential pathway for inhibiting the natural ligand–protein interactions that maintain normal activity in the steroid target organs [34,35,36]. Therefore, natural SHBG inhibitors could be used to treat male infertility. Virtual screening methods, such as the molecular docking and molecular dynamics (MD) simulation, are reliable high-throughput screening approaches that identify candidate inhibitors from a diverse phytochemical library through their better binding energy and bond stability in simulated molecular interactions with a target protein substrate. In addition, computational assessments of binding modes and bonds are preferred for rapidly identifying phytochemical-like ligand inhibitors for a specific target protein. Therefore, in this study, we aimed to identify effective inhibitors and plausible therapeutic targets to block SHBG function and prevent male sterility by calculating binding affinities and modes and the protein–ligand complex stability between the targeted SHBG protein and numerous phytochemical-based ligands. ## 2.1. Protein Preparation The SHBG protein’s three-dimensional (3D) structure was retrieved from the RCSB Protein Data Bank (PDB; ID: 1KDM) [37]. Pymol (version 2.5.4) [38] and Discovery Studio (version 4.5.0) [39] were used to dispel and clean the heteroatoms and water molecules from its crystal structure. The Swiss-PDB Viewer (version 4.1) [40] minimized the missing hydrogens, sidechain geometry, improper bond order, and other obligatory factors using the GROMOS (GROningen Molecular Simulation) 43B1 force field. ## 2.2. Ligand Preparation The ligands were identified from a comprehensive review of 200 articles on severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), identifying 345 phytochemicals as promising inhibitors. These studies indicated that these 345 phytochemicals were the most potent molecules that successfully inhibited SARS-CoV-2. Following their identification, the PubChem Database [41] was used to collect the 3D structures of these lead phytochemicals and the standard drug, anastrazole. Structure optimization and ligand cleaning, preparation, and minimization were performed using the mmff94 force field [42] with 2000 minimization steps and the perpendicular gradient-optimization algorithm. ## 2.3. Molecular Docking The PyRx (version 0.9) [43] virtual-screening approach was used to better understand the binding affinity and interaction of candidate ligands with SHBG and standard drug anastrazole with SHBG. Molecular docking was performed in association with the AutoDock Vina protocol. Every potential ligand was converted into PDBQT format to make it suitable for molecular docking, and a universal force field was used to minimize energies. Every bond could be rotated during this process since all of the docking configurations were protein-fixed and ligand-flexible. In AutoDock Vina, a grid box with a center point set of $x = 2.4963$, $y = 39.1902$, and $z = 29.4898$ and the dimensions (in Å) $x = 44.4875$, $y = 39.2403$, and $z = 41.2118$ was formed. The ligand-binding affinity values were shown in negative kcal/mole units, where the best confirmation had the lowest binding-affinity scores. In addition, non-bonding interactions were identified using PyMol, Discovery Studio, and UCSF ChimeraX (version 1.5) [44]. ## 2.4. ADMET Prediction The online servers, pKCSM [45], SwissADME [46], and admetSAR [47] were used to assess the pharmacokinetic properties through ADMET (adsorption, distribution, metabolism, and excretion) predictions. The phytochemicals’ canonical SMILES were retrieved from the PubChem database, and their ADMET properties were estimated by these web servers using SMILES. ## 2.5. MD Simulation The MD software, YASARA (version 22.9.24) [48,49], and force field, AMBER14 [50,51], were used to perform the MD simulation of the protein–ligand and the protein-standard drug complex. Firstly, the docked complexes were cleaned before their hydrogen-bond network was optimized and oriented. The simulation’s cubic cell with periodic boundary conditions was created using the TIP3P solvation model [52,53,54]. Furthermore, the simulation cell was stretched to 20 Å from the protein–ligand complexes in each direction. The simulation cell’s physiological parameters included $0.9\%$ sodium chloride, a temperature of 298 K, and a pH of 7.4. The steepest gradient algorithm, with 5000 cycles, was used to initially minimize energy in the stimulated annealing system [55,56]. The simulation system’s time step was adjusted to 1.25 femtoseconds (fs). Long-range electrostatic interactions were calculated using the particle-mesh Ewald (PME) system with a cutoff radius of 8.0 Å [57,58,59,60]. Simulation-trajectory data were saved at every 100 picoseconds (ps). Simulations were performed using the Berendsen thermostat, accompanied by fixed pressure and temperature for 100 nanoseconds (ns) [61,62]. The root-mean-square-fluctuation (RMSF), root-mean-square deviation (RMSD), radius of gyration (Rg), solvent-accessible surface area (SASA), and hydrogen bond were examined through the simulation trajectory data [63,64,65,66]. Additionally, using the following equation, the binding free energies of the simulation snapshots were estimated using MM-PBSA (Molecular Mechanics-Poisson Boltzmann Surface Area) techniques. Binding Energy = EpotRecept + EsolvRecept + EpotLigand + EsolvLigand − EpotComplex − EsolvComplex The YASARA macro was utilized to calculate the binding free energy for the MM-PBSA system, where a greater positive energy denotes a stronger binding [67,68,69,70,71] affinity. The stepwise procedure in terms of the materials and methods is depicted in Figure 1. ## 3.1. Molecular Docking The top ten molecules with the highest binding affinity were identified among the 345 candidate phytochemicals (Supplementary File). The three ligands with the best binding affinities (Figure 2) were cryptomisrine, dorsilurin E, and isoiguesterin at −9.2, −9, and −8.8 kcal/mole, respectively (Table 1). The common medication, anastrozole, had a binding affinity of −7.0 kcal/mole to the SHBG protein. Then, PyMol, Discovery Studio, and UCSF ChimeraX were used to look into their non-bond interactions with the SHBG protein. Cryptomisrine formed two conventional hydrogen bonds (at MET30 and PRO14), one electrostatic (Pi-Anion) bond (at ASP168), one hydrophobic (Pi-Sigma) bond (at SER169), and four hydrophobic (Pi-Alkyl) bonds (at LYS173, ALA28, VAL16, and LEU185) with the SHBG protein (Table 2; Figure 3A). Dorsilurin E formed one conventional hydrogen bond (at LYS173), one electrostatic (Pi-Anion) bond (at ASP168), five hydrophobic (Alkyl) bonds (at PRO14, LEU185, VAL16, MET30, and LEU143), and one hydrophobic (Pi-Alkyl) bond (at TRP170) with the SHBG protein (Table 2; Figure 3B). Isoiguesterin formed seven hydrophobic (Alkyl) bonds (at ALA28, MET30, ALA179, LEU185, PRO182, VAL16, and LYS173) with the SHBG protein (Table 2; Figure 3C). Three conventional hydrogen bonds (at VAL29, SER180, and ASP168), one electrostatic (Pi-Anion) bond (at GLU176), and four hydrophobic (Alkyl) bonds (at ALA28, ALA179, PRO182, and LEU185) stabilized the anastrozole-SHBG complex (Table 2; Figure 3D). ## 3.2. ADMET Prediction The pharmacokinetics and toxicity properties of the three top ligands were evaluated to ensure their efficiency and safety. The molecular weights of cryptomisrine, dorsilurin E, and isoiguesterin were 462.5, 490.6, and 404.6 g/mol, respectively (Table 3). Moreover, they followed Lipinski’s rule of five, which stipulates that a potent molecule should have a molecular weight of ≤500 g/mol [72,73]. Moreover, cryptomisrine, dorsilurin E, and isoiguesterin had 3, 6, and 2 hydrogen-bond acceptors, respectively, and 2, 1, and 1 hydrogen-bond donors, respectively. The topological polar surface areas (TPSAs) of cryptomisrine, dorsilurin E, and isoiguesterin were 74.43, 74.22, and 37.30 Å2, respectively. Cryptomisrine, dorsilurin E, and isoiguesterin were $96.507\%$, $93.133\%$, and $95.798\%$ absorbed in the human intestinal tract, respectively. In addition, they had no Ames toxicity, skin sensitization, or P-glycoprotein substrates. Cryptomisrine, Dorsilurin E, and Isoiguesterin, each had CNS permeability scores of −1.073, −2.703, and −1.955, respectively. Each of the top three compounds passed the carcinogenicity test and was found to be non-carcinogenic. None of the top three compounds displayed toxicity in the instance of acute oral poisoning. Isoiguesterin, Dorsilurin E, and Cryptomisrine all had BBB permeability ratings of −0.202, −0.368, and −0.7629, respectively. Furthermore, none of the top three compounds exhibited any toxicity during the hepatotoxicity test. Moreover, they followed Lipinski’s rule of five, with one violation for cryptomisrine and isoiguesterin but none for dorsilurin E. However, one lone violation of this rule does not disqualify a candidate from consideration as a viable therapeutic candidate [46]. ## 3.3. MD Simulation The top three ligand–protein complexes as well as the standard drug–protein complex were subjected to 100 ns MD simulations to explore their structural firmness and confirm their docking scenarios. The stability of protein–ligand complexes was evaluated by measuring the RMSD of C-alpha atoms. First, the cryptomisrine, dorsilurin E, and isoiguesterin–SHBG complexes’ RMSD increased for the first few seconds of the simulation, indicating their preliminarily higher instability. The RMSD of the isoiguesterin–SHBG complex was generally greater than those of the dorsilurin E–SHBG and cryptomisrine–SHBG complexes (Figure 4a). The dorsilurin E–SHBG showed a lower average RMSD than that of the isoiguesterin–SHBG and cryptomisrine–SHBG complexes. While the RMSD of the dorsilurin E–SHBG complex suddenly increased after 45 ns, it stabilized at around 80 ns and remained stable for the final 20 ns. The RMSD of the isoiguesterin–SHBG complex decreased appreciably after 40 ns but stabilized at around 65 ns and remained stable for the remaining simulation time with only minor fluctuations. The RMSD of the cryptomisrine–SHBG complex fluctuated until it stabilized at 85 ns. RMSD instability was present in the complex containing the reference drug anastrozole throughout the simulation period, with the peak occurring at 75 ns. Nevertheless, all three complexes remained stable throughout the simulation since their RMSD remained <2.5 Å [64]. To examine how the SHBG’s surface changes in response to ligands, SASAs were determined for the three top complexes since this parameter is crucial for understanding protein stability and folding [74]. Higher SASAs imply an enlarged protein surface area, while lower SASAs imply a reduced protein surface area [60]. Between 30–50 ns, the SASAs of the dorsilurin E–SHBG complex were greater than those of the cryptomisrine–SHBG and isoiguesterin–SHBG complexes, indicating that it had a greater surface area (Figure 4b). In addition, the cryptomisrine–SHBG complex had the lowest average of SASAs among these complexes, indicating that it had the smallest surface area. While the cryptomisrine–SHBG, dorsilurin E–SHBG, and isoiguesterin–SHBG complexes showed fluctuating SASAs until 70 ns, they remained relatively stable over the final 30 ns, indicating that they were stable. The SASA value for the complex containing anastrozole initially increased, but after 25 ns of simulation, the value drastically decreased. This complex showed the lowest SASA value of all the complexes after 70 ns, which indicates that the complex’s protein had been truncated. The protein–ligand complexes were determined as either more rigid or labile by measuring their Rg values. Lower Rg values indicate a more rigid protein–ligand complex, and higher Rg values indicate a more labile protein–ligand complex [56]. All three complexes showed an initial increase in their Rg value. The isoiguesterin–SHBG complex had higher Rg values on average, indicating that it had a more labile nature than the other two complexes during the simulation (Figure 4c). In contrast, the cryptomisrine–SHBG complex had the lowest Rg value between 40 and 65 ns, indicating rigidness. In addition, all three complexes showed very slight fluctuations in Rg values after the initial increase and comparatively lower Rg values, confirming their rigidness throughout the simulation. Compared to the other three complexes, the complex containing the common medicine, anastrazole, had the highest average Rg value across the simulation period, indicating that it was a more labile compound. The complexes’ hydrogen bonds were evaluated since they play a vital role in sustaining the integrity and stability of the docked complexes during the simulation [75]. The complexes of cryptomisrine–SHBG, dorsilurin E–SHBG, isoiguesterin–SHBG, and anastrozole–SHBG produced a significant number of hydrogen bonds, indicating a robust and rigid complex throughout the simulation (Figure 4d). To better understand SHBG’s suppleness across the amino-acid residues, RMSFs were examined for these three protein–ligand complexes. The RMSFs of all amino acids in the top three protein–ligand complexes did not exceed 2.5 Å. While the first few amino acids initially showed higher RMSFs, they dropped drastically after a few ns (Figure 4e). In general, lower RMSFs correspond to a higher rigidness. It is evident from the lower RMSFs of the top three complexes that they remained stable throughout the simulation [75]. With the largest peak occurring at 120 amino acid residues, the RMSF value for the complex containing anastrozole was larger on average than that of the other three complexes. It is evident from the top three complexes’ lower RMSF values that they remained steady longer than the complex with the typical medicine, anastrazole, throughout the simulation time. Lower RMSF values frequently imply a higher level of firmness. A stronger binding is denoted by a greater positive energy, which is shown in Figure 5, along with the results of the MM-PBSA binding free-energy calculation for the top three docked complexes and the common medication, anastrazole. I soiguesterin, dorsilurin E, cryptomisrine, and anastrazole (standard drug) had average binding free energies of 67.64, 71.39, 69.13, and 57.38 KJ/mol, respectively (Table 4). The possible three ligand molecules bind the SHBG protein more effectively than the conventional complex due to their larger average-binding free energies. ## 4. Discussion While numerous studies have used in silico analyses of distinct disease-causing receptor proteins to predict their bioactive molecules and inhibitors, few have examined the role of potent natural inhibitors of the SHBG protein to prevent male infertility [9,76]. The multifunctional SHBG protein, also known as testosterone-estradiol-binding globulin, is amalgamated by hepatocytes [21]. SHBG transports androgens, estrogens, testosterone, and estradiol in the blood and regulates their entry into target tissues [25,77,78]. Sex hormone levels are particularly altered when SHBG binds sex hormones, affecting their bioavailability. Since testosterone is the dominant male sex hormone, SHBG’s binding to it limits its biological action in the male reproductive system, causing reproductive problems [21]. These changes influence normal male reproductive system activity, leading to reproductive system diseases, such as male infertility (mostly by affecting semen quality and sperm count [79]), sexual dysfunction, erectile dysfunction, and prostate cancer [21]. Plasma SHBG levels and sperm count are lower in infertile men [12]. Therefore, natural inhibitors of the SHBG protein could be used to treat male infertility [76]. A similar computational study of the SHBG protein used molecular docking to examine 47 natural phytochemicals. The most potent natural compound was chlorogenic acid, which had a docking score of −7.255 kcal/mol but was only stable at 10 ns in MD simulations [76]. However, in this study, the three compounds identified (crytomisrine, dorsilurin E, and isoiguesterin) had docking scores of −9.2, −9, and −8.8 kcal/mol, respectively, and were stable at 100 ns in MD simulations. Similarly, Ishfaq et al. examined the binding of nine phthalates with SHBG inhibiting activity (BBP, DNHP, DEHP, DMP, DNOP, DINP, DIDP DBP, and DIBP) to human SHBG, finding docking scores between −6 and −10.12 kcal/mol. However, they did not examine protein–ligand complex stability [34]. In addition, another study explored molecular interactions between human-SHBG and chlorpyrifos and its degradation derivatives, including TMP, TCP, CPYO, and DEC. While they had docking scores between −6.097 and −7.662 kcal/mol, protein–ligand complex stability was not assessed through MD simulations [79]. Therefore, this study’s three selected compounds have more stable molecular interactions and more reasonable binding affinity than previously explored inhibitors. An in silico analysis of cryptomisrine, an alkaloid extracted from Cryptolepis sanguinolenta, found it to be a potent inhibitor of RNA-dependent RNA polymerase and SARS-CoV-2′s main protease (Mpro). It had docking scores of −9.80 and −10.60, respectively, the highest among 13 natural compounds from the plant with molecular-interaction stability during the MD simulation [80]. When docked against two different conformations of the Mpro protein of SARS-CoV 2, including typical substrate binding sites and ligand-induced substrate-binding sites, dorsilurin E had binding energies of −9.51 kcal/mol and −11.31 kcal/mol, respectively [81]. One of the top two terpenoids, with high binding scores, that was isolated from African medicinal plants, isoiguesterin, had −9.5 Kcal/mol binding energies when it interacted with the ACE2 and TMPRSS2 proteins to block the SARS-CoV 2 host-cell entry [82]. Therefore, the chosen molecule has previously shown inhibitory effects against viral proteins, encouraging further analysis against other protein targets for inhibition. This study has shown that crytomisrine, dorsilurin E, and isoiguesterin strongly bind to the SHBG protein, and their complexes remain stable and rigid during simulations. Their pharmacokinetics and toxicity properties indicate that all three natural compounds are safe and follow Lipinski’s rule for drug candidates. Anastrazole, a medication used to treat male infertility on a regular basis, had a binding affinity calculated at −7.0 kcal/mole, which was lower than the three top compounds in our analysis. Additionally, for the entire molecular dynamic-simulation period, the control medication revealed a more unstable and labile interaction with the target protein than our lead compounds. Additionally, compared to the standard medication, anastrazole, the three lead compounds showed a much-improved RMSD, Rg, RMSF, and SASA profiles. In addition, each of the top three compounds in our investigation had a lower average binding free energy than the common medication, anastrazole. Therefore, we can conclude that these three compounds, with safe ADMET predictions, can be used as drugs for male infertility by inhibiting SHBG-protein activity. ## 5. Conclusions Phytochemicals are becoming increasingly feasible and more promising therapeutic sources than synthetic constituents due to their broad appearance, wide specificity, and lower side effects. This study integrated many computational approaches to identify effective SHBG inhibitors. A detailed study of 200 articles identified 345 potent phytochemicals with activity against SARS-CoV-2 for screening. Three prospective SHBG inhibitors were cryptomisrine, dorsilurin E, and isoiguesterin, based on virtual screening and molecular docking. These were observed to have higher binding affinities and average binding free energies than the reference drug, anastrazole, and their pharmacokinetic properties satisfied the requirements for promising therapeutic candidates. Additionally, MD simulations verified that the top three docked protein–ligand complexes were more robust and stable than the control medication, anastrazole. Further in vitro experiments are required to establish the precise efficiency of these three drug candidates against SHBG since this combinatorial screening study was exclusively computational. ## References 1. Kumar N., Singh A.. Trends of male factor infertility, an important cause of infertility: A review of literature. J. Hum. Reprod. Sci. (2015) 8 191-196. DOI: 10.4103/0974-1208.170370 2. Zegers-Hochschild F., Adamson G.D., De Mouzon J., Ishihara O., Mansour R., Nygren K., Sullivan E., Vanderpoel S.. International Committee for Monitoring Assisted Reproductive Technology (ICMART) and the World Health Organization (WHO) revised glossary of ART terminology, 2009. Fertil. Steril.* (2009) 92 1520-1524. DOI: 10.1016/j.fertnstert.2009.09.009 3. Boivin J., Bunting L., Collins J.A., Nygren K.G.. Reply: International estimates on infertility prevalence and treatment seeking: Potential need and demand for medical care. Hum. Reprod. (2009) 24 2380-2383. DOI: 10.1093/humrep/dep218 4. Vander Borght M., Wyns C.. Fertility and infertility: Definition and epidemiology. Clin. Biochem. (2018) 62 2-10. DOI: 10.1016/j.clinbiochem.2018.03.012 5. Agarwal A., Mulgund A., Hamada A., Chyatte M.R.. A unique view on male infertility around the globe. Reprod. Biol. Endocrinol. (2015) 13 1-9. DOI: 10.1186/s12958-015-0032-1 6. Ombelet W., Cooke I., Dyer S., Serour G., Devroey P.. Infertility and the provision of infertility medical services in developing countries. Hum. Reprod. Update (2008) 14 605-621. DOI: 10.1093/humupd/dmn042 7. Cao X.W., Lin K., Li C.Y., Yuan C.W.. A review of WHO Laboratory Manual for the Examination and Processing of Human Semen. Zhonghua Nan Ke Xue (2011) 17 1059-1063. PMID: 22235670 8. Lotti F., Maggi M.. Ultrasound of the male genital tract in relation to male reproductive health. Hum. Reprod. Update (2015) 21 56-83. DOI: 10.1093/humupd/dmu042 9. Esther M.Y.J., Subramaniyan V., Kumar A.P., Subramanian M., Palani M.. Molecular docking, ADMET analysis and dynamics approach to potent natural inhibitors against sex hormone binding globulin in male infertility. Pharmacogn. J. (2017) 9 s35-s43. DOI: 10.5530/pj.2017.6s.155 10. Anwar S., Anwar A.. Infertility: A review on causes, treatment and management. Womens Heal. Gynecol. (2016) 5 2-5 11. Nath S., Deb B.. Survey on the effect of plant extract on reproducti ve parameters of mammals: A review. Int. J. Pure App. Biosci. (2015) 3 216-223 12. Safarinejad M.R., Shafiei N., Safarinejad S.. Association of the (TAAAA)n repeat and Asp327Asn polymorphisms in the sex hormone-binding globulin (SHBG) gene with idiopathic male infertility and relation to serum SHBG concentrations. J. Steroid Biochem. Mol. Biol. (2011) 123 37-45. DOI: 10.1016/j.jsbmb.2010.10.005 13. Akingbemi B.T.. Estrogen regulation of testicular function. Reprod. Biol. Endocrinol. (2005) 3 1-13. DOI: 10.1186/1477-7827-3-51 14. Pavlovich C.P., King P., Goldstein M., Schlegel P.N.. Evidence of a treatable endocrinopathy in infertile men. J. Urol. (2001) 165 837-841. DOI: 10.1016/S0022-5347(05)66540-8 15. Hammoud A.O., Gibson M., Peterson C.M., Meikle A.W., Carrell D.T.. Impact of male obesity on infertility: A critical review of the current literature. Fertil. Steril. (2008) 90 897-904. DOI: 10.1016/j.fertnstert.2008.08.026 16. Rosner W., Hryb D.J., Kahn S.M., Nakhla A.M., Romas N.A.. Interactions of sex hormone-binding globulin with target cells. Mol. Cell. Endocrinol. (2010) 316 79-85. DOI: 10.1016/j.mce.2009.08.009 17. Hammond G.L.. Molecular properties of corticosteroid binding globulin and the sex-steroid binding proteins. Endocr. Rev. (1990) 11 65-79. DOI: 10.1210/edrv-11-1-65 18. Siiteri P.K., Murai J.T., Hammond G.L., Nisker J.A., Raymoure W.J., Kuhn R.W.. The Serum Transport of Steroid Hormones. Proceedings of the 1981 Laurentian Hormone Conference (1982) Volume 38 457-510 19. Hong H., Branham W.S., Ng H.W., Moland C.L., Dial S.L., Fang H., Perkins R., Sheehan D., Tong W.. Human sex hormone-binding globulin binding affinities of 125 structurally diverse chemicals and comparison with their binding to androgen receptor, estrogen receptor, and α-Fetoprotein. Toxicol. Sci. (2015) 143 333-348. DOI: 10.1093/toxsci/kfu231 20. Vos M.J., Mijnhout G.S., Rondeel J.M.M., Baron W., Groeneveld P.H.P.. Sex hormone binding globulin deficiency due to a homozygous missense mutation. J. Clin. Endocrinol. Metab. (2014) 99 E1798-E1802. DOI: 10.1210/jc.2014-2055 21. Wan Q., Xie Y., Zhou Y., Shen X.. Research progress on the relationship between sex hormone-binding globulin and male reproductive system diseases. Andrologia (2021) 53 1-8. DOI: 10.1111/and.13893 22. Hammond G.L., Bocchinfuso W.P.. Sex hormone-binding globulin: Gene organization and structure/function analyses. Horm. Res. Paediatr. (1996) 45 197-201. DOI: 10.1159/000184787 23. Avvakumov G.V., Cherkasov A., Muller Y.A., Hammond G.L.. Structural analyses of sex hormone-binding globulin reveal novel ligands and function. Mol. Cell. Endocrinol. (2010) 316 13-23. DOI: 10.1016/j.mce.2009.09.005 24. Grishkovskaya I., Avvakumov G.V., Sklenar G., Dales D., Hammond G.L., Muller Y.A.. Crystal structure of human sex hormone-binding globulin: Steroid transport by a laminin G-like domain. EMBO J. (2000) 19 504-512. DOI: 10.1093/emboj/19.4.504 25. Hammond G.L.. Diverse roles for sex hormone-binding globulin in reproduction. Biol. Reprod. (2011) 85 431-441. DOI: 10.1095/biolreprod.111.092593 26. Grishkovskaya I., Avvakumov G.V., Hammond G.L., Catalano M.G., Muller Y.A.. Steroid ligands bind human sex hormone-binding globulin in specific orientations and produce distinct changes in protein conformation. J. Biol. Chem. (2002) 277 32086-32093. DOI: 10.1074/jbc.M203999200 27. Coviello A.D., Haring R., Wellons M., Vaidya D., Lehtimäki T., Keildson S., Lunetta K.L., He C., Fornage M., Lagou V.. A genome-wide association meta-analysis of circulating sex hormone-binding globulin reveals multiple loci implicated in sex steroid hormone regulation. PLoS Genet. (2012) 8. DOI: 10.1371/journal.pgen.1002805 28. Laurent M.R., Hammond G.L., Blokland M., Jardí F., Antonio L., Dubois V., Khalil R., Sterk S.S., Gielen E., Decallonne B.. Sex hormone-binding globulin regulation of androgen bioactivity in vivo: Validation of the free hormone hypothesis. Sci. Rep. (2016) 6 1-12. DOI: 10.1038/srep35539 29. Goudas V.T., Dumesic D.A.. Polycystic ovary syndrome. Endocrinol. Metab. Clin. North Am. (1997) 26 893-912. DOI: 10.1016/S0889-8529(05)70286-3 30. Haring R., Baumeister S.E., Völzke H., Dörr M., Felix S.B., Kroemer H.K., Nauck M., Wallaschofski H.. Prospective association of low total testosterone concentrations with an adverse lipid profile and increased incident dyslipidemia. Eur. J. Prev. Cardiol. (2011) 18 86-96. DOI: 10.1097/HJR.0b013e32833c1a8d 31. Alkan C., Coe P., Eichler E.. Sex Hormone–Binding Globulin and Risk of Type 2 Diabetes in Women and Men. N. Engl. J. Med. (2011) 23 1-7. DOI: 10.1056/NEJMoa0804381.Sex 32. Thompson D.J., Healey C.S., Baynes C., Kalmyrzaev B., Ahmed S., Dowsett M., Folkerd E., Luben R.N., Cox D., Ballinger D.. Identification of common variants in the SHBG gene affecting sex hormone-binding globulin levels and breast cancer risk in postmenopausal women. Cancer Epidemiol. Biomark. Prev. (2008) 17 3490-3498. DOI: 10.1158/1055-9965.EPI-08-0734 33. Liao W., Cai M., Chen J., Huang J., Liu F., Jiang C., Gao Y.. Hypobaric hypoxia causes deleterious effects on spermatogenesis in rats. Reproduction (2010) 139 1031-1038. DOI: 10.1530/REP-09-0557 34. Sheikh I.A., Turki R.F., Abuzenadah A.M., Damanhouri G.A., Beg M.A.. Endocrine disruption: Computational perspectives on human sex hormone-binding globulin and phthalate plasticizers. PLoS ONE (2016) 11 1-13. DOI: 10.1371/journal.pone.0151444 35. Déchaud H., Ravard C., Claustrat F., De La Perrière A.B., Pugeat M.. Xenoestrogen interaction with human sex hormone-binding globulin (hSHBG). Steroids (1999) 64 328-334. DOI: 10.1016/S0039-128X(98)00114-7 36. Jury H.H., Zacharewski T.R., Hammond G.L.. Interactions between human plasma sex hormone-binding globulin and xenobiotic ligands. J. Steroid Biochem. Mol. Biol. (2000) 75 167-176. DOI: 10.1016/S0960-0760(00)00168-0 37. Berman H.M., Westbrook J., Feng Z., Gilliland G., Bhat T.N., Weissig H., Shindyalov I.N., Bourne P.E.. The Protein Data Bank. Nucleic Acids Res. (2000) 28 235-242. DOI: 10.1093/nar/28.1.235 38. DeLano W.L.. The PyMOL Molecular Graphics System (2020) 39. Studio D.. Dassault Systemes BIOVIA, Discovery Studio Modelling Environment (2015) 40. Kaplan W., Littlejohn T.G.. Swiss-PDB Viewer (Deep View). Brief. Bioinform. (2001) 2 195-197. DOI: 10.1093/bib/2.2.195 41. Kim S., Thiessen P., Bolton E.E., Chen J., Fu G., Gindulyte A., Han L., He J., He S., Shoemaker B.A.. PubChem Substance and Compound databases. Nucleic Acids Res. (2019) 47 D1202-D1213. DOI: 10.1093/nar/gkv951 42. Halgren T.A.. Performance of MMFF94. Scope Parameterization J. Comput. Chem.* (1996) 17 490-519. DOI: 10.1002/(SICI)1096-987X(199604)17:5/6<490::AID-JCC1>3.0.CO;2-P 43. Trott O., Olson A.J.. AutoDock Vina: Improving the Speed and Accuracy of Docking with a New Scoring Function, Efficient Optimization, and Multithreading. J. Comput. Chem. (2010) 31 455-461. DOI: 10.1002/jcc.21334 44. Pettersen E.F., Goddard T.D., Huang C.C., Meng E.C., Couch G.S., Croll T.I., Morris J.H., Ferrin T.E.. UCSF ChimeraX: Structure visualization for researchers, educators, and developers. Protein Sci. (2021) 30 70-82. DOI: 10.1002/pro.3943 45. Pires D.E.V., Blundell T.L., Ascher D.B.. pkCSM: Predicting small-molecule pharmacokinetic and toxicity properties using graph-based signatures. J. Med. Chem. (2015) 58 4066-4072. DOI: 10.1021/acs.jmedchem.5b00104 46. Daina A., Michielin O., Zoete V.. SwissADME: A free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci. Rep. (2017) 7 42717. DOI: 10.1038/srep42717 47. Cheng F., Li W., Zhou Y., Shen J., Wu Z., Liu G., Lee P.W., Tang Y.. AdmetSAR: A comprehensive source and free tool for assessment of chemical ADMET properties. J. Chem. Inf. Model. (2012) 52 3099-3105. DOI: 10.1021/ci300367a 48. Land H., Humble M.S.. YASARA: A tool to obtain structural guidance in biocatalytic investigations. Methods Mol. Biol. (2018) 1685 43-67. DOI: 10.1007/978-1-4939-7366-8_4 49. Krieger E., Vriend G., Spronk C.. YASARA–Yet Another Scientific Artificial Reality Application. Yasara (2013) 993 51-78 50. Wang J., Wolf R.M., Caldwell J.W., Kollman P.A., Case D.A.. Development and testing of a general Amber force field. J. Comput. Chem. (2004) 25 1157-1174. DOI: 10.1002/jcc.20035 51. Mahmud S., Mita M.A., Biswas S., Paul G.K., Promi M.M., Afrose S., Hasan R., Shimu S.S., Zaman S., Uddin S.. Molecular docking and dynamics study to explore phytochemical ligand molecules against the main protease of SARS-CoV-2 from extensive phytochemical datasets. Expert Rev. Clin. Pharmacol. (2021) 14 1305-1315. DOI: 10.1080/17512433.2021.1959318 52. Harrach M.F., Drossel B.. Structure and dynamics of TIP3P, TIP4P, and TIP5P water near smooth and atomistic walls of different hydroaffinity. J. Chem. Phys. (2014) 140 174501. DOI: 10.1063/1.4872239 53. Mahmud S., Paul G.K., Biswas S., Afrose S., Mita M.A., Hasan M.R., Shimu M.S.S., Hossain A., Promi M.M., Ema F.K.. Prospective Role of Peptide-Based Antiviral Therapy Against the Main Protease of SARS-CoV-2. Front. Mol. Biosci. (2021) 8 628585. DOI: 10.3389/fmolb.2021.628585 54. Biswas S., Mahmud S., Mita M.A., Afrose S., Hasan M.R., Sultana Shimu M.S., Saleh M.A., Mostafa-Hedeab G., Alqarni M., Obaidullah A.J.. Molecular Docking and Dynamics Studies to Explore Effective Inhibitory Peptides Against the Spike Receptor Binding Domain of SARS-CoV-2. Front. Mol. Biosci. (2022) 8 1-10. DOI: 10.3389/fmolb.2021.791642 55. Mahmud S., Hasan M.R., Biswas S., Paul G.K., Afrose S., Mita M.A., Sultana Shimu M.S., Promi M.M., Hani U., Rahamathulla M.. Screening of Potent Phytochemical Inhibitors Against SARS-CoV-2 Main Protease: An Integrative Computational Approach. Front. Bioinforma. (2021) 1 1-15. DOI: 10.3389/fbinf.2021.717141 56. Mahmud S., Biswas S., Paul G.K., Mita M.A., Promi M.M., Afrose S., Hasan M.R., Zaman S., Uddin M.S., Dhama K.. Plant-based phytochemical screening by targeting main protease of sars-cov-2 to design effective potent inhibitors. Biology (2021) 10. DOI: 10.3390/biology10070589 57. Essmann U., Perera L., Berkowitz M.L., Darden T., Lee H., Pedersen L.G.. A smooth particle mesh Ewald method. J. Chem. Phys. (1995) 103 8577-8593. DOI: 10.1063/1.470117 58. Krieger E., Nielsen J.E., Spronk C.A.E.M., Vriend G.. Fast empirical pKa prediction by Ewald summation. J. Mol. Graph. Model. (2006) 25 481-486. DOI: 10.1016/j.jmgm.2006.02.009 59. Krieger E., Vriend G.. New ways to boost molecular dynamics simulations. J. Comput. Chem. (2015) 36 996-1007. DOI: 10.1002/jcc.23899 60. Mahmud S., Paul G.K., Afroze M., Islam S., Gupt S.B.R., Razu M.H., Biswas S., Zaman S., Uddin M.S., Khan M.. Efficacy of phytochemicals derived from avicennia officinalis for the management of covid-19: A combined in silico and biochemical study. Molecules (2021) 26. DOI: 10.3390/molecules26082210 61. Kumar Paul G., Mahmud S., Aldahish A.A., Afroze M., Biswas S., Briti Ray Gupta S., Hasan Razu M., Zaman S., Salah Uddin M., Nahari M.H.. Computational screening and biochemical analysis of Pistacia integerrima and Pandanus odorifer plants to find effective inhibitors against Receptor-Binding domain (RBD) of the spike protein of SARS-Cov-2. Arab. J. Chem. (2022) 15 103600. DOI: 10.1016/j.arabjc.2021.103600 62. Mahmud S., Rafi O., Paul G.K., Promi M.M., Sharmin M., Shimu S., Biswas S., Emran T., Dhama K.. Designing a multi-epitope vaccine candidate to combat MERS—CoV by employing an immunoinformatics approach. Sci. Rep. (2021) 11 1-21. DOI: 10.1038/s41598-021-92176-1 63. Mahmud S., Afrose S., Biswas S., Nagata A., Paul G.K., Mita M.A., Hasan M.R., Shimu M.S.S., Zaman S., Uddin M.S.. Plant-derived compounds effectively inhibit the main protease of SARS-CoV-2: An in silico approach. PLoS ONE (2022) 17. DOI: 10.1371/journal.pone.0273341 64. Mahmud S., Biswas S., Kumar Paul G., Mita M.A., Afrose S., Robiul Hasan M., Sharmin Sultana Shimu M., Uddin M.A.R., Salah Uddin M., Zaman S.. Antiviral peptides against the main protease of SARS-CoV-2: A molecular docking and dynamics study. Arab. J. Chem. (2021) 14 103315. DOI: 10.1016/j.arabjc.2021.103315 65. Mahfuz A.M.U., Khan A., Biswas S., Afrose S., Mahmud S., Mohammed Bahadur N., Ahmed F.. In search of inhibitors of anti-cancer drug target Fibroblast Growth Factor Receptors: Insights from virtual screening, molecular docking, and molecular dynamics study. Arab. J. Chem. (2022) 15 103882. DOI: 10.1016/j.arabjc.2022.103882 66. Islam S., Pramanik M.J., Biswas S., Moniruzzaman M., Biswas J.. Biological Efficacy of Compounds from Stingless Honey and Sting Honey against Two Pathogenic Bacteria: An In Vitro and In Silico Study. Molecules (2022) 27. DOI: 10.3390/molecules27196536 67. Kumari R., Kumar R., Consortium O.S.D.D., Lynn A.. g _ mmpbsa—A GROMACS tool for MM-PBSA and its optimization for high-throughput binding energy calculations. J. Chem. Inf. Model. (2014) 54 1951-1962. DOI: 10.1021/ci500020m 68. Gohlke H., Case D.A.. Converging Free Energy Estimates: MM-PB(GB)SA Studies on the Protein-Protein Complex Ras-Raf. J. Comput. Chem. (2004) 25 238-250. DOI: 10.1002/jcc.10379 69. Åqvist J., Medina C., Samuelsson J.E.. A new method for predicting binding affinity in computer-aided drug design. Protein Eng. (2010) 7 1-7. DOI: 10.1093/protein/7.3.385 70. Kuhn B., Gerber P., Schulz-Gasch T., Stahl M.. Validation and use of the MM-PBSA approach for drug discovery. J. Med. Chem. (2005) 48 4040-4048. DOI: 10.1021/jm049081q 71. Dash R., Junaid M., Mitra S., Arifuzzaman M., Hosen S.M.Z.. Structure-based identification of potent VEGFR-2 inhibitors from in vivo metabolites of a herbal ingredient. J. Mol. Model. (2019) 25 98. DOI: 10.1007/s00894-019-3979-6 72. Lipinski C., Lipinski C.A.. Lead- and drug-like compounds: The rule-of-five revolution. Drug Discov. Today Technol. (2004) 1 337-341. DOI: 10.1016/j.ddtec.2004.11.007 73. Mahmud S., Paul G.K., Biswas S., Kazi T., Mahbub S., Mita M.A., Afrose S., Islam A., Ahaduzzaman S., Hasan R.. Phytochemdb: A Platform for Virtual Screening and Computer-Aided Drug Designing. Database (2022) 2022 1-14. DOI: 10.1093/database/baac002 74. Savojardo C., Manfredi M., Martelli P.L., Casadio R.. Solvent Accessibility of Residues Undergoing Pathogenic Variations in Humans: From Protein Structures to Protein Sequences. Front. Mol. Biosci. (2021) 7 626363. DOI: 10.3389/fmolb.2020.626363 75. Islam S., Mahmud L., Almalki W.H., Biswas S., Islam A.. Cell-Free Supernatants ( CFSs ) from the Culture of Bacillus subtilis Inhibit Pseudomonas sp. Biofilm Formation. Microorganisms (2022) 10. DOI: 10.3390/microorganisms10112105 76. Kamiński M., Ding M., Truccolo W.A., Bressler S.L.. Evaluating causal relations in neural systems: Granger causality, directed transfer function and statistical assessment of significance. Biol. Cybern. (2001) 85 145-157. DOI: 10.1007/s004220000235 77. Del Mar Grasa M., Gulfo J., Camps N., Alcalá R., Monserrat L., Moreno-Navarrete J.M., Ortega F.J., Esteve M., Remesar X., Fernández-López J.A.. Modulation of SHBG binding to testosterone and estradiol by sex and morbid obesity. Eur. J. Endocrinol. (2017) 176 393-404. DOI: 10.1530/EJE-16-0834 78. Corvol P.L., Chrambach A., Rodbard D., Bardin C.W.. Physical Properties and Binding Capacity of Testosterone-Estradiol-binding Globulin in Human Plasma, Determined by Polyacrylamide Gel Electrophoresis. J. Biol. Chem. (1971) 246 3435-3443. DOI: 10.1016/S0021-9258(18)62149-0 79. Hazarika J., Ganguly M., Mahanta R.. Molecular interactions of chlorpyrifos and its environmental degradation products with human sex hormone-binding globulin: An in silico study. J. Appl. Toxicol. (2019) 39 1002-1011. DOI: 10.1002/jat.3789 80. Borquaye L.S., Gasu E.N., Ampomah G.B., Kyei L.K., Amarh M.A., Mensah C.N., Nartey D., Commodore M., Adomako A.K., Acheampong P.. Alkaloids from Cryptolepis sanguinolenta as Potential Inhibitors of SARS-CoV-2 Viral Proteins: An in Silico Study. Biomed Res. Int. (2020) 2020 1-14. DOI: 10.1155/2020/5324560 81. Jiménez-Avalos G., Vargas-Ruiz A.P., Delgado-Pease N.E., Olivos-Ramirez G.E., Sheen P., Fernández-Díaz M., Quiliano M., Zimic M., Agurto-Arteaga A., Antiparra R.. Comprehensive virtual screening of 4.8 k flavonoids reveals novel insights into allosteric inhibition of SARS-CoV-2 MPRO. Sci. Rep. (2021) 11 1-19. DOI: 10.1038/s41598-021-94951-6 82. Gyebi G.A., Ogunyemi O.M., Ibrahim I.M., Ogunro O.B., Adegunloye A.P., Afolabi S.O.. SARS-CoV-2 host cell entry: An in silico investigation of potential inhibitory roles of terpenoids. J. Genet. Eng. Biotechnol. (2021) 19 1-22. DOI: 10.1186/s43141-021-00209-z
--- title: 'Higher Intake of Vegetable Protein and Lower Intake of Animal Fats Reduce the Incidence of Diabetes in Non-Drinking Males: A Prospective Epidemiological Analysis of the Shika Study' authors: - Aya Ogawa - Hiromasa Tsujiguchi - Masaharu Nakamura - Koichi Hayashi - Akinori Hara - Keita Suzuki - Sakae Miyagi - Takayuki Kannon - Chie Takazawa - Jiaye Zhao - Yasuhiro Kambayashi - Yukari Shimizu - Aki Shibata - Tadashi Konoshita - Fumihiko Suzuki - Hirohito Tsuboi - Atsushi Tajima - Hiroyuki Nakamura journal: Nutrients year: 2023 pmcid: PMC9966791 doi: 10.3390/nu15041040 license: CC BY 4.0 --- # Higher Intake of Vegetable Protein and Lower Intake of Animal Fats Reduce the Incidence of Diabetes in Non-Drinking Males: A Prospective Epidemiological Analysis of the Shika Study ## Abstract Although nutrient intake and alcohol consumption are both closely associated with the incidence of diabetes, their interrelationships remain unclear. Therefore, we herein have investigated the interrelationships among nutrient intake, alcohol consumption, and the incidence of diabetes using longitudinal data. This study included 969 residents ≥40 years living in Japan. In 2011 and 2012, a baseline study was conducted using questionnaires on basic demographics, diabetes, nutrient intake, and lifestyle habits. In 2018 and 2019, a follow-up study was performed using questionnaires and medical records on diabetes. Two-way analysis of covariance (two-way ANCOVA) was used to test the interactions of drinking habits and diabetes incidence on nutrients intake. The prospective relationship between nutrient intake at baseline and the incidence of diabetes in the follow-up stratified by drinkers and non-drinkers was evaluated using multiple logistic regression analysis. Interactions were observed for vegetable protein intake ($$p \leq 0.023$$) and animal fat intake ($$p \leq 0.016$$) in males. Vegetable protein intake negatively correlated with the incidence of diabetes in non-drinkers (odds ratio (OR): 0.208; $95\%$ confidence interval ($95\%$ CI): 0.046–0.935; $$p \leq 0.041$$). Furthermore, animal fat intake positively correlated with the incidence of diabetes in non-drinkers (OR: 1.625; $95\%$ CI: 1.020–2.589; $$p \leq 0.041$$). Therefore, vegetable protein and animal fat intakes in combination with drinking habits need to be considered for the prevention of diabetes. ## 1. Introduction The prevalence of type 2 diabetes (DM2) is rapidly increasing worldwide and is estimated to reach 592 million individuals by 2035 [1]. According to the National Health and Nutrition Survey reported by the Ministry of Health, Labour and Welfare, Japan, in 2019, $19.7\%$ of Japanese males and $10.8\%$ of Japanese females were suspected to have diabetes [2]. Over the past decade, there has been an increasing trend of a higher proportion in the elderly group. There is a concern that the rapid aging of the Japanese population might lead to a further increase in the prevalence of diabetes. Insulin secretion is lower and also declines at an earlier age in East Asians than in Caucasians in European countries [3]. Furthermore, overnutrition, physical inactivity, obesity, smoking, and heredity factors reduce insulin sensitivity and increase insulin requirements for blood glucose regulation due to insulin resistance. If this condition continues, insulin secretion declines, leading to the development of DM2. Previous studies on the relationship between nutrient intake and diabetes demonstrated that the intake of polyunsaturated fatty acids (PUFAs) [4], dietary fiber [5,6], zinc [7,8,9], magnesium [10], vitamin D [11], and vitamin E [12,13] reduced the risk of diabetes. The nutritional epidemiology of diabetes in the Japanese population is important [14], as their diets differ from those of western countries [15]. In this regard, studies of the Japanese population revealed diets with fiber [16] and magnesium [17] intakes were associated with a reduced risk of DM2 and carbohydrate [18] was associated with an increased risk. Recent studies that focused on the relationship between dietary protein intake and the risk of DM2 reported that the development of DM2 inversely correlated with vegetable protein intake [19,20,21,22], while others found no relationship [23]. The key nutrients associated with the risk of DM2 are fats and carbohydrates, and, thus, the World Health Organization (WHO) recommends controlling the intake of saturated fatty acids (SFAs) and animal fats to prevent DM2 [24]. The Prevention with a Mediterranean Diet (PREDIMED) study, a prospective cohort analysis of the relationship between dietary fat intake and the risk of DM2, showed that the intake of animal fat was associated with a higher risk of DM2 [25], whereas other studies did not [26]. Similarly, the intake of SFAs was related to the risk of DM2 in some studies [27] but not in others [28,29]. Therefore, the relationship between nutrient intake and the risk of diabetes remains unclear. This may be attributed to the involvement of other factors such as lifestyle habits, including alcohol consumption, which may affect nutrient intake and diabetes. Since alcohol produces more than twice as much energy as the same amount of carbohydrates or protein, excessive alcohol consumption has been proposed as a risk factor for diabetes [24,25]. Many studies have shown that alcohol consumption increases the risk of DM2 [25]. However, it has not yet been established whether alcohol consumption increases or decreases the risk of diabetes in relation to the involvement of other nutrients. Although nutrient intake and alcohol consumption are both considered to be closely associated with the incidence of diabetes, this interrelationship has not yet been examined in detail, particularly using longitudinal data. Therefore, the present study investigated the interrelationships among nutrient intake—including protein, fat, and carbohydrates—alcohol consumption, and the incidence of diabetes, and compared prospective relationships between nutrient intake and the incidence of diabetes between non-drinkers and drinkers using longitudinal data from a general population in Japan. ## 2.1. Study Design and Subjects The Shika study, which examines the health status of community-dwelling residents living in Shika town, has been conducted since 2011. The town is located in the center of the Noto Peninsula in Ishikawa Prefecture, Japan, and had a total population of 23,208 in July 2011. The present longitudinal study was conducted utilizing data from the Shika study. The target population of the present study was residents ≥40 years living in model districts in the town. In 2011 and 2012, a baseline study was conducted using questionnaires on basic demographics, diabetes, nutrient intake, and lifestyle habits. In 2018 and 2019, a follow-up study was performed using questionnaires and medical records from hospitals on diabetes. Among 2264 eligible residents in the baseline study, 1948 were analyzed (316 non-respondents; response rate of $86.0\%$). Of these, 1670 individuals were eligible for the follow-up study after the exclusion of 200 deaths, 35 relocations in the town, and 43 moving out of the town. Following the exclusion of 554 individuals who did not respond to the follow-up study or had no medical information (follow-up rate of $66.8\%$), 1116 individuals were eligible. A total of 969 individuals were ultimately included in analyses after the exclusion of 46 individuals with diabetes at baseline, 10 with an over- or under-reported nutritional intake, and 91 with missing data. The present study was conducted with the approval of the Ethics Committee of Kanazawa University (No. 1491). Written informed consent was obtained from all participants prior to data collection. ## 2.2. Assessment of Diabetes Incidence Regarding diabetes in the baseline study, participants were asked if they were currently being treated for diabetes. Concerning the incidence of diabetes in the follow-up period, medical record information on diabetes was collected thorough a medical record collecting system, which we constructed in four hospitals in the study area. In parallel with the system, we used questionnaires asking whether participants had been diagnosed with diabetes by a physician. ## 2.3. Assessment of Nutrient Intake Nutrient intake was estimated using the brief self-administered dietary history questionnaire (BDHQ). The BDHQ provides information on the average daily frequency of consumption of 58 food and beverage items in the past month. These 58 items were selected from foods or drinks commonly consumed in Japan, mostly based on the food list of the Japanese National Health and Nutrition Survey [30,31]. The questionnaire listed items on all forms of food preparation, including raw, cooked, processed, and pickled foods. In addition, seasonings and cooking oils consumed were included. For eating and drinking habits, the intake of noodle soup and the intensity of seasoning were evaluated. The reported intakes of food and beverages were converted into energy, macronutrient, and micronutrient values using a computer algorithm for BDHQ. Nutrient intake was reported in terms of energy density. Drinking habits were calculated as the average daily net alcohol intake in the past month and classified into two groups: non-drinkers (0 g/day of net alcohol) and drinkers (>0 g/day of net alcohol). The reproducibility and validity of the BDHQ has been already demonstrated [30,31]. Details on the BDHQ are described elsewhere [30,31]. Participants with a reported nutritional intake of less than 600 kcal or more than 4000 kcal energy per day were excluded from the analyses as they were considered to be under- or over-reporters. ## 2.4. Basic Demographics Data on sex, age, height, weight, smoking, drinking habits, leisure-time physical activities, and the diagnosis of hypertension were obtained using questionnaires. Body mass index (BMI) was calculated by dividing weight (kg) by height squared (m2). Smoking was categorized into two groups: current non-smokers (non-smokers and ex-smokers) and current smokers. Regarding leisure-time physical activities, we used the following question: “How many times a week do you exercise?”. The response options included: [1] daily; [2] 5–6 days a week; [3] 3–4 days a week; [4] 1–2 days a week; [5] none. ## 2.5. Statistical Analysis All analyses were stratified by sex. Between males and females, unpaired Student’s t-tests were used for comparisons of the means of continuous variables and chi-square tests for comparisons of the percentages of categorical variables. Regarding basic characteristics in the diabetes-incidence and non-diabetes-incidence groups, unpaired Student’s t-tests were used for comparisons of the means of continuous variables and chi-square tests for comparisons of the proportions of categorical variables. In comparisons of mean nutrient intake between the diabetes and non-diabetes incidence groups, a one-way analysis of covariance (ANCOVA), adjusted for age, BMI, leisure-time physical activities, hypertension, magnesium, and alpha-tocopherol, was used. Two-way ANCOVA was used to test the interactions of drinking habits and diabetes incidence on nutrients intake. Age, BMI, leisure-time physical activities, hypertension, magnesium, and α-tocopherol were used as covariates. The Bonferroni post hoc test was used for nutrients that showed interactions to test differences in nutrient intake between the diabetes and non-diabetes incidence groups stratified by drinkers and non-drinkers. The prospective relationship between nutrient intake and the incidence of diabetes stratified by drinkers and non-drinkers was examined using multiple logistic regression analysis. The covariates were age, BMI, leisure-time physical activities, hypertension, magnesium, and alpha-tocopherol, and each nutrient was entered separately for each analysis. Statistical Package for Social Science (SPSS, IBM Corp., Tokyo, Japan) version 27 was used for these analyses. The significance of differences was set at $p \leq 0.05$ for all analyses. ## 3.1. Participant Characteristics Table 1 shows participants’ characteristics. The mean age of male subjects was 60.04 ± 10.80 years, whereas that of female subjects was 61.72 ± 11.69 years, with a significant difference ($$p \leq 0.021$$). BMI was significantly higher in males than in females ($p \leq 0.001$). The percentages of subjects who were smokers ($p \leq 0.001$), drinkers ($p \leq 0.001$), and with frequent leisure-time physical activities ($$p \leq 0.047$$) were significantly higher in males than in females. No significant differences were observed in the incidence of diabetes in the follow-up study between males and females, although it was slightly higher in males ($$p \leq 0.053$$). The prevalence of hypertension did not significantly differ between males and females. Regarding comparisons of nutrient intake, the intakes of animal protein ($p \leq 0.001$), vegetable protein ($p \leq 0.001$), animal fat ($p \leq 0.001$), vegetable fat ($p \leq 0.001$), and carbohydrates ($p \leq 0.001$) were significantly higher in females than in males. ## 3.2. Comparison of Subjects with and without Diabetes Incidence Table 2 shows comparisons of subjects with and without diabetes incidence stratified by sex. In males, age was significantly higher in the diabetes incidence group than in the non-diabetes incidence groups ($$p \leq 0.035$$), while no significant differences were observed in females. In males, BMI was slightly higher in the diabetes incidence group than in the non-diabetes incidence group ($$p \leq 0.092$$), while no significant differences were noted in females. The percentages of smokers, drinkers, those with frequent leisure-time physical activities, and the prevalence of hypertension did not significantly differ between the diabetes and non-diabetes incidence groups in both sexes. Furthermore, mean nutrient intake did not significantly differ between the diabetes and non-diabetes incidence groups in both sexes. ## 3.3. Interactions of Drinking Habits and Diabetes Incidence on Nutrients Intake Table 3 shows the interactions of drinking habits and diabetes incidence on nutrients intake. Mean nutrient intake did not significantly differ between the diabetes and non-diabetes incidence groups, and no main effect was found. The main effect of mean nutrient intake by drinkers and non-drinkers was noted for carbohydrate intake ($$p \leq 0.003$$) in males. Interactions of diabetes incidence and drinking habits on vegetable protein ($$p \leq 0.023$$) and animal fat intake ($$p \leq 0.016$$) were found in males. The results of Bonferroni post hoc tests for the non-drinking group showed that the intake of vegetable protein was significantly lower in the diabetes incidence group than in the non-diabetes incidence group ($$p \leq 0.040$$). In contrast, no significant differences were observed between the two groups for the drinking group. In the non-drinking group, the intake of animal fats was significantly higher ($$p \leq 0.048$$) in the diabetes incidence group than in the non-diabetes incidence group, while no significant differences were observed between the two groups for the drinking group. ## 3.4. Prospective Relationship between Nutrient Intake and Diabetes Incidence According to Drinking Habits Table 4 shows the prospective relationship between nutrient intake and the incidence of diabetes according to drinking habits. A negative correlation was observed between vegetable protein intake and the incidence of diabetes in the non-drinking group (odds ratio (OR): 0.208; $95\%$ confidence interval ($95\%$ CI): 0.046–0.935; $$p \leq 0.041$$). Likewise, a positive correlation was noted between animal fat intake and the incidence of diabetes in the non-drinking group (OR: 1.625; $95\%$ CI: 1.020–2.589; $$p \leq 0.041$$). In other words, the incidence of diabetes was found to be frequent when vegetable protein intake was lower and animal fat intake was higher only in non-drinkers. ## 4. Discussion The present study examined the interrelationships among nutrient intake, alcohol intake, and the incidence of diabetes utilizing longitudinal data from a general population of middle-aged and elderly Japanese individuals. We also compared the prospective relationship between nutrient intake and the incidence of diabetes between non-drinkers and drinkers. The results obtained showed an interaction of drinking habits and the incidence of diabetes on nutrient intake, demonstrating the preventive effects of vegetable protein intake and the promotive effects of animal fat on the incidence of diabetes in male non-drinkers. Previous studies reported that a higher intake of vegetable protein was associated with a reduced risk of DM2 [19,20]. A longitudinal study in USA, which followed 165,080 female nurses and 40,722 male health professionals for 22 years, revealed a slight reduction in the risk of diabetes with a higher vegetable protein intake [19]. A cohort study that followed 21,523 people in Australia for approximately 12 years and a meta-analysis integrating 11 cohort studies found an inverse relationship between vegetable protein intake and the development of DM2 in females [20,21]. Another cohort study that followed 2332 Finnish males for approximately 20 years showed that replacing $1\%$ of energy from animal protein with that from vegetable protein reduced the risk of DM2 by $18\%$ ($95\%$ CI 0, 32) [22]. A 10-year prospective cohort study examining the association between the intake of soya protein and the risk of diabetes in Japan reported that a higher intake of soya protein reduced the risk of diabetes [32]. In contrast, another study reported no relationship between vegetable protein intake and the incidence of DM2 [23]. These discrepancies among studies may be attributed to racial differences, differences in research methods, or bias caused by subject selection or dropouts, particularly in longitudinal and cohort studies. Bias in the present study was considered to be relatively small because subjects were representative of a community-dwelling population with a high response rate. Furthermore, differences in assessment methods for nutrient intake markedly affect the data obtained. Various types of dietary assessment methods are currently available, with dietary record methods, food frequency questionnaires (FFQ), and dietary history methods mainly used. Due to the burden on participants, dietary record methods are generally only recorded for a few days, which increases the difficulty associated with estimating habitual intake over a long period of time. It is challenging to quantify the actual amount of nutrients consumed by FFQ because information is only available for the limited number of foods listed, and questions are based on food units (ingredients before cooking) rather than menu items. In the present study, we used the BDHQ, which consists of approximately 80 questions and calculates the intake of 58 foods and more than 100 nutrients. Additionally, the BDHQ asks about the meals eaten in the past month and gathers qualitative and quantitative information on an individual’s cooking and seasoning habits, thereby providing a more realistic picture of their eating habits. Even though the BDHQ is not a direct assessment method, certain dietary habits may be assessed in detail with accuracy. Therefore, vegetable protein intake appears to reduce the risk of diabetes incidence, as shown in the present study. More importantly, the present study categorized subjects as non-drinkers and drinkers and found that a higher intake of vegetable protein reduced the risk of diabetes incidence in male non-drinkers only. Numerous studies have reported a relationship between alcohol consumption and the incidence of diabetes. Since alcohol produces more than twice as much energy as the same weight of carbohydrates or protein, excessive alcohol consumption has been proposed as a risk factor for diabetes [33,34]. However, moderate alcohol consumption has been shown to exert protective effects against the development of DM2 and had no adverse effects on insulin sensitivity in several studies [35,36]. Therefore, it remains unclear whether alcohol consumption increases or decreases the risk of DM2. Given this controversial relationship between alcohol consumption and diabetes, the relationship between vegetable protein intake and the incidence of diabetes needs to be examined in subjects stratified by alcohol consumption, as in the present study. To the best of our knowledge, this is the first study to investigate the interactive effects of alcohol consumption and nutrient intake on the incidence of diabetes. This may support the validity of the present results showing the preventive effects of vegetable protein intake on the development of diabetes incidence in non-drinkers. The results obtained on non-drinkers are consistent with the aforementioned studies showing the preventive effects of vegetable protein against the development of diabetes incidence [19,22], while those on drinkers are in agreement with the findings of studies showing no such effects [23]. Animal fats mainly contain SFAs. The PREDIMED study, a prospective cohort study that followed 3349 Spanish individuals for 3 to 4 years, reported that higher intakes of animal fats and SFAs were associated with an increased risk of DM2 [25]. In addition, a European prospective study that followed 23,631 participants for 3 to 7 years revealed that increased PUFAs/SFAs ratio was associated with a reduced risk of diabetes, regardless of age, sex, a family history of diabetes, and other lifestyle factors [27]. Moreover, a cohort study in the Netherlands and Finland showed that a higher SFAs intake contributed to the risk of impaired glucose tolerance and non-insulin-dependent DM utilizing 20-year follow-up data on 1260 individuals [37]. On the contrary, a study of the Japanese population reported that higher animal fat intake was inversely associated with diabetes [18]. However, a systematic review and meta-analysis of observational studies found no relationship between animal fat intake and the incidence of DM2, whereas diets containing vegetable fat, but not animal fat, were beneficial for the prevention of DM2 [26]. Similarly, a meta-analysis of cohort studies reported no relationship between SFAs intake and the development of DM2 [28,29,38]. The present results demonstrated the interactions of drinking habits and animal fat intake with the incidence of diabetes, and a logistic regression analysis stratifying drinkers and non-drinkers showed that a low animal fat intake was protective against the development of diabetes in male non-drinkers only. By taking drinking habits into consideration as a third factor, our results were consistent with the former studies for non-drinkers and with the latter studies for drinkers. Almost the same discussion as to the association between vegetable protein intake and diabetes incidence can be applied for animal fat intake. The lack of similar results for vegetable protein and animal fat intakes in females may be due to the small number of female drinkers examined in the present study. The mechanisms by which vegetable protein, animal fat, and alcohol intake affect the development of diabetes remain unclear. Previous findings have suggested that the amino acid composition of vegetable protein increased the efficiency of amino acid metabolism, thereby reducing the risk of DM2 [39]. However, drinkers consume amino acids when taking alcohol, which may decrease the efficiency of amino acid metabolism. It is possible that a combination of higher vegetable protein intake and non-drinking habits increases the efficiency of amino acids metabolism, thereby decreasing the risk of diabetes incidence. Several mechanisms have been proposed for the increased risk of diabetes with a high animal fat intake. The fatty acid composition of structural membrane lipids may affect insulin sensitivity. For example, the higher SFAs content of membrane phospholipids was previously shown to increase insulin resistance and the risk of developing diabetes [40,41,42]. Furthermore, alcohol may increase the SFAs content of structural membrane lipids and change fatty acid compositions, thereby reducing insulin sensitivity and consequently increasing insulin resistance. Therefore, a lower intake of animal fats in non-drinkers may effectively reduce the SFAs content of membrane phospholipids and decrease the risk of diabetes. Foods containing vegetable protein include beans and their processed products, cereals, and vegetables [43,44]. In addition to reducing alcohol consumption, the active consumption of these foods may contribute to the prevention of diabetes. Animal fat is present in fatty meat, lard, beef fat, bacon, cream, cheese, butter, chocolate, and ice cream [44]. Reducing the intake of these foods in addition to alcohol consumption may be important for preventing the development of diabetes. It may be also beneficial to consume vegetable fats, such as olive oil and sesame oil, as alternatives. The strengths of the present study include its longitudinal design, which allowed prospective and causal relationships between nutrient intake and the incidence of diabetes to be established. In addition, evaluations of the incidence of diabetes were supplemented by the medical records of hospitals to ensure objectivity. Moreover, the participation rate at baseline was high, which may have excluded selection biases. However, this study had several limitations that need to be addressed. Dietary data were subjective because the intake frequencies of foods or food groups were based on a self-administered questionnaire, which may lead to recall biases. In addition, the calculation of nutrients intake did not account for changes in nutrients composition due to cooking, such as heating. Furthermore, protein data from the nutrient calculations were not analyzed in detail other than vegetable and animal categories. Likewise, we could not use more detailed categorized data for fats because they have not been adequately validated. Moreover, supplements were not included in the calculation, although they were included as a question item. The influence of bioactive substances (e.g., antioxidants), which are abundant in vegetables, was also not examined. Furthermore, for some participants, source information on the incidence of diabetes was only self-reported and not corroborated by medical information. Additionally, the questionnaire for the assessment of diabetes has not been tested for reproducibility or validity, although we utilized medical information whenever possible. Finally, the follow-up rate was low, which may have led to withdrawal biases. ## 5. Conclusions The present study showed that a higher intake of vegetable protein and a lower intake of animal fat were prospectively associated with a lower incidence of diabetes in male non-drinkers. Considering vegetable protein and animal fat intakes in combination with drinking habits might be important for the prevention of diabetes incidence. ## References 1. Guariguata L., Whiting D.R., Hambleton I., Beagley J., Linnenkamp U., Shaw J.E.. Global Estimates of Diabetes Prevalence for 2013 and Projections for 2035. Diabetes Res. Clin. Pract. (2014.0) 103 137-149. DOI: 10.1016/j.diabres.2013.11.002 2. National Health and Nutrition Survey the Ministry of Health 3. Yoon K.H., Lee J.H., Kim J.W., Cho J.H., Choi Y.H., Ko S.H., Zimmet P., Son H.Y.. Epidemic Obesity and Type 2 Diabetes in Asia. Lancet (2006.0) 368 1681-1688. DOI: 10.1016/S0140-6736(06)69703-1 4. Imamura F., Micha R., Wu J.H.Y., De Oliveira Otto M.C., Otite F.O., Abioye A.I., Mozaffarian D.. Effects of Saturated Fat, Polyunsaturated Fat, Monounsaturated Fat, and Carbohydrate on Glucose-Insulin Homeostasis: A Systematic Review and Meta-Analysis of Randomised Controlled Feeding Trials. PLoS Med. (2016.0) 13. DOI: 10.1371/journal.pmed.1002087 5. Reynolds A.N., Akerman A.P., Mann J.. Dietary Fibre and Whole Grains in Diabetes Management: Systematic Review and Meta-Analyses. PLoS Med. (2020.0) 17. DOI: 10.1371/journal.pmed.1003053 6. Zhang W., Tang Y., Huang J., Yang Y., Yang Q., Hu H.. Efficacy of Inulin Supplementation in Improving Insulin Control, HbA1c and HOMAAIR in Patients with Type 2 Diabetes: A Systematic Review and Meta-analysis of Randomized Controlled Trials. J. Clin. Biochem. Nutr. (2020.0) 66 176-183. DOI: 10.3164/jcbn.19-103 7. Singh R.B., Niaz M.A., Rastogi S.S., Bajaj S., Gaoli Z., Shoumin Z.. Current Zinc Intake and Risk of Diabetes and Coronary Artery Disease and Factors Associated with Insulin Resistance in Rural and Urban Populations of North India. J. Am. Coll. Nutr. (1998.0) 17 564-570. DOI: 10.1080/07315724.1998.10718804 8. Do D., Marreiro N., Geloneze B., Tambascia M.A., Lerário A.C., Halpern A., Maria S., Cozzolino F.. Effect of Zinc Supplementation on Serum Leptin Levels and Insulin Resistance of Obese Women. Biol. Trace Elem. Res. (2006.0) 109 109-118 9. Sun Q., Van Dam R.M., Willett W.C., Hu F.B.. Prospective Study of Zinc Intake and Risk of Type 2 Diabetes in Women. Diabetes Care (2009.0) 32 629-634. DOI: 10.2337/dc08-1913 10. Simental-Mendía L.E., Sahebkar A., Rodríguez-Morán M., Guerrero-Romero F.. A Systematic Review and Meta-Analysis of Randomized Controlled Trials on the Effects of Magnesium Supplementation on Insulin Sensitivity and Glucose Control. Pharmacol. Res. (2016.0) 111 272-282. DOI: 10.1016/j.phrs.2016.06.019 11. Li X., Liu Y., Zheng Y., Wang P., Zhang Y.. The Effect of Vitamin D Supplementation on Glycemic Control in Type 2 Diabetes Patients: A Systematic Review and Meta-Analysis. Nutrients (2018.0) 10. DOI: 10.3390/nu10030375 12. Likidlilid A., Sriratanasathavorn C.. Plasma Lipid Peroxidation and Antioxidiant Nutrients in Type 2 Diabetic Patients. J. Med. Assoc. Thai. (2006.0) 89 S147-S155 13. Reunanen A., Knekt P., Aaran R.K., Aromaa A.. Serum Antioxidants and Risk of Non-Insulin Dependent Diabetes Mellitus. Eur. J. Clin. Nutr. (1998.0) 52 89-93. DOI: 10.1038/sj.ejcn.1600519 14. Nanri A., Shimazu T., Takachi R., Ishihara J., Mizoue T., Noda M., Inoue M., Tsugane S.. Dietary patterns and type 2 diabetes in Japanese men and women: The Japan Public Health Center-based Prospective Study. Eur. J. Clin. Nutr. (2013.0) 67 18-24. DOI: 10.1038/ejcn.2012.171 15. Suzuki N., Goto Y., Ota H., Kito K., Mano F., Joo E., Ikeda K., Inagaki N., Nakayama T.. Characteristics of the Japanese Diet Described in Epidemiologic Publications: A Qualitative Systematic Review. J. Nutr. Sci. Vitaminol. (2018.0) 64 129-137. DOI: 10.3177/jnsv.64.129 16. Kimura Y., Yoshida D., Hirakawa Y., Hata J., Honda T., Shibata M., Sakata S., Uchida K., Kitazono T., Ninomiya T.. Dietary Fiber Intake and Risk of Type 2 Diabetes in a General Japanese Population: The Hisayama Study. J. Diabetes Investig. (2021.0) 12 527-536. DOI: 10.1111/jdi.13377 17. Konishi K., Wada K., Tamura T., Tsuji M., Kawachi T., Nagata C.. Dietary Magnesium Intake and the Risk of Diabetes in the Japanese Community: Results from the Takayama Study. Eur. J. Nutr. (2017.0) 56 767-774. DOI: 10.1007/s00394-015-1122-8 18. Nanri A., Mizoue T., Kurotani K., Goto A., Oba S., Noda M., Sawada N., Tsugane S.. Low-Carbohydrate Diet and Type 2 Diabetes Risk in Japanese Men and Women: The Japan Public Health Center-Based Prospective Study. PLoS ONE (2015.0) 10. DOI: 10.1371/journal.pone.0118377 19. Malik V.S., Li Y., Tobias D.K., Pan A., Hu F.B.. Dietary Protein Intake and Risk of Type 2 Diabetes in US Men and Women. Am. J. Epidemiol. (2016.0) 183 715-728. DOI: 10.1093/aje/kwv268 20. Shang X., Scott D., Hodge A.M., English D.R., Giles G.G., Ebeling P.R., Sanders K.M.. Dietary Protein Intake and Risk of Type 2 Diabetes: Results from the Melbourne Collaborative Cohort Study and a Meta-Analysis of Prospective Studies. Am. J. Clin. Nutr. (2016.0) 104 1352-1365. DOI: 10.3945/ajcn.116.140954 21. Tian S., Xu Q., Jiang R., Han T., Sun C., Na L.. Dietary Protein Consumption and the Risk of Type 2 Diabetes: A Systematic Review and Meta-Analysis of Cohort Studies. Nutrients (2017.0) 9. DOI: 10.3390/nu9090982 22. Virtanen H.E.K., Koskinen T.T., Voutilainen S., Mursu J., Tuomainen T.P., Kokko P., Virtanen J.K.. Intake of Different Dietary Proteins and Risk of Type 2 Diabetes in Men: The Kuopio Ischaemic Heart Disease Risk Factor Study. Br. J. Nutr. (2017.0) 117 882-893. DOI: 10.1017/S0007114517000745 23. Van Nielen M., Feskens E.J.M., Mensink M., Sluijs I., Molina E., Amiano P., Ardanaz E., Balkau B., Beulens J.W.J., Boeing H.. Dietary Protein Intake and Incidence of Type 2 Diabetes in Europe: The EPIC-InterAct Case-Cohort Study. Diabetes Care (2014.0) 37 1854-1862. DOI: 10.2337/dc13-2627 24. 5 Tips for a Healthy Diet This New Year 25. Guasch-Ferre M., Becerra-Tomas N., Ruiz-Canela M., Corella D., Schroder H., Estruch R., Ros E., Aros F., Gomez-Gracia E., Fiol M.. Total and Subtypes of Dietary Fat Intake and Risk of Type 2 Diabetes Mellitus in the Prevencí on Con Dieta Mediterŕanea (PREDIMED) Study. Am. J. Clin. Nutr. (2017.0) 105 723-735. DOI: 10.3945/ajcn.116.142034 26. Neuenschwander M., Barbaresko J., Pischke C.R., Iser N., Beckhaus J., Schwingshackl L., Schlesinger S.. Intake of Dietary Fats and Fatty Acids and the Incidence of Type 2 Diabetes: A Systematic Review and Dose-Response Meta-Analysis of Prospective Observational Studies. PLoS Med. (2020.0) 17. DOI: 10.1371/journal.pmed.1003347 27. Harding A.H., Day N.E., Khaw K.T., Bingham S., Luben R., Welsh A., Wareham N.J.. Dietary Fat and the Risk of Clinical Type 2 Diabetes: The European Prospective Investigation of Cancer-Norfolk Study. Am. J. Epidemiol. (2004.0) 159 73-82. DOI: 10.1093/aje/kwh004 28. Van Dam R.M., Willett W.C., Rimm E.B., Stampfer M.J., Hu F.B.. Dietary Fat and Meat Intake in Relation to Risk of Type 2 Diabetes in Men. Diabetes Care (2002.0) 25 417-424. DOI: 10.2337/diacare.25.3.417 29. Salmerón J., Frank B.H., Manson J.A.E., Stampfer M.J., Colditz G.A., Rimm E.B., Willett W.C.. Dietary fat intake and risk of type 2 diabetes in women. Am. J. Clin. Nutr. (2001.0) 73 1019-1026. PMID: 11382654 30. Kobayashi S., Honda S., Murakami K., Sasaki S., Okubo H., Hirota N., Notsu A., Fukui M., Date C.. Both Comprehensive and Brief Self-Administered Diet History Questionnaires Satisfactorily Rank Nutrient Intakes in Japanese Adults. J. Epidemiol. (2012.0) 22 151-159. DOI: 10.2188/jea.JE20110075 31. Kobayashi S., Murakami K., Sasaki S., Okubo H., Hirota N., Notsu A., Fukui M., Date C.. Comparison of Relative Validity of Food Group Intakes Estimated by Comprehensive and Brief-Type Self-Administered Diet History Questionnaires against 16 d Dietary Records in Japanese Adults. Public Health Nutr. (2011.0) 14 1200-1211. DOI: 10.1017/S1368980011000504 32. Konishi K., Wada K., Yamakawa M., Goto Y., Mizuta F., Koda S., Uji T., Tsuji M., Nagata C.. Dietary Soy Intake Is Inversely Associated with Risk of Type 2 Diabetes in Japanese Women but Not in Men. J. Nutr. (2019.0) 149 1208-1214. DOI: 10.1093/jn/nxz047 33. Tsumura K., Hayashi T., Suematsu C., Endo G., Fujii S., Okada K.. Daily alcohol consumption and the risk of type 2 diabetes in Japanese men: The Osaka Health Survey. Diabetes Care (1999.0) 22 1432-1437. DOI: 10.2337/diacare.22.9.1432 34. Seike N., Noda M., Kadowaki T.. Alcohol Consumption and Risk of Type 2 Diabetes Mellitus in Japanese: A Systematic Review. Asia Pac. J. Clin. Nutr. (2008.0) 17 545-551. PMID: 19114388 35. Baliunas D.O., Taylor B.J., Irving H., Roerecke M., Patra J., Mohapatra S., Rehm J.. Alcohol as a Risk Factor for Type 2 Diabetes: A Systematic Review and Meta-Analysis. Diabetes Care (2009.0) 32 2123-2132. DOI: 10.2337/dc09-0227 36. Polsky S., Akturk H.K.. Alcohol Consumption, Diabetes Risk, and Cardiovascular Disease Within Diabetes. Curr. Diab. Rep. (2017.0) 17 1-12. DOI: 10.1007/s11892-017-0950-8 37. Feskens E.J.M., Vlrtanen S.M., Rasanen L., Tuomilehto J., Stengard J., Pekkanen J., Nlssinen A.. Dietary Factors Determining Diabetes and Impaired Glucose Tolerance A 20-Year Follow-up of the Finnish and Dutch Cohorts of the Seven Countries Study. Feskens, E.J.; Virtanen, S.M.; Räsänen, L.; et al. Dietary factors determining diabetes and impaired glucose tolerance: A 20-year follow-up of the Finnish and Dutch cohorts of the Seven Countries Study. Diabetes Care (1995.0) 18 1104-1112. PMID: 7587845 38. De Souza R.J., Mente A., Maroleanu A., Cozma A.I., Ha V., Kishibe T., Uleryk E., Budylowski P., Schünemann H., Beyene J.. Intake of Saturated and Trans Unsaturated Fatty Acids and Risk of All Cause Mortality, Cardiovascular Disease, and Type 2 Diabetes: Systematic Review and Meta-Analysis of Observational Studies. BMJ (2015.0) 351. DOI: 10.1136/bmj.h3978 39. Wu G.. Amino Acids: Metabolism, Functions, and Nutrition. Amino. Acids (2009.0) 37 1-17. DOI: 10.1007/s00726-009-0269-0 40. Storlien L.H., Pan D.A., Kriketos A.D., O’Connor J., Caterson I.D., Cooney G.J., Jenkins A.B., Baur L.A.. Skeletal Muscle Membrane Lipids and Insulin Resistance. Lipids (1996.0) 31 S261-S265. DOI: 10.1007/BF02637087 41. Borkman M., Storlien L.H., Pan D.A., Jenkins A.B., Chisholm D.J., Campbell L.V.. The Relation between Insulin Sensitivity and the Fatty-Acid Composition of Skeletal-Muscle Phospholipids. N. Engl. J. Med. (1993.0) 328 238-244. DOI: 10.1056/NEJM199301283280404 42. Odegaard J.I., Chawla A.. Pleiotropic Actions of Insulin Resistance and Inflammation in Metabolic Homeostasis. Science (2013.0) 339 172-177. DOI: 10.1126/science.1230721 43. Langyan S., Yadava P., Khan F.N., Dar Z.A., Singh R., Kumar A.. Sustaining Protein Nutrition Through Plant-Based Foods. Front. Nutr. (2022.0) 8 1237. DOI: 10.3389/fnut.2021.772573 44. MEXT: Standard Tables of Food Composition in Japan-2015-(Seventh Revised Version)
--- title: Oxidized Mitochondrial DNA Engages TLR9 to Activate the NLRP3 Inflammasome in Myelodysplastic Syndromes authors: - Grace A. Ward - Robert P. Dalton - Benjamin S. Meyer - Amy F. McLemore - Amy L. Aldrich - Nghi B. Lam - Alexis H. Onimus - Nicole D. Vincelette - Thu Le Trinh - Xianghong Chen - Alexandra R. Calescibetta - Sean M. Christiansen - Hsin-An Hou - Joseph O. Johnson - Kenneth L. Wright - Eric Padron - Erika A. Eksioglu - Alan F. List journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966808 doi: 10.3390/ijms24043896 license: CC BY 4.0 --- # Oxidized Mitochondrial DNA Engages TLR9 to Activate the NLRP3 Inflammasome in Myelodysplastic Syndromes ## Abstract Myelodysplastic Syndromes (MDSs) are bone marrow (BM) failure malignancies characterized by constitutive innate immune activation, including NLRP3 inflammasome driven pyroptotic cell death. We recently reported that the danger-associated molecular pattern (DAMP) oxidized mitochondrial DNA (ox-mtDNA) is diagnostically increased in MDS plasma although the functional consequences remain poorly defined. We hypothesized that ox-mtDNA is released into the cytosol, upon NLRP3 inflammasome pyroptotic lysis, where it propagates and further enhances the inflammatory cell death feed-forward loop onto healthy tissues. This activation can be mediated via ox-mtDNA engagement of Toll-like receptor 9 (TLR9), an endosomal DNA sensing pattern recognition receptor known to prime and activate the inflammasome propagating the IFN-induced inflammatory response in neighboring healthy hematopoietic stem and progenitor cells (HSPCs), which presents a potentially targetable axis for the reduction in inflammasome activation in MDS. We found that extracellular ox-mtDNA activates the TLR9-MyD88-inflammasome pathway, demonstrated by increased lysosome formation, IRF7 translocation, and interferon-stimulated gene (ISG) production. Extracellular ox-mtDNA also induces TLR9 redistribution in MDS HSPCs to the cell surface. The effects on NLRP3 inflammasome activation were validated by blocking TLR9 activation via chemical inhibition and CRISPR knockout, demonstrating that TLR9 was necessary for ox-mtDNA-mediated inflammasome activation. Conversely, lentiviral overexpression of TLR9 sensitized cells to ox-mtDNA. Lastly, inhibiting TLR9 restored hematopoietic colony formation in MDS BM. We conclude that MDS HSPCs are primed for inflammasome activation via ox-mtDNA released by pyroptotic cells. Blocking the TLR9/ox-mtDNA axis may prove to be a novel therapeutic strategy for MDS. ## 1. Introduction Myelodysplastic Syndromes (MDSs) are bone marrow (BM) failure diseases typified by chronic BM inflammation, ineffective hematopoiesis, and peripheral blood (PB) cytopenias [1,2,3]. We and others demonstrated that the danger-associated molecular pattern (DAMP) protein S100A9 plays a critical role in the pathogenesis of MDS by creating an inflammatory microenvironment [4,5,6,7]. S100A9 engages Toll-like receptor (TLR)-4 to initiate pyroptosis in hematopoietic stem and progenitor cells (HSPCs) through the Nod-like receptor 3 (NLRP3) inflammasome complex, leading to the induction of clinically evident ineffective hematopoiesis [5]. Constitutive inflammasome activation prevents HPSCs from differentiating, causing cytopenias and contributing to expansion of the malignant clone. Upon pyroptosis execution, cells expel their intracellular contents, including DAMPs, into the extracellular space triggering a feed-forward process that propagates inflammasome and innate immune activation to neighboring cells [8]. Elucidating those signals is important for understanding why the inflammasome is active in MDS. Increased cell-free DNA levels have been reported in chronic inflammatory disorders [9,10], where there is also increased reactive oxygen species (ROS) induced during pyroptotic events, which not only further contribute to inflammasome activation but to the release of cell-free DAMPs into the extracellular space [11,12,13,14,15]. Among the DAMPs released by pyroptotic cytolysis, and mitochondrial membrane depolarization, is ROS-oxidized mitochondrial DNA (ox-mtDNA) which is then recognized by pattern recognition receptors (PRRs) [16,17]. Ox-mtDNA can amplify pyroptosis through direct engagement of the NLRP3 inflammasome, in addition to DNA-recognition receptors such as TLR9 [18]. Recently we demonstrated the diagnostic importance of ox-mtDNA release in MDS by showing a profound elevation of this DAMP in the PB of MDS specimens and in MDS murine models [19], suggesting that DAMPs such as ox-mtDNA aid in the maintenance of the evolutionary pressures that give rise to the malignant clone. Therefore, a better mechanistic understanding of the role of ox-mtDNA in MDS pathogenesis warrants further investigation. Our previous studies also demonstrated the importance of the activation of the NLRP3 inflammasome in the initiation and development of MDS pathogenesis [5]. This work has led to understanding that the inflammatory microenvironment’s activation of pyroptosis leads to the phenotypic anemia in the disease. The NLRP3 protein, after an inflammatory stimulus, multimerizes, leading to the recruitment of the adaptor protein apoptosis-associated speck-like protein containing a CARD (PYCARD, ASC), which also polymerizes, creating ‘specks’ which can be used as predictors of MDS [20,21]. This further demonstrates that the NLRP3-mediated pyroptosis in MDS is not only a biproduct of disease pathogenesis but is linked to the development and evolution of disease progression. Due to the importance of this process in MDS, we investigated whether the pyroptotically released DAMP ox-mtDNA acts as a catalyst of inflammasome activation to perpetuate BM failure in MDS. We found ox-mtDNA aids in maintaining the pyroptotic microenvironment, via engagement and re-localization of TLR9, from the endoplasmic reticulum (ER) to the cellular surface in MDS cells, allowing the formation of functionally active Myddosome (MyD88) signaling complex [22]. Importantly, specific therapeutic interruption of the TLR9/ox-mtDNA axis decreases pyroptosis and improves hematopoiesis, demonstrating a novel targetable axis in MDS, a disease with few therapeutic options. ## 2.1. Extracellular Ox-mtDNA Is a DAMP That Triggers Inflammasome Activation Previous studies demonstrate that ox-mtDNA triggers sterile inflammation and has been implicated as an indispensable effector of NLRP3 inflammasome activation [13,14,23,24], and we reported that the pyroptosis-induced extracellular ox-mtDNA can serve as an MDS diagnostic marker [19]. We validated the increased levels of ox-mtDNA in lower risk (LR) disease, both in PB and BM plasma compared to healthy BM plasma and found increased levels of ox-mtDNA in both (Figure 1A). To evaluate its role as a non-canonical extracellular DAMP in pyroptosis, we treated SKM1 or U937 cells with increasing doses of synthetic ox-mtDNA (5, 50, and 500 ng/mL, Supplemental Figure S1) and observed increased phosphorylation of NFκB (p65) and cleavage of caspase-1 at 50 ng/mL, indicating the induction of inflammasome assembly (Figure 1B). While ox-mtDNA also activated IL-1β, there was no dose response since the lower band (17 kDa) was visible among all doses (Supplemental Figure S2A,B). The 50 ng/mL dosage is comparable to the average LR BM ox-mtDNA observed in Figure 1A, demonstrating that this level is physiologically relevant. Inflammasome activation was confirmed through the activation of IL-1β at 50 ng/mL ox-mtDNA (Figure 1C) and two-fold increase in caspase-1 activity (Figure 1D). Additionally, ox-mtDNA treatment resulted in lactate dehydrogenase (LDH) release indicating lytic cell death (Figure 1E). To validate the role of the NLRP3 inflammasome in ox-mtDNA pyroptosis activation, we either CRISPR knocked out (KO) NLRP3 with a pool of guides or treated cells with MCC950, an established NLRP3 inhibitor [5]. Caspase-1 activation by ox-mtDNA was abrogated by blockage of the NLRP3 inflammasome, confirming its role in our observations (Figure 1F). Similar results were observed with U937 cells (Supplemental Figure S3). ASC specks are polymerized during inflammasome activation, diagnostically relevant [20], serve as a platform for caspase-1 binding and are also released upon lytic death [5,20,25]. We assessed whether ox-mtDNA induce ASC specks and observed a significant increase in ASC speck formation by immunofluorescence (IF) (Figure 1G,H), which was confirmed by Western blot demonstrating increased oligomerization and decreased monomer subunits in the treated cells (Figure 1I). Finally, to demonstrate ox-mtDNA treatment results in activation of pyroptosis, as opposed to apoptosis, we probed for PARP and caspase-3 activation, which was absent upon treatment (Figure 1J). Having confirmed that MDS BM plasma has significantly high levels of ox-mtDNA [5,19] and that this excess is sufficient to induce pyroptosis in leukemic cell lines, we assessed if ox-mtDNA can also induce pyroptosis in healthy BM-MNCs. As expected, treatment with 50 ug/mL ox-mtDNA induced a significant caspase-1 activation and lytic cell death (LDH) in healthy BM (Figure 1K), and significantly decreased hematopoiesis evidenced by decreased colonies (Figure 1L,M). This confirms that ox-mtDNA is a DAMP capable of triggering the NLRP3 inflammasome, in a time- and dose-dependent manner, in cell lines and primary cells affecting hematopoietic potential. ## 2.2. MDS HSPC and Leukemic Cell Lines Have Increased Expression of TLR9 The observation that ox-mtDNA can induce direct effects on hematopoiesis indicates recognition by the target cells. An inflammatory receptor of cell free DNA is TLR9, which has also been shown to play a role in pyroptosis activation [26]. We observed abundant co-localization and increased expression of cytosolic ox-mtDNA and TLR9 in MDS HSPCs, compared to healthy BM-MNCs (Figure 2A, Supplemental Figure S4). This colocalization was quantifiable with MDS cells having an average of $92\%$ of ox-mtDNA bound to TLR9 compared to $72\%$ in normal BM-MNCs (Figure 2B) with cytosolic concentration of ox-mtDNA, and TLR9 surface expression, significantly increased in the cytoplasm of MDS BM-MNCs, compared to healthy controls (Figure 2C). Moreover, there was a comparable 2.5-fold increase in TLR9 gene expression in MDS HSPCs (Figure 2D), confirming previous results showcasing higher TLR9 gene expression in MDS [27]. There was an increase in the percentage of TLR9+ cells, particularly CD34+TLR9+ cells, in MDS compared to normal BM (Figure 2E, Supplemental Figure S5), and particularly in the stem (CD34+CD71−CD14−, CD34+CD33−) and progenitor (CD34+CD38−) populations (Figure 2F). While the overall common myeloid progenitor (CMP, CD34+CD38−) population was not significantly different between MDS and healthy specimens within this small sample size, stratification by risk showed that LR samples were significantly higher than healthy and HR samples (Figure 2G), similar to our observations that circulating ox-mtDNA is higher in lower risk specimens [19] and those of others showing TLR9 reduction during progression to transformation [27]. Moreover, tSNE analysis of TLR9+CD34+ cells showed both an increase in the patient population and, importantly, in their expression of myeloid markers denoted by CD14+ and CD33+ denoting myeloid skewing (Figure 2H, Supplemental Figure S6). TLR9 circulates through the cytoplasm to the surface prior to entering lysosomes, where it’s proximity with MyD88 directs activation [28]. We observed a strong lysosome induction, as read by LC3 and Lysotracker® Deep Red, in MDS (Figure 2I,J). This increased lysosome activation in the MDS HSPC suggests that the TLR9 pathway might be triggered and functional in this disease. We further linked this phenomenon to excess ox-mtDNA in the plasma of MDS patients by demonstrating that treatment of normal BM-MNCs with ox-mtDNA results in increased lysosomes with internalized ox-mtDNA and TLR9 phenocopying MDS (Figure 2K). ## 2.3. TLR9 Is Necessary for Ox-mtDNA Directed Pyroptosis Next, we investigated the impact of ox-mtDNA in the activation of the TLR9/IRF7 signaling axis. Upon incubation of SKM1 and U937 cells with ox-mtDNA, we observed a time-dependent internalization of ox-mtDNA and co-localization with TLR9 (Figure 3A). To confirm the binding of ox-mtDNA with TLR9, we immunoprecipitated TLR9 and confirmed its binding to oxDNA by immunoblotting and vice versa (Figure 3B). To further characterize the effect of TLR9 in the ox-mtDNA-mediated activation, we assessed NLRP3 inflammasome activation by ox-mtDNA, by caspase-1 cleavage, relative to the density of TLR9 expression: high TLR9 expression (SKM1), and medium (U937) and low to absent expression (THP1) (Figure 3C). Receptor density is an important determinant of the time interval to caspase-1 cleavage, with SKM1 cells responding within 1 h of treatment, U937 cells responding within 2 h, and THP1 cells showing no caspase-1 cleavage in response to ox-mtDNA exposure up to 4 h after ox-mtDNA incubation (Figure 3D). Importantly, after 4 h in SKM1 cells, most of the caspase-1 bands are depleted, compared to U937 and THP1, a mechanism used to restrict excessive inflammation [29]. Activation of IL-1β was also time-dependent and a later response in both types of cells, while still TLR9 density dependent, 4 h in SKM1 cells compared to 24h in U937 cells (Supplemental Figure S8A). Additionally, TLR9 KO cells showed observed an abrogation of IL-1β activation induced by ox-mtDNA (Supplemental Figure S8B). To determine if TLR9 is responsible for NLRP3 inflammasome activation after ox-mtDNA exposure, we created a TLR9 knockout in SKM1 (high TLR9) and U937 (medium TLR9) cells using CRISPR/Cas 9 gene editing (Supplemental Figure S9A). Ox-mtDNA treatment induces phosphorylation of NFκB, and the maturation of caspase-1 and IL-1β (Figure 3E and Supplemental Figure S8), as well as caspase-1 and LDH, in the scrambled control but not in TLR9 CRISPR KO SKM1 cells (Figure 3F,G, Supplemental Figure S9B,C) indicating that TLR9 is indispensable for ox-mtDNA-dependent inflammasome activation. To confirm the necessity of TLR9 to sensitize bystander cells to ox-mtDNA, we overexpressed TLR9 in THP1 (no TLR9) with a lentiviral vector prior to treatment with ox-mtDNA, which restored sensitivity to ox-mtDNA (Figure 3H,I). To assess if ox-mtDNA is directly being internalized by TLR9 via its lysosomal trafficking, we assessed ox-mtDNA-treated cells by LC3 and Lysotracker® Deep Red co-localization of lysosome and oxDNA (Figure 3J, Supplemental Figure S9D). These data strengthen the conclusion that TLR9 is a main receptor for ox-mtDNA dependent inflammasome activation, and the significant plasma membrane translocation of TLR9 upon exposure to ox-mtDNA (Figure 3K, Supplemental Figure S9E) indicates that ox-mtDNA leads into a feed-forward loop of pyroptosis. ## 2.4. IRF7 Signaling Is Activated by Ox-mtDNA/TLR9 Engagement To confirm TLR9 activation by ox-mtDNA, we tested potential downstream mediators (TBK1, IRF7, IRF3, and NF-κB) and found that IRF7 was activated by ox-mtDNA treatment of SKM1 cells (Figure 4A), including nuclear translocation upon ox-mtDNA treatment, corroborating its activation (Figure 4B,C). This IRF7 activation occurred rapidly (nuclear translocation within 30 min, Figure 4D, Supplemental Figure S10A) and required TLR9 expression, as TLR9 CRISPR KO prevented ASC formation and IRF7 nuclear translocation (Figure 4E,F, Supplemental Figure S10B,C). Moreover, this activation correlated with the level of TLR9, as SKM1 cells had a faster IRF7 translocation (30 min) compared to U937 cells (4 h), which have comparatively less TLR9 (Supplemental Figure S10A, Figure 3D,F). Lastly, taking advantage of our previously published RNA-seq dataset [30] comparing healthy versus LR MDS specimens, we found increased TLR9 pathway activation, including of CTSB and IRF7, which were accordingly activated by ox-mtDNA (Supplemental Figure S10D). At baseline, MDS patients have significantly higher levels of type 1 interferons (IFNs) [31] and, accordingly, gene expression of the interferon-stimulated genes (ISGs) IFNα1, IFNα10, IFNβ1, CXCL10, ISG15, SAMD9L, and IFI27L2 were elevated in MDS in data obtained from 213 WHO-defined MDS patient specimens at time of diagnosis, as well as from 20 healthy donors from the National Taiwan University Hospital (Figure 4G, Supplemental Figure S11A; additionally, IFNα2, α4, α5, α8, α14, and α21 all significantly elevated). This activation correlated with increased expression of TLR9-mediators, including IRF7 and IRAK in MDS, although there was no difference with regard to risk (Figure 4H, Supplemental Figure S11B). To confirm that the ox-mtDNA is involved in ISG induction through the TLR9/IRF7 axis, we analyzed the gene expression of type I IFN genes after treatment with ox-mtDNA and found a time-dependent increase in their activation (Figure 4I, Supplemental Figure S12A). This activation was mediated through TLR9 activation, as silencing this receptor abrogated increased ISG expression after ox-mtDNA treatment (Figure 4J, Supplemental Figure S12B). These findings show that ox-mtDNA engages and activates TLR9 through IRF7 nuclear translocation. ## 2.5. Ox-mtDNA/TLR9 Signaling Can Be Therapeutically Targeted in MDS Recently, studies have shown that mtDNA/TLR9 ligation is linked to anemia development during inflammation [32]. Having established that MDS BM plasma is sufficient to induce pyroptosis in normal BM-MNCs and that MDS BM plasma has significantly high levels of ox-mtDNA [5,19], we assessed the specific impact of ox-mtDNA and TLR9 on hematopoietic potential. Primary healthy BM-MNCs transfected with lentivirus vectors to overexpress (OE) or knock out (KO) TLR9 with a pool of specific CRISPR guides (Supplemental Figure S13A) were treated with synthetic ox-mtDNA and their colony-forming capacity was assessed. As expected, ox-mtDNA treatment of healthy BM-MNCs significantly reduced colony-forming capacity (Figure 5A,B). TLR9 OE increased sensitization of healthy BM-MNCs to ox-mtDNA, particularly in the erythroid compartment (BFU-E, Figure 5A,B), while TLR9 KO disrupted the effect of ox-mtDNA with hematopoiesis matching untreated levels (Figure 5A,B). These data indicate that ox-mtDNA accumulates in the microenvironment in MDS where it binds TLR9 sensitizing cells to inflammasome activation affecting hematopoiesis. Next, we evaluated the therapeutic potential of targeting the ox-mtDNA/TLR9 axis by testing the ability of several compounds to block caspase-1 activation in SKM1 or U937 cells (Figure 5C and Supplemental Figure S13B). To test the role of cGAS, another nucleic acid sensor that we have recently shown to recognize intracellular DAMPs, we used a non-TLR9 targeting cGAS inhibitor (RU.521) or CRISPR KO cGAS cells (Figure 5C). We used TLR9 KO cells as a negative control. We found that only TLR9 KO cells were able to significantly reduce ox-mtDNA activation of caspase-1, but not cGAS inhibition or KO. To validate the therapeutic potential of these inhibitors in MDS, we tested hematopoietic potential in primary MDS BM-MNCs. Caspase-1 activation was prevented by either blocking TLR9 signaling with an IRAK inhibitor that prevents downstream activation of Myd88, oligodeoxynucleotide (ODN)-F (TLR9 antagonist), blocking ox-DNA with IRS954 (an inert non-sense ODN), preventing lysosome internalization with HCQ, the inflammasome inhibitor MCC950, or depleting excess ox-mtDNA with a soluble TLR9-IgG4 chimeric molecule (the ectodomain of TLR9 fused to the Fc domain of human IgG4) developed by us to serve as a decoy receptor or ligand trap (Supplemental Figure S13C). Blocking TLR9 signaling with IRAKi, trapping excess ox-mtDNA with TLR9 chimera, blocking binding to TLR9 with ODN-F, or preventing lysosomal internalization significantly improved the colony-forming capacity of MDS BM-MNCs (Figure 5D), confirming the therapeutic potential of targeting the ox-mtDNA/TLR9 axis in MDS. ## 3. Discussion Ox-mtDNA is a novel diagnostic biomarker [19] that may represent a therapeutic target for MDS. Here, we demonstrate the role that the diagnostically evident excess of ox-mtDNA plays in hematopoietic potential by inducing the overexpression and engagement of TLR9. Our findings identify it as a key DAMP contributing to both medullary HSPC pyroptosis and propagation of sterile inflammation in MDS. We show that incubation with ox-mtDNA provides a secondary signal that is sufficient to induce activation of the canonical NLRP3 inflammasome pathway and pyroptotic cell death, as evidenced by cleavage of caspase-1 and IL-1β, and ASC speck formation and release. Our studies also suggest that strategies that effectively neutralize extracellular ox-mtDNA/TLR9 may suppress DNA-sensor-directed inflammation in the BM niche and possibly improve hematopoiesis. Indeed, strategies that mitigate mitochondrial membrane depolarization, through activation of the Nrf2-antioxidant pathway, or the binding of ox-mtDNA to its cognate TLR9 receptor, may offer promising therapeutic potential [33,34]. Moreover, these findings may be extended to other disorders in which ox-mtDNA has been implicated in innate immune activation [14,35,36,37]. The important role that pyroptosis plays in the development of the phenotypic cytopenias of MDS demonstrates the pathogenetic involvement of the innate immune microenvironment which provides the selective pressure for evolution of malignant/defective cells [4,5]. The role of extracellular ox-mtDNA in this process is supported by TLR9’s reduction after transformation, when the biological pressure for survival and evolution is no longer needed [27]. Our investigation demonstrates that TLR9 translocates to the plasma membrane in MDS HSPCs, where it can engage nucleic acid-based DAMPs, such as ox-mtDNA, further enhancing pyroptosis and reinforcing surface TLR9 translocation. In this way, we expect that ox-mtDNA contributes to accelerating the death of healthy HSPC and selection of increasingly aggressive malignant clones. Hence, another potential consequence of ox-mtDNA/TLR9 targeting is to therapeutically prevent the evolution towards leukemia. Early in the disease progression, abundance of ox-mtDNA in MDS reinforces surface TLR9 translocation, initiates transcription of ISGs, inflammatory cytokines, and further induction of inflammasome activation. Pellagati et al. demonstrated that ISG transcription is the most upregulated pathway in MDS [31]; our research suggests that this could be a result of ox-mtDNA/TLR9 pathway activation. Inflammasome activation then results in both cell death and proliferation via β-catenin activation [5,38] and IL-1β, which has been implicated in immuno-senescence and myeloid skewing with aging (reviewed in [39]). This activation also results in degradation of the erythroid transcription factor GATA1, through caspase-1 activation [40], changing the ratio between GATA1 and the myeloid transcription factor PU.1 favoring myeloid commitment, maturation arrest, and anemia [40,41]. Additionally, chronic TLR activation in HSPC causes loss of quiescence with recruitment into the cell cycle and HSPC depletion [42,43,44,45]. Having established the importance of ox-mtDNA/TLR9 in MDS, a next step will be to understand its contribution to the phenotype: ISG overexpression, myeloid skewing, immune-senescence, and BM failure. Recent publications show a novel role for red blood cells (RBCs) as an anti-inflammatory sink for mtDNA [32,46,47,48]. MDS is especially sensitive to the loss of this natural ox-mtDNA removal system due to HSPC pyroptosis, which contributes to both the inflammatory milieu and cytopenias. In this manner, the loss of RBCs, coupled with increased ox-mtDNA, results in further HSPC death and feed-forward BM failure. Importantly, we also demonstrate that TLR9 is indispensable for ox-mtDNA-initiated inflammasome activation proportionate to cellular TLR9 density, highlighting the viability of therapeutically targeting this ligand/receptor interaction. We showcase that TLR9 agonists, or a chimeric TLR9-IgG trap we developed, improved hematopoietic potential in MDS BM explants through the reduction in inflammasome activity. Despite the role of cGAS as another potential receptor for ox-mtDNA, our work demonstrates that inhibiting this pathway did not affect the signaling studied here. However, considering pyroptotic release of other nucleic acid-based DAMPs, it will be important to dissect their contribution to this phenotype to assess the hierarchy of this axis in MDS pathogenesis. However, our data demonstrate that TLR9 is critical, so the use of a soluble TLR9 trap could be beneficial at clearing other potential DAMPs that use TLR9 for signaling. ## 4.1. Patient Samples Normal samples were obtained from Stem Express (Folsom, CA, USA). MDS specimens were acquired from consented MDS patient specimens through Moffitt’s Total Cancer Care™ Protocol [49]. De-identified tissues were released for this study, stratified according to the International Prognostic Scoring System (IPSS). The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of the University of South Florida. ## 4.2. Cells THP1 (TIB-202), HEL 92.1.7 (TIB-180), and U937 (CRL-1593.2) cells were obtained from American Type Culture Collection (ATCC) and SKM1 (ACC 547) from the Leibniz-Institute DSMZ–German Collection of Human & Animal Cell Lines. THP1 and HEL overexpressing TLR9 were transduced via lentiviral infection, as previously described, with pcDNA3-TLR9-YFP, which was a gift from Doug Golenbock (Addgene plasmid # 13642) and selected with Neomycin 3 days post-transfection. ## 4.3. Colony Forming Capacity MDS BM mononuclear cells (BM-MNCs), treated as described in the results, were plated in duplicate (1 × 105 cells/dish) in MethoCult™ (H4434, StemCell Technologies, Vancouver, BC, Canada) as previously described [5,20]. After incubating for 14 days, colonies were counted using the StemVision microscope and software (Catalog # 22006, StemCell Technologies) and counts were validated manually. ## 4.4. CRISPR TLR9-deficient cells were created using CRISPR CRIPSR/Cas9 gene editing with RNA guides TLR9 (F- CACCGTTGCAGTTCACCAGGCCGTG R- AAACCACGGCCTGGTGAACTGCAAC), or Scrambled control (F- GACGGAGGCTAAGCGTCGCA, R- TGCGACGCTTAGCCTCCGTC) into a puromycin resistance pL-CRISPR.SFFV plasmid (a gift from Benjamin Ebert, Addgene plasmid # 57829) [22]. Forward and reverse guide oligonucleotides were purchased from Integrated DNA Technologies (Coralville, IA, USA). CACC on forward and AAAC on reverse oligonucleotides were added 5′ for plasmid ligation. CRISPR plasmids were packaged into lentivirus and transduced as previously described [5]. ## 4.5. Inflammasome Activation Ox-mtDNA for treatment was synthesized from mtDNA extracted using the Mitochondrial Extraction Kit according to the manufacture’s protocol (Active Motif, Carlsbad CA, USA) and amplified by ND1 primers (ND1 Forward: 5′-CCCTAAAACCCGCCACATCT-3′; ND1 Reverse: 5′-GAGCGATGGTGAGAGCTAAGGT-3′) with the addition of oxidized guanosine to the master mix and mtDNA amplified, as described in [13] (Supplemental Figure S1). The pyroptotic TLR4 signaling pathway was activated by incubation with LPS, ATP, and nigericin (LAN) [23]. Caspase-Glo® 1 Inflammasome and LDH-Glo® Cytotoxicity assays (Promega Corporation, Madison, WI, USA) were used to assess inflammasome activation following manufacturer’s protocols. Ox-mtDNA levels were quantified using the DNA/RNA Oxidative Damage (High Sensitivity) ELISA Kit (Cayman Chemical Company, Ann Arbor, MI, USA) [19]. ## 4.6. Inhibitors Cells were treated with inhibitors for 1 h prior to the addition of ox-mtDNA. Inhibitors used: our developed TLR9-IgG4 chimera at 50 ng/ul, and corresponding isotype; 0.5 uM IRAK $\frac{1}{4}$ inhibitor (Caymen 17540, CAS 509093-47-4), 1 uM ODN 4048-F (TLR9 antagonist) and ODN 2395 control (Invivogen, San Diego, CA, USA), 30 uM Hydroxychloroquine (HCQ, Sigma-Aldrich, St. Louis, MO, USA) to accumulate autophagosomes for imaging and 10 uM for CFAs, 20 ug/mL IRS95424, 1 µM RU.521 (cGAS inhibitor, Aobious, Gloucester, MA, USA). ## 4.7. Immunofluorescence Immunofluorescent staining was performed as undertaken previously [5] and stained with conjugated primary antibodies. Nuclei were stained with ProLong® Gold anti-fade reagent with DAPI (Thermo Fisher Scientific, Waltham, MA, USA). Primary antibodies used were anti-oxDNA/RNA FITC (Abcam), CD289 (TLR9) APC (Thermo Fisher Scientific), IRF7 Alexa Fluor 647 (Thermo Fisher Scientific). Lysotracker Deep Red (Invitrogen) staining was performed following the manufacturer’s protocol, with the addition of 30 uM HCQ (Sigma) to ensure accumulation of autophagosomes for imaging. Slides were imaged with a Leica SP8 laser scanning confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany). Images were captured through a 63×/1.4 NA objective lens. Images were analyzed with LAS X software version 3.1.5 (Leica Microsystems GmbH, Wetzlar, Germany). Quantification ASC specks, stained with anti-ASC (Santa Cruz, 1:200) and Alexa 488 goat anti-mouse (1:400), were counted from at least 200 cells per group from the MFI on the TIF images with Definiens Tissue Studio version 4.7 (Definiens AG, Munich, Germany). Nucleus detection and cell growth algorithms were used to segment individual cells within each image. ## 4.8. Western Immunoblot Following treatment, the harvested cells were lysed in RIPA buffer with phosphatase and protease inhibitors as previously described [5]. Immunoprecipitations from specimen plasma were isolated using Protein A and G Agarose Fast Flow beads (Millipore Sigma, Burlington, MA, USA) according to the manufacturer’s protocol and using 4 µg of anti-ox-DNA antibody. The following antibodies were used: caspase-1, NF-κB p65, LC3, TLR9, Histone H3, caspase-3, PARP, IRF7, phospho-TBK1, phospho-IRF7, phospho-IRF3 (Cell Signaling Technology, Inc., Danvers, MA, USA), Human IL-1 beta/IL-1F2 (R&D Systems, Inc., Minneapolis, MN, USA), anti-oxDNA/RNA (Abcam), Anti-NLRP3 Antibody (Millipore Sigma), ASC (Santa Cruz), β-actin (Sigma-Aldrich), and appropriate Amersham ECL HRP Conjugated Antibodies (Thomas Scientific, Swedesboro, NJ, USA). Apoptosis positive control was A431 Whole Cell Lysate EGF Stimulated (Rockland Immunochemicals, Inc., Limerick, PA, USA). To assess ASC oligomerization, cell pellets were harvested and lysed. The pellet was then resuspended in PBS and fresh 2 mM Disuccinimidyl suberate (DSS, Thermo Scientific), and incubated on a rotator for 30 min at room temperature. Following this DSS-crosslinking, the resultant was again pelleted for 10 m at 5000 rpm at 4 °C prior to blotting. ## 4.9. Flow Cytometry Cryopreserved normal and MDS BM-MNCs were thawed, washed with $2\%$ BSA-PBS, blocked with Human FcR Blocking reagent (Miltenyi Biotec; Bergisch Gladbach, Germany), stained for CD14, CD33, CD34, CD38, (Becton, Dickinson and Company, Franklin Lakes, NJ, USA), TLR9/CD289, CD71 (Thermo Fisher Scientific; Waltham, MA, USA), resuspended in 0.1 uM DAPI (Thermo Fisher Scientific), and run on the BD LSRII (Becton, Dickinson and Company) at Moffitt’s Flow Cytometry core facility. FCS files were analyzed using FlowJo v10 (FlowJo LLC. Ashland, Oregon). ## 4.10. Gene Expression Analysis RNA extraction was performed with Qiagen RNeasy Mini Kit (Hilden, Germany) following the manufacturer protocol. RNA concentration and integrity were verified using the ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE, USA). RNA was reverse transcribed using qScript XLT cDNA Supermix (QuantaBio, Beverly, MA, USA) according to the manufacturer’s protocol. INFB1/Infb1 and GATA1 expression was quantified using TaqMan Advanced Gene Expression Assays, TaqMan Fast Advanced Mastermix, and 7900 HT Fast Real-Time PCR System (Life Technologies, Foster City, CA, USA). The rest were amplified using the primers in Supplemental Table S1 with SYBR Green PCR Master Mix (Life Technologies). Data were analyzed using the ΔΔCt methodology with ACTB/Actb expression for TaqMan assays and GAPDH/Gapdh for SYBR assays as internal control. Gene expression profiling data were also obtained from 213 WHO-defined MDS patient specimens at time of diagnosis, as well as from 20 healthy donors from the National Taiwan University Hospital using the Human HT-12 v4 Expression BeadChip (Illumina, San Diego, CA, USA) [30]. For each sample, 1.5 μg cDNA was hybridized to Illumina HumanHT-12 v4 Expression BeadChip according to the manufacturer’s instructions. Intensities of bead fluorescence were detected by Illumina BeadArray Reader, and the results were transformed to numeric values using GenomeStudio v2010.1 Software (Illumina). ## 4.11. Sandwich ELISA The binding efficacy of the TLR9-IgG chimera was assessed using a Bethyl Labs (Montgomery, TX, USA) Sandwich ELISA kit (ELISA Starter Accessory Kit I), following the manufacture’s protocol for Indirect ELISA. Briefly, a 96-well plate was coated with Coating Buffer and TLR9-IgG chimera or IgG4 was added in a 2-fold dilution curve and, after blocking, the ODN 2006 Biotin (Invivogen) was added followed by streptavidin-HRP (Cell Signaling Technology). Binding of the chimera to CpG was assessed by colorimetric quantification of TMB substrate development. ## 4.12. Statistical Analysis Statistical analyses were performed using GraphPad Prism Software: two treatment groups by unpaired Student t tests; fold change by paired Student t tests; and one-way ANOVA was applied to other data with multiple comparison analysis. All analyses and graphics show standard error of the mean bars (SEM) and p values < 0.05 are statistically significant. p values are shown as asterisk: * p ≤ 0.05, p ≤ 0.01, * p ≤ 0.001, ** p ≤ 0.0001. ## References 1. Nimer S.D.. Myelodysplastic syndromes*. Blood* (2008) 111 4841-4851. DOI: 10.1182/blood-2007-08-078139 2. Estey E.. Acute myeloid leukemia and myelodysplastic syndromes in older patients. J. Clin. Oncol. (2007) 25 1908-1915. DOI: 10.1200/JCO.2006.10.2731 3. Look A.T.. Molecular Pathogenesis of MDS. Hematology (2005) 2005 156-160. DOI: 10.1182/asheducation-2005.1.156 4. Chen X., Eksioglu E.A., Zhou J., Zhang L., Djeu J., Fortenbery N., Epling-Burnette P., Van Bijnen S., Dolstra H., Cannon J.. Induction of myelodysplasia by myeloid-derived suppressor cells. J. Clin. Investig. (2013) 123 4595-4611. DOI: 10.1172/JCI67580 5. Basiorka A.A., McGraw K.L., Eksioglu E.A., Chen X., Johnson J., Zhang L., Zhang Q., Irvine B.A., Cluzeau T., Sallman D.A.. The NLRP3 inflammasome functions as a driver of the myelodysplastic syndrome phenotype. Blood (2016) 128 2960-2975. DOI: 10.1182/blood-2016-07-730556 6. Schneider R.K., Schenone M., Ferreira M.V., Kramann R., Joyce C.E., Hartigan C., Beier F., Brummendorf T.H., Germing U., Platzbecker U.. Rps14 haploinsufficiency causes a block in erythroid differentiation mediated by S100A8 and S100A9. Nat. Med. (2016) 22 288-297. DOI: 10.1038/nm.4047 7. Zambetti N.A., Ping Z., Chen S., Kenswil K.J., Mylona M.A., Sanders M.A., Hoogenboezem R.M., Bindels E.M., Adisty M.N., Van Strien P.M.. Mesenchymal Inflammation Drives Genotoxic Stress in Hematopoietic Stem Cells and Predicts Disease Evolution in Human Pre-leukemia. Cell Stem Cell (2016) 19 613-627. DOI: 10.1016/j.stem.2016.08.021 8. Jorgensen I., Zhang Y., Krantz B.A., Miao E.A.. Pyroptosis triggers pore-induced intracellular traps (PITs) that capture bacteria and lead to their clearance by efferocytosis. J. Exp. Med. (2016) 213 2113-2128. DOI: 10.1084/jem.20151613 9. Giles B.M., Boackle S.A.. Linking complement and anti-dsDNA antibodies in the pathogenesis of systemic lupus erythematosus. Immunol. Res. (2013) 55 10-21. DOI: 10.1007/s12026-012-8345-z 10. Hajizadeh S., DeGroot J., TeKoppele J.M., Tarkowski A., Collins L.V.. Extracellular mitochondrial DNA and oxidatively damaged DNA in synovial fluid of patients with rheumatoid arthritis. Arthritis. Res. Ther. (2003) 5 R234-R240. DOI: 10.1186/ar787 11. Zhong Z., Liang S., Sanchez-Lopez E., He F., Shalapour S., Lin X.-j., Wong J., Ding S., Seki E., Schnabl B.. New mitochondrial DNA synthesis enables NLRP3 inflammasome activation. Nature (2018) 560 198-203. DOI: 10.1038/s41586-018-0372-z 12. Lotze M.T., Zeh H.J., Rubartelli A., Sparvero L.J., Amoscato A.A., Washburn N.R., Devera M.E., Liang X., Tor M., Billiar T.. The grateful dead: Damage-associated molecular pattern molecules and reduction/oxidation regulate immunity. Immunol. Rev. (2007) 220 60-81. DOI: 10.1111/j.1600-065X.2007.00579.x 13. Collins L.V., Hajizadeh S., Holme E., Jonsson I.M., Tarkowski A.. Endogenously oxidized mitochondrial DNA induces in vivo and in vitro inflammatory responses. J. Leukoc. Biol. (2004) 75 995-1000. DOI: 10.1189/jlb.0703328 14. Zhang Q., Raoof M., Chen Y., Sumi Y., Sursal T., Junger W., Brohi K., Itagaki K., Hauser C.J.. Circulating mitochondrial DAMPs cause inflammatory responses to injury. Nature (2010) 464 104-107. DOI: 10.1038/nature08780 15. Chen G.Y., Nunez G.. Sterile inflammation: Sensing and reacting to damage. Nat. Rev. Immunol. (2010) 10 826-837. DOI: 10.1038/nri2873 16. Shimada K., Crother T.R., Karlin J., Dagvadorj J., Chiba N., Chen S., Ramanujan V.K., Wolf A.J., Vergnes L., Ojcius D.M.. Oxidized mitochondrial DNA activates the NLRP3 inflammasome during apoptosis. Immunity (2012) 36 401-414. DOI: 10.1016/j.immuni.2012.01.009 17. Vollmer J., Weeratna R.D., Jurk M., Samulowitz U., McCluskie M.J., Payette P., Davis H.L., Schetter C., Krieg A.M.. Oligodeoxynucleotides lacking CpG dinucleotides mediate Toll-like receptor 9 dependent T helper type 2 biased immune stimulation. Immunology (2004) 113 212-223. DOI: 10.1111/j.1365-2567.2004.01962.x 18. Grishman E.K., White P.C., Savani R.C.. Toll-like receptors, the NLRP3 inflammasome, and interleukin-1beta in the development and progression of type 1 diabetes. Pediatr. Res. (2012) 71 626-632. DOI: 10.1038/pr.2012.24 19. Ward G.A., McGraw K.L., Abbas-Aghababazadeh F., Meyer B.S., McLemore A.F., Vincelette N.D., Lam N.B., Aldrich A.L., Al Ali N.H., Padron E.. Oxidized mitochondrial DNA released after inflammasome activation is a disease biomarker for myelodysplastic syndromes. Blood Adv. (2021) 5 2216-2228. DOI: 10.1182/bloodadvances.2020003475 20. Basiorka A.A., McGraw K.L., Abbas-Aghababazadeh F., McLemore A.F., Vincelette N.D., Ward G.A., Eksioglu E.A., Sallman D.A., Ali N.A., Padron E.. Assessment of ASC specks as a putative biomarker of pyroptosis in myelodysplastic syndromes: An observational cohort study. Lancet. Haematol. (2018) 5 e393-e402. DOI: 10.1016/S2352-3026(18)30109-1 21. Broz P., Dixit V.M.. Inflammasomes: Mechanism of assembly, regulation and signalling. Nat. Rev. Immunol. (2016) 16 407-420. DOI: 10.1038/nri.2016.58 22. Paludan S.R.. Activation and regulation of DNA-driven immune responses. Microbiol. Mol. Biol. Rev. MMBR (2015) 79 225-241. DOI: 10.1128/MMBR.00061-14 23. Garcia-Martinez I., Santoro N., Chen Y., Hoque R., Ouyang X., Caprio S., Shlomchik M.J., Coffman R.L., Candia A., Mehal W.Z.. Hepatocyte mitochondrial DNA drives nonalcoholic steatohepatitis by activation of TLR9. J. Clin. Investig. (2016) 126 859-864. DOI: 10.1172/JCI83885 24. Liu L., Mao Y., Xu B., Zhang X., Fang C., Ma Y., Men K., Qi X., Yi T., Wei Y.. Induction of neutrophil extracellular traps during tissue injury: Involvement of STING and Toll-like receptor 9 pathways. Cell Prolif. (2019) 52 e12579. DOI: 10.1111/cpr.12579 25. Fernandes-Alnemri T., Wu J., Yu J.W., Datta P., Miller B., Jankowski W., Rosenberg S., Zhang J., Alnemri E.S.. The pyroptosome: A supramolecular assembly of ASC dimers mediating inflammatory cell death via caspase-1 activation. Cell Death Differ. (2007) 14 1590-1604. DOI: 10.1038/sj.cdd.4402194 26. Kumar V.. The Trinity of cGAS, TLR9, and ALRs Guardians of the Cellular Galaxy Against Host-Derived Self-DNA. Front. Immunol. (2020) 11 624597. DOI: 10.3389/fimmu.2020.624597 27. Kuninaka N., Kurata M., Yamamoto K., Suzuki S., Umeda S., Kirimura S., Arai A., Nakagawa Y., Suzuki K., Kitagawa M.. Expression of Toll-like receptor 9 in bone marrow cells of myelodysplastic syndromes is down-regulated during transformation to overt leukemia. Exp. Mol. Pathol. (2010) 88 293-298. DOI: 10.1016/j.yexmp.2010.01.009 28. Martinez-Garcia E.A., Zavala-Cerna M.G., Lujano-Benitez A.V., Sanchez-Hernandez P.E., Martin-Marquez B.T., Sandoval-Garcia F., Vazquez-Del Mercado M.. Potential Chronotherapeutic Optimization of Antimalarials in Systemic Lupus Erythematosus: Is Toll-Like Receptor 9 Expression Dependent on the Circadian Cycle in Humans?. Front. Immunol. (2018) 9 1497. DOI: 10.3389/fimmu.2018.01497 29. Walsh J.G., Logue S.E., Luthi A.U., Martin S.J.. Caspase-1 promiscuity is counterbalanced by rapid inactivation of processed enzyme. J. Biol. Chem. (2011) 286 32513-32524. DOI: 10.1074/jbc.M111.225862 30. Cheng P., Eksioglu E.A., Chen X., Kandell W., Le Trinh T., Cen L., Qi J., Sallman D.A., Zhang Y., Tu N.. S100A9-induced overexpression of PD-1/PD-L1 contributes to ineffective hematopoiesis in myelodysplastic syndromes. Leukemia (2019) 33 2034-2046. DOI: 10.1038/s41375-019-0397-9 31. Pellagatti A., Cazzola M., Giagounidis A.A., Malcovati L., Porta M.G., Killick S., Campbell L.J., Wang L., Langford C.F., Fidler C.. Gene expression profiles of CD34+ cells in myelodysplastic syndromes: Involvement of interferon-stimulated genes and correlation to FAB subtype and karyotype. Blood (2006) 108 337-345. DOI: 10.1182/blood-2005-12-4769 32. Lam L.K.M., Murphy S., Kokkinaki D., Venosa A., Sherrill-Mix S., Casu C., Rivella S., Weiner A., Park J., Shin S.. DNA binding to TLR9 expressed by red blood cells promotes innate immune activation and anemia. Sci. Transl. Med. (2021) 13 eabj1008. DOI: 10.1126/scitranslmed.abj1008 33. Zhang C., Hao X., Chang J., Geng Z., Wang Z.. Mn-TAT PTD-Ngb attenuates oxidative injury by an enhanced ROS scavenging ability and the regulation of redox signaling pathway. Sci. Rep. (2019) 9 20103. DOI: 10.1038/s41598-019-56595-5 34. Sekheri M., El Kebir D., Edner N., Filep J.G.. 15-Epi-LXA4 and 17-epi-RvD1 restore TLR9-mediated impaired neutrophil phagocytosis and accelerate resolution of lung inflammation. Proc. Natl. Acad. Sci. USA (2020) 117 7971-7980. DOI: 10.1073/pnas.1920193117 35. Lindqvist D., Fernstrom J., Grudet C., Ljunggren L., Traskman-Bendz L., Ohlsson L., Westrin A.. Increased plasma levels of circulating cell-free mitochondrial DNA in suicide attempters: Associations with HPA-axis hyperactivity. Transl. Psychiatry (2016) 6 e971. DOI: 10.1038/tp.2016.236 36. Pinti M., Cevenini E., Nasi M., De Biasi S., Salvioli S., Monti D., Benatti S., Gibellini L., Cotichini R., Stazi M.A.. Circulating mitochondrial DNA increases with age and is a familiar trait: Implications for "inflamm-aging". Eur. J. Immunol. (2014) 44 1552-1562. DOI: 10.1002/eji.201343921 37. Mehta N.N., Teague H.L., Swindell W.R., Baumer Y., Ward N.L., Xing X., Baugous B., Johnston A., Joshi A.A., Silverman J.. IFN-gamma and TNF-alpha synergism may provide a link between psoriasis and inflammatory atherogenesis. Sci. Rep. (2017) 7 13831. DOI: 10.1038/s41598-017-14365-1 38. Huang L., Luo R., Li J., Wang D., Zhang Y., Liu L., Zhang N., Xu X., Lu B., Zhao K.. beta-catenin promotes NLRP3 inflammasome activation via increasing the association between NLRP3 and ASC. Mol. Immunol. (2020) 121 186-194. DOI: 10.1016/j.molimm.2020.02.017 39. Mantovani A., Dinarello C.A., Molgora M., Garlanda C.. Interleukin-1 and Related Cytokines in the Regulation of Inflammation and Immunity. Immunity (2019) 50 778-795. DOI: 10.1016/j.immuni.2019.03.012 40. Tyrkalska S.D., Perez-Oliva A.B., Rodriguez-Ruiz L., Martinez-Morcillo F.J., Alcaraz-Perez F., Martinez-Navarro F.J., Lachaud C., Ahmed N., Schroeder T., Pardo-Sanchez I.. Inflammasome Regulates Hematopoiesis through Cleavage of the Master Erythroid Transcription Factor GATA1. Immunity (2019) 51 50-63.e55. DOI: 10.1016/j.immuni.2019.05.005 41. De Maria R., Zeuner A., Eramo A., Domenichelli C., Bonci D., Grignani F., Srinivasula S.M., Alnemri E.S., Testa U., Peschle C.. Negative regulation of erythropoiesis by caspase-mediated cleavage of GATA-1. Nature (1999) 401 489-493. DOI: 10.1038/46809 42. Kennedy D.E., Knight K.L.. Inflammatory Changes in Bone Marrow Microenvironment Associated with Declining B Lymphopoiesis. J. Immunol. (2017) 198 3471-3479. DOI: 10.4049/jimmunol.1601643 43. Holl T.M., Kelsoe G.. Outside influence: TLRs direct hematopoietic cell fates. Immunity (2006) 24 667-669. DOI: 10.1016/j.immuni.2006.06.007 44. Esplin B.L., Shimazu T., Welner R.S., Garrett K.P., Nie L., Zhang Q., Humphrey M.B., Yang Q., Borghesi L.A., Kincade P.W.. Chronic exposure to a TLR ligand injures hematopoietic stem cells. J. Immunol. (2011) 186 5367-5375. DOI: 10.4049/jimmunol.1003438 45. O’Hagan-Wong K., Nadeau S., Carrier-Leclerc A., Apablaza F., Hamdy R., Shum-Tim D., Rodier F., Colmegna I.. Increased IL-6 secretion by aged human mesenchymal stromal cells disrupts hematopoietic stem and progenitor cells’ homeostasis. Oncotarget (2016) 7 13285-13296. DOI: 10.18632/oncotarget.7690 46. Hotz M.J., Qing D., Shashaty M.G.S., Zhang P., Faust H., Sondheimer N., Rivella S., Worthen G.S., Mangalmurti N.S.. Red Blood Cells Homeostatically Bind Mitochondrial DNA through TLR9 to Maintain Quiescence and to Prevent Lung Injury. Am. J. Respir. Crit. Care Med. (2018) 197 470-480. DOI: 10.1164/rccm.201706-1161OC 47. Akilesh H.M., Buechler M.B., Duggan J.M., Hahn W.O., Matta B., Sun X., Gessay G., Whalen E., Mason M., Presnell S.R.. Chronic TLR7 and TLR9 signaling drives anemia via differentiation of specialized hemophagocytes. Science (2019) 363 eaao5213. DOI: 10.1126/science.aao5213 48. Darbonne W.C., Rice G.C., Mohler M.A., Apple T., Hebert C.A., Valente A.J., Baker J.B.. Red blood cells are a sink for interleukin 8, a leukocyte chemotaxin. J. Clin. Invest. (1991) 88 1362-1369. DOI: 10.1172/JCI115442 49. Fenstermacher D.A., Wenham R.M., Rollison D.E., Dalton W.S.. Implementing personalized medicine in a cancer center. Cancer J. (2011) 17 528-536. DOI: 10.1097/PPO.0b013e318238216e
--- title: 'Prognostic Value of Magnesium in COVID-19: Findings from the COMEPA Study' authors: - Anna La Carrubba - Nicola Veronese - Giovanna Di Bella - Claudia Cusumano - Agnese Di Prazza - Stefano Ciriminna - Antonina Ganci - Liliana Naro - Ligia J. Dominguez - Mario Barbagallo journal: Nutrients year: 2023 pmcid: PMC9966815 doi: 10.3390/nu15040830 license: CC BY 4.0 --- # Prognostic Value of Magnesium in COVID-19: Findings from the COMEPA Study ## Abstract Magnesium (Mg) plays a key role in infections. However, its role in coronavirus disease 2019 (COVID-19) is still underexplored, particularly in long-term sequelae. The aim of the present study was to examine the prognostic value of serum Mg levels in older people affected by COVID-19. Patients were divided into those with serum Mg levels ≤1.96 vs. >1.96 mg/dL, according to the Youden index. A total of 260 participants (mean age 65 years, $53.8\%$ males) had valid Mg measurements. Serum Mg had a good accuracy in predicting in-hospital mortality (area under the curve = 0.83; $95\%$ CI: 0.74–0.91). Low serum Mg at admission significantly predicted in-hospital death (HR = 1.29; $95\%$ CI: 1.03–2.68) after adjusting for several confounders. A value of Mg ≤ 1.96 mg/dL was associated with a longer mean length of stay compared to those with a serum Mg > 1.96 (15.2 vs. 12.7 days). Low serum Mg was associated with a higher incidence of long COVID symptomatology (OR = 2.14; $95\%$ CI: 1.30–4.31), particularly post-traumatic stress disorder (OR = 2.00; $95\%$ CI: 1.24–16.40). In conclusion, low serum Mg levels were significant predictors of mortality, length of stay, and onset of long COVID symptoms, indicating that measuring serum Mg in COVID-19 may be helpful in the prediction of complications related to the disease. ## 1. Introduction Magnesium (Mg) is the most frequent divalent cation present intracellularly in the human body and the second most common intracellular ion after potassium [1]. Mg is essential for numerous cellular processes because it is a cofactor of over 600 intracellular enzymatic reactions [2]. Magnesium is essential for energy production, oxidative phosphorylation, glycolysis, protein synthesis, and nucleic acid synthesis and stability [3,4]. Magnesium modulates muscle contraction, normal heart rhythm and neuron excitability, as it is necessary for the transport of other ions across cellular membranes [5]. This essential ion is involved in all ATP-dependent biochemical processes as part of the activated MgATP complex, as well as in RNA expression, DNA synthesis, muscular and neural cellular signaling, glucose metabolism and blood pressure control [6,7]. Low serum Mg concentrations, rather common in the Western world, are frequently observed in older people, because of poor intake in the diet, some comorbidities (e.g., diabetes), and polypharmacy [8,9,10,11]. Older adults, together with a higher frequency of Mg deficiency, suffer alterations to the immune system that make them more susceptible to infections and their complications [12]. This includes the increased risk of major complications and especially mortality, which has been observed in SARS-CoV-2 infection [13]. Low Mg levels play an important role in several chronic diseases affecting older people, including respiratory conditions, such as chronic obstructive pulmonary disease (COPD) and asthma [14,15]. A former retrospective study reported that serum Mg was an independent predictor of frequent readmissions for acute COPD exacerbations [16]. More recently, a study showed that low Mg status was predictive of the risk of bacterial pneumonia in old age [17]. Furthermore, other studies indicated that low serum Mg levels were present in asthmatic patients [18], and that Mg supplementation may be useful in the treatment of acute asthma [19]. This is based on the evidence that bronchial hyperreactivity is inversely proportional to the serum level of this cation [20,21]. Mg plays a role in the immune system, in both innate and acquired immune response, being involved as a cofactor for immunoglobulin (Ig) synthesis, C3 convertase, immune cell adherence, antibody-dependent cytolysis, IgM lymphocyte binding, macrophage response to lymphokines, and T helper–B cell adherence [22]. A recent study showed that altered Mg status seems to have a prognostic role in older people affected by bacterial pneumonia [17]. Of interest, both hypomagnesemia and hypermagnesemia were associated with an increased short-term mortality rate compared to normal values of serum Mg in these patients with community acquired pneumonia [17]. Vitamin D, which seems to play a key role in the immune response [23], requires an adequate level of Mg for its proper transport and activation [5,24,25]. Thus, a Mg deficiency can exacerbate susceptibility to infections, including COVID-19, by reducing the availability or adequate functional levels of vitamin D. Recently, some works explored the possible role of Mg in COVID-19 to predict the prognosis in these patients [26,27,28,29,30]. However, no study so far has examined the role of Mg for long COVID prediction, an increasing entity with limited therapeutical options [31]. The present study, exploring the prognostic value of serum Mg in patients with COVID-19, follows one of the original aims of the COMEPA study, which was to explore possible prognostic factors for COVID-19 complications during or after hospitalization based on our real-life experience [32]. Validated markers capable of predicting the variable trajectory of the disease from completely asymptomatic to mild, moderate or several clinical manifestation and rapidly progressive forms that can lead to multiorgan failure and death have not yet fully identified. In fact, although several proposed prognostic models have been reported, most of them are of poor quality [33], highlighting the need to continue research on useful tools with this specific aim. In consideration of the role of Mg in the most frequent respiratory diseases, including infections, and due to the lack of literature in this regard, the purpose of our study was to examine the prognostic value of serum Mg in patients affected by COVID-19, in terms of in-hospital mortality, length of stay, and the occurrence of long COVID. ## 2.1. Study Population All patients aged ≥18 years hospitalized in the internal medicine or geriatrics wards from the 1st of September 2020 at the University Hospital ‘p. Giaccone’ in Palermo, Italy with a diagnosis of SARS-CoV-2 infection confirmed by the observation of SARS-CoV-2 nucleic acid on a nasopharyngeal swab by means of RT-PCR were enrolled [32]. No other inclusion criteria were considered to better represent a real-life scenario. The study was approved by the Local Ethical Committee during the session of the 28th of April 2021 (protocol number $\frac{04}{2021}$). For hygiene reasons, the informed consent to participate in the study was collected orally and reported in the medical records. ## 2.2. Exposure: Serum Mg Levels Serum Mg levels at the baseline were measured in the first four days of the hospital admission, including measurements at the emergency department. The values of serum Mg assessed during hospitalization were also recorded and the maximum value during the hospital stay was used as covariate for the analyses. The normal range of our laboratory for serum Mg was 1.7 to 2.5 mg/dL. ## 2.3. Outcomes The primary outcome was mortality during hospital stay. This information was collected using dates of death according to the clinical records and death certificates. As secondary outcomes, we considered the length of stay in hospital and the incidence of long COVID symptomatology. In October 2021, the World Health Organization (WHO) defined long COVID as “a condition that occurs in individuals with a history of probable or confirmed SARS-CoV-2 infection, usually 3 months from the onset of COVID-19 with symptoms that last for at least 2 months and cannot be explained by an alternative diagnosis”. [ 34] Accordingly, the presence of long COVID was assessed after a median of 17 months (range: 13–22) from hospital discharge through phone calls similar to other works using the same approach [35,36,37,38]. We considered as signs or symptoms of long COVID those indicated in recent systematic reviews [35,36,37,38], i.e., neurological, respiratory, mobility impairment, heart, digestive, skin, or general signs and symptoms that can be attributable to COVID-19 infection. All of the questions were posed as yes/no questions by phone. Psychiatric conditions were assessed using the Post-traumatic Stress Disorder (PTSD) Checklist (PCL)-5 [39] and the Hospital Anxiety and Depression Scale (HADS) for detecting anxiety and depression [40]. ## 2.4. Covariates Among the parameters that were collected in the COVID-19 Palermo (COMEPA) study [32], for the aim of the present study, we used the information potentially affecting the association between serum Mg levels and the outcomes of interest, i.e., age, gender, smoking status (actual vs. previous or never), and alcohol abuse (yes vs. no). Among laboratory measurements, we considered creatinine clearance according to the Modification of Diet in Renal Disease (MDRD) formula, hemoglobin, serum parameters of inflammation (white blood cells, C reactive protein (CRP), interleukin (IL)-6, procalcitonin), parameters of arterial blood gas exchange expressed as partial pressure of oxygen/fraction of inspired oxygen (PaO2/FiO2) ratio (with a value below 150 indicative of acute respiratory failure) [41], serum 25 hydroxyvitamin D (25OHD), hepatic function, fasting plasma glucose, sodium, potassium, and albumin. The presence and the severity of comorbidities were investigated using the Cumulative Illness Rating Scale (CIRS) [42] that estimates the severity of pathology in each of 13 systems, with a grade from 0 to 4 (severity index: CIRS-SI). ## 2.5. Statistical Analyses All patient records and information were anonymized and de-identified prior to the analyses. We selected the cut-off value of 1.96 mg/dL of serum Mg since it was the best in terms of sensitivity and specificity (Youden’s index) [43] for testing the prediction of our primary outcome. Data on continuous variables were normally distributed according to the Kolmogorov–Smirnov test and then reported as means and standard deviation (SD) values for quantitative measures and percentages for the categorical variables, by serum Mg status. Levene’s test was used to test the homoscedasticity of variances and, if its assumption was violated, Welch’s ANOVA was used. p values were calculated using Student’s t-test for continuous variables and the Mantel–Haenszel chi-square test for categorical ones. The accuracy of serum Mg in predicting in-hospital mortality during follow-up was calculated in terms of area under the curve (AUC) with its $95\%$ confidence intervals (CIs). The association between serum Mg at baseline being less or more than 1.96 mg/dL and in-hospital mortality was assessed using Cox’s regression analysis, adjusted for potential confounders that were introduced in the model if they did differ between low and high serum Mg levels (p-value < 0.05) or if they were associated with in-hospital death using a p-value threshold of 0.10. Collinearity among factors was analyzed using a variance inflation factor (VIF) of two as a reason for exclusion. The results, considering participants with serum Mg over 1.96 mg/dL as a reference, were reported as hazard ratios (HRs) with their $95\%$ confidence interval (CI). Data regarding long COVID were reported using an adjusted logistic regression and reported as odds ratios (ORs) with their $95\%$ CI. All analyses were performed using the SPSS 26.0 for Windows (SPSS Inc., Chicago, IL, USA) and STATA 14.0. All statistical tests were two-tailed and statistical significance was assumed for a p-value < 0.05. ## 3. Results Among 530 patients initially included in the COMEPA study, 270 were excluded: 250 had not any serum Mg measurement in the first four days from admission and the other 20 did not register any of the outcomes of interest. Consequently, a total of 260 participants (mean age 65.4 ± 15.4, range: 21–96 years; $53.8\%$ men) were included in the analyses. The mean serum Mg level was 2.07 ± 0.23 mg/dL (range: 1.32–2.50), with 26 patients ($10.0\%$) reporting hypomagnesemia identified as a value <1.85 mg/dL. In cases of hypomagnesemia, the patients were supplemented using intravenous Mg sulfate, with the dose depending on the severity of hypomagnesemia until the normalization of Mg serum concentrations. No one reported hypermagnesemia (serum Mg > 2.5 mg/dL). Table 1 shows the baseline characteristics according to the serum Mg levels. The 74 patients with a value less than 1.96 mg/dl were significantly older ($$p \leq 0.01$$), but they did not differ in terms of the percentage of males or in their smoking prevalence or alcohol abuse, compared to their counterparts with higher serum Mg levels. Among the laboratory parameters assessed, patients with low serum Mg levels displayed significantly lower hemoglobin and albumin levels, but they did not differ in any of the inflammatory parameters investigated (Table 1). Finally, patients with low serum Mg levels reported a significantly lower prevalence of any COVID-19 symptomatology, but a higher severity of medical conditions, according to the CIRS-SI. As shown in Figure 1, a model including serum Mg, adjusted for age and sex, had a good accuracy in predicting in-hospital mortality (AUC = 0.83; $95\%$ CI: 0.74–0.91; $p \leq 0.0001$). A value of serum Mg = 1.96 during the first four days of hospitalization had a good sensitivity ($75\%$) and a modest specificity ($58\%$) in predicting mortality during hospital stay. Low serum Mg at admission significantly predicted in-hospital death (HR = 1.29; $95\%$ CI: 1.03–2.68) after adjusting for age, sex, comorbidities, renal function, presence of respiratory failure, CRP, hemoglobin, and maximum Mg serum levels during hospitalization (Figure 2). A value of Mg ≤ 1.96 was associated with a longer mean length of stay compared to those with a serum Mg > 1.96 (15.2 vs. 12.7 days; $$p \leq 0.048$$). Finally, we investigated the association between low serum Mg levels and the presence of long COVID among 95 patients with available data. Among all the signs and symptoms investigated, low serum Mg was associated with a higher incidence of overall long COVID symptomatology (OR = 2.14; $95\%$ CI: 1.30–4.31), particularly PTSD (OR = 2.00; $95\%$ CI: 1.24–16.40), whilst no significant association was found for the other single long COVID signs/symptoms investigated in our questionnaire. ## 4. Discussion Our study including 260 participants hospitalized for COVID-19 indicates the important role of Mg in the prognosis of these patients. Patients with lower serum Mg levels, in fact, had an increased risk not only of in-hospital mortality, but also a longer length of stay and higher incidence of long COVID symptomatology, with this study being, to the best of our knowledge, the first to confirm these significant associations. In our study, at hospital admission the prevalence of patients with hypomagnesemia was high, i.e., $10\%$. Patients with lower serum Mg levels reported some baseline characteristics that could increase the risk of mortality, including older age, significantly lower hemoglobin and albumin levels, and a higher comorbidity and severity of medical conditions. However, after adjusting for all these parameters, the association between lower serum Mg and in-hospital mortality remained statistically significant, indicating an independent role of Mg in poor prognosis among patients hospitalized for COVID-19. Previous studies have reported the prognostic importance of low serum Mg in COVID-19. Of interest, in a large North American population, it was reported that the infection risk for COVID-19 of the populations distributed in low-Mg areas was higher than those introducing a higher intake of dietary Mg [44]. Moreover, other studies reported an important role of Mg in the prognosis of patients affected by COVID-19. For example, Guerrero-Romero et al. analyzed 1064 patients with COVID-19, showing a significant association between serum magnesium-calcium ratio and mortality in severe forms of the disease [27]. Similarly, a retrospective cohort study analyzing 390 hospitalized patients with COVID-19 showed that reduced kidney function and lower serum Mg levels were associated with increased mortality in obese patients affected by COVID-19 [30]. In addition, Zeng et al. performed a retrospective study, analyzing 306 patients with COVID-19 for their whole blood levels of essential minerals, including Mg, and found that severe cases showed significantly lower levels of Mg than mild and moderate cases [29]. All these reported findings, associated with the results of our study, indicates that Mg might play a key role in maintaining proper immune, vascular and lung function. This strongly supports the hypothesis on which several studies have been based, that serum Mg status may influence susceptibility and response to SARS-CoV-2 infection [5]. As mentioned, our study confirmed the significant prognostic value in COVID-19 patients already reported in previous studies [26,27,28,29,30]. There are several mechanisms that may help to explain the link between a low Mg status and an increased risk of severe forms of the disease and mortality. COVID-19 is now considered a potential systematic disease due to the possibility not only of leading to acute respiratory distress syndrome requiring hyperoxic ventilation, but also of impacting other organs and systems, comprising the cardiovascular, hepatic, intestinal, renal and nervous systems [45]. Older adults are more susceptible to severe illness, ICU admission, and mortality from COVID-19 [46]. This trend has been confirmed since the beginning of the pandemic and it is particularly high in older adults with multimorbidity. Although the ultimate mechanisms of COVID-19 clinical manifestations and mortality are not completely clear, the cytokine storm seems to contribute significantly to the pathogenesis of the most severe manifestations of the disease [45]. Cytokine storm refers to the overproduction of soluble markers of inflammation that maintain an aberrant response of systemic inflammation. It seems that the collateral damage caused by the excessive production of inflammatory mediators, in an attempt to eliminate the pathogen, may be more damaging than the pathogen itself. Indeed, this exuberant inflammatory response may initially be appropriate to control the infection, but if uncontrolled and persistent, it can fuel the multi-organ dysfunction that may follow, increasing the risk of mortality. The cascade of inflammatory mediators during cytokine storm includes immunoactive molecules, e.g., interferon, chemokines, interleukins, TNF-alpha, and colony-stimulating factors [47]. There is extensive evidence in experimental investigations [48,49,50,51] as well as in observational studies in humans [52,53,54,55,56,57] confirming that a low Mg status is associated with a state of chronic inflammation with increased inflammatory markers, particularly IL-6, TNF-alpha, and the complex IL-33/ST2. Furthermore, some studies have reported anti-inflammatory actions of Mg supplementation and suppression of cytokine release [58,59]. A meta-analysis including eight RCTs reported a significant reduction in serum CRP after Mg supplementation, which was independent of Mg dosage or the length of follow-up [60]. A well-known action of Mg is its antagonistic effects on calcium channels [61,62]. Indeed, Mg is considered a natural calcium blocker, similar to those of chemical synthesis [63]. Interestingly, calcium channel blocking effects of Mg can lead to the suppression of NF-kB, IL-6, and CRP [59], which may limit systemic inflammation. Patients with severe forms of COVID-19 may need ICU admission. Remarkably, up to $60\%$ of critically ill patients in the ICU have some degree of Mg deficiency [64,65], which makes them more susceptible to potentially fatal effects, also associated with the consequent hypokalemia and hypocalcemia. Perhaps the lack of attention to paid Mg in COVID-19 may be due to the fact that it is not routinely measured in most databases and studies [66]. In addition, serum concentrations that are clinically available represent only $1\%$ of the total body Mg and do not accurately reflect the whole Mg status, being this ion predominantly intracellular [9]. Thus, the preceding Mg deficiency associated with conditions that favor a detrimental course of COVID-19, including age, diabetes, hypertension [6,9,15] and the Mg deficiency frequently observed in critically ill patients [64,65], can contribute to exacerbate the inflammatory response induced by SARS-CoV-2, which in turn can determine an increased Mg consumption, resulting in a further reduction in its intracellular levels, maintaining and propagating an uncontrolled inflammatory response. The evidence that COVID-19 pneumonia and multi-organ dysfunction has a vascular basis is robust [67,68]. The vascular endothelium is crucial in the maintenance of homeostasis and the control of fibrinolysis, inflammation, vasomotion, oxidative stress, vascular permeability and structure. All of these functions acting in concert regulate many of the defense mechanisms against external noxae, but they can also contribute to disease at different levels when the usual homeostatic functions are overwhelmed and turn against the host, as has been reported in COVID-19 [69]. There is also convincing evidence that Mg has antithrombotic effects [70], while low Mg concentrations have been associated with endothelial dysfunction [71,72]. A systematic review and meta-analysis of RCTs exploring the effects of Mg supplementation on vascular function showed that oral Mg supplementation significantly improved flow-mediated dilation in studies lasting longer than 6 months, including healthy people, older than 50 years, or with BMI greater than 25 kg/m2 [73]. Hence, it is possible that a chronic Mg deficiency, common in older adults [9], may generate a favorable environment for SARS-CoV-2 to promote thrombosis [66], a fundamental characteristic of COVID-19. It is widely known that Mg plays a role in the immune system, in both innate and acquired immune response [22], and this effect is probably of importance in COVID-19, often characterized by a decreased immune response. In fact, Mg is a cofactor for the synthesis of immunoglobulins (Ig), as well as for C3 convertase, antibody-dependent cytolysis, immune cell adherence, macrophage response to lymphokines, IgM lymphocyte binding, and T helper–B cell adherence [22,74]. Mg induces the reduction in proinflammatory molecule release, such as P, by controlling nuclear factor kappa-light-chain-enhancer of activated B cell NF-kB activity [75]. In addition, Mg affects acquired immunity by regulating lymphocyte development and proliferation [76]. There is evidence that experimental animals fed Mg-deficient diets showed altered polymorphonuclear cell number and function, as well as increased phagocytosis [77]. Mast cell proliferation and function are also modified by Mg deficiency [78]. In addition, Mg deficit has been involved in mast cell-dependent hepatic fibrosis and steatosis [79]. In addition, Fas-induced B cell apoptosis is a Mg-dependent process [80]. Other studies have confirmed that Mg-deficient experimental animals exhibited high rates of inflammation and reduced specific immune responses [51,81,82,83]. The increased inflammation associated with Mg deficiency in old age [9] has been linked to several mechanisms, including opening of calcium channels, activation of phagocytic cells, activation of N-methyl-d-aspartate (NMDA) receptor and of NF-kB [48]. The best evidence of the fundamental role of Mg as a second messenger in immunity was the discovery of a genetic disease, X-linked immunodeficiency with magnesium defect (XMEN), which can lead to severe and chronic Epstein–Barr virus infections and neoplasia [84,85,86]. Moreover, due to its vasodilatory, anti-inflammatory and anti-thrombotic effects, the role of Mg was recently explored in COVID-19 patients [26]. All these effects, in fact, might contribute to the reduction in the ventilation-perfusion mismatch, which is one of the most important reasons for hypoxemia in COVID-19, and to the improvement of oxygenation in these patients [87]. Additionally, because of the emerging role of mastocytes in driving diffuse alveolar injury in COVID-19 [88], it should be recalled that Mg may reduce mastocyte degranulation and, subsequently, prevent the release of inflammatory, pro-thrombotic and fibrotic mediators [89]. We believe that our study adds novel information to the current literature regarding Mg in COVID-19 debate. Low serum Mg levels not only were associated with a higher mortality risk during hospitalization and improved the accuracy of the prediction of this outcome among hospitalized patients, but also predicted a longer length of stay in hospital and a higher incidence of long COVID. To the best of our knowledge, our study is the first to show the impact of low serum Mg status for long COVID and, in particular, for PTSD. Since long COVID may affect more than $50\%$ of the patients previously hospitalized for COVID-19 [90], our study suggests the need to early identify and correct poor Mg status in order to help prevent this complication. Of importance, our study suggests that a peculiar association with psychiatric disorders may exist, confirming the previous literature in this direction [91]. The findings of our study must be interpreted within its limitations. First, a consistent part of the initially considered population was not included, since data regarding serum Mg were not always available. Therefore, a selection bias cannot be ruled out. Second, long COVID was detected using phone calls and not using other more validated tests, such as medical records. We have recently had the opportunity to review systematically and perform a meta-analysis of the incidence and frequency of signs and symptoms of long COVID according to the definition of the World Health Organization among 120,979 patients from 196 studies, as shown below [90]. In the [Supplementary Table S2 of the article, we report the characteristics of the 196 studies included, comprising the methods of follow-up assessing the symptomatology for the formulation of long COVID diagnosis. Among the 196 studies, 51 ($26\%$) used phone calls, 90 ($45.9\%$) used an outpatient visit, 18 ($9.2\%$) used an online electronic survey, 13 ($6.6\%$) used an in-person interview, 15 ($7.7\%$) used a mixed method, 5 ($2.6\%$) used other methods, and 4 ($2\%$) did not specify any method. Therefore, about one-quarter of published studies used phone calls, the method we used in our study. The use of these methods, which in almost half of the cases did not involve a classic outpatient visit, is understandable due to the conditions of the pandemic and the measures for containing its spread in accordance with WHO and with all the health systems worldwide in unique conditions. This was the way to be able to continue with the investigations. Third, even if we clearly asked if a sign or symptom could be independent of COVID-19 during the follow-up, we cannot exclude the possibility that the symptomatology could be attributed to other concurrent issues. It must also be considered that in clinical practice, only serum Mg assessment is available, which may not accurately reflect the total body Mg status, with Mg being a prevalently intracellular ion. Finally, even if highly prevalent in percentage ($10\%$), only 26 patients had hypomagnesemia at the baseline, making the research of potential risk factors associated with this condition very difficult. ## 5. Conclusions Our study indicates the importance of low serum Mg levels in the prognosis of COVID-19 complications, not only for predicting mortality and a longer length of stay in hospital, but also for the prediction of a higher presence of long COVID, even if this latter condition was ascertained using phone calls. Therefore, we warmly recommend that serum Mg be determined in all patients admitted for COVID-19. Further studies involving Mg supplementation are needed to determine if this intervention can indeed alter the course of the disease in a selected cohort. ## References 1. Faryadi Q.. The magnificent effect of magnesium to human health: A critical review. Int. J. Appl. (2012.0) 2 118-126 2. Caspi R., Altman T., Billington R., Dreher K., Foerster H., Fulcher C.A., Holland T.A., Keseler I.M., Kothari A., Kubo A.. The MetaCyc database of metabolic pathways and enzymes and the BioCyc collection of Pathway/Genome Databases. Nucleic Acids Res. (2014.0) 42 D459-D471. DOI: 10.1093/nar/gkt1103 3. Saris N.-E.L., Mervaala E., Karppanen H., Khawaja J.A., Lewenstam A.. Magnesium: An update on physiological, clinical and analytical aspects. Clin. Chim. Acta (2000.0) 294 1-26. DOI: 10.1016/S0009-8981(99)00258-2 4. Barbagallo M., Gupta R.K., Dominguez L.J., Resnick L.M.. Cellular ionic alterations with age: Relation to hypertension and diabetes. J. Am. Geriatr. Soc. (2000.0) 48 1111-1116. DOI: 10.1111/j.1532-5415.2000.tb04788.x 5. Dominguez L.J., Veronese N., Guerrero-Romero F., Barbagallo M.. Magnesium in infectious diseases in older people. Nutrients (2021.0) 13. DOI: 10.3390/nu13010180 6. Barbagallo M., Dominguez L.J., Galioto A., Ferlisi A., Cani C., Malfa L., Pineo A., Paolisso G.. Role of magnesium in insulin action, diabetes and cardio-metabolic syndrome X. Mol. Asp. Med. (2003.0) 24 39-52. DOI: 10.1016/S0098-2997(02)00090-0 7. Gröber U., Schmidt J., Kisters K.. Magnesium in prevention and therapy. Nutrients (2015.0) 7 8199-8226. DOI: 10.3390/nu7095388 8. Whang R.. Magnesium deficiency: Pathogenesis, prevalence, and clinical implications. Am. J. Med. (1987.0) 82 24-29. DOI: 10.1016/0002-9343(87)90129-X 9. Barbagallo M., Dominguez L.. Magnesium and aging. Curr. Pharm. Des. (2010.0) 16 832-839. DOI: 10.2174/138161210790883679 10. Barbagallo M., Veronese N., Dominguez L.J.. Magnesium in Type 2 Diabetes Mellitus, Obesity, and Metabolic Syndrome. Nutrients (2022.0) 14. DOI: 10.3390/nu14030714 11. Barbagallo M., Veronese N., Dominguez L.J.. Magnesium in aging, health and diseases. Nutrients (2021.0) 13. DOI: 10.3390/nu13020463 12. 12. World Health Organization Available online: www.who.int/ageing/events/world-report-2015-launch/en(accessed on 10 January 2023) 13. Zhang J.-J., Dong X., Liu G.-H., Gao Y.-D.. Risk and protective factors for COVID-19 morbidity, severity, and mortality. Clin. Rev. Allergy Immunol. (2022.0) 64 90-107. DOI: 10.1007/s12016-022-08921-5 14. Singh J., Kohli S., Devi A., Mahajan S.. Serum magnesium level in COPD patients attending a tertiary hospital—A cross sectional study. JK Sci. (2012.0) 14 185 15. Barbagallo M., Belvedere M., Dominguez L.J.. Magnesium homeostasis and aging. Magnes. Res. (2009.0) 22 235. DOI: 10.1684/mrh.2009.0187 16. Bhatt S.P., Khandelwal P., Nanda S., Stoltzfus J.C., Fioravanti G.T.. Serum magnesium is an independent predictor of frequent readmissions due to acute exacerbation of chronic obstructive pulmonary disease. Respir. Med. (2008.0) 102 999-1003. DOI: 10.1016/j.rmed.2008.02.010 17. Nasser R., Naffaa M.E., Mashiach T., Azzam Z.S., Braun E.. The association between serum magnesium levels and community-acquired pneumonia 30-day mortality. BMC Infect. Dis. (2018.0) 18. DOI: 10.1186/s12879-018-3627-2 18. Lee S.-H., Song W.-J., Park H.-W., Kim S.-H., Park H.-K., Kim S.-H., Kwon Y.E., Kim T.-B., Lee B.-J., Jee Y.-K.. Serum micronutrients levels and clinical features of elderly asthmatics. Allergy Asthma Respir. Dis. (2017.0) 5 223-227. DOI: 10.4168/aard.2017.5.4.223 19. Veronese N., Demurtas J., Pesolillo G., Celotto S., Barnini T., Calusi G., Caruso M.G., Notarnicola M., Reddavide R., Stubbs B.. Magnesium and health outcomes: An umbrella review of systematic reviews and meta-analyses of observational and intervention studies. Eur. J. Nutr. (2020.0) 59 263-272. DOI: 10.1007/s00394-019-01905-w 20. Emelyanov A., Fedoseev G., Barnes P.. Reduced intracellular magnesium concentrations in asthmatic patients. Eur. Respir. J. (1999.0) 13 38-40. DOI: 10.1183/09031936.99.13103899 21. Dominguez L.J., Barbagallo M., Lorenzo G.D., Drago A., Scola S., Morici G., Caruso C.. Bronchial reactivity and intracellular magnesium: A possible mechanism for the bronchodilating effects of magnesium in asthma. Clin. Sci. (1998.0) 95 137-142. DOI: 10.1042/CS19980082 22. Tam M., Gomez S., Gonzalez-Gross M., Marcos A.. Possible roles of magnesium on the immune system. Eur. J. Clin. Nutr. (2003.0) 57 1193-1197. DOI: 10.1038/sj.ejcn.1601689 23. Charoenngam N., Holick M.F.. Immunologic effects of vitamin D on human health and disease. Nutrients (2020.0) 12. DOI: 10.3390/nu12072097 24. Uwitonze A.M., Razzaque M.S.. Role of magnesium in vitamin D activation and function. J. Osteopath. Med. (2018.0) 118 181-189. DOI: 10.7556/jaoa.2018.037 25. Zittermann A.. Magnesium deficit-overlooked cause of low vitamin D status?. BMC Med. (2013.0) 11. DOI: 10.1186/1741-7015-11-229 26. Trapani V., Rosanoff A., Baniasadi S., Barbagallo M., Castiglioni S., Guerrero-Romero F., Iotti S., Mazur A., Micke O., Pourdowlat G.. The relevance of magnesium homeostasis in COVID-19. Eur. J. Nutr. (2021.0) 61 625-636. DOI: 10.1007/s00394-021-02704-y 27. Guerrero-Romero F., Mercado M., Rodriguez-Moran M., Ramírez-Renteria C., Martínez-Aguilar G., Marrero-Rodríguez D., Ferreira-Hermosillo A., Simental-Mendía L.E., Remba-Shapiro I., Gamboa-Gómez C.I.. Magnesium-to-Calcium Ratio and Mortality from COVID-19. Nutrients (2022.0) 14. DOI: 10.3390/nu14091686 28. Tan C.W., Ho L.P., Kalimuddin S., Cherng B.P.Z., Teh Y.E., Thien S.Y., Wong H.M., Tern P.J.W., Chandran M., Chay J.W.M.. Cohort study to evaluate the effect of vitamin D, magnesium, and vitamin B12 in combination on progression to severe outcomes in older patients with coronavirus (COVID-19). Nutrition (2020.0) 79 111017. DOI: 10.1016/j.nut.2020.111017 29. Zeng H.L., Yang Q., Yuan P., Wang X., Cheng L.. Associations of essential and toxic metals/metalloids in whole blood with both disease severity and mortality in patients with COVID-19. FASEB J. (2021.0) 35 e21392. DOI: 10.1096/fj.202002346RR 30. Pulido Perez P., Póndigo de Los Angeles J.A., Perez Peralta A., Ramirez Mojica E., Torres Rasgado E., Hernandez-Hernandez M.E., Romero J.R.. Reduction in Serum Magnesium Levels and Renal Function Are Associated with Increased Mortality in Obese COVID-19 Patients. Nutrients (2022.0) 14. DOI: 10.3390/nu14194054 31. Veronese N., Bonica R., Cotugno S., Tulone O., Camporeale M., Smith L., Trott M., Bruyere O., Mirarchi L., Rizzo G.. Interventions for improving long COVID-19 symptomatology: A systematic review. Viruses (2022.0) 14. DOI: 10.3390/v14091863 32. Barbagallo M., Citarrella R., Dominguez L., Giannitrapani L., Licata A., Mansueto P., Soresi M., Veronese N.. COMEPA (COVID-19 Medicina Policlinico Palermo): A study in hospitalized patients. Geriatr. Care (2021.0) 7 9895 33. Wynants L., Van Calster B., Collins G.S., Riley R.D., Heinze G., Schuit E., Bonten M.M., Dahly D.L., Damen J.A., Debray T.P.. Prediction models for diagnosis and prognosis of covid-19: Systematic review and critical appraisal. Bmj (2020.0) 369 m1328. DOI: 10.1136/bmj.m1328 34. 34. World Health Organization A Clinical Case Definition of Post COVID-19 Condition by a Delphi Consensus, 6 October 2021World Health OrganizationGeneva, Switzerland2021. A Clinical Case Definition of Post COVID-19 Condition by a Delphi Consensus, 6 October 2021 (2021.0) 35. Chen C., Haupert S.R., Zimmermann L., Shi X., Fritsche L.G., Mukherjee B.. Global Prevalence of Post COVID-19 Condition or Long COVID: A Meta-Analysis and Systematic Review. J. Infect. Dis. (2022.0) 226 1593-1607. DOI: 10.1093/infdis/jiac136 36. Stavem K., Ghanima W., Olsen M.K., Gilboe H.M., Einvik G.. Prevalence and determinants of fatigue after COVID-19 in non-hospitalized subjects: A population-based study. Int. J. Environ. Res. Public Health (2021.0) 18. DOI: 10.3390/ijerph18042030 37. Natarajan A., Shetty A., Delanerolle G., Zeng Y., Zhang Y., Raymont V., Rathod S., Halabi S., Elliot K., Phiri P.. A systematic review and meta-analysis of Long COVID symptoms. Medrxiv Prepr. Serv. Health Sci. (2022.0). DOI: 10.1101/2022.03.08.22272091 38. Han Q., Zheng B., Daines L., Sheikh A.. Long-Term sequelae of COVID-19: A systematic review and meta-analysis of one-year follow-up studies on post-COVID symptoms. Pathogens (2022.0) 11. DOI: 10.3390/pathogens11020269 39. Blevins C.A., Weathers F.W., Davis M.T., Witte T.K., Domino J.L.. The posttraumatic stress disorder checklist for DSM-5 (PCL-5): Development and initial psychometric evaluation. J. Trauma. Stress (2015.0) 28 489-498. DOI: 10.1002/jts.22059 40. Bjelland I., Dahl A.A., Haug T.T., Neckelmann D.. The validity of the Hospital Anxiety and Depression Scale: An updated literature review. J. Psychosom. Res. (2002.0) 52 69-77. DOI: 10.1016/S0022-3999(01)00296-3 41. Covelli H.D., Nessan V.J., Tuttle W.. Oxygen derived variables in acute respiratory failure. Crit. Care Med. (1983.0) 11 646-649. DOI: 10.1097/00003246-198308000-00012 42. Linn B.S., Linn M.W., Gurel L.. Cumulative illness rating scale. J. Am. Geriatr. Soc. (1968.0) 16 622-626. DOI: 10.1111/j.1532-5415.1968.tb02103.x 43. Fluss R., Faraggi D., Reiser B.. Estimation of the Youden Index and its associated cutoff point. Biom. J. (2005.0) 47 458-472. DOI: 10.1002/bimj.200410135 44. Tian J., Tang L., Liu X., Li Y., Chen J., Huang W., Liu M.. Populations in Low-Magnesium Areas Were Associated with Higher Risk of Infection in COVID-19’s Early Transmission: A Nationwide Retrospective Cohort Study in the United States. Nutrients (2022.0) 14. DOI: 10.3390/nu14040909 45. Fajgenbaum D.C., June C.H.. Cytokine storm. N. Engl. J. Med. (2020.0) 383 2255-2273. DOI: 10.1056/NEJMra2026131 46. Perrotta F., Corbi G., Mazzeo G., Boccia M., Aronne L., D’Agnano V., Komici K., Mazzarella G., Parrella R., Bianco A.. COVID-19 and the elderly: Insights into pathogenesis and clinical decision-making. Aging Clin. Exp. Res. (2020.0) 32 1599-1608. DOI: 10.1007/s40520-020-01631-y 47. Kalinina O., Golovkin A., Zaikova E., Aquino A., Bezrukikh V., Melnik O., Vasilieva E., Karonova T., Kudryavtsev I., Shlyakhto E.. Cytokine Storm Signature in Patients with Moderate and Severe COVID-19. Int. J. Mol. Sci. (2022.0) 23. DOI: 10.3390/ijms23168879 48. Mazur A., Maier J.A., Rock E., Gueux E., Nowacki W., Rayssiguier Y.. Magnesium and the inflammatory response: Potential physiopathological implications. Arch. Biochem. Biophys. (2007.0) 458 48-56. DOI: 10.1016/j.abb.2006.03.031 49. Kramer J.H., Mak I.T., Phillips T.M., Weglicki W.B.. Dietary magnesium intake influences circulating pro-inflammatory neuropeptide levels and loss of myocardial tolerance to postischemic stress. Exp. Biol. Med. (2003.0) 228 665-673. DOI: 10.1177/153537020322800604 50. Stankovic M.S., Janjetovic K., Velimirovic M., Milenkovic M., Stojkovic T., Puskas N., Zaletel I., De Luka S.R., Jankovic S., Stefanovic S.. Effects of IL-33/ST2 pathway in acute inflammation on tissue damage, antioxidative parameters, magnesium concentration and cytokines profile. Exp. Mol. Pathol. (2016.0) 101 31-37. DOI: 10.1016/j.yexmp.2016.05.012 51. Malpuech-Brugère C., Nowacki W., Daveau M., Gueux E., Linard C., Rock E., Lebreton J.-P., Mazur A., Rayssiguier Y.. Inflammatory response following acute magnesium deficiency in the rat. Biochim. Et Biophys. Acta (BBA)-Mol. Basis Dis. (2000.0) 1501 91-98. DOI: 10.1016/S0925-4439(00)00018-1 52. King D.E., Mainous III A.G., Geesey M.E., Woolson R.F.. Dietary magnesium and C-reactive protein levels. J. Am. Coll. Nutr. (2005.0) 24 166-171. DOI: 10.1080/07315724.2005.10719461 53. Guerrero-Romero F., Bermudez-Peña C., Rodríguez-Morán M.. Severe hypomagnesemia and low-grade inflammation in metabolic syndrome. Magnes. Res. (2011.0) 24 45-53. DOI: 10.1684/mrh.2011.0281 54. Song Y., Li T.Y., Van Dam R.M., Manson J.E., Hu F.B.. Magnesium intake and plasma concentrations of markers of systemic inflammation and endothelial dysfunction in women. Am. J. Clin. Nutr. (2007.0) 85 1068-1074. DOI: 10.1093/ajcn/85.4.1068 55. Song Y., Ridker P.M., Manson J.E., Cook N.R., Buring J.E., Liu S.. Magnesium intake, C-reactive protein, and the prevalence of metabolic syndrome in middle-aged and older US women. Diabetes Care (2005.0) 28 1438-1444. DOI: 10.2337/diacare.28.6.1438 56. Mazidi M., Kengne A.P., Mikhailidis D.P., Cicero A.F., Banach M.. Effects of selected dietary constituents on high-sensitivity C-reactive protein levels in US adults. Ann. Med. (2018.0) 50 1-6. DOI: 10.1080/07853890.2017.1325967 57. Konstari S., Sares-Jäske L., Heliövaara M., Rissanen H., Knekt P., Arokoski J., Sundvall J., Karppinen J.. Dietary magnesium intake, serum high sensitivity C-reactive protein and the risk of incident knee osteoarthritis leading to hospitalization—A cohort study of 4,953 Finns. PLoS ONE (2019.0) 14. DOI: 10.1371/journal.pone.0214064 58. Su N.-Y., Peng T.-C., Tsai P.-S., Huang C.-J.. Phosphoinositide 3-kinase/Akt pathway is involved in mediating the anti-inflammation effects of magnesium sulfate. J. Surg. Res. (2013.0) 185 726-732. DOI: 10.1016/j.jss.2013.06.030 59. Lin C., Tsai P., Hung Y., Huang C.. L-type calcium channels are involved in mediating the anti-inflammatory effects of magnesium sulphate. Br. J. Anaesth. (2010.0) 104 44-51. DOI: 10.1093/bja/aep336 60. Mazidi M., Rezaie P., Banach M.. Effect of magnesium supplements on serum C-reactive protein: A systematic review and meta-analysis. Arch. Med. Sci. (2018.0) 14 707-716. DOI: 10.5114/aoms.2018.75719 61. Altura B.M., Altura B.T., Gebrewold A., Ising H., Gunther T.. Magnesium deficiency and hypertension: Correlation between magnesium-deficient diets and microcirculatory changes in situ. Science (1984.0) 223 1315-1317. DOI: 10.1126/science.6701524 62. Iseri L.T., French J.H.. Magnesium: Natures physiologic calcium blocker. Am. Heart J. (1984.0) 108 188-193. DOI: 10.1016/0002-8703(84)90572-6 63. Louvet L., Büchel J., Steppan S., Passlick-Deetjen J., Massy Z.A.. Magnesium prevents phosphate-induced calcification in human aortic vascular smooth muscle cells. Nephrol. Dial. Transplant. (2013.0) 28 869-878. DOI: 10.1093/ndt/gfs520 64. Escuela M.P., Guerra M., Añón J.M., Martínez-Vizcaíno V., Zapatero M.D., García-Jalón A., Celaya S.. Total and ionized serum magnesium in critically ill patients. Intensive Care Med. (2005.0) 31 151-156. DOI: 10.1007/s00134-004-2508-x 65. Hansen B.-A., Bruserud Ø.. Hypomagnesemia in critically ill patients. J. Intensive Care (2018.0) 6 21. DOI: 10.1186/s40560-018-0291-y 66. Iotti S., Wolf F., Mazur A., Maier J.. The COVID-19 pandemic: Is there a role for magnesium? Hypotheses and perspectives. Magnes. Res. (2020.0) 33 21-27. DOI: 10.1684/mrh.2020.0465 67. Ackermann M., Verleden S.E., Kuehnel M., Haverich A., Welte T., Laenger F., Vanstapel A., Werlein C., Stark H., Tzankov A.. Pulmonary vascular endothelialitis, thrombosis, and angiogenesis in COVID-19. N. Engl. J. Med. (2020.0) 383 120-128. DOI: 10.1056/NEJMoa2015432 68. Libby P., Lüscher T.. COVID-19 is, in the end, an endothelial disease. Eur. Heart J. (2020.0) 41 3038-3044. DOI: 10.1093/eurheartj/ehaa623 69. Evans P.C., Rainger G.E., Mason J.C., Guzik T.J., Osto E., Stamataki Z., Neil D., Hoefer I.E., Fragiadaki M., Waltenberger J.. Endothelial dysfunction in COVID-19: A position paper of the ESC Working Group for Atherosclerosis and Vascular Biology, and the ESC Council of Basic Cardiovascular Science. Cardiovasc. Res. (2020.0) 116 2177-2184. DOI: 10.1093/cvr/cvaa230 70. Maier J.A., Malpuech-Brugère C., Zimowska W., Rayssiguier Y., Mazur A.. Low magnesium promotes endothelial cell dysfunction: Implications for atherosclerosis, inflammation and thrombosis. Biochim. Et Biophys. Acta (BBA)-Mol. Basis Dis. (2004.0) 1689 13-21. DOI: 10.1016/j.bbadis.2004.01.002 71. Barbagallo M., Dominguez L.J., Galioto A., Pineo A., Belvedere M.. Oral magnesium supplementation improves vascular function in elderly diabetic patients. Magnes. Res. (2010.0) 23 131-137. PMID: 20736142 72. Shechter M., Sharir M., Labrador M.J.P., Forrester J., Silver B., Bairey Merz C.N.. Oral magnesium therapy improves endothelial function in patients with coronary artery disease. Circulation (2000.0) 102 2353-2358. DOI: 10.1161/01.CIR.102.19.2353 73. Marques B.C.A.A., Klein M.R.S.T., da Cunha M.R., de Souza Mattos S., de Paula Nogueira L., de Paula T., Corrêa F.M., Oigman W., Neves M.F.. Effects of oral magnesium supplementation on vascular function: A systematic review and meta-analysis of randomized controlled trials. High Blood Press. Cardiovasc. Prev. (2020.0) 27 19-28. DOI: 10.1007/s40292-019-00355-z 74. Galland L.. Magnesium and immune function: An overview. Magnesium (1988.0) 7 290-299. PMID: 3075245 75. Bernstein H., Sugimoto J., Suzuki-Kakisaka H., Romani A.. 823: Magnesium decreases inflammatory cytokine production: A novel innate immunomodulatory mechanism. Am. J. Obstet. Gynecol. (2012.0) 206 S361. DOI: 10.1016/j.ajog.2011.10.841 76. Feske S., Skolnik E.Y., Prakriya M.. Ion channels and transporters in lymphocyte function and immunity. Nat. Rev. Immunol. (2012.0) 12 532-547. DOI: 10.1038/nri3233 77. Bussière F.I., Tridon A., Zimowska W., Mazur A., Rayssiguier Y.. Increase in complement component C3 is an early response to experimental magnesium deficiency in rats. Life Sci. (2003.0) 73 499-507. DOI: 10.1016/S0024-3205(03)00291-1 78. Kraeuter S.L., Schwartz R.. Blood and mast cell histamine levels in magnesium-deficient rats. J. Nutr. (1980.0) 110 851-858. DOI: 10.1093/jn/110.5.851 79. TAkEMOTO S., Yamamoto A., Tomonaga S., Funaba M., Matsui T.. Magnesium deficiency induces the emergence of mast cells in the liver of rats. J. Nutr. Sci. Vitaminol. (2013.0) 59 560-563. DOI: 10.3177/jnsv.59.560 80. Chien M.M., Zahradka K.E., Newell M.K., Freed J.H.. Fas-induced B cell apoptosis requires an increase in free cytosolic magnesium as an early event. J. Biol. Chem. (1999.0) 274 7059-7066. DOI: 10.1074/jbc.274.11.7059 81. Bussière F., Gueux E., Rock E., Mazur A., Rayssiguier Y.. Protective effect of calcium deficiency on the inflammatory response in magnesium-deficient rats. Eur. J. Nutr. (2002.0) 41 197-202. DOI: 10.1007/s00394-002-0376-0 82. Petrault I., Zimowska W., Mathieu J., Bayle D., Rock E., Favier A., Rayssiguier Y., Mazur A.. Changes in gene expression in rat thymocytes identified by cDNA array support the occurrence of oxidative stress in early magnesium deficiency. Biochim. Et Biophys. Acta (BBA)-Mol. Basis Dis. (2002.0) 1586 92-98. DOI: 10.1016/S0925-4439(01)00089-8 83. Zimowska W., Girardeau J., Kuryszko J., Bayle D., Rayssiguier Y., Mazur A.. Morphological and immune response alterations in the intestinal mucosa of the mouse after short periods on a low-magnesium diet. Br. J. Nutr. (2002.0) 88 515-522. DOI: 10.1079/BJN2002696 84. Li F.-Y., Chaigne-Delalande B., Kanellopoulou C., Davis J.C., Matthews H.F., Douek D.C., Cohen J.I., Uzel G., Su H.C., Lenardo M.J.. Second messenger role for Mg. Nature (2011.0) 475 471-476. DOI: 10.1038/nature10246 85. Chaigne-Delalande B., Li F.-Y., O’Connor G.M., Lukacs M.J., Jiang P., Zheng L., Shatzer A., Biancalana M., Pittaluga S., Matthews H.F.. Mg. Science (2013.0) 341 186-191. DOI: 10.1126/science.1240094 86. Li F.-Y., Chaigne-Delalande B., Su H., Uzel G., Matthews H., Lenardo M.J.. XMEN disease: A new primary immunodeficiency affecting Mg. Blood J. Am. Soc. Hematol. (2014.0) 123 2148-2152. DOI: 10.1182/blood-2013-11-538686 87. Mason R.J.. Pathogenesis of COVID-19 from a cell biology perspective. Eur. Respiratory Soc. (2020.0) 55 2000607. DOI: 10.1183/13993003.00607-2020 88. Theoharides T.C.. COVID-19, pulmonary mast cells, cytokine storms, and beneficial actions of luteolin. Biofactors (2020.0) 46 306-308. DOI: 10.1002/biof.1633 89. Veronese N., Pizzol D., Smith L., Dominguez L.J., Barbagallo M.. Effect of Magnesium Supplementation on Inflammatory Parameters: A Meta-Analysis of Randomized Controlled Trials. Nutrients (2022.0) 14. DOI: 10.3390/nu14030679 90. Di Gennaro F., Belati A., Tulone O., Diella L., Fiore Bavaro D., Bonica R., Genna V., Smith L., Trott M., Bruyere O.. Incidence of long COVID-19 in people with previous SARS-Cov2 infection: A systematic review and meta-analysis of 120,970 patients. Intern. Emerg. Med. (2022.0) 1-9. DOI: 10.1007/s11739-022-03164-w 91. Gushcha S., Oleshko A.Y., Bakholdina E., Badiuk N., Zabolotna I.. Correction of disturbances of functional activity of the central nervous system in rats with the post-traumatic stress disorder model using remedy with a high magnesium content. (2021.0) Volume 1 12-19
--- title: Anlotinib Alleviates Renal Fibrosis via Inhibition of the ERK and AKT Signaling Pathways authors: - Donglin Sun - Jing Guo - Weifei Liang - Yangxiao Chen - Xiangqiu Chen - Li Wang journal: Oxidative Medicine and Cellular Longevity year: 2023 pmcid: PMC9966823 doi: 10.1155/2023/1686804 license: CC BY 4.0 --- # Anlotinib Alleviates Renal Fibrosis via Inhibition of the ERK and AKT Signaling Pathways ## Abstract ### Purpose We examined whether anlotinib can attenuate folic acid-induced and unilateral ureteral obstruction-induced renal fibrosis and explored the underlying antifibrotic mechanism. ### Materials and Methods We have evaluated the effects of anlotinib on folic acid-induced and unilateral ureteral obstruction-induced renal fibrosis in mice through in vivo experiments of unilateral ureteral obstruction or folic acid-induced interstitial fibrosis and in vitro models of transforming growth factor-β1 induced HK-2 human renal proximal tubule cells. Serum renal function parameters and inflammatory cytokine levels were measured, and histological changes of renal injury and fibrosis were analyzed by HE staining and immunohistochemistry. Immunohistochemistry and Western blotting were used to determine the mechanism of action of anlotinib in ameliorating renal fibrosis. ### Results Anlotinib improved proteinuria and reduced renal impairment in folic acid-induced mouse models of renal fibrosis. Anlotinib reduced tubular injury, deposition of tubular extracellular matrix, and expression of alpha-smooth muscle actin, transforming growth factor-β1, and cytosolic inflammatory factors compared with controls. ### Conclusions Anlotinib ameliorated renal function, improved extracellular matrix deposition, reduced protein levels of epithelial-mesenchymal transition markers, and decreased cellular inflammatory factors. Anlotinib reduced renal injury and fibrosis by inhibiting the transforming growth factor-β1 signaling pathway through AKT and ERK channels. ## 1. Introduction Chronic kidney disease (CKD), a global public health concern, often ends in serious conditions like renal failure and cardiovascular disease, posing a serious risk to human health and leading to premature death in $82\%$ of patients [1]. As of today, the only options available to patients with end-stage renal disease are kidney transplantation and dialysis, imposing a huge economic burden on patients and society alike [2]. Therefore, it is crucial to find effective treatments to prolong the life span of CKD patients and reduce corresponding medical costs. Progressive tubulointerstitial fibrosis is a common pathological change in chronic kidney disease [3] that eventually leads to renal failure. As the main component (about $90\%$) of the kidney is tubulointerstitium, the degree of tubulointerstitial fibrosis is a keen indicator of CKD prognosis. Given this, effective prevention or handling of tubulointerstitial fibrosis will contribute to improving the prognosis of CKD. Nevertheless, there are currently no effective drugs for the treatment of tubulointerstitial fibrosis. The identification of new therapeutic agents for patients with chronic kidney disease is therefore of utmost importance. Renal tubular interstitial fibrosis is characterized by inflammation, extracellular matrix (ECM) deposition, loss of tubular cells, accumulation of fibroblast activation, and reduction of peritubular microvasculature [4, 5]. Pathological deposition of ECM, primarily collagen I and collagen IV, is a hallmark of renal fibrosis [6]. Epithelial-mesenchymal transition (EMT) is a major cause of renal interstitial fibrosis [7, 8] and manifests as the acquisition of a mesenchymal phenotype and myofibroblast function by renal tubular epithelial cells [9]. EMT reduces the expression of cell adhesion proteins like E-cadherin by renal epithelial cells while inducing fibroblasts to express wave proteins and alpha-smooth muscle actin (α-SMA) [7]. Renal fibrosis essentially evolves from inflammatory cell infiltration, and activated inflammatory cells not only produce several proinflammatory cytokines and chemokines such as chemokine ligand 2 (CCL-2)/monocyte chemotactic protein-1 (MCP-1) but also produce profibrotic cytokines like TGF-β1, which are regarded as major mediators in renal fibrosis pathogenesis [10]. Therefore, inhibiting the inflammatory response and reducing the fibrotic response and ECM deposition would be potential targets for the treatment of renal fibrosis [4]. Anlotinib hydrochloride (AL3818) is a novel multitargeted tyrosine kinase inhibitor (TKI) that has been reported to be effective in the treatment of cancers like non-small-cell lung cancer, osteosarcomas, and endometriomas [11, 12]. More importantly, it has been found to ameliorate the symptoms of mouse lung fibrosis by interfering with the TGF-β signaling pathway [13]. AL3818's target receptors include vascular endothelial growth factor receptors (VEGFR) 1 to 3, epidermal growth factor receptor (EGFR), fibroblast growth factor receptors (FGFR) 1 to 4, platelet-derived growth factor receptors (PDGFR) α and β, and stem cell factor receptor [14, 15]. TGF-β1 and fibroblast activation is essential in the pathogenesis of renal fibrosis, so we speculated that anlotinib could improve and prevent renal fibrosis, and no studies have been conducted on the effects of anlotinib on renal fibrosis and the associated mechanism of action. Our study examined anlotinib's antifibrotic effects on fibroblast transdifferentiation induced by TGF-1, folic acid (FA), and unilateral ureteral obstruction (UUO) in healthy subjects. Anlotinib therapy was found to assist in reducing both in vitro and in vivo markers of fibrosis, and its antifibrotic effect is mediated by inhibition of AKT, ERK signaling channels, and TGF-β1 transduction. This study lays the groundwork for further exploring the therapeutic potential of anlotinib intervention in renal fibrosis. ## 2.1. Animals Male C57BL/6 mice of 8 weeks were available at Rise Mice Biotechnology Co., Ltd. (Zhaoqing, China). Body weight was approximately 20-25 g. The mice were free to walk and drink under stable conditions of 25°C and 12 hours of light/dark. Experimental animals were cared for and used by the Guide for the Care and Use of Laboratory Animals, which has been approved by Southern Medical University Affiliated Longhua People's Hospital's Institutional Biomedical Research Ethics Committee. ## 2.2. FA Mouse Models The FA nephropathy mouse models were established and randomly divided into 4 groups. In the control group, DMSO was injected intraperitoneally; in the control+anlotinib group, DMSO+1 mg/kg anlotinib was injected intraperitoneally; in the FA group, 25 μg/g folic acid was injected intraperitoneally; in the anlotinib treatment group, 25 μg/g folic acid and 1 mg/kg anlotinib were injected intraperitoneally. DMSO or anlotinib was administered consistently daily for 34 days, with FA administered on day 1 only. After 34 days, 24 h urine was collected and the mice were euthanized to obtain serum. ## 2.3. UUO Mouse Models Prior to surgery, mice were anesthetized with $1.25\%$ 2,2,2-tribromoethanol through intraperitoneal injection. Then, UUO was established by double ligation of the left ureter with 4-0 silk after the abdominal incision. The ureter of the sham-operated mice was exposed without ligation. Anlotinib was injected intraperitoneally at 1 mg/kg in the anlotinib-treated group daily for 7 days. After that, mice were executed, and urine, serum, and kidney tissues were collected for further analysis. ## 2.4. Cell Culture ATCC (Manassas, VA) provided HK-2 human renal proximal tubule cells. An HK2 cell line was grown in a keratinocyte serum-free medium (Invitrogen) and tested at generations 10-13. Following pretreatment with/without anlotinib (2 μM) for 4 h, cells were incubated with TGF-β1 for 48 h. Thereafter, cell lysates and supernatants were prepared as described below. ## 2.5. CCK8 Assay The Cell Counting Kit 8 (Beyotime) was used in accordance with provided protocols. Briefly, 2000 cells were added into 96-well plates per well. 10 μL CCK8 solution was added per cell. The cells were then cultured for 2 h at 37°C while protected from light. Data were collected by reading the optical density (OD) at 450 nm at Thermo Scientific Microplate Reader. ## 2.6. RNA Extraction and Real-Time RT-PCR Kidney tissue was homogenized using TRIzol reagent (Life Technologies, 15596-026), and total RNA was extracted as directed by the manufacturer. The RNA was then converted to cDNA according to the instructions of the PrimeScript™ RT reagent Kit with gDNA Eraser (TaKaRa, RR047A). Real-time RT-PCR was performed by the Q7 RT-PCR detection system (Life Technologies) using SYBR® premix Ex Taq (TaKaRa, RR420A). Relative quantification (RQ) was derived from the cycling threshold (Ct) using the equation RQ = 2−ΔΔCt. The results were normalized to the expression of β-actin mRNA. The primers used are listed in Additional file 1. ## 2.7. Western Blotting Kidney tissues were washed with precooled PBS, homogenized by a homogenizer, and then lysed in RIPA (Beyotime, P0013F), protease inhibitor mixture (Roche Diagnostics, 4693116001), phosphatase inhibitor mixture (Roche Diagnostics, 4906845001), and PMSF (Beyotime, ST506), then incubated on ice for 30 min, and the supernatant was collected by centrifugation at 13,000g for 15 min and stored at -80°C. Total protein concentration was estimated using the BCA Protein Assay Kit (Thermo Fisher Scientific, 23225). Protein samples (10-20 μg) were diluted with 5x loading buffer, bathed in metal for 10-15 min, separated by $12.5\%$ SDS polyacrylamide gel electrophoresis (SDS-PAGE), and then transferred to PVDF membranes (Millipore) for detection of specific antibody probes. The immunoreactive proteins were visualized by an enhanced chemiluminescence detection system (Millipore). The following primary antibodies were used: anti-collagen type I (anti-collagen I; rabbit; Abcam, ab260043), anti-collagen type IV (anti-collagen IV; rabbit; Abcam, ab6586), anti-TGF-β1 (rabbit; Abcam, ab215715), anti-ERK (rabbit; Abcam, ab184699), anti-p-ERK (rabbit; Abcam, ab201015), anti-AKT (rabbit; Abcam, ab8805), anti-p-AKT (rabbit; Abcam, ab38449), anti-α smooth muscle actin (anti-α-SMA; rabbit; Abcam, ab5694), and anti-GAPDH (KANGCHEN, KC-5G4). The secondary antibodies were used: peroxidase-conjugated goat anti-mouse IgG (H+L) (33201ES60, Yeasen) and peroxidase-conjugated rabbit anti-goat IgG (H+L) (33701ES60, Yeasen). ## 2.8. Histological and Immunohistochemical Analyses The kidneys were fixed in $4\%$ paraformaldehyde (Beyotime, P0099-500 mL) and embedded in paraffin. Paraffin-embedded mouse kidney sections (3 μm thickness) were prepared following a routine procedure, where the sections were stained with hematoxylin and eosin (H&E) and Masson trichrome reagent (Service, G1006) separately. For immunohistochemical staining, the sections were deparaffinized and rehydrated, followed by antigen retrieval and blocking. Then, the tissue sections were incubated at 4°C overnight with diluted primary antibodies: anti-α-SMA (Abcam, ab5694), anti-F$\frac{4}{80}$ (Abcam, ab6640), and anti-CD163 (Abcam, ab182422). The secondary antibodies were applied, and a diaminobenzidine (DAB) solution was used as a chromogen. In addition, the sections were counterstained with hematoxylin (Sigma) to identify nuclei. The images were photographed at 200x and 400x by a general optical microscope (Carl Zeiss) and analyzed using Image-Pro Plus 6.0 software (Media Cybernetics, Inc.). ## 2.9. Statistical Analysis Each experiment was repeated at least three times. Values are expressed as mean ± SD, and statistical analyses were performed using unpaired Student's t-test (two groups) and Kruskal-Wallis one-way analysis of variance (more than two groups). P values below 0.05 indicate a significant difference between groups. ## 3.1. Anlotinib Ameliorates Renal Function in FA-Induced Mouse Models We examined the effect of anlotinib on renal function and fibrosis phenotype in vivo; we determined the dose to be 1 mg/kg for mice. Then, the cytotoxicity of anlotinib was evaluated by the CCK8 assay in vivo. Combined with cytotoxicity and mRNA levels of fibrosis-related indicators, the optimal concentration of anlotinib in vitro is 2 μM (Supplementary Figure S1). In order to evaluate anlotinib's effects on urinary albumin, we established a control group and a FA group (Figure 1(a)) and measured 24-hour urinary albumin and serum creatinine levels at the time of execution. In the control group, as shown in Figure 1(b), there was no significant difference in urinary albumin and creatinine in mice with or without anlotinib; however, in the FA group, urinary albumin and creatinine were significantly reduced after anlotinib treatment (P value < 0.05). Meanwhile, we established UUO mouse models (Figure 1(c)) to investigate whether anlotinib could improve the renal function of mice. In the UUO group, anlotinib was found to exert no significant effect on the renal function of mouse models, as shown in Figure 1(d). Considering only the left ureter was ligated, the reason may be that the healthy contralateral kidney compensated and maintained the normal function of the whole kidney. According to these findings, anlotinib improves the recovery of renal function after toxic injury. ## 3.2. Anlotinib Inhibits Renal ECM Gene Expression and Reduces Tubulointerstitial Fibrosis Collagen (collagen I and collagen IV) is an essential part of the renal ECM, α-SMA is a key secretory protein of fibroblasts, and TGF-β1 is an underlying factor in renal fibrosis. In order to determine whether anlotinib regulates renal fibrosis, we examined the impact of anlotinib on type I and IV collagen expression, α-SMA, and TGF-β1. As shown in Figures 2(a) and 2(b), in the UUO mouse models, type I and type IV collagen and α-SMA and TGF-β1 expressions were observed in the kidneys of UUO-treated animals compared with the control group, and it was found to be significantly reduced in the anlotinib-treated group, hence its significant inhibition by anlotinib. In addition, we examined the effects of anlotinib on the expression of interstitial collagen fibers with HE and Masson staining and analyzed α-SMA with immunohistochemical staining and the expression of type I and IV collagen and α-SMA with Western blotting, whose results showed that the expression of type I and IV collagen and α-SMA protein in the anlotinib-treated group was significantly reduced compared with the UUO group ($P \leq 0.05$, Figure 2(b)). In addition, collagen deposition was significantly increased in the interstitial region of the kidney in UUO mice compared with the anlotinib-treated group (Figure 2(c)). Based on these data, anlotinib inhibits ECM production and inhibits renal fibrosis in the UUO mouse models. ## 3.3. Anlotinib Reduces Inflammatory Responses in Mouse Models of UUO To investigate the effect of anlotinib on renal inflammation in UUO mice, we examined three proinflammatory cytokines CCL-2, CCL-5, and IL-6 in renal tissues along with two macrophage markers, F$\frac{4}{80}$ and CD163. As shown in Figure 3(a), CCL-2, CCL-5, and IL-6 expressions in the UUO group were significantly increased compared to the anlotinib-treated group, and mRNA expression of the three proinflammatory cytokines was attenuated by anlotinib. To further determine whether macrophage infiltration contributed to UUO-induced renal injury, F$\frac{4}{80}$ and CD163 expressions were assessed by immunohistochemical staining (Figure 3(b)), and anlotinib significantly inhibited the expression of F$\frac{4}{80}$ ($P \leq 0.05$) and CD163 ($P \leq 0.01$) (Figure 3(c)). These results indicate that anlotinib may reduce inflammation. ## 3.4. Anlotinib Suppresses TGF-β1-Stimulated Collagen and α-SMA Expression in Human Proximal Renal Tubular Cells TGF-β1, a key player in renal fibrosis [16], stimulates the upregulation of collagen and α-SMA expressions in renal tubular cells. To investigate the effect of anlotinib on ECM production after TGF-β1 exposure, we used real-time PCR and Western blotting to detect the expression of type I and type IV collagen and α-SMA. As shown in Figures 4(a)–4(c), TGF-β1 induced the mRNA expression of type I and type IV collagen and α-SMA in human proximal renal tubular cells, which decreased significantly in the presence of anlotinib ($P \leq 0.05$, Figure 4(a)). Further validation using Western blotting demonstrated that anlotinib reduced type I and type IV collagen and α-SMA expression in TGF-β1-exposed human proximal tubular cells ($P \leq 0.01$, Figures 4(b) and 4(c)). ## 3.5. Anlotinib Attenuates Renal Fibrosis by Inhibiting TGF-β1-Mediated ERK and AKT Pathways To explore anlotinib's potential mechanisms for treating renal fibrosis, we estimated the most probable molecular targets of anlotinib and obtained 100 potential targets by SwissTargetPrediction (Additional file 2) [17]. A total of 6549 targets (Additional file 3) associated with renal fibrosis were obtained from the GeneCards database. The intersection of targets was mapped by drawing Venn diagrams and constructing target networks (Figures 5(a) and 5(b)) for the purpose of elucidating the interactions between potential targets of anlotinib and renal fibrosis-related targets. Results indicated that the potential targets of anlotinib and renal fibrosis-related targets share a total of 70 common targets (Additional file 3) (Figure 5(a)). Protein interaction (PPI) analysis was conducted using STRING (version 11.0) (Figure 5(b)). Bioinformatics data suggests that among the 70 common target genes, the mitogen-activated protein kinase (MAPK) and phosphatidylinositol 3-kinase (PI3K)/protein kinase B (AKT) signaling pathways are most closely related (Figure 5(c)). Given that the MAPK and PI3K-PKB/Akt pathways are thought to be associated with fibrosis in many organs [18–20], we used GO analysis to enrich these two signaling pathways, which suggested an association with protein tyrosine kinase activity (Figure 5(d)). We hypothesized that the reduction of renal fibrosis by anlotinib is transduced through the MAPK and PI3K-PKB/Akt pathways. We further validated our hypothesis using Western blotting, and as expected, anlotinib significantly reduced ERK ($P \leq 0.01$) and AKT ($P \leq 0.05$) in TGF-β1-exposed human proximal tubular cells, as shown in Figures 5(e) and 5(f). Taken together, anlotinib reduces renal fibrosis by inhibiting ERK and AKT pathways through TGF-β1 signaling transduction. ## 4. Discussion The purpose of this study was to investigate whether anlotinib could ameliorate functional and pathological renal injury in fibrotic nephropathy mouse models. Therefore, we studied the FA-induced and UUO-induced renal injury models and attempted to illustrate the potential mechanism of action of anlotinib against renal fibrosis. We first demonstrated that anlotinib improved renal function after FA-induced renal injury and reduced proteinuria and serum creatinine. Anlotinib was found to reduce renal fibrosis by significantly attenuating inflammation and matrix protein expression in UUO mouse models of FA-induced kidney injury. In addition, we also found that TGF-β1 expression was inhibited and decreased in mice treated with anlotinib for renal fibrosis. The results of in vitro experiments further support that the renoprotective effect of anlotinib may be mediated through the TGF-β1 signaling pathway. Anlotinib is effective against a wide range of tumors, and inhibition of inflammatory and fibrotic processes in mouse models of bleomycin-induced pulmonary fibrosis has already been reported [13]. Other than that, no literature has been published regarding anlotinib's antifibrotic effects. Our study provides the first evidence that anlotinib has similar anti-inflammatory and antifibrotic effects on the kidney after FA-induced kidney injury in mice and anti-inflammatory and antifibrotic effects in the UUO models. Anlotinib has been shown to exert a nephroprotective effect. According to previous studies, TGF-β1 is associated with renal fibrosis, inflammation, and progression of kidney disease [16, 21]. CKD and renal fibrosis can result from several underlying etiologies. Regardless of the cause, TGF-β1 is significantly upregulated both in animal experiments and in human kidney disease. Meanwhile, TGF-β1 has been identified as the most potent EMT inducer, inducing the conversion of renal tubular epithelial cells (TECs) to myofibroblasts [22, 23]. A number of pathways are involved in the regulation of renal fibrosis meridian [24], and the TGF-β1 pathway plays a crucial role among them; therefore, regulation of TGF-β1 expression is considered effective and important for the treatment of renal fibrosis/chronic kidney disease [25]. Recently, it has been reported that anlotinib has a profibrotic effect by inhibiting TGF-β1 in lung fibroblasts [13]. Inspired by their findings, we showed a significant downregulation of TGF-β1 in the anlotinib-treated group compared to the untreated group after establishing UUO mouse models. These studies indicated a close relationship between TGF-β1 overexpression and renal fibrosis. It is therefore suggested that anlotinib improves renal function and antifibrosis by inhibiting TGF-β1 expression. A key function of TGF-β1 is to promote renal fibrosis, mainly through Smad and non-Smad signaling pathways. The Akt signaling pathway in the non-Smad pathway regulates EMT, and the activation of Akt is a key link in the induction of EMT by TGF-β1 [26]. In addition, extracellular signal-regulated protein kinases (ERK), one of the mitogen-activated protein kinase (MAPK) signaling pathway family, may affect renal tubular interstitial fibrosis through independent mechanisms [27]. The main mechanism is that angiotensin II (AngII) acts on the AT1 receptor in proximal tubular epithelial cells and stimulates the phosphorylation of caveolin-1 (Cav-1) and epidermal growth factor receptor (EGFR), which induces the phosphorylation of these proteins in the proximal tubular epithelium. Phosphorylated proteins cross-linked on trap protein-rich lipid membrane rafts, causing sustained EGFR- ERK signaling and promoting EMT of tubular epithelial cells. In the present study, HK-2 cells showed significantly elevated inflammatory cytokines, collagen, and α-SMA in response to TGF-β1 stimulation. Further studies suggested that EKR and AKT phosphorylation levels were significantly increased, and all of the above expressions were significantly decreased after anlotinib treatment. Our study shows that TGF-β1 stimulates ERK and AKT signaling channels and promotes renal tubular interstitial fibrosis, and anlotinib is used to reduce the inflammatory response, delay EMT, and improve renal fibrosis by inhibiting TGF-β1. ## 5. Conclusions In conclusion, we confirmed through in vivo and in vitro experiments that anlotinib inhibits inflammatory cell infiltration and reduces inflammatory cytokines, decreases ECM accumulation, and improves renal function and tubulointerstitial fibrosis by inhibiting TGF-β1-mediated ERK and AKT signaling pathways. This study provides the basis for basic research related to the treatment of CKD with anlotinib. ## Data Availability The data that support the findings of this study are available from the corresponding authors upon reasonable request. ## Conflicts of Interest The authors declare no conflict of financial interest or benefit. ## Authors' Contributions Donglin Sun, Jing Guo, and Weifei Liang contributed equally to this work. ## References 1. Jha V., Garcia-Garcia G., Iseki K.. Chronic kidney disease: global dimension and perspectives. (2013) 382 260-272. DOI: 10.1016/S0140-6736(13)60687-X 2. Webster A. C., Nagler E. V., Morton R. L., Masson P.. Chronic kidney disease. (2017) 389 1238-1252. DOI: 10.1016/S0140-6736(16)32064-5 3. Panizo S., Martínez-Arias L., Alonso-Montes C.. Fibrosis in chronic kidney disease: pathogenesis and consequences. (2021) 22 p. 408. DOI: 10.3390/ijms22010408 4. Ferenbach D. A., Bonventre J. V.. Mechanisms of maladaptive repair after AKI leading to accelerated kidney ageing and CKD. (2015) 11 264-276. DOI: 10.1038/nrneph.2015.3 5. Wick G., Grundtman C., Mayerl C.. The immunology of fibrosis. (2013) 31 107-135. DOI: 10.1146/annurev-immunol-032712-095937 6. Bülow R. D., Boor P.. Extracellular matrix in kidney fibrosis: more than just a scaffold. (2019) 67 643-661. DOI: 10.1369/0022155419849388 7. Kriz W., Kaissling B., Le Hir M.. Epithelial-mesenchymal transition (EMT) in kidney fibrosis: fact or fantasy?. (2011) 121 468-474. DOI: 10.1172/JCI44595 8. Nieto M. A., Huang R. Y.-J., Jackson R. A., Thiery J. P.. EMT: 2016. (2016) 166 21-45. DOI: 10.1016/j.cell.2016.06.028 9. Liu Y.. New insights into epithelial-mesenchymal transition in kidney fibrosis. (2010) 21 212-222. DOI: 10.1681/ASN.2008121226 10. Zeisberg M., Neilson E. G.. Mechanisms of tubulointerstitial fibrosis. (2010) 21 1819-1834. DOI: 10.1681/ASN.2010080793 11. Liang L., Hui K., Hu C.. Autophagy inhibition potentiates the anti-angiogenic property of multikinase inhibitor anlotinib through JAK2/STAT3/VEGFA signaling in non-small cell lung cancer cells. (2019) 38. DOI: 10.1186/s13046-019-1093-3 12. Taurin S., Yang C.-H., Reyes M.. Endometrial cancers harboring mutated fibroblast growth factor receptor 2 protein are successfully treated with a new small tyrosine kinase inhibitor in an orthotopic mouse model. (2018) 28 152-160. DOI: 10.1097/IGC.0000000000001129 13. Ruan H., Lv Z., Liu S.. Anlotinib attenuated bleomycin-induced pulmonary fibrosis via the TGF-. (2019) 72 44-55. DOI: 10.1111/jphp.13183 14. Sun Y., Niu W., Du F.. Safety, pharmacokinetic, and antitumor activity of anlotinib, an oral multi-target tyrosine kinase inhibitor, in patients with advanced refractory solid tumors. (2015) 33, article e13586. DOI: 10.1200/jco.2015.33.15_suppl.e13586 15. Xie C., Wan X., Quan H.. Preclinical characterization of anlotinib, a highly potent and selective vascular endothelial growth factor receptor-2 inhibitor. (2018) 109 1207-1219. DOI: 10.1111/cas.13536 16. Tang P. M. K., Zhang Y. Y., Mak T. S. K., Tang P. C. T., Huang X. R., Lan H. Y.. Transforming growth factor-. (2018) 596 3493-3503. DOI: 10.1113/JP274492 17. Gfeller D., Grosdidier A., Wirth M., Daina A., Michielin O., Zoete V.. SwissTargetPrediction: a web server for target prediction of bioactive small molecules. (2014) 42 W32-W38. DOI: 10.1093/nar/gku293 18. An J. N., Yang S. H., Kim Y. C.. Periostin induces kidney fibrosis after acute kidney injury via the p38 MAPK pathway. (2019) 316 F426-F437. DOI: 10.1152/ajprenal.00203.2018 19. El-Agamy D. S., El-Harbi K. M., Khoshhal S.. Pristimerin protects against doxorubicin-induced cardiotoxicity and fibrosis through modulation of Nrf2 and MAPK/NF-kB signaling pathways. (2019) 11 47-61. DOI: 10.2147/CMAR.S186696 20. Liu N., Feng J., Lu X.. Isorhamnetin inhibits liver fibrosis by reducing autophagy and inhibiting extracellular matrix formation via the TGF-. (2019) 2019 14. DOI: 10.1155/2019/6175091 21. Ma T.-T., Meng X.-M., Liu B. C., Lan H. Y., Lv L. L.. TGF-. (2019) 1165 347-364. DOI: 10.1007/978-981-13-8871-2_16 22. Liu Q., Liu S., Shi Y.. Suppressors of cytokine signaling inhibit tubular epithelial cell-myofibroblast transdifferentiation. (2011) 34 142-151. DOI: 10.1159/000329325 23. Yao Y., Chen R., Wang G., Zhang Y., Liu F.. Exosomes derived from mesenchymal stem cells reverse EMT via TGF-. (2019) 10. DOI: 10.1186/s13287-019-1332-8 24. Voelker J., Berg P. H., Sheetz M.. Anti–TGF-. (2017) 28 953-962. DOI: 10.1681/ASN.2015111230 25. Meng X.-m., Nikolic-Paterson D. J., Lan H. Y.. TGF-. (2016) 12 325-338. DOI: 10.1038/nrneph.2016.48 26. Cho H. J., Baek K. E., Saika S., Jeong M.-J., Yoo J.. Snail is required for transforming growth factor-. (2007) 353 337-343. DOI: 10.1016/j.bbrc.2006.12.035 27. Chen J., Chen J.-K., Harris R. C.. Angiotensin II induces epithelial-to-mesenchymal transition in renal epithelial cells through reactive oxygen species/Src/caveolin-mediated activation of an epidermal growth factor receptor–extracellular signal-regulated kinase signaling pathway. (2012) 32 981-991. DOI: 10.1128/MCB.06410-11
--- title: Empagliflozin-Pretreated Mesenchymal Stem Cell-Derived Small Extracellular Vesicles Attenuated Heart Injury authors: - Boyu Chi - Ailin Zou - Lipeng Mao - Dabei Cai - Tingting Xiao - Yu Wang - Qingjie Wang - Yuan Ji - Ling Sun journal: Oxidative Medicine and Cellular Longevity year: 2023 pmcid: PMC9966826 doi: 10.1155/2023/7747727 license: CC BY 4.0 --- # Empagliflozin-Pretreated Mesenchymal Stem Cell-Derived Small Extracellular Vesicles Attenuated Heart Injury ## Abstract ### Objective Small extracellular vesicles derived from mesenchymal stem cells (MSCs) play important roles in cardiac protection. Studies have shown that the cardiovascular protection of sodium-glucose cotransporter 2 inhibitor (SGLT2i) is independent of its hypoglycemic effect. This study is aimed at investigating whether small extracellular vesicles derived from MSCs pretreated with empagliflozin (EMPA) has a stronger cardioprotective function after myocardial infarction (MI) and to explore the underlying mechanisms. ### Methods and Results We evaluated the effects of EMPA on MSCs and the effects of EMPA-pretreated MSCs-derived small extracellular vesicles (EMPA-sEV) on myocardial apoptosis, angiogenesis, and cardiac function after MI in vitro and in vivo. The small extracellular vesicles of control MSCs (MSC-sEV) and EMPA-pretreated MSCs were extracted, respectively. Small extracellular vesicles were cocultured with apoptotic H9c2 cells induced by H2O2 or injected into the infarcted area of the Sprague-Dawley (SD) rat myocardial infarction model. EMPA increased the cell viability, migration ability, and inhibited apoptosis and senescence of MSCs. In vitro, EMPA-sEV inhibited apoptosis of H9c2 cells compared with the control group (MSC-sEV). In the SD rat model of MI, EMPA-sEV inhibited myocardial apoptosis and promoted angiogenesis in the infarct marginal areas compared with the MSC-sEV. Meanwhile, EMPA-sEV reduced infarct size and improved cardiac function. Through small extracellular vesicles (miRNA) sequencing, we found several differentially expressed miRNAs, among which miR-214-3p was significantly elevated in EMPA-sEV. Coculture of miR-214-3p high expression MSC-derived small extracellular vesicles with H9c2 cells produced similar protective effects. In addition, miR-214-3p was found to promote AKT phosphorylation in H9c2 cells. ### Conclusions Our data suggest that EMPA-sEV significantly improve cardiac repair after MI by inhibiting myocardial apoptosis. miR-214-3p at least partially mediated the myocardial protection of EMPA-sEV through the AKT signaling pathway. ## 1. Introduction Acute myocardial infarction (AMI) is the major cause of morbidity and mortality all over the world [1]. Percutaneous transluminal coronary intervention (PCI), as the main treatment for acute myocardial infarction, can improve the prognosis but cannot avoid the occurrence of myocardial cell death, ventricular remodeling, and heart failure [2]. Bone marrow-derived MSCs are a promising approach to the treatment of cardiac injury after myocardial infarction (MI). However, therapeutic effect on MI by MSCs is limited due to the poor survival rate given a hostile microenvironment in an ischemic heart. At the same time, some studies have shown that the beneficial effect of MSCs is mainly due to the paracrine action of secretory factors, rather than the direct muscle regeneration of MSCs or MSC derived cells [3, 4]. Currently, cell-free therapy based on MSCs derived small extracellular vesicles has made remarkable progress in myocardial protection [4, 5]. Small extracellular vesicles are lipid bilayered structures, 30–150 nm in size, enclosing cargo containing messenger ribonucleic acid (mRNA), miRNAs, growth factors, and proteins for transfer into recipient cells [6, 7]. Small extracellular vesicles have great advantages, such as high cyclic stability, dose control of transplantation, lack of immunogenicity, low toxicity, and easy functionalization. However, the therapeutic effect of small extracellular vesicles lack modification is not as effective as expected. Therefore, small extracellular vesicles can be optimized by various engineering methods to improve the targeting and therapeutic effect of small extracellular vesicles, such as pharmacological compound-pretreated MSCs and gene-modified MSCs [8]. MSCs pretreated with drugs such as atorvastatin [9], rosuvastatin [10], and curcumin [11] have been shown to improve the survival of local cardiomyocytes and promote angiogenesis. At present, small extracellular vesicles released from MSCs pretreated with atorvastatin can significantly improve cardiac function in the MI model of SD rats, and lncRNA H19 is partially involved in cardiac repair mediated by angiogenesis [12]. So, the development of new strategies to optimize cell-free therapy is expected to play an important role in future clinical applications. EMPA, a sodium glucose cotransporter 2 inhibitor (SGLT2i), is a new oral hypoglycemic agent for the treatment of type 2 diabetes mellitus (T2DM). SGLT2i were shown to decrease mortality from cardiovascular diseases in the EMPA-REG trial [13]. Inhibition of the SGLT2 reduces cardiovascular morbidity and mortality in patients with T2DM with atherosclerotic, cardiovascular disease [14, 15]. SGLT2i significantly improve cardiovascular outcomes including cardiovascular and all-cause mortality in patients with HF without an increased risk of serious adverse events [16]. Combined with the benefits of SGLT2i in cardiovascular disease and the potential benefits of biomimetic small extracellular vesicles for stem cell self-therapy [17, 18], it becomes critical to determine the effects of EMPA pretreatment on the functions of MSCs and MSC-derived small extracellular vesicles. In this study, we investigated the cardioprotective effects of EMPA-pretreated MSC-derived small extracellular vesicles in vitro and in vivo. Compared with MSC-sEV, EMPA-sEV significantly enhanced cardiac function and promoted angiogenesis. The apoptosis rate of EMPA-sEV was also significantly reduced. To explore the underlying molecular mechanisms, we found that EMPA-sEV showed significantly increased miR-214-3p expression levels. The high expression of miR-214-3p could simulate the improvement effect of EMPA-sEV. Therefore, our study suggests that EMPA preconditioning may strengthen the therapeutic effect of the small extracellular vesicles derived from MSCs on MI by inhibiting myocardial apoptosis through a paracrine mechanism. In addition, miR-214-3p appears to mediate the protective effect of EMPA-sEV on the heart after AMI. ## 2.1. Cell Culture and EMPA Pretreatment Human bone marrow MSCs are commercially purchased (CRK Pharam, CRKP-H166, BMSC). MSCs were cultured in MEM alpha media (Gibco, C12571500BT) with $10\%$ fetal bovine serum (FBS, BI, 04-001-1ACS) in a humidified chamber at 37°C with $5\%$ CO2. Human bone marrow MSCs were characterized by the expression of cell surface markers. They were washed with PBS and incubated at 4°C for 30 min with 1 μl of a monoclonal antibody specific for CD73, CD44, CD105, CD31, CD34, and CD45 (1: 100, eBioscience, Waltham, MA, USA). And 1 ul of isotype antibody (1: 100, eBioscience, Waltham, MA, USA) was used as controls. FACS CantoII (BD Biosciences, San Jose, CA, USA) was used for cytometry analysis. In addition, human bone marrow MSCs were tested for the multiple differentiation potential by Oil red staining, Alizarin red staining and Alcian blue staining according to the product instructions (BGsciences, BGM-0122, BGM-0133, BGM-0144). H9c2 cells (BNCC, BNCC337726, and H9c2 [2-1]) were cultured in DMEM (Gibco, C11995500BT) containing $10\%$ FBS at 37°C in a humidified atmosphere with $5\%$ CO2. When the concentration reached $70\%$-$80\%$, the adherent cells were harvested by trypsin digestion (Gibco, A12859-01). MSCs at passage 5-7 were used for further testing. When the concentration reached $70\%$-$80\%$, MSCs were treated with EMPA (Beyotime, SD2411, the purity of $98\%$). Briefly, this reagent was consisted of 5 mg EMPA and 1.11 ml of dimethyl sulfoxide (DMSO). The EMPA concentration and treatment time were determined by CCK-8 assay. MSCs treated with 500 nM EMPA and cultured for 48 h were determined as the uniform experimental conditions for in vitro analysis. ## 2.2. Small Extracellular Vesicle Isolation and Characterization MSCs were cultured in complete medium to $80\%$ confluence and then incubated with extracellular vesicle-free medium for 48 h. After 48 h, conditioned medium was collected and centrifuged at 1500 g for 30 min to remove apoptotic bodies and cell debris, followed by incubation with Ribo™ Exosome Isolation Reagent (for cell culture media, RiboBio, C10130-2) for 12 h at 4°C. The supernatant was centrifuged at 2000 g for 30 min. The supernatant was discarded, and the pellet was suspended in PBS (100 μl) and stored at -80°C. The particle size and concentration of small extracellular vesicles were analyzed using nanoparticle tracking analysis (NTA). The collected small extracellular vesicles were fixed on carbon-coated copper grids with $1\%$ glutaraldehyde and then stained with $1\%$ phosphotungstic acid. The samples were examined using a JEM-2100 transmission electron microscope (TEM). And western blotting for CD63, CD81, and TGS101 was used to characterize the collected EMPA-sEV and MSC-sEV. To verify whether small extracellular vesicles could be absorbed by H9c2 cells, small extracellular vesicles were labeled with PHK26 and cocultured with H9c2 cells for 24 hours. And the cell nucleus were stained with DAPI after coculture. ## 2.3. Cell Counting Kit Assay Cell viability was detected using Cell Counting Kit (CCK-8, Yeasen, 40203ES76) according to the manufacturer's instructions. In brief, cells were cultured in 96-well plates in the medium of different treatment groups for 24 h and 48 h. Then, 10 μl of CCK-8 solution was added to each well and incubated for 1.5 h. Finally, the absorbance was measured at 450 nm by using a microplate reader. ## 2.4. Scratch Wound Assay MSCs were cultured in 6-well plate with complete culture medium. When the MSCs reached $100\%$ confluence, a 200 μl pipetting tip was used to make a scratch of the same width across the entire well. Subsequently, MSCs were carefully washed with PBS to remove cell debris and then incubated with 500 nM EMPA or 500 nM DMSO in an incubator with $5\%$ CO2 at 37°C. After 24 h of incubation, the migration of MSCs into the wound was observed. The experiment was repeated at least three times. ## 2.5. EdU Assay Cell proliferation was detected using EdU (5-ethynyl-2′-deoxyuridine) Detection Kit (RiboBio, R10034.5). According to the instructions, proliferating cells were stained with Apollo-positive nuclei and cell nucleus were stained with Hoechst 33342. The percentage of proliferating cells in each group was calculated. Percentage of increment cells was used for subsequent statistical analysis. ## 2.6. Senescence β-Galactosidase Staining Assay Senescence-associated β-galactosidase (SA-β-Gal) activity was upregulated during aging. Cell senescence was detected by SA-β-Gal Staining kit (Beyotime, C0602). The cells in each group were fixed and stained with SA-β-Gal Staining solution after fixation. We incubate at 37°C overnight. It should be noted that incubation cannot be carried out in a CO2 incubator. ## 2.7. Caspase-3/7 Analysis of Apoptosis Caspase-$\frac{3}{7}$ assays were used for cell apoptosis. Caspase-$\frac{3}{7}$ apoptosis detection kit (RiboBio, R11094.2) was used to detect apoptosis. The cell nucleus were stained with Hoechst 33342. Early apoptotic cells were stained with caspase-$\frac{3}{7.}$ PI staining was used for late apoptosis. The apoptosis rate was calculated for subsequent analysis. ## 2.8. Total RNA Isolation and Quantitative Real-Time PCR Total RNA was isolated from H9c2 cells by TRIzol (Vazyme, R401-01) extraction method. The cDNA libraries of mRNA were synthesized using HiScript II 1st-Strand cDNA Synthesis kit (Vazyme, R211-01). And the cDNA libraries of miRNA were synthesized using the miRNA 1st-Strand cDNA Synthesis Kit (Vazyme, MR101-01). Quantitative real-time PCR (RT-qPCR) was performed using HiScript II One step qRT-PCR SYBR Green kit (Vazyme, Q221-01) and miRNA Universal SYBR qPCR Master Mix kit (Vazyme, MQ101-01). The results were normalized to U6 (miRNA) expression levels. All specific primers were obtained from Shangya Bioengineering. Primer sequences are shown in supplementary table 1. RT-qPCR data were analyzed using the comparative Ct (ΔΔCt) method to quantify relative gene expression. ## 2.9. Western Blotting Whole Cell Lysis Assay kit (Keygen, KGP2100) was used to extract proteins from cells and small extracellular vesicles. Total protein concentration was analyzed using the BCA Protein Quantitation Assay kit (Keygen, KGPBCA). Western blotting was performed as previously described [19]. Antibodies used were as follows: TSG101 (1: 1000, AF8259, Beyotime), CD63 (1: 1000, AF1471, Beyotime), CD81 (1: 1000, AF2428, Beyotime), P21 (1: 1000, 10355-1-AP, Proteintech), BAX (1: 1000, 50599-2-Ig, Proteintech), Bcl-2 (1: 1000, 26593-1-AP, Proteintech), phosphorylated-AKT (p-AKT) (1: 1000, 28731-1-AP, Proteintech), AKT (1: 1000, AF1789, Beyotime), glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (1: 1000, 10494-1-AP, Proteintech), and beta-actin (β-actin, 1: 1000, 20536-1-AP, Proteintech). The bands were visualized by using enhanced chemiluminescence (Vazyme, E412-01) and analyzed with a gel documentation system. ImageJ was used for gray analysis of strips. ## 2.10. MI Model Induction and Small Extracellular Vesicle Injection The Institutional Animal Care and Use Committee of the Nanjing Medical University for Laboratory Animal Number to authorize the animal experiment protocol of this study. And all procedures were in accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health (NIH) Bethesda, MD, USA). SD rats (6-8 weeks-old; male) were procured from Animal Core Facility of Nanjing Medical University. The rats were acclimated to sterile animal management conditions for 7 days. The animals were randomly assigned to different treatment groups ($$n = 6$$). Anesthesia was performed by intraperitoneal injection of 100 mg/kg ketamine, and assisted ventilation was performed by endotracheal intubation on a small animal ventilator (ALC-V, ALCBIO). Myocardial infarction was caused by ligation of the left anterior descending coronary artery, as previously reported [20]. After MI, 50 μl PBS or small extracellular vesicles (from 1 × 106 cultured MSCs) in a total volume of 50 μl PBS was injected into the myocardium of each SD rat at three different locations in the infarcted area. Left ventricular function was evaluated 23 and 36 days after infarction. To avoid experimental bias and variation, the person performing the surgery is unaware of the grouping. ## 2.11. Assessment of Cardiac Function The rats were intraperitoneal anesthesia with sodium pentobarbital (50 mg/kg) at 23 and 36 days after MI for echocardiography (Vevo 2000 high-resolution microimaging system equipped with 30 MHz transducer). Left ventricular ejection fraction (LVEF) and left ventricular fractional shortening (LVFS) were measured and calculated. ## 2.12. Histological Analysis Hearts were taken 36 days after MI and subjected to paraffin embedding, sectioning, and staining. The extent of fibrosis and infarct size in left ventricular were quantified by Masson's trichrome stain and Sirius Red stain. Hematoxylin-Eosin (HE) stain was used to evaluate the degree of inflammatory cell infiltration. Blood vessel density was analyzed by CD31 immunofluorescence. The number of apoptotic cells in the tissue sections was quantified by terminal deoxynucleotidyl transferase dUTP nick end labelling (TUNEL) assay. ## 2.13. Small Extracellular Vesicle miRNA Sequencing and Bioinformatics Analyses The miRNA sequencing was performed in both EMPA-sEV and MSC-sEV. Small extracellular vesicles' miRNA sequence was analyzed by RiboBio (Guangzhou, China) using the Illumina HiSeq™ 2500 instrument. Differentially expressed miRNA were identified by fold change (\|log2(FoldChange)\| > 1) and significance levels (P value < 0.05). Bioinformatics analysis was performed, including differential expression miRNA analysis, miRNA target gene prediction, GO analysis, and KEGG pathway enrichment analysis. ## 2.14. miRNA Transfection miR-214-3p mimics (miR-214 mimics) (50 nmol/l) and negative controls (NC mimics) (50 nmol/l) were transfected using riboFECT™ CP Reagent according to the manufacturer's instructions. In short, MSCs were cultured at $50\%$ confluence. miR-214 mimics and the NC mimics were mixed with transfection reagents and then added to cell culture medium. The transfection efficiency was determined by RT-qPCR. The transfected miRNA sequences are shown in supplementary table 1. ## 2.15. Statistical Analysis Data were shown in bar graphs as mean ± standard error of mean (SEM). Continuous variables were compared using Student's t-tests. When there were more than two groups of data, one-way analysis of variance (ANOVA) was used for analysis. Data were analyzed with GraphPad Prism 8 for Windows, version 8.01 (GraphPad Software, Inc.). Graphs were assembled in GraphPad Prism 8. A P value < 0.05 was considered significant. ## 3.1. EMPA Enhances Mesenchymal Stem Cell Viability Consistent with previous studies, MSCs rapidly adhered to tissue culture containers and took on a fibroblast-like morphology in culture [20] (Figure 1(a)). The multiple differentiation potential of human bone marrow MSCs into adipogenesis, osteogenesis, and chondrogenesis was confirmed by oil red staining, alizarin red staining, and Alcian blue staining (Figure 1(a)). MSCs were also characterized by the expression of CD44, CD73, and CD105. Meanwhile, MSCs were also negative for CD45, CD31, and CD34 (Figure 1(b)). To evaluate the optimal therapeutic dose for EMPA, MSCs were treated at different concentrations. The cell viability of MSCs was determined by CCK-8 assay after 24 h and 48 h culture, respectively. The data showed that MSCs exhibited better cell viability at 500 nM and 48 h compared with other concentrations and control (Supplementary figure 1). Therefore, we treated MSCs with EMPA at 500 nM concentration and cultured them for 48 h as unified experimental conditions for in vitro analysis. We evaluated the proliferation of MSCs by EdU staining. Compared with the control group, the cell proliferation ability of EMPA group was significantly increased (Figures 1(c) and 1(d)). In addition, the wound healing assay showed that the EMPA group showed higher migration compared to the control group (Figure 1(e)). As is exhibited in Figures 1(f) and 1(g), the levels of SA-β-gal activity were significantly decreased in the EMPA group compared with the control. Furthermore, in contrast to the control, the EPMA group demonstrated a reduced expression level of senescence-associated protein 21 (Figure 1(h)). In conclusion, EMPA-pretreated MSCs showed better proliferation, stronger migration, and reduced senescence of the MSCs. ## 3.2. Characterization of Small Extracellular Vesicles Derived from MSCs TEM analysis confirmed that the EMPA-pretreated MSC-derived small extracellular vesicles were morphologically similar to a typical cup-shaped structure with a diameter of about 100 nm [21] (Figure 2(a)). Similarly, the MSC-derived small extracellular vesicles in the control group also had the same structure and size (Figure 2(a)). Next, NTA was used to analyze the small extracellular vesicles' size distribution and concentration. As shown in Figures 2(b) and 2(d) and supplementary table 2, EMPA-sEV and MSC-sEV have similar particle sizes and concentrations. Western blot assay showed that there were specific marker proteins of small extracellular vesicles in both groups, including TSG101, CD81, and CD63 (Figure 2(e)). In addition, H9c2 cells was incubated with pKH26-labeled small extracellular vesicles for 24 h to assess whether MSC-derived small extracellular vesicles could be taken up by H9c2 cells. Microscopic analysis showed that MSC-derived small extracellular vesicles could be absorbed by H9c2 cells (Figure 2(f)). ## 3.3. EMPA-sEV Protects H9c2 Cells by Inhibiting Apoptosis A large number of studies have shown that MSC-derived small extracellular vesicles can protect cardiomyocytes from ischemia and hypoxia by inhibiting apoptosis [22]. Through the above experiments, we confirmed that EMPA can improve the activity of MSCs and has no adverse effect on the secretion of small extracellular vesicles by MSCs. To verify whether EMPA-sEV enhances the protective effect on cardiomyocytes, H9c2 cells were induced by H2O2 (120 μM) under serum deprivation conditions. EMPA-sEV, MSC-sEV, or PBS were then treated with the apoptotic cardiomyocytes for 12 h. By caspase-$\frac{3}{7}$ staining, we found that the caspase-$\frac{3}{7}$-positive cells were significantly reduced in the EMPA-sEV group compared with the PBS group and MSC-sEV group (Figures 3(a) and 3(b)). PI staining showed that the PI-positive cells in the EMPA-sEV group were also significantly lower than those in the other two groups (Figures 3(c) and 3(d)). In addition, compared with the other groups, the expression of proapoptotic proteins (BAX) was significantly decreased and the expression of antiapoptotic proteins (Bcl-2) was significantly increased in the EMPA-sEV group (Figure 3(e)). These results suggest that EMPA-pretreated MSC-derived small extracellular vesicles can better protect cardiomyocytes by inhibiting apoptosis compared with normal small extracellular vesicles from MSCs. ## 3.4. EMPA-sEV Promotes Cardiomyocyte Survival and Angiogenesis in MI Rats To demonstrate whether EMPA-sEV has a more significant cardioprotective effect in vivo, we established a SD rat model of myocardial infarction by ligation of the left anterior descending coronary artery. To observe the effect of EMPA-sEV on cardiac function, cardiac ultrasound was performed on rats in each group at 23 and 36 days after MI surgery. Treatment with small extracellular vesicles significantly improved cardiac function compared with the MI group. At the same time, the EMPA-sEV group showed a greater improvement in cardiac function (Figures 4(a)–4(c)). In addition, Masson staining showed that small extracellular vesicles therapy reduced infarct size. EMPA-sEV was superior to MSC-sEV in reducing infarct size (Figure 4(d)). Sirius red staining also showed that EMPA-sEV inhibited myocardial fibrosis better than MSC-sEV (Figure 5(b)). As can be seen from Figure 5(a), there are obvious inflammatory cell infiltration in the post-MI infarct area and the infarct border area. However, the small extracellular vesicle treatment groups effectively reduced inflammatory cell infiltration, and EMPA-sEV significantly improved inflammatory cell infiltration in the infarct area and infarct margin area. In addition, the heart tissue sections were stained with immunofluorescence and immunohistochemistry. Compared with the MI group and MSC-sEV group, the EMPA-sEV group had lower TUNEL-positive rate (Figures 6(a) and 6(c)), higher CD31-positive rate (Figures 6(a) and 6(d)), and higher Bcl-2-positive rate (Figures 6(b) and 6(e)). These data suggest that EMPA-sEV could inhibit myocardial cell apoptosis and promote angiogenesis better than MSC-sEV (Figures 6(a)–6(e)). In summary, small extracellular vesicle therapy can effectively improve cardiac function after infarction by inhibiting myocardial apoptosis, myocardial fibrosis, and promoting angiogenesis. Furthermore, EMPA-pretreated MSC-derived small extracellular vesicles have more significant inhibitory effects on cardiomyocyte apoptosis and myocardial fibrosis, as well as more significant angiogenesis promotion and cardiac function improvement. ## 3.5. EMPA Pretreatment Increased the Expression of MSCs miR-214-3p and Its Release through Small Extracellular Vesicles Previous studies have shown that miRNAs carried by small extracellular vesicles play a very important role in regulating the cellular function of recipient cells. To investigate the mechanisms of cardiac protective effect of EMPA-sEV, miRNA sequencing was performed on small extracellular vesicles of two groups (MSC-sEV and EMPA-sEV) of MSCs (Figure 7(a)). Compared with MSC-sEV, EMPA-sEV detected a total of 42 upregulated miRNAs, while EMPA-sEV detected 59 downregulated miRNAs (Figure 7(b)). Using RT-qPCR analysis, we confirmed that 4 miRNAs were upregulated. miR-214-3p was significantly elevated in EMPA-sEV (Figure 7(c)). Candidate target genes of miRNA were predicted by miRTarBase, miRDB, miRWalk, and TargetScan (Figure 7(e)). Candidate target genes of miR-214-3p include BAX, BCL2L11, PTEN, and TWF1. It has been confirmed that BAX, BCL2L11, PTEN, and TWF1 are related to myocardial cell apoptosis [23–26]. However, further studies are needed to determine whether EMPA-sEV targets these genes through miR-214-3p to mediate cardiac protection. In addition, KEGG analysis of miRNA target genes showed that 273 differentially expressed genes were enriched into the AKT signaling pathway, including miR-214-3p (Figure 7(d)). These data suggest that the cardioprotective effect of EMPA-pretreated small extracellular vesicles may be mediated by miR-214-3p. Next, whether miR-214-3p mediates the cardioprotective effect of EMPA-pretreated MSC-derived small extracellular vesicles was further verified. MSCs were transfected with NC mimics and miR-214 mimics, respectively. Then, small extracellular vesicles were extracted and cocultured with H9c2 cells. RT-qPCR verified that mR-214-3p expression was increased in H9c2 cells compared with the control groups (Figure 8(a)). Compared with the negative control group, the expression of antiapoptotic protein (Bcl-2) was increased and the expression of proapoptotic protein (BAX) was decreased in the miR-214 mimics group under ischemia and hypoxia (Figures 8(b) and 8(c)). These results suggest that miR-214-3p can mediate the antiapoptotic effect of EMPA-pretreated MSC-derived small extracellular vesicles. Based on the results of miRNA target gene KEGG analysis, we focused on the AKT signaling pathway. Studies have confirmed that the AKT signaling pathway plays a key role in inhibiting myocardial apoptosis. To verify whether miR-214-3p activates the AKT signaling pathway, we observed the expression levels of the AKT signaling pathway-related proteins. As can be seen from Figures 8(b) and 8(c), p-AKT expression was significantly increased in the miR-214 mimics group. Therefore, EMPA-pretreated MSC-derived small extracellular vesicle-mediated inhibition of cardiomyocyte apoptosis, angiogenesis, and improvement of cardiac function, and most importantly, small extracellular vesicles-mediated effects are partially mediated through miR-214-3p-mediated AKT signaling activation (Figure 8(d)). ## 4. Discussion The current study has several major findings. First, EMPA can enhance cell viability and promote proliferation and migration of MSCs. EMPA also can inhibit the senescence of MSCs. Second, compared with MSC-sEV, EMPA-sEV has a more significant protective effect on cardiac function, including promoting angiogenesis and inhibiting fibrosis. Finally, the cardioprotective effect of EMPA-sEV is mediated, at least in part, by miR-214-3p via the AKT signaling pathway. In recent years, research on MSC transplantation has made great progress and become a potential treatment for ischemic heart disease [27]. MSCs are a pluripotent stem cell population that can be isolated from the bone marrow, fat, umbilical cord blood, etc. [ 28]. MSCs are characterized by multidirectional differentiation potential and low immunogenicity [29, 30]. Studies have shown that intravenous injection of MSCs can be tolerated, while repeated intracardial or coronary cell delivery cannot [31]. Nevertheless, there are still many security issues with MSC transplantation [32, 33]. The aging of MSCs reduces their cardioprotective effect [34]. The complex microenvironment and poor localization control of infarcted hearts after transplantation also limit the therapeutic effect of MSC transplantation [35]. To improve the therapeutic effect and survival of MSC transplantation in the ischemic and hypoxic microenvironment, many studies have been reported. These include genetic engineering and pretreatment of pharmacological compounds. A recent study demonstrated that high expression of miR-221-3p can enhance the cardioprotective effects of aging MSCs after myocardial infarction through the PTEN/AKT pathway, including promoting angiogenesis and inhibiting apoptosis [20]. Another study showed that the nonadherent culture of MSCs promoted angiogenesis and reduced cardiac remodeling after MSC transplantation by increasing the secretion of hydrolytic growth factor A (VEGFA) [36]. In this study, MSCs were pretreated with SGLT2i (EMPA). As a new hypoglycemic agent, SGLT2i has a cardiovascular protective effect independent of the hypoglycemic effect [37]. The cardioprotective effect of SGLT2i may be mediated by its ability to reduce cardiac inflammation, oxidative stress, apoptosis, and ion imbalance [38–41]. EMPA pretreatment increased the cell viability, proliferation, and migration of MSC and decreased the senescence of MSC. Further studies will clarify whether the beneficial effect of EMPA on MSCs can partially explain the cardiovascular protective mechanism of SGLT2i. More and more evidence showed that MSCs played a cardioprotective role mainly by secreting paracrine factors including small extracellular vesicles [4]. Compared with stem cell transplantation, it has more advantages. Firstly, small extracellular vesicles therapeutic effect was not affected by the complex microenvironment and poor localization. Secondly, as a cell-free therapy, small extracellular vesicles therapy has shown the advantages of low tumorigenic potential and minimal immunogenicity. Finally, small extracellular vesicles therapy is characterized by high circulatory stability [42, 43]. In the past few years, MSC-derived small extracellular vesicles have been shown to improve cardiac function after MI [5]. To improve the targeting and therapeutic effect of small extracellular vesicles, optimizing small extracellular vesicles through various engineering methods has become a promising therapeutic strategy [8]. Studies have shown that miR-25-3p overexpression in MSCs inhibits apoptosis of cardiomyocytes by targeting proapoptotic proteins and EZH2 [44]. Another study showed that overexpression of SDF1 in MSC-derived small extracellular vesicles can inhibit autophagy of ischemic cardiomyocytes and promote endothelial micro-angiogenesis [45]. Wei et al. reported that the overexpression of miRNA-181a in MSC-derived small extracellular vesicles affects the inflammatory response after myocardial ischemia-reperfusion injury [46]. Although these genetic approaches can directly alter the expression of cytokines or genes related to cardiac protection in MSC-derived small extracellular vesicles, they are not currently feasible in clinical practice. In contrast, we pretreated MSCs with EMPA with higher clinical feasibility. Our study found that EMPA-sEV inhibited H9c2 cells apoptosis better than MSC-sEV. Compared with MSC-sEV, EMPA-sEV had a better protective effect on cardiac function, promoting angiogenesis and inhibiting fibrosis. It is well known that small extracellular vesicles perform biological functions by containing proteins, mRNAs, miRNAs, etc [47]. In recent years, there has been increasing evidence that small extracellular vesicles' miRNAs mediate many biological functions of small extracellular vesicles. miRNAs including miR-22, miR-199a, and miR-214 play a key role in antiapoptosis effects [22]. In this study, the miRNA expression profile in EMPA-sEV showed that miR-214-3p in EMPA-sEV was significantly higher than that in MSC-sEV. In addition, the expression level of miR-214-3p in H9c2 cells was also increased after treatment with EMPA-sEV. miR-214-3p has been reported to have cardioprotective effects. miR-214-3p plays an antiapoptotic role by regulating sodium/calcium recovery 1, cyclophilin D, and Bcl-2 like protein 11 [48]. Small extracellular vesicles derived from miR-214-enriched MSCs reduce cardiac stem cell death by inhibiting reactive oxygen species production. In contrast, miR-214 inhibitors attenuate these effects [49, 50]. Based on these and other observations, we propose that miR-214-3p mediates the role of EMPA-sEV in cardiac protection. By overexpressing miR-214-3p in MSCs, the same antiapoptotic effect as that of EMPA-sEV was obtained. These results support the important role of miR-214-3p in anti-apoptosis in EMPA-sEV-mediated cardiovascular protection. In addition, we found that miR-214-3p promotes the phosphorylation of AKT in H9c2 cells, suggesting a possible molecular mechanism of miR-214-3p's role in cardiac protection. It has been reported that activation of the AKT signaling pathway is a key target for cardiac protection [51]. Current evidence suggests that miR-214-3p exerts cardioprotective effects by enhancing the AKT signaling pathway in cardiomyocytes. However, further studies are needed to elucidate the downstream mechanisms of the cardioprotective effects of EMPA-sEV mediated by miR-214-3p in physiology and pathophysiology. ## 4.1. Limitation The current study has some limitations. Firstly, H9c2 cells do not represent cardiomyocytes well. Cardiac primary cardiomyocytes are a good source for in vitro research. Secondly, our study suggests that the cardioprotective effects of EMPA-sEV were at least partly mediated through miR-214. Other miRNAs, proteins, lipids, and mRNAs which were also contained in small extracellular vesicles may also be functionally involved with the cardioprotective effects. However, these molecules have not been further investigated in this study. Thirdly, EMPA-sEV is at least partially protected by miR-214-3p through the AKT signaling pathway, but its specific mechanism has not been elucidated. Finally, although EMPA-sEV has a more significant therapeutic effect compared to MSC-sEV, further preclinical studies are needed to validate its efficacy and safety. ## 5. Conclusion EMPA pretreatments promoted the effect of MSC-derived small extracellular vesicles on inhibiting myocardial apoptosis, enhancing angiogenesis, and improving cardiac function after myocardial infarction. MiR-214-3p at least partly mediated the cardioprotective effects of EMPA-sEV via activating the AKT signaling pathway. ## Data Availability The datasets and materials used in the study are available from the corresponding author. ## Ethical Approval Animal experiments were conducted according to the Guidelines for the Care and Use of Laboratory Animals and were approved by the Ethics Committee of Nanjing medical university (No. IACUC-2205071). ## Consent All the coauthors consent to publishing the paper in Oxidative Medicine and Cellular Longevity. ## Conflicts of Interest The authors have declared that no competing interest exists. ## Authors' Contributions BYC, ALZ, YJ, QJW, and LS participated in the design of the study. BYC, ALZ, LPM, and TTX performed the experiments. BYC, ALZ, LPM, DBC, YW, LS, and TTX analyzed the data. BYC, YJ, QJW, and LS conceived of the study and participated in its design and coordination and helped to draft the manuscript. All authors read and approved the final manuscript. Boyu Chi and Ailin Zou contributed equally to this work. ## References 1. Wang H., Abajobir A. A., Abate K. H.. Global, regional, and national under-5 mortality, adult mortality, age- specific mortality, and life expectancy, 1970-2016: a systematic analysis for the Global Burden of Disease Study 2016. (2017) 390 1084-1150. DOI: 10.1016/S0140-6736(17)31833-0 2. Heusch G., Gersh B. J.. The pathophysiology of acute myocardial infarction and strategies of protection beyond reperfusion: a continual challenge. (2016) 38 ehw224-ehw784. DOI: 10.1093/eurheartj/ehw224 3. Makridakis M., Roubelakis M. G., Vlahou A.. Stem cells: insights into the secretome. (2013) 1834 2380-2384. DOI: 10.1016/j.bbapap.2013.01.032 4. Gnecchi M., Zhang Z., Ni A., Dzau V. J.. Paracrine mechanisms in adult stem cell signaling and therapy. (2008) 103 1204-1219. DOI: 10.1161/CIRCRESAHA.108.176826 5. Yu H., Cheng J., Shi W.. Bone marrow mesenchymal stem cell-derived exosomes promote tendon regeneration by facilitating the proliferation and migration of endogenous tendon stem/progenitor cells. (2020) 106 328-341. DOI: 10.1016/j.actbio.2020.01.051 6. Raposo G., Stoorvogel W.. Small extracellular vesicles: exosomes, microvesicles, and friends. (2013) 200 373-383. DOI: 10.1083/jcb.201211138 7. Théry C., Zitvogel L., Amigorena S.. Exosomes: composition, biogenesis and function. (2002) 2 569-579. DOI: 10.1038/nri855 8. Moghadasi S., Elveny M., Rahman H. S.. A paradigm shift in cell-free approach: the emerging role of MSCs-derived exosomes in regenerative medicine. (2021) 19 p. 302. DOI: 10.1186/s12967-021-02980-6 9. Xu J., Xiong Y., Li Q.. Optimization of timing and times for administration of atorvastatin-pretreated mesenchymal stem cells in a preclinical model of acute myocardial infarction. (2019) 8 1068-1083. DOI: 10.1002/sctm.19-0013 10. Xu H., Yang Y., Qian H., Tang Y., Wang H., Zhang Q.. Rosuvastatin treatment activates JAK-STAT pathway and increases efficacy of allogeneic mesenchymal stem cell transplantation in infarcted hearts. (2011) 75 1476-1485. DOI: 10.1253/circj.cj-10-1275 11. Liu J., Zhu P., Song P.. Pretreatment of adipose derived stem cells with curcumin facilitates myocardial recovery via antiapoptosis and angiogenesis. (2015) 2015 12. DOI: 10.1155/2015/638153 12. Huang P., Wang L., Li Q.. Atorvastatin enhances the therapeutic efficacy of mesenchymal stem cells-derived exosomes in acute myocardial infarction via up-regulating long non-coding RNA H19. (2020) 116 353-367. DOI: 10.1093/cvr/cvz139 13. Jhuo S., Liu I., Tsai W.. Effects of secretome from fat tissues on ion currents of cardiomyocyte modulated by sodium-glucose transporter 2 inhibitor. (2020) 25 p. 3606. DOI: 10.3390/molecules25163606 14. Zinman B., Wanner C., Lachin J. M.. Empagliflozin, cardiovascular outcomes, and mortality in type 2 diabetes. (2015) 373 2117-2128. DOI: 10.1056/NEJMoa1504720 15. Zinman B., Lachin J. M., Inzucchi S. E.. Empagliflozin, cardiovascular outcomes, and Mo-rtality in type 2 diabetes. (2016) 374 1092-1094. DOI: 10.1056/NEJMc1600827 16. Giugliano D., Longo M., Scappaticcio L., Bellastella G., Maiorino M. I., Esposito K.. SGL-T-2 inhibitors and cardiorenal outcomes in patients with or without type 2 diabetes: a meta-analysis of 11 CVOTs. (2021) 20 p. 236. DOI: 10.1186/s12933-021-01430-3 17. Luo L., Tang J., Nishi K.. Fabrication of synthetic mesenchymal stem cells for the treatment of acute myo-cardial infarction in mice. (2017) 120 1768-1775. DOI: 10.1161/CIRCRESAHA.116.310374 18. Tang J., Shen D., Caranasos T. G.. Therapeutic microparticles functionalized with biomimetic cardiac stem cell membranes and secretome. (2017) 8. DOI: 10.1038/ncomms13724 19. Sun L., Zhu W., Zhao P.. Long noncoding RNA UCA1 from hypoxia-conditioned hMSC-derived exosomes: a novel molecular target for cardioprotection through miR-873-5p/XIAP axis. (2020) 11 p. 696. DOI: 10.1038/s41419-020-02783-5 20. Sun L., Zhu W., Zhao P.. Down-regulated exosomal microRNA-221-3p derived from senescent mesenchymal stem cells impairs heart repair. (2020) 8 p. 263. DOI: 10.3389/fcell.2020.00263 21. Wang Q., Zhang L., Sun Z.. HIF-1. (2021) 12, article 100171. DOI: 10.1016/j.mtbio.2021.100171 22. Sun S., Wei R., Li F., Liao S., Tse H.. Mesenchymal stromal cell-derived exosomes in cardiac regeneration and repair. (2021) 16 1662-1673. DOI: 10.1016/j.stemcr.2021.05.003 23. Hochhauser E., Cheporko Y., Yasovich N.. Bax deficiency reduces infarct size and improves long-term function after myocardial infarction. (2007) 47 11-19. DOI: 10.1385/CBB:47:1:11 24. Wan J., Lin S., Yu Z.. Protective effects of microRNA-200b-3p encapsulated by mesenchymal stem cells–secreted small extracellular vesicles in myocardial infarction via regulating BCL2L11. (2022) 11. DOI: 10.1161/JAHA.121.024330 25. Feng Q., Li X., Qin X., Yu C., Jin Y., Qian X.. PTEN inhibitor improves vascular remodeling and cardiac function after myocardial infarction through PI3k/Akt/VEGF signaling pathway. (2020) 26 p. 111. DOI: 10.1186/s10020-020-00241-8 26. Mu L., Yu H., Xu X., Liu Y., Hui C.. The effect of Twf1 gene on apoptosis and NF-. (2018) 12 10-12 27. Williams A. R., Hare J. M.. Mesenchymal stem cells: biology, pathophysiology, translational findings, and therapeutic implications for cardiac disease. (2011) 109 923-940. DOI: 10.1161/CIRCRESAHA.111.243147 28. Pittenger M. F., Mackay A. M., Beck S. C.. Multilineage potential of adult human me-senchymal stem cells. (1999) 284 143-147. DOI: 10.1126/science.284.5411.143 29. Kim S. H., Cho J. H., Lee Y. H.. Improvement in left ventricular function with intracoronary mesenchymal stem cell therapy in a patient with anterior wall ST-segment elevation myocardial infarction. (2018) 32 329-338. DOI: 10.1007/s10557-018-6804-z 30. Bartolucci J., Verdugo F. J., González P. L.. Safety and efficacy of the intr-avenous infusion of umbilical cord mesenchymal stem cells in patients with heart failure. (2017) 121 1192-1204. DOI: 10.1161/CIRCRESAHA.117.310712 31. Mäkelä T., Takalo R., Arvola O.. Safety and biodistribution study of bone marrow-derived mesenchymal stromal cells and mononuclear cells and the impact of the administration route in an intact porcine model. (2015) 17 392-402. DOI: 10.1016/j.jcyt.2014.12.004 32. Gazdic M., Volarevic V., Arsenijevic N., Stojkovic M.. Mesenchymal stem cells: a fri-end or foe in immune-mediated diseases. (2015) 11 280-287. DOI: 10.1007/s12015-014-9583-3 33. Meier R. P. H., Müller Y. D., Morel P., Gonelle-Gispert C., Bühler L. H.. Transplantation of mesenchymal stem cells for the treatment of liver diseases, is there enough evidence?. (2013) 11 1348-1364. DOI: 10.1016/j.scr.2013.08.011 34. Nguyen P. K., Rhee J., Wu J. C.. Adult stem cell therapy and heart failure, 2000 to 2016: a systematic review. (2016) 1 831-841. DOI: 10.1001/jamacardio.2016.2225 35. Van D. A., Naaijkens B. A., Jurgens W. J. F. M.. Reduction of infarct size by intravenous injection of uncultured adipose derived stromal cells in a rat model is dependent on the time point of application. (2011) 7 219-229. DOI: 10.1016/j.scr.2011.06.003 36. Deng B., Zhang X., Liang Y.. Nonadherent culture method promotes MSC-mediated vascularization in myocardial infarction via miR-519d/VEGFA pathway. (2020) 11 p. 266. DOI: 10.1186/s13287-020-01780-x 37. Zelniker T. A., Braunwald E.. Cardiac and renal effects of sodium-glucose co-transporter 2 inhibitors in diabetes:. (2018) 72 1845-1855. DOI: 10.1016/j.jacc.2018.06.040 38. Ye Y., Bajaj M., Yang H., Perez-Polo J. R., Birnbaum Y.. SGLT-2 inhibition with dapag-liflozin reduces the activation of the Nlrp 3/ASC inflammasome and attenuates the development of diabetic cardiomyopathy in mice with type 2 diabetes. Further aug-mentation of the effects with saxagliptin, a DPP4 inhibitor. (2017) 31 119-132. DOI: 10.1007/s10557-017-6725-2 39. Lee T., Chang N., Lin S.. Dapagliflozin, a selective SGLT2 Inhibitor, attenuated cardiac fibrosis by regulating the macrophage polarization via STAT3 signaling in infarcted rat hearts. (2017) 104 298-310. DOI: 10.1016/j.freeradbiomed.2017.01.035 40. Tanajak P., Sanguanmoo P., Sivasinprasasn S.. Cardioprotection of dapagliflozin and vildagliptin in rats wit-h cardiac ischemia-reperfusion injury. (2018) 236 69-84. DOI: 10.1530/JOE-17-0457 41. Baartscheer A., Schumacher C. A., Wüst R. C.. Empagliflozin decreases myocardial cytoplasmic Na. (2017) 60 568-573. DOI: 10.1007/s00125-016-4134-x 42. Vandergriff A., Huang K., Shen D.. Targeting regenerative exosomes to myocardial infarction using cardiac homing pe-ptide. (2018) 8 1869-1878. DOI: 10.7150/thno.20524 43. Vandergriff A. C., de Andrade J. B. M., Tang J.. Intravenous cardiac stem cell-derived exosomes ameliorate cardiac dysfunction in doxorubicin induced dilated cardiomyopathy. (2015) 2015 8. DOI: 10.1155/2015/960926 44. Peng Y., Zhao J., Peng Z., Xu W., Yu G.. Exosomal miR-25-3p from mesenchymal stem cells alleviates myocardial infarction by targeting pro-apoptotic proteins and EZH2. (2020) 11 p. 317. DOI: 10.1038/s41419-020-2545-6 45. Gong X., Liu H., Wang S., Liang S., Wang G.. Exosomes derived from SDF1-overexpressing mesenchymal stem cells inhibit ischemic myocardial cell apoptosis and promote cardiac endothelial microvascular regeneration in mice with myocardial infarction. (2019) 234 13878-13893. DOI: 10.1002/jcp.28070 46. Wei Z., Qiao S., Zhao J.. miRNA-181a over-expression in mesenchymal stem cell-derived exosomes influenced inflammatory response after myocardial ischemia-reperfusion injury. (2019) 232, article 116632. DOI: 10.1016/j.lfs.2019.116632 47. Dhanabal M., Wu F., Alvarez E.. Recombinant semaphorin 6A-1 ectodomain inhibits in vivo growth factor and tumor cell line-induced angiogenesis. (2005) 4 659-668. DOI: 10.4161/cbt.4.6.1733 48. Aurora A. B., Mahmoud A. I., Luo X.. MicroRNA-214 protects the mou-se heart from ischemic injury by controlling Ca. (2012) 122 1222-1232. DOI: 10.1172/JCI59327 49. Wang Y., Zhao R., Liu D.. Exosomes derived from miR-214-enriched bone marrow-derived mesenchymal stem cells regulate oxidative damage in cardiac stem cells by targeting CaMKII. (2018) 2018 21. DOI: 10.1155/2018/4971261 50. Lv G., Shao S., Dong H., Bian X., Yang X., Dong S.. MicroRNA-214 protects cardiac myocytes against H. (2014) 115 93-101. DOI: 10.1002/jcb.24636 51. Zhang Q., Wang L., Wang S.. Signaling pathways and targeted therapy for myocardial infarction. (2022) 7 p. 78. DOI: 10.1038/s41392-022-00925-z
--- title: In Vitro Photoinactivation of Fusarium oxysporum Conidia with Light-Activated Ammonium Phthalocyanines authors: - Sara R. D. Gamelas - Isabel N. Sierra-Garcia - Augusto C. Tomé - Ângela Cunha - Leandro M. O. Lourenço journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966838 doi: 10.3390/ijms24043922 license: CC BY 4.0 --- # In Vitro Photoinactivation of Fusarium oxysporum Conidia with Light-Activated Ammonium Phthalocyanines ## Abstract Antimicrobial photodynamic therapy (aPDT) has been explored as an innovative therapeutic approach because it can be used to inactivate a variety of microbial forms (vegetative forms and spores) without causing significant damage to host tissues, and without the development of resistance to the photosensitization process. This study assesses the photodynamic antifungal/sporicidal activity of tetra- and octasubstituted phthalocyanine (Pc) dyes with ammonium groups. Tetra- and octasubstituted zinc(II) phthalocyanines (1 and 2) were prepared and tested as photosensitizers (PSs) on Fusarium oxysporum conidia. Photoinactivation (PDI) tests were conducted with photosensitizer (PS) concentrations of 20, 40, and 60 µM under white-light exposure at an irradiance of 135 mW·cm–2, applied during 30 and 60 min (light doses of 243 and 486 J·cm−2). High PDI efficiency corresponding to the inactivation process until the detection limit was observed for both PSs. The tetrasubstituted PS was the most effective, requiring the lowest concentration and the shortest irradiation time for the complete inactivation of conidia (40 µM, 30 min, 243 J·cm−2). Complete inactivation was also achieved with PS 2, but a longer irradiation time and a higher concentration (60 µM, 60 min, 486 J·cm−2) were necessary. Because of the low concentrations and moderate energy doses required to inactivate resistant biological forms such as fungal conidia, these phthalocyanines can be considered potent antifungal photodynamic drugs. ## 1. Introduction The Fusarium genus corresponds to a filamentous fungus commonly found in soil and plants, and it includes pathogen species to plants, animals, and humans [1]. Fusarium diseases affect many crop plants, imposing significant economic losses on fruit, vegetable, cereal, and cellulose production [2,3]. In humans, Fusarium oxysporum is an opportunistic pathogen that causes keratitis, onychomycosis, and invasive infection in immunocompromised and immunocompetent patients [4,5]. The asexual life cycle of F. oxysporum involves the production of chlamydospores, macroconidia, and microconidia, which ensures highly efficient dissemination in the environment [6]. Plant infection occurs through the roots. After the dispersal of conidia by wind or rain, germination begins in rhizosphere soil and growing hyphae penetrate root tissues, initiating infection [6,7]. Conventional fungicides usually target the conidial germination and early development stages [8]. In order to control infection by Fusarium species, the most effective choice is the use of resistant plant varieties. However, the tolerance level depends on the conditions in which they grow. In some regions, even with the use of resistant plant varieties, if the temperatures are elevated, the colonization by Fusarium spp. can be severe [9]. Besides the use of resistant plant varieties, some of the applied strategies are the use of seeds and seedlings already treated with fungicides before planting, and the use of fungicides through crop development [9,10]. However, the rise of tolerant fungal strains renders these strategies unreliable. The intensive use of the current fungicides is also considered a potential risk to humans and the environment [11,12]. To control the possibility of fungal inoculum in postharvest crops, it is common to treat them with chlorine or organic acids, even though they can be toxic to the environment [13]. The increased development of resistance to currently used fungicides and the progressive ban of the most popular pesticides in the EU imposes a drastic limitation on the chemical options to control fungal diseases in crops [14], so developing more effective technologies to control pathogenic fungi has gained interest. Interest in antimicrobial photodynamic therapy (aPDT) as an alternative for the inactivation of microorganisms in environmental matrices has been growing [15,16]. In fact, it was successfully used to inactivate or kill bacteria or fungi in animal hosts and environments [17,18,19,20,21]. aPDT is based on three nontoxic elements: a photosensitizer (PS), visible light, and oxygen (3O2). The combination of these three elements generates reactive oxygen species (ROS, e.g., singlet oxygen (1O2) and free radicals) that are responsible for the lethal oxidative damage of microbial targets (lipids, proteins, and nucleic acids), leading to the death of the target cells without causing significant damage to the host cells [22,23]. The application of aPDT for plant pathogens represents one of the latest developments for this technique aiming at alternatives to toxic agrochemicals. It was efficient in controlling bacterial diseases such as kiwifruit cancer [24] and citrus cancer [25]. This therapy was also used successfully against phytopathogenic fungi such as Lasiodiplodia theobromae (causes vine trunk disease), Botrytis cinerea (causes plant necrosis), and Colletotrichum sp. ( causes anthracnose in various fruit trees) [26,27,28]. As already mentioned, spores are crucial to spreading fungal diseases, and fungal conidia are relevant targets for photosensitization [29]. Studying the structure–activity relationship is crucial in designing powerful PSs capable of inducing lethal damage to plant pathogens within short irradiation periods while preserving the integrity of host plant tissues. Porphyrin (Por) [21,30,31,32,33,34], chlorin (Chl) [16,35,36], and phthalocyanine (Pc) [37,38,39] dyes have been extensively used for the aPDT approach. Pcs display unique UV-vis spectra, typically with a Soret band at a wavelength maximum of 350 nm and intense Q bands in the red/near-IR region (600–800 nm) [37,39,40]. Structural features and particular functionalities greatly determine the physicochemical properties and biological activities of Pcs. The modification of the Pc macrocycle with ammonium peripheral moieties or the introduction of metallic ions (e.g., Zn(II)), which may improve the triplet excited state features and the 1O2 quantum yield, is considered a reliable strategy to finetune the physicochemical properties of Pcs in relation to the intended microbial targets [37,39]. Regarding fungal spores, the affinity of a PS is affected by the overall hydrophobicity of the spore coating and by the charged units in the PS structure [41]. So, a PS might continue to be adsorbed to the outer layers of spores (or vegetative hyphae) or reach the intercellular compartment. In this case, there is an increase in the spectra of both physiological and biochemical targets [26,42]. Fungal conidia have eukaryotic genomes and are less susceptible to oxidative stress than prokaryote cells are, but they have a more comprehensive array of subcellular targets. Therefore, if there is a multitarget photosensitization capacity of a PS, there is a reduced ability of these spores to develop resistance [41]. This work assesses the antifungal photodynamic activity of two quaternized zinc(II) Pc derivatives, and determines the relations between the number of cationic peripheral substituents using [1] tetra- and [2] octasubstituted ammonium phthalocyanines [39]. ## 2.1. Synthesis and Photophysical Characterization of Phthalocyanine Derivatives PS 1 and 2 were synthesized and characterized by NMR (Figure S1–S3) according to the literature [39,43]. The absorption and emission spectra of 1 and 2 were determined in N,N-dimethylformamide (DMF) at low concentrations (10−6 M) [39]. The absorption spectra (Figure 1a) show the characteristic absorption features of zinc (II) Pcs with a high absorption band in the range of 300–450 nm (Soret band), and strong Q bands between 630 and 750 nm. Considering the excited state, after excitation at different wavelengths, both ZnPcs showed an emission band with the maximum between 670–691 nm (Figure 1b). The fluorescence quantum yields (ΦF) in DMF were less than 0.01 [39]. Given the potential use of PSs 1 and 2 as agents against Fusarium oxysporum conidia, it was essential to assess their ability to generate singlet oxygen. A previous assessment of the 1O2 generation capacity of these Pc derivatives, as determined with the indirect method of the 9,10-dimethylanthracene (9,10-DMA) absorption decay, showed that the tetrasubstituted PS 1 (ΦΔ = 0.14) produced more 1O2 than the fluorinated octasubstituted PS 2 did (ΦΔ = 0.03) [39]. ## 2.2. Photodynamic Inactivation of Fusarium oxysporum Conidia Figure 2 shows the logarithmic reduction in the concentration of viable F. oxysporum conidia after 60 min of irradiation with artificial white light at a fluence rate of 135 mW·cm−2 (light dose of 486 J·cm−2) in the presence of PS 1 and PS 2 at 20, 40, and 60 μM. After 60 min of irradiation, the highest concentration (60 μM) of either PS 1 or PS 2 caused a significant reduction in the concentration of viable F. oxysporum conidia. At the highest tested concentration (60 µM), none of the PSs caused lethal damage in the absence of light (dark controls: DC PS 1 and DC PS 2, Figure 2). Moreover, PS 1 caused the complete inactivation of conidia (>5 log10 reduction) upon 60 min of irradiation at both the tested concentrations (40 and 60 µM). Within the light dose, a lower concentration of 20 μM of PS 1 caused only a 2 log10 reduction in conidia viability. PS 2 caused a steady decrease in the viability of conidia with increasing concentrations: reduction of 2, 3, and 5 log10 with 20, 40, or 60 µM, respectively, at 60 min of light irradiation. To understand the inactivation kinetics with PS 1 or 2, the viability of conidia after 30 and 60 min of irradiation with white light was determined as represented in Figure 3. PS 1 was more effective than PS 2, as either 40 or 60 μM caused the complete inactivation of F. oxysporum conidia within 30 min of irradiation. On the other hand, PS 2, at the highest tested concentration (60 μM), showed a reduction of approximately 2 logs in the concentration of viable conidia after 30 min of irradiation, and complete inactivation after 60 min. No significant inactivation was evident with 40 μM of PS 2 after 30 min of irradiation, but after 60 min of irradiation, a reduction of 2 logs was observed. ## 3. Discussion The development of the resistance of conidia to chemical and physical fungal agents motivated research for new effective, sustainable, and environmentally friendly methods for their control, such as the aPDT approach. The structure of PS molecules is one of the major determinants of the efficiency of photosensitization [37,44,45,46]. In particular, positive charges are essential to make the PSs more water-soluble and attain the efficient photosensitization of fungal targets [36,37,47]. For this purpose, quaternized ammonium tetra- and octasubstituted phthalocyanines 1 (Figure 4—green) and 2 (Figure 4—blue) were synthesized [39] and tested against the conidia of Fusarium oxysporum, taken as the fungal pathogen model. aPDT efficiency was estimated as the logarithmic reduction in the concentration of viable conidia (Figure 2) and variation in the concentration of viable conidia (Figure 3). Irradiation in the absence of PS did not induce a decrease in the concentration of viable F. oxysporum conidia; similarly, no decrease was observed in the absence of light and the presence of PS (LC, DC 1 and DC 2, Figure 2). After 60 min of irradiation (light dose of 486 J·cm−2), with concentrations of 5 and 10 µM, the inactivation caused by either PS 1 or PS 2 was negligible (<1 log10; Figure S4). In order to attain complete inactivation, higher concentrations of PS (20, 40, and 60 μM) were tested. In this case, with the same irradiation time and 40 or 60 μM of PS 1, it was possible to reduce the concentration of viable spores to the method’s quantification limit, taking this result as complete inactivation (Figure 2). With 20 μM and the equivalent energy dose, the inactivation corresponded to a 2 log10 reduction in the concentration of viable conidia. In the case of PS 2 (Figure 2), an even higher concentration may be required, since the concentrations of 20, 40, or 60 μM caused reductions of 2, 3, and 5 log10 in the concentration of viable conidia, respectively, with a light dose of 486 J·cm−2. Overall, our results confirmed that ammonium PSs above concentrations of 40 μM could efficiently inactivate F. oxysporum conidia. In PS 1, despite being considered an antimicrobial agent, the introduction of ammonium substituents to the β-position seemed to reduce the antifungal activity when compared with phthalocyanine with the same group in the α-position (complete inactivation of C. albicans with 1.46 μM and a light dose of 27 J·cm−2) [48]. When compared to other ammonium phthalocyanines with four and eight charges, the increase in charge number in PS 1 (12 charges) increased the minimal concentration needed to achieve a minimal 4 log10 inactivation of C. albicans (from 0.5, 1, 10, and 20 μM to 40 μM) with a light dose of 30 J·cm−2 [49]. On the other hand, the concentration of PS 2 (24 charges) needed to increase to 60 μM to achieve the same effect. However, when compared with other [50] ammonium phthalocyanines (100 μM, 10 J·cm−2), the synthesized PSs 1 and 2, tested in the present study, seemed more effective. This is the first study concerning the use of ammonium phthalocyanines against F. oxysporum conidia. In order to better compare the photoinactivation efficiency of PSs 1 and 2, the effect of the energy dose was assessed by testing two exposure periods: 30 min corresponding to an energy dose of 243 J·cm−2, and 60 min corresponding to 486 J.cm−2. The results show that the tetrasubstituted PS 1 caused the complete inactivation of F. oxysporum conidia (>5 log10 reduction, with 40 or 60 μM) with the lowest energy dose (243 J·cm−2); with the same energy dose, the octasubstituted PS 2 caused a 2 log10 reduction (60 μM). These differences in efficiency could have been due to the difference in 1O2 quantum yields (ΦΔ 1 > ΦΔ 2) [39]. So, the increase in the number of charges did not seem to inherently improve photodynamic efficiency. Besides the difference in 1O2 generation, PS 2 seemed to aggregate, as seen in the absorption spectrum (Figure 1); for that reason, a decrease in the photodynamic efficiency may have resulted. Experiments involving the inactivation of a Gram-negative bacterial model (Escherichia coli) also indicated that PS 1 was more effective than PS 2 [39]. Thus, although the photosensitization of fungal structures requires higher PS concentrations, the cellular targets of photosensitization with PS 1 may be sufficiently diverse to support the prospect of a broad-spectrum (multiorganism) phytosanitary approach applicable to plant nurseries. The effectiveness of the two PSs against pathogenic fungal microbes in vitro showed their promising application in an in vivo approach, such as the one presented by Plaetzer and coworkers, who used strawberry leaves and solar light as a green irradiation source [51]. ## 4.1. Synthesis and Photophysical Characterization of the Photosensitizers The structures of the cationic PSs with ammonium groups (1 and 2) are depicted in Figure 4. Phthalocyanines 1 and 2 were prepared according to previously described experimental procedures [39], using reagents with high-level purity (purchased from Merck, Steinheim, Germany). Analytical TLC was carried out on precoated silica gel sheets (Merck, 0.2 mm, Darmstadt, Germany). According to the literature, solvents were used as received or distilled and dried using standard procedures [52]. 1H and 19F NMR spectra were recorded on a Bruker Avance-300 spectrometer (Wissembourg, France) at 300.13 and 282.38 MHz, where tetramethylsilane (TMS) was used as an internal reference. Absorption and steady-state fluorescence spectra were recorded using a Shimadzu UV-2501PC (Shimadzu, Kyoto, Japan) and a Horiba Jobin-Yvon FluoroMax Plus spectrofluorometer (Horiba Ltd., Kisshoint, Japan), respectively. The absorption and emission spectra of PS 1 and 2 were measured in DMF in 1 × 1 cm quartz optical cells at 298.15 K. The fluorescence quantum yield (ΦF) of 1 and 2 were calculated in DMF by comparing the area below the corrected emission spectra using ZnPcF16 as the standard (ΦF = 0.04 in acetone) [53]. ## 4.2. Photosenstizer Stock Solutions The stock solutions of PS at 500 µM were prepared in DMF or dimethyl sulfoxide (DMSO) for photophysical analyses or photodynamic inactivation assays, respectively, stored in the dark and previously sonicated for 30 min to each assay. ## 4.3. Light Source All photodynamic inactivation assays were performed by exposing the samples and light controls to a white light (400–800 nm) from a compatible fiber-optic probe attached to a 150 W quartz/halogen lamp (model LC122, LumaCare™ MBG Technologies Inc., New Port Beach, CA, USA) with an irradiance of 135 mW·cm–2, measured with a Coherent FieldMaxII-Top energy meter combined with a Coherent PowerSensPS19Q energy sensor. ## 4.4. Preparation of Stock Suspensions of Fusarium oxysporum Conidia Cultures of Fusarium oxysporum grown, for 7 days at 25 °C in Potato Dextrose Agar (PDA, Merck, KGaA, Darmstadt, Germany), were used to prepare the conidia suspension as described in the literature [36]. The absence of hyphae in the suspensions was checked via light microscopy (Leitz Laborlux K, Ernst Leitz GmbH, Wetzlar, Germany). The concentration of the viable conidia was determined with the serial dilutions of an aliquot in phosphate saline buffer (PBS, pH 7.4) and spread-plated on Rose Bengal Chloramphenicol Agar (Merck, KGaA, Darmstadt, Germany). Colonies were counted after 2 days of incubation at 25 °C, and the concentration of conidia is expressed as colony forming units per milliliter (CFU·mL−1) of suspension. ## 4.5. Photodynamic Inactivation Assay The photoinactivation assays were performed on PBS suspensions containing approx. 4 × 105 CFU·mL−1 in the presence of final concentrations of 20, 40, or 60 μM of PS 1 or PS 2. The assays were carried out in 24-well plates with a final volume of 1.5 mL. Conidia suspensions were preincubated with the PS solutions in the dark for 30 min at room temperature under stirring. After this period, irradiation was conducted for 1 h of continuous exposure. During irradiation, the suspension was kept under stirring on melting ice to prevent heating. Aliquots of 100 μL were collected at the beginning ($t = 0$ min), in the middle ($t = 30$ min), and at the end of irradiation ($t = 60$ min), serially diluted in PBS and spread-plated on Rose Bengal Chloramphenicol Agar in triplicate for the determination of the concentration of viable spores. Colonies were counted in the most convenient dilution after 48 h incubation at 25 °C. The average of the colonies in the replicates was used to estimate the concentration of viable conidia in the suspension and is expressed as CFU·mL−1. Two controls were included in each irradiation experiment: a light control (LC) submitted to the same irradiation conditions as the samples but without PS, and a dark control (DC) containing 60 μM of PS but kept in the dark. Three independent assays were conducted for each PS. The inactivation efficiency was calculated as the logarithmic (log10) reduction in the concentration of viable F. oxysporum conidia during the period corresponding to the irradiation of each independent assay. ## 4.6. Statistical Analysis Statistical analysis was performed, and the significance of the F. oxysporum conidia inactivation was assessed via a two-way univariate analysis of variance (ANOVA) model with Turkey’s multiple-comparisons post hoc test. A value of $p \leq 0.05$ was considered significant. ## 5. Conclusions The relations between structural features and the efficiency of photosensitization of F. oxysporum conidia indicated that cationic tetrasubstituted PS 1 was more efficient than octasubstituted PS 2, most probably because the ability to generate 1O2 species was considerably higher in the former. In this study, doubling the number of charges did not improve the photoinactivation process. The obtained results support the prospects of using these cationic phthalocyanines as new phytosanitary drugs on the basis of the photodynamic control of fungal spores. ## References 1. van Diepeningen A.D., Feng P., Ahmed S., Sudhadham M., Bunyaratavej S., de Hoog G.S.. Spectrum of. Mycoses (2015) 58 48-57. DOI: 10.1111/myc.12273 2. Ortoneda M., Guarro J., Madrid M.P., Caracuel Z., Roncero M.I.G., Mayayo E., di Pietro A.. Infect. Immun. (2004) 72 1760-1766. DOI: 10.1128/IAI.72.3.1760-1766.2004 3. Gordon T.R., Swett C.L., Wingfield M.J.. Management of. Crop Prot. (2015) 73 28-39. DOI: 10.1016/j.cropro.2015.02.018 4. Wu C.Y., Chen G.S., Lan C.C.E.. Onychomycosis Caused by. Clin. Exp. Dermatol. (2009) 34 772-774. DOI: 10.1111/j.1365-2230.2009.03498.x 5. Guilhermetti E., Takahachi G., Shinobu C.S., Svidzinski T.I.E.. Int. J. Dermatol. (2007) 46 822-826. DOI: 10.1111/j.1365-4632.2007.03120.x 6. Nelson P.E., Dignani M.C., Anaissie E.J.. Taxonomy, Biology, and Clinical Aspects of Fusarium Species. Clin. Microbiol. Rev. (1994) 7 479-504. DOI: 10.1128/CMR.7.4.479 7. de Menezes H.D., Massola N.S., Flint S.D., Silva G.J., Bachmann L., Rangel D.E.N., Braga G.U.L.. Growth under Visible Light Increases Conidia and Mucilage Production and Tolerance to UV-B Radiation in the Plant Pathogenic Fungus. Photochem. Photobiol. (2015) 91 397-402. DOI: 10.1111/php.12410 8. Deng G.M., Yang Q.S., He W.D., Li C.Y., Yang J., Zuo C.W., Gao J., Sheng O., Lu S.Y., Zhang S.. Proteomic Analysis of Conidia Germination in. Appl. Microbiol. Biotechnol. (2015) 99 7189-7207. DOI: 10.1007/s00253-015-6768-x 9. Koike S.T., Gordon T.R.. Management of. Crop Prot. (2015) 73 67-72. DOI: 10.1016/j.cropro.2015.02.003 10. McGovern R.J.. Management of Tomato Diseases Caused by. Crop Prot. (2015) 73 78-92. DOI: 10.1016/j.cropro.2015.02.021 11. Amini J., Sidovich D.. The Effects of Fungicides on. J. Plant Prot. Res. (2010) 50 172-178. DOI: 10.2478/v10045-010-0029-x 12. Amarasinghe C.C., Tamburic-Ilincic L., Gilbert J., Brûlé-Babel A.L., Dilantha Fernando W.G.. Evaluation of Different Fungicides for Control of Fusarium Head Blight in Wheat Inoculated with 3ADON and 15ADON Chemotypes of. Can. J. Plant Pathol. (2013) 35 200-208. DOI: 10.1080/07060661.2013.773942 13. Ölmez H., Kretzschmar U.. Potential Alternative Disinfection Methods for Organic Fresh-Cut Industry for Minimizing Water Consumption and Environmental Impact. Food Sci. Technol. (2009) 42 686-693. DOI: 10.1016/j.lwt.2008.08.001 14. Lykogianni M., Bempelou E., Karamaouna F., Aliferis K.A.. Do Pesticides Promote or Hinder Sustainability in Agriculture? The Challenge of Sustainable Use of Pesticides in Modern Agriculture. Sci. Total Environ. (2021) 795 148625. DOI: 10.1016/j.scitotenv.2021.148625 15. Alves E., Faustino M.A.F., Neves M.G.P.M.S., Cunha Â., Nadais H., Almeida A.. Potential Applications of Porphyrins in Photodynamic Inactivation beyond the Medical Scope. J. Photochem. Photobiol. C Photochem. Rev. (2015) 22 34-57. DOI: 10.1016/j.jphotochemrev.2014.09.003 16. Mesquita M.Q., Menezes J.C.J.M.D.S., Neves M.G.P.M.S., Tomé A.C., Cavaleiro J.A.S., Cunha Â., Almeida A., Hackbarth S., Röder B., Faustino M.A.F.. Photodynamic Inactivation of Bioluminescent. Bioorg. Med. Chem. Lett. (2014) 24 808-812. DOI: 10.1016/j.bmcl.2013.12.097 17. Dai T., Fuchs B.B., Coleman J.J., Prates R.A., Astrakas C., St. Denis T.G., Ribeiro M.S., Mylonakis E., Hamblin M.R., Tegos G.P.. Concepts and Principles of Photodynamic Therapy as an Alternative Antifungal Discovery Platform. Front. Microbiol. (2012) 3 120. DOI: 10.3389/fmicb.2012.00120 18. de Menezes H.D., Tonani L., Bachmann L., Wainwright M., Braga G.Ú.L., von Zeska Kress M.R.. Photodynamic Treatment with Phenothiazinium Photosensitizers Kills Both Ungerminated and Germinated Microconidia of the Pathogenic Fungi Fusarium Oxysporum,. J. Photochem. Photobiol. B (2016) 164 1-12. DOI: 10.1016/j.jphotobiol.2016.09.008 19. Almeida J., Tomé J.P.C., Neves M.G.P.M.S.P.M.S., Tomé A.C., Cavaleiro J.A.S.S., Cunha Â., Costa L., Faustino M.A.F.F., Almeida A.. Photodynamic Inactivation of Multidrug-Resistant Bacteria in Hospital Wastewaters: Influence of Residual Antibiotics. Photochem. Photobiol. Sci. (2014) 13 626-633. DOI: 10.1039/c3pp50195g 20. Beirão S., Fernandes S., Coelho J., Faustino M.A.F.F., Tomé J.P.C.C., Neves M.G.P.M.S.P.M.S., Tomé A.C., Almeida A., Cunha A.. Photodynamic Inactivation of Bacterial and Yeast Biofilms with a Cationic Porphyrin. Photochem. Photobiol. (2014) 90 1387-1396. DOI: 10.1111/php.12331 21. Gomes M.C., Woranovicz-Barreira S.M., Faustino M.A.F., Fernandes R., Neves M.G.P.M.S., Tomé A.C., Gomes N.C.M., Almeida A., Cavaleiro J.A.S., Cunha Â.. Photodynamic Inactivation of. Photochem. Photobiol. Sci. (2011) 10 1735-1743. DOI: 10.1039/c1pp05174a 22. Rodrigues G.B., Dias-Baruffi M., Holman N., Wainwright M., Braga G.U.L.. In Vitro Photodynamic Inactivation of. Photodiagn. Photodyn. Ther. (2013) 10 141-149. DOI: 10.1016/j.pdpdt.2012.11.004 23. Almeida A., Faustino M.A.F., Tomé J.P.C.. Photodynamic Inactivation of Bacteria: Finding the Effective Targets. Future Med. Chem. (2015) 7 1221-1224. DOI: 10.4155/fmc.15.59 24. Lopes M.M., Bartolomeu M., Gomes A.T.P.C., Figueira E., Pinto R., Reis L., Balcão V.M., Faustino M.A.F., Neves M.G.P.M.S., Almeida A.. Antimicrobial Photodynamic Therapy in the Control of. Microorganisms (2020) 8. DOI: 10.3390/microorganisms8071022 25. Ndemueda A., Pereira I., Faustino M.A.F.. Cunha Photodynamic Inactivation of the Phytopathogenic Bacterium. Lett. Appl. Microbiol. (2020) 71 420-427. DOI: 10.1111/lam.13350 26. Ambrosini V., Issawi M., Sol V., Riou C.. Photodynamic Inactivation of Botrytis Cinerea by an Anionic Porphyrin: An Alternative Pest Management of Grapevine. Sci. Rep. (2020) 10 17438. DOI: 10.1038/s41598-020-74427-9 27. Menezes J.C.J.M.D.S., Faustino M.A.F., de Oliveira K.T., Uliana M.P., Ferreira V.F., Hackbarth S., Röder B., Teixeira Tasso T., Furuyama T., Kobayashi N.. Synthesis of New Chlorin E6 Trimethyl and Protoporphyrin IX Dimethyl Ester Derivatives and Their Photophysical and Electrochemical Characterizations. Chem. Eur. J. (2014) 20 13644-13655. DOI: 10.1002/chem.201403214 28. Garcia M., David B., Sierra-Garcia I.N., Faustino M.A.F., Alves A., Esteves A.C., Cunha A.. Photodynamic Inactivation of. Biol. Lett. (2021) 17 20200820. DOI: 10.1098/rsbl.2020.0820 29. Sharma I., Sharma V.K., Shah M., Parmar S., Kumar A.. Phytopathogenic Fungi and Their Biocontrol Applications. Fungi Bio-Prospects in Sustainable Agriculture, Environment and Nano-Technology (2021) 155-188 30. Almeida A., Cunha A., Faustino M.A.F., Tomé A.C., Neves M.G.P.M.S., Hamblin M.R., Jori G.. Porphyrins as Antimicrobial Photosensitizing Agents. Photodynamic Inactivation of Microbial Pathogens: Medical Environmental Applications (2011) 83-160 31. Ribeiro C.P.S., Gamelas S.R.D., Faustino M.A.F., Gomes A.T.P.C., Tomé J.P.C., Almeida A., Lourenço L.M.O.. Unsymmetrical Cationic Porphyrin-Cyclodextrin Bioconjugates for Photoinactivation of. Photodiagn. Photodyn. Ther. (2020) 31 101788. DOI: 10.1016/j.pdpdt.2020.101788 32. Santos I., Gamelas S.R.D., Vieira C., Faustino M.A.F., Tomé J.P.C., Almeida A., Gomes A.T.P.C., Lourenço L.M.O.. Pyrazole-Pyridinium Porphyrins and Chlorins as Powerful Photosensitizers for Photoinactivation of Planktonic and Biofilm Forms of. Dyes Pigm. (2021) 193 109557. DOI: 10.1016/j.dyepig.2021.109557 33. Calmeiro J.M.D., Gamelas S.R.D., Gomes A.T.P.C., Faustino M.A.F., Neves M.G.P.M.S., Almeida A., Tomé J.P.C., Lourenço L.M.O.. Versatile Thiopyridyl/Pyridinone Porphyrins Combined with Potassium Iodide and Thiopyridinium/Methoxypyridinium Porphyrins on. Dyes Pigm. (2020) 181 108476. DOI: 10.1016/j.dyepig.2020.108476 34. Ribeiro C.P.S., Faustino M.A.F., Almeida A., Lourenço L.M.O.. The Antimicrobial Photoinactivation Effect on. Microorganisms (2022) 10. DOI: 10.3390/microorganisms10040718 35. Calmeiro J.M.D., Dias C.J., Ramos C.I.V., Almeida A., Tomé J.P.C., Faustino M.A.F., Lourenço L.M.O.. Comparative Photodynamic Inactivation of Bioluminescent. Dyes Pigm. (2020) 173 107410. DOI: 10.1016/j.dyepig.2019.03.021 36. Sierra-Garcia I.N., Cunha Â., Lourenço L.M.O.. J. Photochem. Photobiol. A Chem. (2022) 432 114081. DOI: 10.1016/j.jphotochem.2022.114081 37. Ribeiro C.P.S., Lourenço L.M.O.. Overview of Cationic Phthalocyanines for Effective Photoinactivation of Pathogenic Microorganisms. J. Photochem. Photobiol. C Photochem. Rev. (2021) 48 100422. DOI: 10.1016/j.jphotochemrev.2021.100422 38. Lourenço L.M.O., Rocha D.M.G.C., Ramos C.I.V., Gomes M.C., Almeida A., Faustino M.A.F., Almeida Paz F.A., Neves M.G.P.M.S., Cunha Â., Tomé J.P.C.. Photoinactivation of Planktonic and Biofilm Forms of. ChemPhotoChem (2019) 3 251-260. DOI: 10.1002/cptc.201900020 39. Gamelas S.R.D., Gomes A.T.P.C., Faustino M.A.F., Tomé A.C., Tomé J.P.C., Almeida A., Lourenço L.M.O.. Photoinactivation of. ACS Appl. Bio Mater. (2020) 3 4044-4051. DOI: 10.1021/acsabm.0c00128 40. Lourenço L.M.O., Sousa A., Gomes M.C., Faustino M.A.F., Almeida A., Silva A.M.S., Neves M.G.P.M.S., Cavaleiro J.A.S., Cunha Â., Tomé J.P.C.. Inverted Methoxypyridinium Phthalocyanines for PDI of Pathogenic Bacteria. Photochem. Photobiol. Sci. (2015) 14 1853-1863. DOI: 10.1039/c5pp00145e 41. Calzavara-Pinton P.G., Venturini M., Sala R.. A Comprehensive Overview of Photodynamic Therapy in the Treatment of Superficial Fungal Infections of the Skin. J. Photochem. Photobiol. B (2005) 78 1-6. DOI: 10.1016/j.jphotobiol.2004.06.006 42. Al-Asmari F., Mereddy R., Sultanbawa Y.. A Novel Photosensitization Treatment for the Inactivation of Fungal Spores and Cells Mediated by Curcumin. J. Photochem. Photobiol. B (2017) 173 301-306. DOI: 10.1016/j.jphotobiol.2017.06.009 43. Sesalan B.Ş., Koca A., Gül A.. Water Soluble Novel Phthalocyanines Containing Dodeca-Amino Groups. Dyes Pigm. (2008) 79 259-264. DOI: 10.1016/j.dyepig.2008.03.006 44. Hamblin M.R., Sabino C.P.. Photosensitizers. Photodynamic Therapy in Veterinary Medicine: From Basics to Clinical Practice (2016) 25-44 45. Abrahamse H., Hamblin M.R.. New Photosensitizers for Photodynamic Therapy. Biochem. J. (2016) 473 347-364. DOI: 10.1042/BJ20150942 46. Ion R.-M., Ylmaz Y.. Porphyrins and Phthalocyanines: Photosensitizers and Photocatalysts. Phthalocyanines and Some Current Applications (2017) Volume 9 189-221 47. Costa D.C.S., Gomes M.C., Faustino M.A.F., Neves M.G.P.M.S., Cunha Â., Cavaleiro J.A.S., Almeida A., Tomé J.P.C.. Comparative Photodynamic Inactivation of Antibiotic Resistant Bacteria by First and Second Generation Cationic Photosensitizers. Photochem. Photobiol. Sci. (2012) 11 1905-1913. DOI: 10.1039/c2pp25113b 48. Li X.S., Guo J., Zhuang J.J., Zheng B.Y., Ke M.R., Huang J.D.. Highly Positive-Charged Zinc(II) Phthalocyanine as Non-Aggregated and Efficient Antifungal Photosensitizer. Bioorg. Med. Chem. Lett. (2015) 25 2386-2389. DOI: 10.1016/j.bmcl.2015.04.004 49. Dei D., Chiti G., de Filippis M.P., Fantetti L., Giuliani F., Giuntini F., Soncin M., Jori G., Roncucci G.. Phthalocyanines as Photodynamic Agents for the Inactivation of Microbial Pathogens. J. Porphyr. Phthalocyanines (2006) 10 147-159. DOI: 10.1142/S1088424606000181 50. Aroso R.T., Calvete M.J.F., Pucelik B., Dubin G., Arnaut L.G., Pereira M.M., Dąbrowski J.M.. Photoinactivation of Microorganisms with Sub-Micromolar Concentrations of Imidazolium Metallophthalocyanine Salts. Eur. J. Med. Chem. (2019) 184 111740. DOI: 10.1016/j.ejmech.2019.111740 51. Hamminger C., Glueck M., Fefer M., Ckurshumova W., Liu J., Tenhaken R., Plaetzer K.. Photodynamic Inactivation of Plant Pathogens Part II: Fungi. Photochem. Photobiol. Sci. (2022) 21 195-207. DOI: 10.1007/s43630-021-00157-0 52. Armarego W.L.F.. Purification of Laboratory Chemicals (2017) 53. Beveridge A.C., Bench B.A., Gorun S.M., Diebold G.J.. Evaluation of Photodynamic Therapy Agents through Transient Grating Measurements. J. Phys. Chem. A (2003) 107 5138-5143. DOI: 10.1021/jp022646e
--- title: Antioxidant Effect of a Dietary Supplement Containing Fermentative S-Acetyl-Glutathione and Silybin in Dogs with Liver Disease authors: - Elisa Martello - Francesca Perondi - Donal Bisanzio - Ilaria Lippi - Giorgia Meineri - Valeria Gabriele journal: Veterinary Sciences year: 2023 pmcid: PMC9966841 doi: 10.3390/vetsci10020131 license: CC BY 4.0 --- # Antioxidant Effect of a Dietary Supplement Containing Fermentative S-Acetyl-Glutathione and Silybin in Dogs with Liver Disease ## Abstract ### Simple Summary The use of a dietary supplement containing S-acetyl-glutathione (SAG), silybin, and other antioxidant ingredients increased the level of erythrocyte glutathione (GSH) and improved key biochemical parameters in dogs with liver disease. ### Abstract Oxidative stress is often involved in liver disease progression. Liver is the primary site for the synthesis of glutathione (GSH), the major intracellular antioxidant. GSH erythrocyte concentration can decrease in case of liver damage. So, the use of food supplements with antioxidant capacity has been reported in the veterinary literature. In this case–control study, we tested a new supplement containing S-acetyl-glutathione (SAG), silybin, and other antioxidant ingredients in dogs affected by liver disease. After two weeks of supplement administration, we were able to report a significant increase in the level of erythrocyte GSH in the treated (TRT) group, nearly reaching the physiological limit at the end of the study. In addition, most of the key liver parameters are significantly reduced in the TRT group by the end of the trial. The results of this study support the effectiveness of the tested complementary feed, which may be helpful in managing dogs with liver conditions. ## 1. Introduction The liver plays a role in regulating the endogenous antioxidant status, being the primary site for the synthesis of glutathione (GSH), the major intracellular antioxidant [1]. GSH is a thiol tripeptide formed by cysteine, glycine and glutamate. GSH is synthesized in all mammalian cells (primarily formed and stored in the liver) and it has different functions, such as the maintenance of the cellular redox state [2,3,4]. The important role of GSH as major antioxidant is to allow detoxification from the products of the aerobic metabolism (such as superoxide, hydrogen peroxide, and toxic oxygen radicals) and to protect tissues from cell damage [4,5]. GSH deficiency has been reported in people affected by many diseases, including liver disease, diabetes mellitus, renal failure, sepsis, and acute pancreatitis [3]. In veterinary medicine, dogs and cats can have decreased liver and blood GSH concentrations in cases of liver disease [1,5,6]. Consequently, therapeutic interventions aiming to overcome the GSH deficiency are recommended by clinicians. The use of several food supplements with antioxidant activities (such as S-adenosyl-methionine (SAME), curcumin, phosphatidylcholine, ursodeoxycholic acid, glycine, N-acetylcysteine (NAC), and silymarin) has been reported in the veterinary literature [7,8,9]. In particular, silymarin is a relevant natural ingredient with hepatoprotective function and antioxidant, immunomodulatory, anti-inflammatory, and choleretic proprieties in both human and veterinary medicine [7,10,11]. Silymarin is a flavonoid extracted from the milk thistle Silybum marianum, and it is composed of multiple flavonolignans, including the most active constituent—silybin [12]. In human medicine a valid approach to increase the endogenous GSH is the use of the S-acetyl-glutathione (SAG), which is a GSH precursor and originates from the fermentation of Saccharomyces cerevisiae [13]. SAG is also a potential hepatoprotective agent preventing oxidative liver damage [13]. However, no studies on dogs and cats regarding the use of SAG have been reported yet. This case–control trial has been performed to test the hepatoprotective effect and the variation of erythrocyte GSH level when administering a new dietary supplement containing SAG and silybin associated with other well-known antioxidant components (orange bioflavonoid, vitamin B2, vitamin B12, vitamin E) in dogs with a diagnosis of liver disease. ## 2. Materials and Methods This case–control study included a total of 24 adult dogs with a diagnosis of liver disease (cholangitis/cholangiohepatitis). The inclusion of the subjects was based on clinical and ultrasonographic signs, hepatic needle aspirate cytology, and hematobiochemical analysis. Dogs were excluded who presented other concomitant metabolic diseases or disorders potentially impacting the liver function (such as diabetes mellitus, gastritis, inflammatory bowel disease, chronic kidney disease, Cushing’s syndrome). Dogs administered with hepatoprotective products during the 30 days before the enrolment were also excluded. All dogs were treated at the beginning of the study with antibiotics (oral amoxicillin-clavulanate BID 12.5–25 mg/kg), anti-inflammatory drugs (oral prednisolone SID 0.5 mg/Kg), and intravenous fluid therapy, as a supportive care for rehydration and the correction of electrolyte concentration if needed. All the animals enrolled had a diet based on Vetsolution monge epatic ($50\%$) and Vetsolution digest ($50\%$) from at least 14 days before starting the trial. Twelve dogs were randomly assigned to the treatment group (TRT; male $$n = 5$$, female $$n = 7$$, mean age 6.8 years) and received the tested supplement (Table 1) at a dose of one tablet/15 kg BW, while the other 12 dogs were designated as the control group (CTR; male $$n = 6$$, female $$n = 6$$, mean age 6.7 years) and did not receive the supplemented diet. For determining the concentration of the erythrocyte GSH, heparinized whole blood aliquots were frozen at −80 °C immediately after collection and then analyzed using a commercial assay kit (Ransel e Ransod, Randox Laboratories Ltd., Crumlin, UK) on an automated analyzer (RX Daytona™; Randox Laboratories Ltd., Crumlin, UK). The effect of the supplement on the erythrocyte GSH and other blood parameters (total proteins (TP), albumins (ALB), alanine transaminases (ALT), alanine aminotransferases (AST), alkaline phosphatases (ALP), gamma–glutamyl transferase (GGT), bilirubin (BIL), and triglycerides (TRI)) was tested using a regression model built as a generalized linear mixed model (GLMM) with Gaussian likelihood (R software). The model included a non-linear variable describing the link between each time point within and between the CTR and TRT group, sex, body weight (BW), and age. The model account accounts for repeated measurements (random effect) and the heterogeneity of individuals. ## 3. Results and Discussion The product under study was well tolerated by all the animals. During the trial no adverse effects (such as vomiting and diarrhea) were observed, confirming the safety of the product, and no dogs were excluded during the trial. No significant alteration of blood counts (values within normal ranges) and no evidence of other diseases was reported based on the biochemical parameters evaluated at the baseline (Supplementary Table S1). Results from the performed statistical analysis show that the supplement had no or limited effects on some of the biochemical values relevant to monitor liver disease (TP, ALB, GLU, TRI, and PCR) (Table 2). Interestingly, most of the key liver parameters (ALT, AST, ALP, GGT, and BIL) are significantly reduced from T4 in the TRT group but not in the CTR group, as showed by the model results (Table 2). Indeed, the hepatoprotective activity of our tested supplement could be the result of the synergic effect of the included ingredients, the therapeutic proprieties of which had also pointed out in other research studies in animals. First of all, the efficacy of a supplement based on silybin with hepatoprotective proprieties was reported in a previous study on cats [12]. Then, a recently published review confirmed the reduction of ALT and GPT enzymes in dogs with hepatopathy following the administration of silybin [7]. In addition, the use of silybin for the treatment of hepatobiliary disease has also been reported in a study on cats, highlighting its antioxidant, anti-inflammatory, and anti-fibrotic capacities [12]. Marchegiani and colleagues [7] also reported a few case studies in which a combination of different active ingredients were selected and used (including silybin) to treat liver disease in dogs and cats. Moreover, Di Paola and colleagues [13] described the effect of another of our ingredients, SAG, as its administration was functional to attenuating liver damage, to reduce liver fibrosis, and to improve hepatic function with a decrease in ALT and AST levels in humans. It has to be noted that the use of SAG derived from the fermentation of Saccharomyces cerevisiae in our supplement is a novelty in veterinary medicine as, to our knowledge, no data on its use in companion animals with liver conditions have yet been published. SAG not only has hepatoprotective proprieties, but it is also involved in GSH regulation. It is a GSH precursor and is more stable than GSH itself in plasma, being taken up directly by cells and later converted to GSH [13,14]. The oral administration of GSH itself is not significantly enhanced in plasma, while SAG is taken up by cells and later converted to GSH [13], and its absorption happens via the intestinal wall [13,15]. GSH is an important endogenous antioxidant that has been demonstrated to be significantly lower in the livers of dogs and cats with hepatic diseases [1,5]. This is in agreement with our findings as all our enrolled dogs showed low erythrocyte GSH levels (under the minimum physiological level, 300 Ug/Hb) at the beginning of the trial. In our study, results from the GLMM model showed a significant increase in erythrocyte GSH levels, even from T2 in the TRT group, and nearly reached the minimum physiological limit (300 Ug/Hb) at the end of the treatment. This could be the result of a joint effect of the previously mentioned SAG and of silybin, vitamin B2, vitamin B12, and vitamin E, which could all have potentially contributed to the increase in erythrocyte GSH [16,17,18,19]. In particular, silymarin, which has silybin as one of its flavonolignans, induced the hepatic synthesis of GSH by increasing cysteine availability in mice [16]. Moreover, vitamin B2 was demonstrated to play a role in the activities of GSH reductase and related antioxidant enzymes [17]. Regarding vitamin B12, it is a cofactor for the enzyme GSH reductase, a catalyst for converting the oxidized form of GSH back to its reduced active form [18]. Finally, Vitamin E was demonstrated to increase cellular GSH concentrations, as it displays radical-scavenging antioxidant activity [19]. This study has some limitations. First, the overall number of dogs included is relatively small and no placebo was given in the control group, thus affecting the quality of the study design. ## 4. Conclusions To our knowledge, this is the first study on dogs with liver disease that tested a dietary supplement with a combination of natural antioxidants (SAG, silybin, orange bioflavonoid, vitamin B2, vitamin B12, and vitamin E) and was able to show hepatoprotective effects and a significant increase in erythrocyte GSH levels. The tested complementary feed may represent an effective aid in managing liver disease in dogs. ## References 1. Heffernan C.B., Ekena J., Dowling S., Pinkerton M.E., Viviano K.. Biomarkers of oxidative stress as an assessment of the redox status of the liver in dogs. J. Vet. Intern. Med. (2019) 33 611-617. DOI: 10.1111/jvim.15443 2. Lushchak V.I.. Glutathione Homeostasis and Functions: Potential Targets for Medical Interventions. J. Amin. Acid. (2012) 2012 736837. DOI: 10.1155/2012/736837 3. Viviano K.R., Lavergne S.N., Goodman L., Vanderwielen B., Grundahl L., Padilla M., Trepanier L.A.. Glutathione, Cysteine, and Ascorbate Concentrations in Clinically Ill Dogs and Cats. J. Vet. Intern. Med. (2009) 23 250-257. DOI: 10.1111/j.1939-1676.2008.0238.x 4. Woolcock A.D., Serpa P.B.S., Santos A.P., Christian J.A., Moore J.E.. Reactive oxygen species, glutathione, and vitamin E concentrations in dogs with hemolytic or nonhemolytic anemia. J. Vet. Intern. Med. (2020) 34 2357-2364. DOI: 10.1111/jvim.15926 5. Center S.A., Warner K.L., Erb H.N.. Liver Glutathione Concentrations in Dogs and Cats with Naturally Occurring Liver Disease. Am. J. Vet. Res. (2002) 63 1187-1197. DOI: 10.2460/ajvr.2002.63.1187 6. Spee B., Arends B., van den Ingh T.S.G.A.M., Penning L.C., Rothuizen J.. Copper Metabolism and Oxidative Stress in Chronic Inflammatory and Cholestatic Liver Diseases in Dogs. J. Vet. Intern. Med. (2006) 20 1085-1092. DOI: 10.1892/0891-6640(2006)20[1085:cmaosi]2.0.co;2 7. Marchegiani A., Fruganti A., Gavazza A., Mangiaterra S., Candellone A., Fusi E., Rossi G., Cerquetella M.. Evidences on Molecules Most Frequently Included in Canine and Feline Complementary Feed to Support Liver Function. Vet. Med. Int. (2020) 2020 9185759. DOI: 10.1155/2020/9185759 8. Kumar P., Osahon O., Sekhar R.. GlyNAC (Glycine and N-Acetylcysteine) Supplementation in Mice Increases Length of Life by Correcting Glutathione Deficiency, Oxidative Stress, Mitochondrial Dysfunction, Abnormalities in Mitophagy and Nutrient Sensing, and Genomic Damage. Nutr. J. (2022) 14. DOI: 10.3390/nu14051114 9. Viviano K.R., VanderWielen B.. Effect of N-acetylcysteine supplementation on intracellular glutathione, urine isoprostanes, clinical score, and survival in hospitalized ill dogs. J. Vet. Intern. Med. (2013) 27 250-258. DOI: 10.1111/jvim.12048 10. Serçe A., Toptancı B.Ç., Tanrıkut S.E., Altaş S., Kızıl G., Kızıl S., Kızıl M.. Assessment of the Antioxidant Activity of. Food Technol. Biotechnol. (2016) 54 455-461. DOI: 10.17113/ftb.54.04.16.4323 11. Guerrini A., Tedesco D.E.A.. Use of Milk Thistle in Farm and Companion Animals: A Review. Planta Med. (2022) 36302565. DOI: 10.1055/a-1969-2440 12. Biasbetti E., Martello E., Cocca T., Bigliati M., Mioletti S., Bruni N., Capucchio M.T.. Methionine Fermented Associated with Silybum Marianum is a Potential Hepatoprotective and Antioxidant in Cats with In ammatory Liver Disease. Int. J. Anim. Sci. (2018) 2 1018 13. Di Paola R., Modafferi S., Siracusa R., Cordaro M., D’Amico R., Ontario M.L., Interdonato L., Salinaro A.T., Fusco R., Impellizzeri D.. S-Acetyl-Glutathione Attenuates Carbon Tetrachloride-Induced Liver Injury by Modulating Oxidative Imbalance and Inflammation. Int. J. Mol. Sci. (2022) 23. DOI: 10.3390/ijms23084429 14. Fanelli S., Francioso A., Cavallaro R.A., d’Erme M., Putignano P., Miraglia N., Mosca L.. Oral Administration of S-acetyl-glutathione: Impact on the Levels of Glutathione in Plasma and in Erythrocytes of Healthy Volunteers. Int. J. Clin. Nutr. Diet. (2018) 4 134. DOI: 10.15344/2456-8171/2018/134 15. Santos L.O., Gonzales T.A., Úbeda B.T., Monte Alegre R.. Influence of culture conditions on glutathione production by Saccharomyces cerevisiae. Appl. Microbiol. Biotechnol. (2007) 77 763-769. DOI: 10.1007/s00253-007-1211-6 16. Kwon D.Y., Jung Y.S., Kim S.J., Kim Y.S., Choi D.W., Kim Y.C.. Alterations in sulfur amino acid metabolism in mice treated with silymarin: A novel mechanism of its action involved in enhancement of the antioxidant defense in liver. Planta Med. (2013) 79 997-1002. DOI: 10.1055/s-0032-1328704 17. Ashoori M., Saedisomeolia A.. Riboflavin (vitamin B. Br. J. Nutr. (2014) 111 1985-1991. DOI: 10.1017/S0007114514000178 18. Van de Lagemaat E.E., de Groot L.C.P.G.M., van den Heuvel E.G.H.M.. Vitamin B. Nutr. J. (2019) 11. DOI: 10.3390/nu11020482 19. Jain S.K., McVie R., Smith T.. Vitamin E supplementation restores glutathione and malondialdehyde to normal concentrations in erythrocytes of type 1 diabetic children. Diab. Care (2000) 23 1389-1394. DOI: 10.2337/diacare.23.9.1389
--- title: Multi-Omics Analyses Reveal the Mechanisms of Early Stage Kidney Toxicity by Diquat authors: - Huazhong Zhang - Jinsong Zhang - Jinquan Li - Zhengsheng Mao - Jian Qian - Cheng Zong - Hao Sun - Beilei Yuan journal: Toxics year: 2023 pmcid: PMC9966843 doi: 10.3390/toxics11020184 license: CC BY 4.0 --- # Multi-Omics Analyses Reveal the Mechanisms of Early Stage Kidney Toxicity by Diquat ## Abstract Diquat (DQ), a widely used bipyridyl herbicide, is associated with significantly higher rates of kidney injuries compared to other pesticides. However, the underlying molecular mechanisms are largely unknown. In this study, we identified the molecular changes in the early stage of DQ-induced kidney damage in a mouse model through transcriptomic, proteomic and metabolomic analyses. We identified 869 genes, 351 proteins and 96 metabolites that were differentially expressed in the DQ-treated mice relative to the control mice ($p \leq 0.05$), and showed significant enrichment in the PPAR signaling pathway and fatty acid metabolism. Hmgcs2, Cyp4a10, Cyp4a14 and Lpl were identified as the major proteins/genes associated with DQ-induced kidney damage. In addition, eicosapentaenoic acid, linoleic acid, palmitic acid and (R)-3-hydroxybutyric acid were the major metabolites related to DQ-induced kidney injury. Overall, the multi-omics analysis showed that DQ-induced kidney damage is associated with dysregulation of the PPAR signaling pathway, and an aberrant increase in Hmgcs2 expression and 3-hydroxybutyric acid levels. Our findings provide new insights into the molecular basis of DQ-induced early kidney damage. ## 1. Introduction Pesticides are the leading cause of poisoning-related accidental deaths in China. Following the discontinuation of paraquat, diquat (DQ) has become the preferred bipyridyl herbicide. However, cases of DQ poisoning have continued to increase in recent years, and the predominant route of exposure is the gastrointestinal tract [1]. The kidney is the main excretory organ as well as the primary target of DQ, and the toxic effects of the latter mainly involve the renal tubules, eventually leading to acute kidney injury (AKI) [2]. The incidence of AKI in patients with DQ poisoning is $73.3\%$, which is significantly higher compared to that caused by paraquat or other pesticides. Previous studies have shown that DQ is selectively toxic to the kidneys, and has a similar chemical structure to that of the highly nephrotoxic orellanine [2]. Renal tubular dysfunction is the initial manifestation of DQ toxicity [3], and obvious renal tubular epithelial cell damage has been observed during autopsy [4]. The offspring of DQ-intoxicated rats exhibit renal duct damage. Furthermore, the prognosis of patients with DQ poisoning is closely related to AKI, which is usually reversible in the early stage. However, given the narrow time window for treatment, the incidence of endpoint events (death or uremia) exceeds $30\%$. Therefore, early detection and prevention of AKI are crucial in cases of DQ poisoning [5,6,7]. The clinical diagnosis of AKI is currently based on elevated blood creatinine (Scr) and blood urea nitrogen (BUN), along with low urine output [7]. However, the rise in Scr and BUN is increased when renal function has already declined by nearly $50\%$, while the urine output is susceptible to multiple factors such as diuretics and blood volume. Moreover, Scr and BUN are easily cleared by continuous renal replacement therapy (CRRT) and the urine volume varies with the ultrafiltration volume of CRRT. Thus, none of these indicators can accurately reflect the changes in renal function during CRRT [8]. Therefore, it is unclear whether using high Scr and oliguria as the clinical criteria for the initiation of CRRT delays the clearance of nephrotoxic substances such as DQ, and whether hemoperfusion (HP) combined with early CRRT improves prognosis [2,8]. Therefore, it is crucial to identify novel biomarkers and effector molecules for early detection and progression of kidney injury, and to guide hemodialysis treatment. In this study, we used integrated metabolomics, transcriptomics and proteomics to explore the molecular mechanisms underlying DQ-induced nephrotoxicity at the very early stage. Based on multi-omics analyses, we found that DQ induced aberrant gene expression at the mRNA, protein, and metabolite levels. Our findings provide novel insights into DQ-induced kidney injury and identify novel biomarkers. ## 2.1. Animals and Chemical Reagents Treatments Male C57BL/6 J mice aged 28 weeks and weighing 25–30 g were bought from Nanjing Medical University (NYD-L-2020082601). The mice were kept in a specialized pathogen-free environment (22–26 °C, $40\%$–$60\%$ humidity, and 12 h light/dark cycles) with food and water provided ad libitum. The feed used in this experiment meets the national standard. The feed mainly contains energy, protein, fat, amino acid, minerals, etc. All mice were given the same food. The mice were randomly divided into the control, low-dose DQ (200 mg/kg) and high-dose DQ (350 mg/kg) groups after one week of acclimatization ($$n = 30$$ per group). DQ and saline (control) were administered via the intragastric route. The mice were euthanized on days 1, 3 and 7 after induction, and kidney tissue samples were collected from 10 mice of each group. Ten kidney samples were used for metabolomics analysis, three were used for proteomics analysis, and three for transcriptomic analysis. Diquat (DQ) was purchased from Aladdin (D101258-100 mg). ## 2.2. Histopathologic Examination The kidney tissues were fixed in $4\%$ paraformaldehyde for 24 h, dehydrated in an ethanol gradient and embedded in paraffin. The paraffin blocks were cut into 5 µm-thick slices, which were stained using hematoxylin and eosin (H&E). Instrument information: Tissue-tekvip6 automatic tissue processor (Sakura, Japan), HistoStar tissue burying machine (Thermo, US), Thermo Finesse E+ paraffin microtome (Thermo, US), Gemini AS automatic dyeing machine (Thermo, US), Olympus BX53 optical microscope (Olympus, Japan), and DP72 image analysis system (Olympus, Japan). ## 2.3. Transcriptome Analysis RNA sequencing (RNA-seq) was performed on three biological replicates of the DQ-treated and control group kidney tissues by Biotree Biotech Co., Ltd. (Shanghai, China). Briefly, total RNA was extracted and reverse transcribed, and the double-stranded cDNA was used to construct libraries. After quality control, the libraries are pooled and sequenced on the Illumina Novaseq 6000 platform (Thermo, US). The clean reads were filtered from the raw sequencing data after checking for the sequencing error rate and the distribution of GC content. The gene expression levels were calculated as the number of fragments per kilobase of transcript per million reads (FPKM). The expression matrix of all samples was generated, and differentially expressed genes (DEGs) between the control and DQ-treated samples were screened using the edgeR program with Padj < 0.05 as the criterion. The DEGs were then functionally annotated by gene ontology (GO) analysis in terms of molecular functions (MF), biological processes (BP) and cellular components (CC), as well as Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment analyses using the clusterProfiler (http://www.bioconductor.org/packages/release/bioc/html/clusterProfiler.html) program, (accessed on 31 December 2021). The GO terms related to molecular function, biological process and cellular component were analyzed. ## 2.4. Proteomics Analysis Total protein was extracted from the kidney tissues of three biological replicates from the control and DQ-treated groups, quantified and stored at −80 °C. Proteomic sequencing and analysis were conducted by Biotree Biotech Co., Ltd. (Shanghai, China). Briefly, the extracted proteins were first quantified by the BCA assay, precipitated using acetone, and then subjected to reduction, alkylation, digestion, TMT labeling, SDC cleanup, peptide desalting and high-pH pre-fractionation. For nanoLC–MS/MS analysis, 2 µg total peptides from each sample was separated and analyzed using a nano-UPLC (EASY-nLC1200) coupled to Orbitrap Exploris 480 (Thermo Fisher Scientific) with a nano-electrospray ion source. Data-dependent acquisition (DDA) was performed in profile and the positive mode with Orbitrap analyzer for 90 min. The Tandem Mass Tag (TMT) was used to identify the proteins and screen for unique peptides with p-Value < 0.05 (Student’s t test) and fold change > 1.5 as the criteria. The proteins were subjected to principal component analysis (PCA), volcano plot analysis, hierarchical clustering analysis, GO and KEGG analyses, and protein–protein interaction (PPI) network analysis. ## 2.5. Untargeted LC–MS Metabolomics Analysis The kidney tissue samples from the control and DQ-treated groups (10 biological replicates per group) were prepared as previously described [9]. Metabolomic sequencing and analysis were performed by Biotree Biotech Co. Ltd. (Shanghai, China). The metabolic profiles were acquired using Quadrupole-Electrostatic Field Orbitrap Mass Spectrometer (Thermo Fisher Scientific). The single peak corresponding to each metabolite was filtered, and the missing values in the original data were reproduced. The internal standard was utilized for normalization, and the outliers were filtered based on the relative standard deviation. Partial least squares discriminant analysis (PLS-DA) and unsupervised principal component analysis (PCA) were used to identify the differential metabolites between two groups, with VIP > 1 and $p \leq 0.05$ as the criteria. The differential metabolites were subjected to correlation analysis, KEGG pathway analysis, and hierarchical clustering. ## 2.6. Statistical Analysis Data visualization was performed using GraphPad Prism 5. The data were expressed as the mean ± standard deviation of the mean (SD). Data were processed by GraphPad Prism 5. The mean values were statistically analyzed by unpaired t-tests and the significant differences among different groups were assessed by a non-parametric test. Differences were considered statistically significant at $p \leq 0.05.$ ## 3.1. Establishment and Validation of DQ-Treated Mouse Model We established a mouse model of DQ-induced kidney injury to study the early stages of AKI (Figure 1a). While DQ did not affect serum Scr levels on day 1, serum BUN levels were not affected by 200 mg/kg or 350 mg/kg DQ. The serum UREA levels were significantly higher in mice treated with 350 mg/kg DQ compared to the control group. In contrast, 200 mg/kg DQ had no significant effect on the urea level. Subsequently, both Scr and BUN continued to rise, and significant differences were observed on the 3rd and 7th days (Figure 1b,c). Furthermore, while no substantial lesions were observed in the kidney tissues of the DQ-treated mice in the first day of exposure, the renal tubules exhibited vacuolation and necrosis 3 days later (Figure S1). Based on these results, we selected the dose of 200 mg/kg to simulate the early stage DQ-induced kidney damage. ## 3.2. Transcriptomic Analysis of DQ-Treated Mice As shown in the UpSet graph in Figure 2a, 16,927 genes were expressed in all samples. Furthermore, 869 genes were differentially expressed in the DQ-treated samples relative to the control, of which 473 genes were downregulated and 396 genes were upregulated (Figure 2b and Table S1). The DEGs were enriched in GO terms related to fatty acid metabolism, extracellular structure organization, sulfur compound metabolism (Figure 2c), extracellular matrix, collagen-containing extracellular matrix (Figure 2d), extracellular matrix structural constituent, and sulfur compound binding (Figure 2e). Furthermore, KEGG analysis revealed that these DEGs were significantly associated with pathways of drug metabolism, drug metabolism-cytochrome P450, glutathione metabolism and retinol metabolism (Figure 2f). These results indicate that DQ might dysregulate numerous pathways in the kidneys. ## 3.3. Proteomic Analysis of DQ-Treated Mice We used TMT-based quantitative proteomics analysis to identify the differentially expressed proteins (DEPs) that might be linked to DQ-induced kidney damage. PCA revealed notable differences in protein abundance between the DQ and control groups (Figure 3a). There were 351 DEPs between the two groups, of which 133 proteins were upregulated and 218 proteins were downregulated in the DQ-treated mice (Figure 3b and Table S2). The DEPs were mainly enriched in pathways associated with Parkinson’s disease, Salmonella infection, chemical carcinogenesis, PPAR signaling, phagosome, tuberculosis, ribosome, bile secretion and retinol metabolism (Figure 3c). According to the GO enrichment analysis, DEPs were primarily associated with terms such as intracellular, intracellular part, organelle, intracellular organelle, cytoplasm, membrane-bounded organelle, intracellular membrane-bounded organelle, cytoplasm part, organelle part and intracellular organelle part (Figure 3d). Furthermore, a protein–protein interaction (PPI) network was constructed using the STRING database. As shown in the network in Figure 3e, DQ exposure altered ribonucleoprotein complex biogenesis (Bop1, Tarbp2, Imp4, Pqbp1, Pop4, Snrpf, Las1l, Mrpl1, Utp18, Ddx49, Prpf39), ncRNA processing (Bop1, Mettl1, Tarbp2, Imp4, Pop4, Las1l, Mrpl1, Utp18, Ddx49), ncRNA metabolic process (Bop1, Mettl1, Tarbp2, Imp4, Pop4, Las1l, Mrpl1, Utp18, Ddx49), response to wounding (Aqp1, Fcer1g, Pdpn, Grn, Jak2, Scnn1b, Tarbp2, Map2k1, Arhgap35), mitochondrial protein complex (Cox4i1, Grpel2, Mrps25, Chchd1, Dnajc15, Sdhd, Ndufa11, Mrpl1, Mrpl30), ribosome (Rpl37a, Uba52, Rps26, Mrps25, Chchd1, Rpl37, Mrpl1, Mrpl30, Rpl17), enzyme activator activity (Apoa2, Bcl10, Thy1, Map2k1, Dnajc15, Tab1, Cwf19l1, Arhgap35, Depdc5), organic hydroxy compound transport (Apoa2, Aqp1, Aqp3, Fcer1g, Slc10a2, Apom, Sdhd, Slc51a), fatty acid metabolic process (Adh7, Apoa2, Cyp2a4, Cyp4a10, Pdpn, Lpl, Gstm7, Acsl3), and positive regulation of cell activation (Bcl10, Fcer1g, Pdpn, Jak2, Thy1, Lgals8, Dnaja3, Hamp). In summary, DQ-induced kidney injury is likely mediated by dysregulated proteins involved in metabolism. ## 3.4. Integrated Transcriptome and Proteome Datasets Integration of the transcriptome and proteome datasets revealed that 34 genes were substantially altered by DQ exposure (Table S3). KEGG pathway analysis showed that these genes are significantly associated with the PPAR signaling pathway, retinol metabolism, asthma, cholesterol metabolism, fatty acid degradation, valine/leucine and isoleucine degradation, fatty acid metabolism, and kidney injury caused by DQ (Figure 4a). Furthermore, GSEA consistently demonstrated that these DEGs and DEPs were substantially enriched for metabolism-related pathways, including the drug metabolism cytochrome P450, the PPAR signaling pathway, retinol metabolism, metabolism of lipids, amino acid metabolism, glutathione metabolism and fatty acid metabolism (Figure 4b). Taken together, the aforementioned pathways are likely targeted by DQ during kidney injury. ## 3.5. Metabolomic Analysis of DQ-Treated Mice The metabolic by-products that may contribute to DQ-induced kidney injury were identified by untargeted LC–MS. The results of PCA and OPLS-DA clearly showed distinct metabolic patterns of the control and DQ-treated mice (Figure 5a,b). Overall, 96 metabolites were differentially expressed between the control and DQ-treated groups (adjusted $p \leq 0.05$), of which 40 were elevated and 56 were decreased in the latter (Figure 5c, Table S4). Furthermore, five of these differentially regulated metabolites are involved in purine metabolism, three in biosynthesis of unsaturated fatty acids, two in primary bile acid biosynthesis, one in fatty acid biosynthesis, and one in fatty acid metabolism (Figure 5d). To ascertain which metabolic pathways were most affected by DQ exposure, we performed KEGG pathway enrichment analysis. As shown in Figure 5d, the top 10 pathways were those related to purine metabolism, biosynthesis of unsaturated fatty acids, primary bile acid biosynthesis, nicotinate and nicotinamide metabolism, taurine and hypotaurine metabolism, fatty acid metabolism, amino sugar and nucleotide sugar metabolism, glycine, serine and threonine metabolism, porphyrin and chlorophyll metabolism, fatty acid elongation in mitochondria. ## 3.6. Integrated Transcriptomic, Proteomic and Metabolomics We constructed a correlation network diagram of the metabolites, DEPs and DEGs to gain further insights into the molecular mechanisms underlying DQ-induced nephrotoxicity. As shown in Figure 6, the top 20 co-related genes were Gm3776, Ccl21a, Vgf, Gsta1, Fgf21, Krt20, Ugt1a9, Lrrc55, Areg, 9130409I23Rik, Ccdc180, Edil3, Prss35, Cbr3, Ccr7, Nppb, Cyp2b10, F2rl3, Gm4841, Zfp683. The top 20 co-related proteins were Q3UFS4, Q9D486, Q62314, O70324, O89050, Q62011, Q8K209, Q9CYH5, P61460, Q99JH8, Q80TE3, O70571, Q8BQM4, Q75N73, P97473, P15409, P33174, P70172, P18469 and Q8VDM1. The top 20 metabolites were hippuric acid, 5-methoxyindoleacetate, chenodeoxycholic acid, tetradecanedioic acid, indoxyl sulfate, (R)-3-hydroxybutyric acid, traumatic acid, gamma-aminobutyric acid, alpha-linolenic acid, adipic acid, phenylacetylglycine, hypotaurine, 3-hydroxybutyric acid, palmitoleic acid, caprylic acid, eicosapentaenoic acid, linoleic acid, 2-furoic acid, beta-alanine and N-acetyl-L-phenylalaninex. These genes, proteins and their metabolites are mostly connected to the PPAR signaling pathway and fatty acid metabolism. ## 4. Discussion DQ is a highly nephrotoxic bipyridine herbicide that primarily targets the renal tubules and induces AKI. The molecular basis of DQ-induced kidney injury is cell death due to excessive production of reactive oxygen species (ROS) formed during lipid peroxidation [2,10]. The prognosis of DQ poisoning is highly correlated with AKI. Although AKI is reversible in its early stages, the therapeutic window is narrow. Therefore, it is crucial to identify the biomarkers and effectors of the incipient stages of AKI for early diagnosis of kidney damage. We identified the time window of DQ-induced kidney damage by analyzing different time points and dosages. There was no evident renal parenchymal damage, or any changes in serum Scr or BUN levels after 24 h exposure to 200 mg/kg DQ, which corresponded to the early stage of the DQ-induced kidney damage. To identify the molecular mechanisms of DQ-induced renal damage at this stage, we used an integrated multi-omics approach, which revealed that exposure to DQ significantly affects the PPAR signaling pathway and fatty acid metabolism. According to the integrated multi-omics data, the PPAR signaling pathway and fatty acid metabolism were associated with upregulation of Hmgcs2, Cyp4a10 and Cyp4a14, and the downregulation of Lpl mRNA and proteins in the DQ-treated kidneys. PPAR, a lipid-activated nuclear receptor, is abundantly expressed in tissues with high fatty acid metabolism, such as the kidney [11]. PPAR-deficient mice accumulate more lipids in their kidneys, which increases production of inflammatory mediators, eventually leading to kidney injury [12,13]. In addition, PPAR is also a transcription factor that controls genes involved in lipid metabolism and the mitochondrial fatty acid oxidation pathway [14], which fulfills a significant portion of the body’s energy needs [15,16]. Integrated proteomic and transcriptomic analysis revealed that the fatty acid oxidation pathway, and subsequently fatty acid metabolism, were downregulated in the DQ-treated group. The primary rate-limiting enzyme for ketogenesis is Hmgcs2 (3-hydroxy-3-methylglutaryl-CoA synthase 2). Hmgcs2 is a key rate-regulating enzyme for ketone body formation, which is related to fatty acid metabolism and mainly exists in cell mitochondria. The HMG-CoA generated by it is converted into acetoacetic acid under the action of HMG-CoA lyase, and acetoacetic acid can be converted into hydroxybutyric acid and acetone, which are called ketone bodies. Ketogenesis of cells is an important part of fatty acid metabolism, and acetyl CoA, the product of fatty acid oxidation, is the raw material for the formation of ketosomes. Therefore, Hmgcs2 may regulate the changes in fatty acid metabolism by regulating the ketogenesis process. Upregulation of Hmgcs2 in the glomeruli of high fructose-fed rats and high fructose-treated differentiated podocytes enhanced ketone bodies level, particularly that of hydroxybutyrate (3-OHB), to block histone deacetylase (HDAC) activity [17]. Hmgcs2 is likely upregulated through the PPAR-α pathway [18]. The findings imply that enhanced renal ketogenesis due to Hmgcs2 overexpression may be significant in the pathogenesis of diabetic neuropathy DN in patients with type 2 diabetes, indicating that Hmgcs2 is a potential therapeutic target for the management of diabetic renal complications [19]. We found that Hmgcs2 gene and protein expression levels increased in the kidney tissues after DQ exposure, indicating its role in DQ-induced renal damage as well. CYP4A (cytochrome P450, family 4, subfamily a) catalyzes the hydroxylation of medium- and long-chain fatty acids [20]. One of the pathway for fatty acid degradation is through oxidation, in which dicarboxylic acids are formed and subsequently undergo β-oxidation from the omega end. This pathway is catalyzed by CYP450 enzymes and the peroxisomal β-oxidation pathway which are regulated by PPARα [21] The mouse genome contains four Cyp4a genes: Cyp4a10, Cyp4a12a, Cyp4a12b, and Cyp4a14—all of which are localized in chromosome 4 [22]. Murine Cyp4a10 and Cyp4a14 (homologous to human CYP4A22 and CYP4A11, respectively) are highly expressed in the liver and kidneys, and are known to convert the arachidonic acid to its metabolite 20-hydroxyeicosatetraenoic acid (20-HETE), which regulates the inflammatory response through the generation of ROS [15,22]. As a result, the aberrant expression of Cyp4a10 and Cyp4a14 observed in our study may lead to fatty acid breakdown. LPL (lipoprotein lipase) catalyzes the hydrolysis of triglyceride (TAG), which is the rate-limiting step in the lipolysis of chylomicrons and VLDL. In addition to other cell types, myocytes and adipocytes also synthesize LPL, which is then stored in the Golgi apparatus for either intracellular breakdown or secretion onto the cell surface. Patients with nephrotic syndrome often have hyperlipidemia due to the lack of LPL activators. Furthermore, the high levels of free fatty acids in the bloodstream of these patients upregulates ANGPTL4, which may inactivate LPL by either converting the active LPL dimers into inactive monomers or as a reversible non-competitive inhibitor of LPL [23]. In this study, LPL expression was downregulated in the DQ-treated kidney tissues, indicating its role in DQ-induced nephrotoxicity. We identified eicosapentaenoic acid, linoleic acid, palmitic acid and (R)-3-hydroxybutyric acid as significant metabolites involved in DQ-related kidney injury. Eicosapentaenoic acid, linoleic acid and palmitic acid are polyunsaturated fatty acids (PUFAs), which have been linked to a number of renal disorders. One study showed that retinoic acid signaling mediates production of toxic PUFAs [24]. Increased PUFA peroxidation by ROS initiates ferroptosis, an iron-dependent form of programmed cell death. Fatty acid oxidation in the liver produced high levels of 3-hydroxybutyrate acid, which is then transferred to extrahepatic tissues including the heart, brain and muscle to be used as a fuel. As one of the ketone bodies, 3-hydroxybutyric acid can directly promotes 3-hydroxybutyrylation of some proteins and functions as an endogenous inhibitor of histone deacetylases as well as an agonist of Gpr109a [25]. β-OHB is one of the intermediate metabolites of fatty acid oxidation. In addition to being a functional vector that transfers energy from liver to peripheral tissues under starvation stress, β-OHB is also an important signaling molecule and epigenetic regulatory molecule in vivo, regulating all aspects of life function. This study showed that glomerular podocytes damage and albuminuria production caused by fructose intake showed an increase in β-OHB beginning at week 8 of modeling and continuing until week 16 of the study deadline [17]. Therefore, β-OHB is a key metabolic substance in the occurrence and development of kidney injury. Taken together, dysregulated fatty acid metabolism may induce by the nephrotoxic effects of DQ. Overall, Hmgcs2 upregulated and subsequently may promote 3-hydroxybutyric acid levels, dysregulating the PPAR signaling pathway. Our findings offer a new insight into the mechanisms underlying DQ-induced nephrotoxicity. ## 5. Conclusions Our study is the first to investigate the mechanism of the early stage of DQ-induced kidney injury using a multi-omics approach. Our findings lay the foundation for diagnosing and treating renal damage following DQ exposure, and offer new insights into the molecular basis of DQ-induced kidney damage. ## References 1. Yu G., Jian T., Cui S., Shi L., Kan B., Jian X.. Acute diquat poisoning resulting in toxic encephalopathy: A report of three cases. Clin. Toxicol. (Phila.) (2022) 60 647-650. DOI: 10.1080/15563650.2021.2013495 2. Magalhaes N., Carvalho F., Dinis-Oliveira R.J.. Human and experimental toxicology of diquat poisoning: Toxicokinetics, mechanisms of toxicity, clinical features, and treatment. Hum. Exp. Toxicol. (2018) 37 1131-1160. PMID: 29569487 3. Guck D., Hernandez R., Moore S., Van de Louw A., Haouzi P.. Rapid Glomerulotubular Nephritis as an Initial Presentation of a Lethal Diquat Ingestion. Case Rep. Nephrol. (2021) 2021 4723092. PMID: 34552801 4. Hantson P., Wallemacq P., Mahieu P.. A case of fatal diquat poisoning: Toxicokinetic data and autopsy findings. J. Toxicol. Clin. Toxicol. (2000) 38 149-152. DOI: 10.1081/CLT-100100930 5. Petejova N., Martinek A., Zadrazil J., Teplan V.. Acute toxic kidney injury. Ren. Fail. (2019) 41 576-594. DOI: 10.1080/0886022X.2019.1628780 6. Ronco C., Bellomo R., Kellum J.A.. Acute kidney injury. Lancet (2019) 394 1949-1964. PMID: 31777389 7. Kellum J.A., Romagnani P., Ashuntantang G., Ronco C., Zarbock A., Anders H.J.. Acute kidney injury. Nat. Rev. Dis. Primers. (2021) 7 52. PMID: 34267223 8. Pan H.C., Chen Y.Y., Tsai I.J., Shiao C.C., Huang T.M., Chan C.K., Liao H.W., Lai T.S., Chueh Y., Wu V.C.. Accelerated versus standard initiation of renal replacement therapy for critically ill patients with acute kidney injury: A systematic review and meta-analysis of RCT studies. Crit. Care (2021) 25 5. PMID: 33402204 9. Yuan B., Wu W., Chen M., Gu H., Tang Q., Guo D., Chen T., Chen Y., Lu C., Song L.. From the Cover: Metabolomics Reveals a Role of Betaine in Prenatal DBP Exposure-Induced Epigenetic Transgenerational Failure of Spermatogenesis in Rats. Toxicol. Sci. (2017) 158 356-366. DOI: 10.1093/toxsci/kfx092 10. Yu G., Cui S., Jian T., Kan B., Jian X.. Diquat poisoning in a pregnant woman resulting in a miscarriage and maternal death. Clin. Toxicol. (Phila.) (2021) 59 1275-1277. DOI: 10.1080/15563650.2021.1905164 11. Tahri-Joutey M., Andreoletti P., Surapureddi S., Nasser B., Cherkaoui-Malki M., Latruffe N.. Mechanisms Mediating the Regulation of Peroxisomal Fatty Acid Beta-Oxidation by PPARalpha. Int. J. Mol. Sci. (2021) 22. DOI: 10.3390/ijms22168969 12. Iwaki T., Bennion B.G., Stenson E.K., Lynn J.C., Otinga C., Djukovic D., Raftery D., Fei L., Wong H.R., Liles W.C.. PPARalpha contributes to protection against metabolic and inflammatory derangements associated with acute kidney injury in experimental sepsis. Physiol. Rep. (2019) 7 e14078. PMID: 31102342 13. Luan Z.L., Zhang C., Ming W.H., Huang Y.Z., Guan Y.F., Zhang X.Y.. Nuclear receptors in renal health and disease. EBioMedicine (2022) 76 103855. PMID: 35123268 14. Yin X., Zeng W., Wu B., Wang L., Wang Z., Tian H., Wang L., Jiang Y., Clay R., Wei X.. PPARalpha Inhibition Overcomes Tumor-Derived Exosomal Lipid-Induced Dendritic Cell Dysfunction. Cell Rep. (2020) 33 108278. DOI: 10.1016/j.celrep.2020.108278 15. Marechal L., Sicotte B., Caron V., Brochu M., Tremblay A.. Fetal Cardiac Lipid Sensing Triggers an Early and Sex-related Metabolic Energy Switch in Intrauterine Growth Restriction. J. Clin. Endocrinol. Metab. (2021) 106 3295-3311. PMID: 34245263 16. Zhang H., Yan Q., Wang X., Chen X., Chen Y., Du J., Chen L.. The Role of Mitochondria in Liver Ischemia-Reperfusion Injury: From Aspects of Mitochondrial Oxidative Stress, Mitochondrial Fission, Mitochondrial Membrane Permeable Transport Pore Formation, Mitophagy, and Mitochondria-Related Protective Measures. Oxid Med. Cell Longev. (2021) 2021 6670579. PMID: 34285766 17. Fang L., Li T.S., Zhang J.Z., Liu Z.H., Yang J., Wang B.H., Wang Y.M., Zhou J., Kong L.D.. Fructose drives mitochondrial metabolic reprogramming in podocytes via Hmgcs2-stimulated fatty acid degradation. Signal Transduct. Target Ther. (2021) 6 253. PMID: 34238920 18. Yi W., Xie X., Du M., Bu Y., Wu N., Yang H., Tian C., Xu F., Xiang S., Zhang P.. Green Tea Polyphenols Ameliorate the Early Renal Damage Induced by a High-Fat Diet via Ketogenesis/SIRT3 Pathway. Oxid Med. Cell Longev. (2017) 2017 9032792. PMID: 28814987 19. Zhang D., Yang H., Kong X., Wang K., Mao X., Yan X., Wang Y., Liu S., Zhang X., Li J.. Proteomics analysis reveals diabetic kidney as a ketogenic organ in type 2 diabetes. Am. J. Physiol. Endocrinol. Metab. (2011) 300 E287-E295. DOI: 10.1152/ajpendo.00308.2010 20. Capdevila J.H., Falck J.R.. The arachidonic acid monooxygenase: From biochemical curiosity to physiological/pathophysiological significance. J. Lipid Res. (2018) 59 2047-2062. DOI: 10.1194/jlr.R087882 21. Khalil Y., Carrino S., Lin F., Ferlin A., Lad H.V., Mazzacuva F., Falcone S., Rivers N., Banks G., Concas D.. Tissue Proteome of 2-Hydroxyacyl-CoA Lyase Deficient Mice Reveals Peroxisome Proliferation and Activation of omega-Oxidation. Int. J. Mol. Sci. (2022) 23 987. PMID: 35055171 22. Yang Z., Smalling R.V., Huang Y., Jiang Y., Kusumanchi P., Bogaert W., Wang L., Delker D.A., Skill N.J., Han S.. The role of SHP/REV-ERBalpha/CYP4A axis in the pathogenesis of alcohol-associated liver disease. JCI Insight. (2021) 6 e140687. DOI: 10.1172/jci.insight.140687 23. Agrawal S., Zaritsky J.J., Fornoni A., Smoyer W.E.. Dyslipidaemia in nephrotic syndrome: Mechanisms and treatment. Nat. Rev. Nephrol. (2018) 14 57-70. PMID: 29176657 24. Sidhom E.H., Kim C., Kost-Alimova M., Ting M.T., Keller K., Avila-Pacheco J., Watts A.J., Vernon K.A., Marshall J.L., Reyes-Bricio E.. Targeting a Braf/Mapk pathway rescues podocyte lipid peroxidation in CoQ-deficiency kidney disease. J. Clin. Investig. (2021) 131. DOI: 10.1172/JCI141380 25. Zhang S.J., Li Z.H., Zhang Y.D., Chen J., Li Y., Wu F.Q., Wang W., Cui Z.J., Chen G.Q.. Ketone Body 3-Hydroxybutyrate Ameliorates Atherosclerosis via Receptor Gpr109a-Mediated Calcium Influx. Adv. Sci. (Weinh.) (2021) 8 2003410. PMID: 33977048
--- title: Causes of Moderate and Severe Anaemia in a High-HIV and TB-Prevalent Adult Population in the Eastern Cape Province, South Africa authors: - Don O’Mahony - Sikhumbuzo A. Mabunda - Mbulelo Mntonintshi - Joshua Iruedo - Ramprakash Kaswa - Ernesto Blanco-Blanco - Basil Ogunsanwo - Kakia Anne Faith Namugenyi - Sandeep Vasaikar - Parimalaranie Yogeswaran journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9966846 doi: 10.3390/ijerph20043584 license: CC BY 4.0 --- # Causes of Moderate and Severe Anaemia in a High-HIV and TB-Prevalent Adult Population in the Eastern Cape Province, South Africa ## Abstract Background: Anaemia affects one in four adults in South Africa, with a higher prevalence in persons with HIV and tuberculosis. The aim of this study is to characterise the causes of anaemia in primary care and a district hospital setting. Methods: A cross-sectional study design investigated a purposive sample of adult males and non-pregnant females at two community health centres and a hospital casualty and outpatients. Fingerpick blood haemoglobin was measured with HemoCueHb201+. Those with moderate and severe anaemia underwent clinical examination and laboratory tests. Results: Of 1327 patients screened, median age was 48 years, and $63.5\%$ were female. Of 471 ($35.5\%$) with moderate and severe anaemia on HemoCue, $55.2\%$ had HIV, $16.6\%$ tuberculosis, $5.9\%$ chronic kidney disease, $2.6\%$ cancer, and $1.3\%$ heart failure. Laboratory testing confirmed 227 ($48.2\%$) with moderate and 111 ($23.6\%$) with severe anaemia, of whom $72.3\%$ had anaemia of inflammation, $26.5\%$ iron-deficiency anaemia, $6.1\%$ folate deficiency, and $2.5\%$ vitamin B12 deficiency. Overall, $57.5\%$ had two or more causes of anaemia. Multivariate modelling showed that patients with severe anaemia were three times more likely to have tuberculosis (OR = 3.1, $95\%$ CI = 1.5–6.5; p-value = 0.002). Microcytosis was present in $40.5\%$ with iron deficiency, macrocytosis in $22.2\%$ with folate deficiency, and $33.3\%$ with vitamin B12 deficiency. The sensitivities of the reticulocyte haemoglobin content and % hypochromic red blood cells in diagnosing iron deficiency were $34.7\%$ and $29.7\%$, respectively. Conclusions: HIV, iron deficiency, and tuberculosis were the most prevalent causes of moderate and severe anaemia. The majority had multiple causes. Iron, folate, and vitamin B12 deficiencies should be identified by biochemical testing rather than by red cell volume. ## 1. Introduction Anaemia is a significant public health problem worldwide, with adverse effects on health and socio-economic development [1]. Global anaemia prevalence in 2019 was $23.2\%$, affecting 1.8 billion people, resulting in an estimated 50.3 million years lived with disability [2]. Dietary iron deficiency is the most prevalent cause globally, followed by inherited hemoglobinopathies and haemolytic anaemias [2]. Iron deficiency (ID) is recognised as a significant contributor to anaemia and ill-health in patients with common chronic inflammatory conditions including chronic kidney disease (CKD) [3], heart failure [4], and cancer [5]. There is also evidence of inadequate investigation of anaemia in primary care [6,7,8]. Anaemia is common in South Africa, with community rates between $22\%$ and $31\%$ in women and $12\%$ and $17\%$ in men older than 15 years [9,10]. In addition, South Africa has 8.2 million persons living with human immunodeficiency virus infection (PLWHIV) [11] and an associated anaemia prevalence up to $72\%$ [12]. South Africa also has a high tuberculosis (TB) prevalence estimated at 737 per 100,000 in 2018 [13]. Anaemia is a risk factor for TB [12,14], and TB is associated with a high prevalence of anaemia ($61.5\%$) [15]. ID is probably the most common cause in South Africa [9]. Anaemia of inflammation (AI), which is synonymous with anaemia of chronic disease, is the second most common cause globally in clinical practice [16] and, most probably, in South Africa [17,18]. In persons living with HIV (PLWHIV), causes may be multifactorial, but anaemia of chronic disease is the predominant cause [19]. In Southern Africa, TB is probably the most common cause of moderate-to-severe anaemia in PLWHIV [12,20,21,22]. The relative contributions of all causes of anaemia are not well characterised in patients attending primary care and district hospitals in South Africa. In a community-based study in the Free State Province, AI was the most prevalent cause of anaemia in women with or without HIV [17]; and in a district hospital (GF Jooste) in Cape Town, >$95\%$ of patients with HIV-associated TB had AI and <$3\%$ had IDA [23]. However, no testing was done for other causes of anaemia in either study. The aims of this study were to determine the causes of anaemia in adults in primary care and a district hospital setting in Mthatha, South Africa. To determine the overall causes in a resource-efficient manner, the study focused on moderate and severe anaemia. ## 2.1. Study Design A cross-sectional study design was used. The study was undertaken between September 2017 and March 2018. ## 2.2. Study Population and Setting The study population was patients aged ≥18 years who attended two community health centres (CHCs), one semi-urban (CHC1) and one rural (CHC2), and Mthatha Regional Hospital (MRH) general outpatients and Emergency Department managed by the Department of Family Medicine and Rural Health. MRH functions as both a district and referral hospital. The catchment area is King Sabata Dalindyebo Local Municipality, comprising a socio-economically deprived predominantly rural population [24]. ## 2.3. Sampling The population prevalence of moderate and severe anaemia was estimated to be ±$2\%$ [9]. This study was facility-based, and the prevalence was therefore estimated to be higher at a minimum of $3\%$. Using the equation n = p (100 − p) (1.96)2/d2, where n = minimum sample size, p = anticipated prevalence of moderate and severe anaemia, and d = set precision [25]; the precision was set at $2\%$ and the estimated prevalence at $3\%$, yielding a minimum sample size of ±279 participants. The confidence interval was $95\%$. A further $20\%$ [56] of participants was added to the sample to factor in for anticipated loss of participants during the evaluation process, loss of data, and data entry errors [26] for a total sample size of 335 participants. ## 2.4. Study Procedures A purposive sample was taken. Patients waiting to consult the nurse or doctor were approached by a research assistant (a nurse at each site during normal working hours). If agreeable, they signed a consent form. Exclusion criteria were: [1] pregnancy, [2] a significant bleeding episode in the last three months that needed an urgent visit to the clinic or doctor or a blood transfusion, and [3] unwillingness to disclose or test for HIV. Entry to the study was by identifying moderate and severe anaemia using a HemoCue Hb201+ analyser® on a capillary (finger-prick) sample. Anaemia was defined as mild, 11–11.9 g/dL for women and 11–12.9 g/dL for men; moderate, 8–10.9 g/dL; and severe as <8 g/dL for both sexes [27]. All patients had their HIV status assessed and a rapid HIV test done if not on antiretroviral therapy (ART). ART was endorsed for all PLWHIV from September 2016 [28]. Patients with mild anaemia were referred for routine care [29]. Those with moderate and severe anaemia had the following tests analysed at the National Health Laboratory Service (NHLS) at Nelson Mandela Central Hospital, Mthatha: full blood count, reticulocyte production index (RPI), and blood film; creatinine, eGFR (Modification of Diet in Renal Disease study equation, without the race coefficient), bilirubin (direct and indirect), C-reactive protein, vitamin B12 and folate, iron, transferrin, transferrin saturation (TSAT), and Xpert MTB/RIF test® on (non-induced) sputum. In addition, patients with severe anaemia had urine and blood cultures for TB (BACTEC™ Myco/F Lytic Culture vials) and urine lateral flow lipoarabinomannan assay LF-LAM (Alere®). In PLWHIV, CD4 count and viral load (VL) tests were done according to national guidelines [28,30]. VL suppression was defined as < 50 RNA copies/mL [30]. Patients were also assessed by a doctor and underwent chest-X-ray and focused ultrasound examination for evidence of TB [31] and chronic kidney disease (CKD). Additional tests were ordered at the treating clinicians’ discretion. A final assessment/diagnosis was then made. A Siemens ADVIA 2120 analyser was used initially followed by a Beckman Coulter LH 780 Analyser. The NHLS normal adult Hb values were 13–17 g/L for males and 12–15 g/L for females. Microcytosis and macrocytosis were defined as MCV <80 fL and >100 fL, respectively [32,33]. A % hypochromic red cells (%HRC) value >$6\%$ [34,35] or a reticulocyte haemoglobin content (CHr) of <29 pg/cell indicate ID [35]. Vitamin B12 deficiency was categorised as: <147.6 pmol/L (200 pg/mL), deficiency likely; >258.3 pmol/L (350 pg/mL), deficiency unlikely; while <73.8 pmol/L (100 pg/mL) usually indicates clinical deficiency [36]. With vitamin B12 values between 147.6–258.3 pmol/L, patient records were evaluated to determine if clinical findings were consistent with deficiency. A serum folate level <6.8 nmol/L was defined as deficiency [37]. TB was diagnosed by detection of Mycobacterium tuberculosis (TB) by molecular testing (Xpert MTB/RIF, Genotype® MTBDRsl), urine LF LAM, culture, or demonstration of acid-fast bacilli on clinical specimens. Patients with probable TB diagnosed on X-ray or ultrasound underwent anti-tuberculosis therapy but were excluded from analysis. IDA was defined as anaemia and a ferritin < 30 mcg/L [38,39] and in the presence of inflammation, a ferritin < 100 mcg/L, and ferritin 100–300 mcg/L if TSAT < $20\%$ [39,40]. AI was diagnosed on three criteria: [1] a chronic inflammatory illness including cancer and haematological malignancies, infections, immune-related diseases, inflammatory diseases, CKD, heart failure, chronic pulmonary disease, and anaemia of the elderly [41]; [2] a ferritin 100–300 mcg and TSAT ≥ $20\%$ or ferritin > 300 mcg/L [40]; and [3] the absence of an identifiable cause [16]. Guidelines for heart failure [42], cancer [43], and CKD [44] accept that inflammation is part of the disease process without a requirement for an elevated CRP. A CRP (NHLS normal value < 10 mg/L) was performed primarily to identify inflammation in people without a diagnosis of an inflammatory disorder. CKD was diagnosed on standard criteria [45,46] and is considered the most likely cause of anaemia if the GFR < 60 mL/min/1.72 m2 when no other cause is identified [44,47]. ## 2.5. Statistical Analysis Data were captured in Microsoft Excel and exported to STATA version 17 for analysis. Numerical data were explored for normality using the Shapiro–Wilk test. All numerical data were not normally distributed and were summarised using the median and interquartile range (IQR = 75th percentile − (minus) 25th percentile). Categorical data were summarised using percentages. A comparison of two categorical variables was undertaken using the chi-square test if expected frequencies were ≥5, and if the expected frequencies were <5, then the Fisher’s exact test was used. A comparison of medians used the Wilcoxon rank-sum test for binary categorical variables and the Kruskal–Wallis test for nominal categorical variables (more than two groups). Predictors of anaemia severity were determined using bivariable and multivariable logistic regression analysis. The adjusted multivariable logistic regression model was determined through purposeful selection of variables to determine the best-fitting model. The odds ratio or adjusted odds ratio (OR/aOR) was the measure of association used for categorical predictors, and coefficients were used to predict numerical variables. The $95\%$ confidence interval ($95\%$CI) was used for the precision of estimates, and statistical significance is a p-value of <0.05. Because of the use of a subsample in the analyses, especially for Tables 6 and 7, the p-values are purely descriptive. Missing data were analysed using complete case analysis. ## 3. Results A total of 21 declined participation in the study. The most common reason [5] was unreadiness to know their HIV status. Table 1 summarises the demographic characteristics of patients who participated, and it shows that 1327 participants were enrolled into the study, of which $44.6\%$ [592] were seen at MRH, and $63.5\%$ [842] were female. The combined median age was 48 years (IQR = 31 years). There was statistical homogeneity between males and females (p-value = 0.100), the CD4 count (p-value = 0.539), and viral load (p-value = 0.235). The median ages of patients were statistically different depending on their recruiting health facility (p-value = 0.0003). Even though the combined HIV prevalence was $20.9\%$ ($\frac{277}{1327}$), it ranged from $5.5\%$ ($\frac{22}{399}$) in CHC1 to $35.8\%$ ($\frac{212}{592}$) at MRH, and this was statistically significant (p-value < 0.0001). The HemoCue also showed anaemia prevalences of $48.4\%$ ($\frac{147}{304}$) and $46.1\%$ ($\frac{246}{534}$) for males and females 40 years old and older, respectively. Of the 1327 patients, 471 ($35.5\%$) qualified to have a laboratory evaluation of their anaemia due to HemoCue blood levels of moderate or severe anaemia status (Table 2). However, only $\frac{463}{471}$ ($98.3\%$) participants had laboratory haemoglobin results, of whom $49.0\%$ ($$n = 227$$) had moderate and $24.0\%$ ($$n = 111$$) severe anaemia, comprising $25.5\%$ ($$n = 338$$) of those screened. Chronic diseases that can cause anaemia (HIV, TB, CKD, cancer, and heart failure) and non-communicable diseases (NCDs) of high prevalence were assessed on the 471 participants who had progressed for a laboratory evaluation. The median haemoglobin levels were statistically different between the three health facilities (p-value = 0.002), and the difference was due to the lower median of MRH (9.4) when compared to CHC1 (10.9), which was statistically significant (p-value < 0.001). Disease prevalences were as follows: tuberculosis ($16.6\%$, $$n = 78$$), hypertension ($20.0\%$, $$n = 94$$), chronic kidney disease ($5.9\%$, 28), and cancers ($2.6\%$, $$n = 12$$). There was a statistically significant difference in the hypertension status depending on the health facility of recruitment (p-value < 0.001). Of those with confirmed TB, $\frac{54}{78}$ ($69.2\%$) had pulmonary TB, $\frac{13}{78}$ ($16.7\%$) extra-pulmonary TB, $\frac{7}{78}$ ($9.0\%$) had both, and in $\frac{4}{78}$ ($5.1\%$), the primary site could not be confirmed. In addition to TB, there were only two patients with CDC-Stage-3-defining illnesses [48], namely cervical cancer and cryptococcal antigenaemia. More than two-thirds of the participants with HemoCue severe and moderate anaemia who had HIV were in CDC Stage 3. The CD4 nadir, median, mean, and zenith were 1, 103, 191, and 2440, respectively. Viral replication was not suppressed in the majority of PLWHIV. Figure 1 illustrates the biochemical diagnosis of IDA and AI. The category AI includes those with specific causes of anaemia (e.g., vitamin deficiencies and haematological malignancies) who also had inflammation. Excluding 17 with missing ferritin results, the percentages of AI and IDA were $72.3\%$ ($\frac{232}{321}$) and $26.5\%$ ($\frac{85}{321}$), respectively. In patients with CKD, $\frac{8}{25}$ ($32\%$) had IDA. Of nine patients with cancer, $\frac{1}{6}$ ($16.7\%$) had IDA. Four patients with cardiac failure had AI. Only 399 individuals had a full record of the time from venesection to FBC analysis. The time a specimen was received at the laboratory was used as a proxy for the time to analysis, as reports did not specify analysis times. With a median of 2.6 h (IQR = 2.2 h), all FBC analyses were undertaken within 24 h. Figure 2 shows no statistically significant difference in the median MCV in those analysed <8 h and ≥8 h ($$p \leq 0.766$$) and in the median %HPO for those analysed <6 h and ≥6 h ($$p \leq 0.176$$). Using RHC to diagnose IDA, as shown in Table 3, sensitivity was $34.7\%$, specificity $89.4\%$, and positive predictive value $64.7\%$. Table 4 shows the use of HRC to diagnose IDA. Sensitivity was $29.7\%$, specificity $87.9\%$, and positive predictive value $92.2\%$. Table 5 shows that patients with ID had predominantly normocytic anaemia, as also did patients with folate deficiency. Patients with vitamin B12 deficiency had equal proportions of microcytic, normocytic, and macrocytic anaemia. In patients 65 years old and older with IDA, $35.3\%$ ($\frac{6}{17}$) had microcytosis. In eight patients with vitamin B12 deficiency, two had neurological disorders consistent with deficiency. One had acute confusion (vitamin B12, 50 pmol/L) and one lumbar radiculopathy (vitamin B12, 75 pmol/L). They had normal MCV and no blood film features of megaloblastic anaemia. Another patient was on metformin, which may cause deficiency. Five had no neurological or haematological features to diagnose deficiency. Of three patients with possible vitamin B12 deficiency (range 147.6–258.3 pmol/L), one had acute psychosis with HIV and hepatitis B virus infections, one had paraparesis of unknown cause, and one PLWHIV had TB meningitis. Overall, $2.5\%$ ($\frac{8}{324}$ results) had vitamin B12 deficiency. In patients with moderate and severe anaemias and a vitamin B12 level <147.6 pmol/L, none had hyper-segmented neutrophils or Howell–Jolly bodies. One patient had macrocytosis and ovalocytes with a vitamin B12 level of 109 pmol/L and a folate level of 2.1 pmol/L. The prevalence of folate deficiency was $6.1\%$ ($\frac{18}{294}$ results). None of the patients with folate deficiency were taking a medication causing deficiency. ## 3.1. Macrocytosis Of 25 patients with macrocytosis, 10 had incomplete drug data and 15 were PLWHIV. The putative causes of macrocytosis in 15 were lamivudine, 8; folate deficiency, 3; vitamin B12 deficiency, 2; cotrimoxazole, 2; valproate, 2; zidovudine, 2; liver disease, 1; and multiple myeloma, 1 (four patients had two causes each, and one had three). ## 3.2. Microcytosis Of 78 patients with microcytosis, one had a Mentzer index value < 13. That patient had IDA (TSAT $4\%$, ferritin 46 mcg/L). ## 3.3. Multiple Causes of Anaemia For each of the eight major causes, the numbers (percentage) of patients with moderate/severe anaemia due solely to that cause were iron deficiency $\frac{32}{85}$ (37.6), vitamin B12 deficiency $\frac{4}{8}$ [50], folate $\frac{1}{18}$ (5.6), HIV $\frac{136}{219}$ (62.1), cancer $\frac{2}{12}$ (16.7), CKD $\frac{8}{28}$ (28.6), TB $\frac{6}{75}$ (8.0), and heart failure $\frac{2}{4}$ [50]. Overall, $\frac{191}{449}$ ($42.5\%$) had a single cause of anaemia, ranging from $5.6\%$ for patients with folate deficiency to $62.1\%$ with HIV. The remainder ($57.5\%$) had two or more causes. In the bivariable model, those who had severe anaemia were $70\%$ (OR = 1.7, $95\%$CI = 1.0–2.7; p-value = 0.039) and $90\%$ (OR = 1.9, $95\%$CI = 1.1–3.2; p-value = 0.017) more likely to have HIV and TB, respectively (Table 6). However, the multivariable best-fitting model that was adjusted for other chronic conditions and the anaemia markers did not show an association between the severity of anaemia and an HIV-positive result (p-value = 0.880). Instead, the measure of effect increased for TB, wherein patients with severe anaemia were found to have been three times more likely to have had TB, and this was also statistically significant (OR = 3.1, $95\%$ CI = 1.5–6.5; p-value = 0.002). Similarly, patients with severe anaemia were 5.5 times more likely to present in a hospital setting than a primary care setting, and this was statistically significant as well (OR = 5.5, $95\%$CI = 1.5–19.7; p-value = 0.009). The multivariable analysis further shows that an increase of 1 in the reticulocyte production index increased the odds of severe anaemia by 5.5 (co-efficient = 1.7, $95\%$CI = 0.9–2.5; p-value < 0.0001); and a 1 g/dL increase in the mean corpuscular haemoglobin concentration (MCHC) increases the odds of having severe anaemia by $65\%$ (co-efficient = 0.5, $95\%$CI = 0.4–0.7; p-value < 0.0001), and both these were statistically significant. Even though PLWHIV with moderate anaemia had a statistically lower median Hb than HIV-negative patients (p-value < 0.0004), there was no statistical difference between the median Hb of patients with a negative HIV status and PLWHIV in patients with severe anaemia (p-value = 0.623) (Table 7). TB showed an opposite effect where patients with severe anaemia had a higher median Hb, and this was statistically significant (p-value = 0.005). The median MCHC was lower than normal in patients with moderate and severe anaemia and lower still in patients with severe anaemia and HIV or TB. Other markers of statistical significance included the following: a higher median CRP for PLWHIV with moderate anaemia (p-value = 0.003); higher median CRP values for both TB-positive patients with moderate (median = 146; p-value = 0.005) and severe anaemia (median = 163; p-value = 0.0004); higher median transferrin saturation severe anaemia in both HIV-positive (median = $19.6\%$; p-value = 0.0004) and TB-positive patients (median = $19.2\%$; p-value = 0.015); lower median RPIs for PLWHIV with moderate (median = 0.5; p-value = 0.005) and severe anaemia (median = 0.3; p-value = 0.036); and higher median ferritin levels for all categories of HIV-positive and TB-positive patients (p-value < 0.05). Vitamin B12 values were higher in PLWHIV with moderate or severe anaemia and in TB with moderate anaemia. Folate levels were unchanged with HIV and TB. ## 4. Discussion There was an overall high ($45.9\%$) prevalence of anaemia in the screened population, with the highest prevalence at MRH. Similar to attendees in primary care in South Africa [49], two-thirds of patients screened for anaemia in this study were female. In this study, the age group 40–49 years had the largest number screened. This may be due to the exclusion of pregnant women and the inclusion of sicker patients attending MRH. In older adults (≥40 years old), $48.4\%$ of men and $46.1\%$ of women had anaemia. High rates of anaemia were also found in community-living older adults in Mpumalanga Province, South Africa, i.e., $40.1\%$ of men and $43\%$ of women [50]. Prevalence of HIV infection was similar, at $20.9\%$ in this study and $20.7\%$ in Mpumalanga Province. Overall, the largest category of anaemia on screening was moderate, followed by severe. In contrast, in the national community surveys, the largest category was mild followed by moderate [9,10]. The skewed distribution in this study is most likely due to selection of participants from health care facilities, which will likely have more sicker individuals than those in the general population. The overall prevalence of moderate and severe anaemia was $35.5\%$ on HemoCue testing but $25.5\%$ on laboratory testing. HemoCue 201 overestimates mean Hb concentrations by 0.1–1.2 g/dL [51]. Furthermore, capillary finger-prick usually produces higher Hb values by 0.2–0.9 g/dL compared to venous blood [51]. In patients with laboratory-confirmed moderate and severe anaemia, AI was the predominant type at $72.6\%$, reflecting the high prevalences of HIV ($59.5\%$) and TB ($16.6\%$). In Cape Town, >$95\%$ of patients with anaemia and HIV-associated TB had AI [23]. In patient populations where there are very low levels of HIV and TB diagnoses, AI was the commonest type of anaemia, but prevalences were much lower at $25.7\%$ in elderly inpatients in South Africa [18] and $41.9\%$ of medical inpatients in Italy [52]. In the U.S., one-third of elderly community-dwellers with anaemia had AI, one-third nutrient deficiencies, and one-third unexplained anaemia [53]. The $26.5\%$ prevalence of IDA is similar to prevalences of $24.3\%$ in older inpatients in South Africa [18] and $20\%$ in older community-dwellers in the U.S. [53] but higher than $14.7\%$ in internal medicine inpatients in Italy [52]. IDA in this study comprised $15.9\%$, with absolute ID (ferritin < 100 mcg/L [54]) and $10.5\%$ functional ID (ferritin ≥ 100–300 mcg/L and TSAT < $20\%$) [39,43]. However, absolute and functional ID overlap such that patients who have ferritin > 100 mcg/L may have absolute iron deficiency. In a systematic review utilising 38 studies, the mean ferritin level in absolute iron deficiency was 82.4 mcg/L (range of means 34–158 mcg/L) in patients with inflammatory diseases, using bone marrow iron as the gold standard [55]. While there are various definitions of ID [23], the “pragmatic” definition of Camaschella and Girelli [39] was used in this study, combining ferritin and TSAT. TSAT is considered an accurate test for ID in patients with inflammation [39,56]. Newer tests for ID include hepcidin and soluble transferrin receptor (sTfR) levels [40] but are not routinely available in South Africa. Forty-one per cent ($41\%$) of patients with IDA had microcytosis. In patients 65 years old and older, $35.3\%$ had microcytosis. Studies in older patients with absolute IDA in developed countries show <$30\%$ have microcytosis [57]. These data strongly support biochemical testing for ID in all patients with anaemia irrespective of MCV values. Only one with microcytosis had a Mentzer index value < 13 [58] but had concomitant ID. This suggests that the thalassaemia trait is uncommon in the study population. While the α-thalassaemia trait is present in some communities in South Africa, with a prevalence of $3.8\%$ [59] and $16\%$ in non-random samples [60], testing is mainly done if there is unexplained microcytosis [60,61]. All MCV measurements in the laboratory were done within 24 h of venesection, with no difference between those measured before and after eight hours. While MCV stored at 4 °C remains unchanged for 24 h, it increases significantly after eight hours at room temperature (standard mean difference −0.30, CI −0.50, −0.10)) [62]. RHC and %HYPO had low sensitivity for IDA, implying that the tests are not suitable for ruling out IDA in the study setting. RHC can perform as well as standard tests for the diagnosis of ID [35]. % HPO has mainly been used to diagnose ID in the setting of CKD [35,39]. While %HPO should be analysed within six hours of venesection [34], Figure 2 shows that there was no significant difference between blood samples drawn before or after six hours. Further research is needed into the sub-optimal performance of these tests in similar settings to this study. The major disease categories associated with moderate and severe anaemia were, in order, HIV, TB, CKD, and cancer, as shown in Table 2. Anaemia is common in PLWHIV from multifactorial causes [63]. While worldwide prevalences in PLWHIV are $21.6\%$, $22.6\%$, and $6.2\%$ for mild, moderate, and severe anaemia, respectively [64], in South Africa, prevalences are mostly higher at $26.7\%$, $41.1\%$, and $4.3\%$ [12]. In this study, $59.5\%$ of those with moderate/severe anaemia were PLWHIV. The majority had low mean CD4 counts, unsuppressed VLs, and CDC Stage 3, which are indicative of advanced HIV disease and a high risk of anaemia. The low prevalence of reported OIs other than TB in this study may be due to limited investigations. Compared to low- and middle-income countries in Latin America and Asia, adults in Sub-Saharan Africa have markedly lower incidences of Pneumocystis jirovecii pneumonia and cerebral toxoplasmosis in ART-naïve patients [65], most likely due to a lack of resources for diagnosis rather than a true difference in infection rates [66]. Patients with severe anaemia were three times more likely to have a diagnosis of TB compared to the other major diagnostic categories (HIV, cancer, and CKD) as demonstrated in the multivariate analysis (Table 6, which only considers patients with laboratory moderate or severe anaemia). While Table 7 shows that in patients with TB and severe anaemia, Hb was higher in those with TB compared to those without, the multivariate analysis is the most accurate summation due to accounting for multiple variables. Table 7 is an analysis of indices to assist in characterising findings in HIV and TB, which are the most prevalent causes, compared to all other causes in patients with moderate/severe anaemia. In PLWHIV with moderate (but not severe) anaemia, Hb was significantly lower compared to those without HIV. HIV frequently causes impaired haematopoiesis, and anaemia is the most common manifestation, increasing in frequency and severity with disease progression [63]. PLWHIV with moderate or severe anaemia had overall low CD4 counts (median values < 200 cells/µL) indicative of more advanced disease [67]. PLWHIV with TB and severe anaemia had lower CD4 counts and higher VL compared to those without TB. This is consistent with HIV as the most potent immunosuppressive risk factor for active TB [68]. HIV infection was associated with a slightly lower RPI that is not clinically significant. Patients with or without HIV or TB and moderate and severe anaemia had normal MCV and low MCHC (indicating hypochromia). However, MCHC is considered of little clinical relevance in interpreting anaemias [69]. Median TSAT was also very low (<$16\%$) in both moderate and severe anaemia without HIV or TB, which is suggestive of absolute ID [56]. Serum iron was lower in severe anaemia without HIV or TB. Serum iron is depressed in both ID and AI and cannot differentiate between the two conditions [70]. Serum transferrin was significantly lower in patients with moderate and severe anaemia and HIV or TB. It is an acute-phase reactant that deceases in inflammation but increases in ID [41]. Ferritin and CRP, both inflammatory markers, were significantly higher in patients with either HIV or TB and moderate or severe anaemia. CRP can be used for diagnosis of pulmonary TB [71] and as an indicator of disease severity [72] and higher mortality [73]. Median CRP levels in patients with TB in this study, i.e., 146 mg/L (IQR 214 mg/L) in moderate and 163 mg/L (IQR 122 mg/L) in severe anaemia, were considerably higher than those of 40 mg/L (IQR 83 mg/L) in U.K. [72] and 17.3 (±37.2 mg/L, SD) in Taiwanese [73] studies and indicative of more severe disease, possibly due to delayed diagnoses. CKD stages 3–5 were present in $6.5\%$ of patients with moderate and severe anaemia. Prevalences of CKD stages 3–5 in Africa are $4.6\%$ in the general population and $13\%$ in high-risk populations, e.g., HIV, diabetes, and hypertension [74], compared with $10.6\%$ in the general population worldwide [75]. The low prevalence of CKD may be due to the strict application of the criteria to define CKD, i.e., abnormalities of kidney structure or function, present for >3 months [46], and non-performance of regular renal evaluations, in patients at high risk, e.g., diabetes mellitus as elsewhere in South Africa [76], and the non-availability of prior patient records with creatinine results. The actual prevalence is most likely double this, as all creatinine-based GFR-estimating equations underestimate GFR in African populations [77]. While cystatin C-based equations may be superior, cystatin C is unaffordable in clinical practice in developing countries [78]. In this study, $32\%$ with CKD had IDA. This compares to a $35.3\%$ IDA prevalence in non-dialysis CKD outpatients in Johannesburg [79] and 15–$72.8\%$ in developed countries [80]. It is important that IDA is recognised and treated to improve quality of life [81]. Most patients with stable CKD can be managed in primary care [82]. In patients with cancer, one of six had IDA. The prevalences of anaemia and ID in different cancers was reported as 29–$46\%$ and 7–$42\%$, respectively, in developed countries (ID prevalence at $60\%$ was highest in colorectal cancer) [5]. No patient had colorectal cancer in this study. There was a low overall prevalence ($2.5\%$) of vitamin B12 deficiency (<147.6 pmol/L) compared to $8.8\%$ in non-pregnant women of childbearing age elsewhere in South Africa [83]. While NHLS normal values for both sexes were 133–675 pmol/L, the lower limit of normal cannot be used as a cut-off to define deficiency due to frequent false-positive and false-negative results [36] Methylmalonic acid and homocysteine levels are required to confirm vitamin B12 deficiency [36], but testing was not mandated in the public-sector guidelines [84]. As seen in this study, neuropsychiatric disorders due to deficiency often occur without haematological changes [36]. While vitamin B12 deficiency can be found in up to $30\%$ of PLWHIV [85], there is little evidence of overt clinical disease [86,87]. A minority of patients with vitamin B12 deficiency exhibited macrocytic changes. This is well documented, particularly in patients with iron deficiency or inflammation [88]. Vitamin B12 values were significantly higher in patients at MRH compared to the CHCs. Levels were higher also in moderate and severe anaemia in PLWHIV and patients with TB. There are reports of higher mortality with elevated vitamin B12 levels, namely >250 pmol/L [89] and >700 pmol/L [90] by mechanisms yet to be elucidated. Low community folate-deficiency prevalences of $0.2\%$ [91] and zero in reproductive-age women [83] were reported in South Africa following folate fortification in 2003. The higher prevalence ($6.2\%$) in this study may be due to inadequate nutrition during illness. The majority ($57.5\%$) of patients had multiple causes of anaemia compared with $13.6\%$ of elderly inpatients in South Africa [18] and $32.5\%$ in Italy [52]. This relates to the high levels of infectious diseases (TB and HIV) in this study population. Approaches in the evaluation of anaemia are traditional evaluation using the MCV and RPI [92,93] and biochemical tests [16,23,40]. South African national guidelines advise categorising anaemia based on MCV and, if microcytic, testing for ID and, if macrocytic, testing for vitamin b12 and folate deficiency [33,94]. As shown in this and other studies [95], this approach will miss many patients with nutrient deficiencies and those with multiple causes. With a substantial prevalence of ID, this study supports the recommendation that iron status needs evaluation in every patient with anaemia [16,32] but also adds the evaluation of folate and vitamin B12 deficiencies. In areas of high TB and HIV prevalence, TB (pulmonary and extra-pulmonary) should be excluded as a cause of moderate and severe anaemia [12,22]. Overall, a combination of MCV/RPI, biochemical, HIV, and TB testing would seem the optimal approach for moderate and severe anaemia evaluation in the study population. The strengths of this study are that causes of anaemia were investigated comprehensively, using biochemical criteria and clinical assessment, and in the context of daily practice in primary care and a district hospital setting. An updated pragmatic definition of ID was used compared with previous studies. The cross-sectional study design is suited to estimate the prevalence of anaemia in the study population [96]. Limitations include missing data due to patient, research assistant, and laboratory factors. Patients had multiple non-interoperable records, making it difficult to obtain historical data. However, there was no indication of a systematic pattern that might invalidate the analysis. Some patients were on drugs associated with anaemia. Due to the cross-sectional study design (with absence of follow-up), it is not possible to state with certainty that any drug was responsible for anaemia. TB was likely responsible for more causes of anaemia, as patients with probable TB on clinical assessment were excluded from analysis. In retrospect, our sample size could have been optimised by using the two-proportions formula with a power of 0.8 instead of using the single-proportion formula, which resulted in only $25.5\%$ of the initial participants being eligible for the final analyses. The findings are not generalisable but may be transferrable to similar settings. ## 5. Conclusions In summary, HIV, ID, and TB were the most prevalent causes of moderate and severe anaemia. Anaemia of inflammation was the most common type of anaemia. Patients with severe anaemia were three times more likely to have TB. Changes in RBC indices do not reliably predict nutritional deficiencies. Biochemical tests should routinely be performed to assess iron, folate, and vitamin B12 deficiencies. ## References 1. 1. WHO The Global Prevalence of Anaemia in 2011World Health OrganizationGeneva, Switzerland2015. The Global Prevalence of Anaemia in 2011 (2015) 2. Safiri S., Kolahi A.-A., Noori M., Nejadghaderi S.A., Karamzad N., Bragazzi N.L., Sullman M.J., Abdollahi M., Collins G.S., Kaufman J.S.. Burden of anemia and its underlying causes in 204 countries and territories, 1990–2019: Results from the Global Burden of Disease Study 2019. J. Hematol. Oncol. (2021) 14 185. DOI: 10.1186/s13045-021-01202-2 3. Fishbane S., Pollack S., Feldman H.I., Joffe M.M.. Iron indices in chronic kidney disease in the National Health and Nutritional Examination Survey 1988–2004. Clin. J. Am. Soc. Nephrol. (2009) 4 57-61. DOI: 10.2215/CJN.01670408 4. Anand I.S., Gupta P.. Anemia and iron deficiency in heart failure: Current concepts and emerging therapies. Circulation (2018) 138 80-98. DOI: 10.1161/CIRCULATIONAHA.118.030099 5. Aapro M., Österborg A., Gascón P., Ludwig H., Beguin Y.. Prevalence and management of cancer-related anaemia, iron deficiency and the specific role of i.v. iron. Ann. Oncol. (2012) 23 1954-1962. DOI: 10.1093/annonc/mds112 6. Yates J., Logan E., Stewart R.. Iron deficiency anaemia in general practice: Clinical outcomes over three years and factors influencing diagnostic investigations. Postgrad. Med. J. (2004) 80 405-410. PMID: 15254305 7. McCartney D., Shine B., Hay D., Lasserson D.S.. The evaluation of anaemia in an older primary care population: Retrospective population-based study. BJGP Open (2017) 1 bjgpopen17X101157. DOI: 10.3399/bjgpopen17X101157 8. Boennelykke A., Jensen H., Falborg A.Z., Granfeldt Østgård L.S., Hansen A.T., Christensen K.S., Vedsted P.. Diagnostic workup of cancer in patients with new-onset anaemia: A Danish cohort study in general practice. Scand. J. Prim. Health Care (2021) 39 391-402. DOI: 10.1080/02813432.2021.1970934 9. Shisana O., Labadarios D., Rehle T., Simbayi L., Zuma K., Dhansay A., Reddy P., Parker W., Hoosain E., Naidoo P.. South African National Health and Nutrition Examination Survey (SANHANES-1) (2013) 10. 10. National Department of Health (NDoH) Statistics South Africa (Stats SA) South African Medical Research Council (SAMRC) ICF South Africa Demographic and Health Survey 2016National Department of HealthPretoria, South AfricaRockville, MD, USA2019. South Africa Demographic and Health Survey 2016 (2019) 11. Stats S.. Mid-Year Population Estimates 2021 (2021) 12. Kerkhoff A.D., Wood R., Cobelens F.G., Gupta-Wright A., Bekker L.-G., Lawn S.D.. The predictive value of current haemoglobin levels for incident tuberculosis and/or mortality during long-term antiretroviral therapy in South Africa: A cohort study. BMC Med. (2015) 13. DOI: 10.1186/s12916-015-0320-9 13. 13. SAMRC HSRC NICD The First National TB Prevalence Survey South Africa 2018Short ReportSouth african Medical Research CouncilCape Town, South Africa2020. The First National TB Prevalence Survey South Africa 2018 (2020) 14. Gelaw Y., Getaneh Z., Melku M.. Anemia as a risk factor for tuberculosis: A systematic review and meta-analysis. Environ. Health Prev. Med. (2021) 26 13. DOI: 10.1186/s12199-020-00931-z 15. Barzegari S., Afshari M., Movahednia M., Moosazadeh M.. Prevalence of anemia among patients with tuberculosis: A systematic review and meta-analysis. Indian J. Tuberc. (2019) 66 299-307. PMID: 31151500 16. Weiss G., Ganz T., Goodnough L.T.. Anemia of inflammation. Blood J. Am. Soc. Hematol. (2019) 133 40-50. DOI: 10.1182/blood-2018-06-856500 17. Walsh C.M., Hattingh Z., Veldman F.J., Bester C.J.. Iron status and anaemia of chronic disease in HIV-infected African women in Mangaung, Bloemfontein. S. Afr. Fam. Pract. (2010) 52 55-59. DOI: 10.1080/20786204.2010.10873936 18. Van Staden A.M., Weich D.J.. Retrospective analysis of the prevalence and causes of anaemia in hospitalised elderly patients. S. Afr. Fam. Pract. (2015) 57 297-299. DOI: 10.1080/20786190.2015.1071538 19. Claster S.. Biology of anemia, differential diagnosis, and treatment options in human immunodeficiency virus infection. J. Infect. Dis. (2002) 185 S105-S109. DOI: 10.1086/340202 20. Lewis D.K., Whitty C.J.M., Walsh A.L., Epino H., Broek N.R., Letsky E.A., Munthali C., Mukiibi J.M., Boeree M.J.. Treatable factors associated with severe anaemia in adults admitted to medical wards in Blantyre, Malawi, an area of high HIV seroprevalence. Trans. R. Soc. Trop. Med. Hyg. (2005) 99 561-567. DOI: 10.1016/j.trstmh.2005.01.002 21. Kerkhoff A.D., Lawn S.D., Schutz C., Burton R., Boulle A., Cobelens F.J., Meintjes G.. Anemia, Blood Transfusion Requirements and Mortality Risk in Human Immunodeficiency Virus-Infected Adults Requiring Acute Medical Admission to Hospital in South Africa. Open Forum Infect. Dis. (2015) 2 ofv173. DOI: 10.1093/ofid/ofv173 22. Mntonintshi M., O’Mahony D., Mabunda S., Namugeyi K.A.F.. Undiagnosed tuberculosis in patients with HIV infection who present with severe anaemia at a district hospital. Afr. J. Prm. Health Care Fam. Med. (2017) 9 a1406 23. Kerkhoff A.D., Meintjes G., Opie J., Vogt M., Jhilmeet N., Wood R., Lawn S.D.. Anaemia in patients with HIV-associated TB: Relative contributions of anaemia of chronic disease and iron deficiency. Int. J. Tuberc. Lung Dis. (2016) 20 193-201. DOI: 10.5588/ijtld.15.0558 24. Stats S.A.. Poverty Mapping in South Africa. Applying Small Area Estimation Techniques Using IES 2010/11 and Census 2011 (2018) 25. Joubert G., Ehrlich R., Katzenellenbogen J.M., Karim S.A.. Epidemiology: A Research Manual for South Africa (2007) 26. Whitley E., Ball J.. Statistics review 4: Sample size calculations. Crit Care (2002) 6 335-341. DOI: 10.1186/cc1521 27. Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. Vitamin and Mineral Nutrition Information System (2011) 28. 28. NDoH Implementation of the Universal Test and Treat Strategy for HIV Positive Patients and Differentiated Care for Stable PatientsCircular dated 16 August 2016National Department of HealthPretoria, South Africa2016. Implementation of the Universal Test and Treat Strategy for HIV Positive Patients and Differentiated Care for Stable Patients (2016) 29. Essential Drugs Programme. Primary Healthcare Standard Treatment Guidelines and Essential Medicines List (2014) 30. 30. NDoH National Consolidated Guidelines for the Prevention of Mother-to-Child Transmission of HIV (PMTCT) and the Management of HIV in Children, Adolescents and AdultsDepartment of HealthPretoria, South Africa2015. National Consolidated Guidelines for the Prevention of Mother-to-Child Transmission of HIV (PMTCT) and the Management of HIV in Children, Adolescents and Adults (2015) 31. Heller T., Wallrauch C., Goblirsch S., Brunetti E.. Focused assessment with sonography for HIV-associated tuberculosis (FASH): A short protocol and a pictorial review. Crit. Ultrasound J. (2012) 4 21. DOI: 10.1186/2036-7902-4-21 32. Adamson J.W., Longo D.L., Kasper D., Fauci A., Hauser S., Longo D., Jameson J., Loscalzo J.. Anemia and Polycythemia. Chapter 77. Harrison’s Principles of Internal Medicine, 19e (2015) 33. Essential Drugs Programme. Hospital Level (Adults) Standard Treatment Guidelines and Essential Medicines List (2019) 34. 34. NICE Anaemia Management in People with Chronic Kidney DiseaseNational Institute for Health and Care ExcellenceLondon, UK2015. Anaemia Management in People with Chronic Kidney Disease (2015) 35. Brugnara C.. Diagnostic Approaches to Iron Deficient States and Iron Deficiency Anemia. Haema (2019) 10 2-14 36. Stabler S.P.. Clinical practice. Vitamin B12 deficiency. N. Engl. J. Med. (2013) 368 149-160. DOI: 10.1056/NEJMcp1113996 37. 37. WHO Serum and Red Blood Cell Folate Concentrations for Assessing Folate Status in PopulationsWHO/NMH/NHD/EPG/15.01World Health OrganizationGeneva, Switzerland2015. Serum and Red Blood Cell Folate Concentrations for Assessing Folate Status in Populations (2015) 38. Mast A.E., Blinder M.A., Gronowski A.M., Chumley C., Scott M.G.. Clinical utility of the soluble transferrin receptor and comparison with serum ferritin in several populations. Clin. Chem. (1998) 44 45-51. DOI: 10.1093/clinchem/44.1.45 39. Camaschella C., Girelli D.. The changing landscape of iron deficiency. Mol. Asp. Med. (2020) 75 100861. DOI: 10.1016/j.mam.2020.100861 40. Cappellini M.D., Musallam K.M., Taher A.T.. Iron deficiency anaemia revisited. J. Intern. Med. (2020) 287 153-170. DOI: 10.1111/joim.13004 41. Ganz T.. Anemia of Inflammation. N. Engl. J. Med. (2019) 381 1148-1157. DOI: 10.1056/NEJMra1804281 42. Ponikowski P., Voors A.A., Anker S.D., Bueno H., Cleland J.G., Coats A.J., Falk V., González-Juanatey J.R., Harjola V.-P., Jankowska E.A.. 2016 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur. Heart J. (2016) 37 2129-2200. DOI: 10.1093/eurheartj/ehw128 43. Aapro M., Beguin Y., Bokemeyer C., Dicato M., Gascón P., Glaspy J., Hofmann A., Link H., Littlewood T., Ludwig H.. Management of anaemia and iron deficiency in patients with cancer: ESMO Clinical Practice Guidelines. Ann. Oncol. (2018) 29 iv96-iv110. DOI: 10.1093/annonc/mdx758 44. 44. NICE Chronic Kidney Disease: Assessment and Management (NG203)National Institute for Health and Care ExcellenceLondon, UK2021. Chronic Kidney Disease: Assessment and Management (NG203) (2021) 45. KDIGO clinical practice guidelines for acute kidney injury. Kidney Int. Suppl. (2012) 2 1-138 46. KDIGO 2012. Clinical Practice Guideline for the Evaluation and Management of Chronic Kidney Disease. Kidney Int. (2013) 3 S1-S150. DOI: 10.1038/kisup.2012.73 47. Mikhail A., Brown C., Williams J.A., Mathrani V., Shrivastava R., Evans J., Isaac H., Bhandari S.. Clinical Practice Guideline: Anaemia of Chronic Kidney Disease (2017) 48. Selik R.M., Mokotoff E.D., Branson B.M., Owen M., Whitmore S., Hall H.I.. Revised surveillance case definition for HIV infection—United States, 2014. Morb. Mortal. Wkly. Rep. Recomm. Rep. (2014) 63 1-10 49. Mash B., Fairall L., Adejayan O., Ikpefan O., Kumari J., Mathee S., Okun R., Yogolelo W.. A morbidity survey of South African primary care. PLoS ONE (2012) 7. DOI: 10.1371/annotation/3545077e-aded-4eef-a460-be1edbd1845c 50. Payne C.F., Davies J.I., Gomez-Olive F.X., Hands K.J., Kahn K., Kobayashi L.C., Tipping B., Tollman S.M., Wade A., Witham M.D.. Cross-sectional relationship between haemoglobin concentration and measures of physical and cognitive function in an older rural South African population. J. Epidemiol. Community Health (2018) 72 796-802. DOI: 10.1136/jech-2018-210449 51. Whitehead R.D., Mei Z., Mapango C., Jefferds M.E.D.. Methods and analyzers for hemoglobin measurement in clinical laboratories and field settings. Ann. N. Y. Acad. Sci. (2019) 1450 147-171. DOI: 10.1111/nyas.14124 52. Randi M.L., Bertozzi I., Santarossa C., Cosi E., Lucente F., Bogoni G., Biagetti G., Fabris F.. Prevalence and Causes of Anemia in Hospitalized Patients: Impact on Diseases Outcome. J. Clin. Med. (2020) 9. DOI: 10.3390/jcm9040950 53. Guralnik J.M., Eisenstaedt R.S., Ferrucci L., Klein H.G., Woodman R.C.. Prevalence of anemia in persons 65 years and older in the United States: Evidence for a high rate of unexplained anemia. Blood (2004) 104 2263-2268. DOI: 10.1182/blood-2004-05-1812 54. Guyatt G.H., Oxman A.D., Ali M., Willan A., McIlroy W., Patterson C.. Laboratory diagnosis of iron-deficiency anemia. J. Gen. Intern. Med. (1992) 7 145-153. DOI: 10.1007/BF02598003 55. Garcia-Casal M.N., Pasricha S.-R., Martinez R.X., Lopez-Perez L., Pena-Rosas J.P.. Are current serum and plasma ferritin cut-offs for iron deficiency and overload accurate and reflecting iron status? A systematic review. Arch. Med. Res. (2018) 49 405-417. DOI: 10.1016/j.arcmed.2018.12.005 56. Cacoub P., Vandewalle C., Peoc’h K.. Using transferrin saturation as a diagnostic criterion for iron deficiency: A systematic review. Crit. Rev. Clin. Lab. Sci. (2019) 56 526-532. DOI: 10.1080/10408363.2019.1653820 57. Girelli D., Marchi G., Camaschella C.. Anemia in the elderly. HemaSphere (2018) 2 e40. DOI: 10.1097/HS9.0000000000000040 58. Mentzer W.. Differentiation of iron deficiency from thalassaemia trait. Lancet (1973) 301 882. DOI: 10.1016/S0140-6736(73)91446-3 59. Bird A., Ellis P., Wood K., Mathew C., Karabus C.. Inherited haemoglobin variants in a South African population. J. Med. Genet. (1987) 24 215-219. DOI: 10.1136/jmg.24.4.215 60. Loonat S.B., Naran N.H., Thein S.L., Alli N.A.. Alpha-thalassaemia trait as a cause of unexplained microcytosis in a South African population. S. Afr. Med. J. (2016) 106 276-279. DOI: 10.7196/SAMJ.2016.v106i3.10005 61. Krause A., Wainstein T., Essop F.B., Goodyear Q.. Testing for haemoglobinopathies in Johannesburg, South Africa: A 30-year review. S. Afr. Med. J. (2013) 103 989-993. DOI: 10.7196/SAMJ.7255 62. Wu D.-W., Li Y.-M., Wang F.. How long can we store blood samples: A systematic review and meta-analysis. EBioMedicine (2017) 24 277-285. DOI: 10.1016/j.ebiom.2017.09.024 63. Durandt C., Potgieter J.C., Khoosal R., Nel J.G., Herd C.L., Mellet J., Rossouw T., Pepper M.S.. HIV and haematopoiesis. S. Afr. Med. J. (2019) 109 S41-S46. DOI: 10.7196/SAMJ.2019.v109i8b.13829 64. Cao G., Wang Y., Wu Y., Jing W., Liu J., Liu M.. Prevalence of anemia among people living with HIV: A systematic review and meta-analysis. eClinicalMedicine (2022) 44 101283. DOI: 10.1016/j.eclinm.2022.101283 65. Low A., Gavriilidis G., Larke N., B-Lajoie M.R., Drouin O., Stover J., Muhe L., Easterbrook P.. Incidence of Opportunistic Infections and the Impact of Antiretroviral Therapy Among HIV-Infected Adults in Low- and Middle-Income Countries: A Systematic Review and Meta-analysis. Clin. Infect. Dis. (2016) 62 1595-1603. DOI: 10.1093/cid/ciw125 66. Morris A., Lundgren J.D., Masur H., Walzer P.D., Hanson D.L., Frederick T., Huang L., Beard C.B., Kaplan J.E.. Current epidemiology of Pneumocystis pneumonia. Emerg. Infect. Dis. (2014) 10 1713. DOI: 10.3201/eid1010.030985 67. Ingle S.M., May M.T., Gill M.J., Mugavero M.J., Lewden C., Abgrall S., Fätkenheuer G., Reiss P., Saag M.S., Manzardo C.. Impact of Risk Factors for Specific Causes of Death in the First and Subsequent Years of Antiretroviral Therapy Among HIV-Infected Patients. Clin. Infect. Dis. (2014) 59 287-297. DOI: 10.1093/cid/ciu261 68. Narasimhan P., Wood J., MacIntyre C.R., Mathai D.. Risk factors for tuberculosis. Pulm. Med. (2013) 2013 828939. DOI: 10.1155/2013/828939 69. Merritt B.Y.. Mean Corpuscular Hemoglobin (MCH) and Mean Corpuscular Hemoglobin Concentration (MCHC) (2019) 70. Camaschella C.. Iron-Deficiency Anemia. N. Engl. J. Med. (2015) 372 1832-1843. DOI: 10.1056/NEJMra1401038 71. Yoon C., Chaisson L.H., Patel S.M., Allen I.E., Drain P.K., Wilson D., Cattamanchi A.. Diagnostic accuracy of C-reactive protein for active pulmonary tuberculosis: A meta-analysis. Int. J. Tuberc. Lung Dis (2017) 21 1013-1019. DOI: 10.5588/ijtld.17.0078 72. Brown J., Clark K., Smith C., Hopwood J., Lynard O., Toolan M., Creer D., Barker J., Breen R., Brown T.. Variation in C—Reactive protein response according to host and mycobacterial characteristics in active tuberculosis. BMC Infect. Dis. (2016) 16. DOI: 10.1186/s12879-016-1612-1 73. Huang C.-T., Lee L.-N., Ho C.-C., Shu C.-C., Ruan S.-Y., Tsai Y.-J., Wang J.-Y., Yu C.-J.. High serum levels of procalcitonin and soluble TREM-1 correlated with poor prognosis in pulmonary tuberculosis. J. Infect. (2014) 68 440-447. DOI: 10.1016/j.jinf.2013.12.012 74. Kaze A.D., Ilori T., Jaar B.G., Echouffo-Tcheugui J.B.. Burden of chronic kidney disease on the African continent: A systematic review and meta-analysis. BMC Nephrol. (2018) 19. DOI: 10.1186/s12882-018-0930-5 75. Hill N.R., Fatoba S.T., Oke J.L., Hirst J.A., O’Callaghan C.A., Lasserson D.S., Hobbs F.R.. Global prevalence of chronic kidney disease–a systematic review and meta-analysis. PLoS ONE (2016) 11. DOI: 10.1371/journal.pone.0158765 76. Webb E.M., Rheeder P., Van Zyl D.G.. Diabetes care and complications in primary care in the Tshwane district of South Africa. Prim. Care Diabetes (2015) 9 147-154. DOI: 10.1016/j.pcd.2014.05.002 77. Fabian J., Kalyesubula R., Mkandawire J., Hansen C.H., Nitsch D., Musenge E., Nakanga W.P., Prynn J.E., Dreyer G., Snyman T.. Measurement of kidney function in Malawi, South Africa, and Uganda: A multicentre cohort study. Lancet Glob. Health (2022) 10 e1159-e1169. DOI: 10.1016/S2214-109X(22)00239-X 78. Niang A., Luyckx V.A.. Measuring kidney function: The voice of Africa. Lancet Glob. Health (2022) 10 e1080-e1081. DOI: 10.1016/S2214-109X(22)00279-0 79. Nalado A.M., Mahlangu J.N., Waziri B., Duarte R., Paget G., Olorunfemi G., Naicker S.. Ethnic prevalence of anemia and predictors of anemia among chronic kidney disease patients at a tertiary hospital in Johannesburg, South Africa. Int. J. Nephrol. Renov. Dis. (2019) 12 19-32. DOI: 10.2147/IJNRD.S179802 80. Babitt J.L., Eisenga M.F., Haase V.H., Kshirsagar A.V., Levin A., Locatelli F., Małyszko J., Swinkels D.W., Tarng D.-C., Cheung M.. Controversies in optimal anemia management: Conclusions from a Kidney Disease: Improving Global Outcomes (KDIGO) Conference. Kidney Int. (2021) 99 1280-1295. DOI: 10.1016/j.kint.2021.03.020 81. Kidney Disease: Improving Global Outcomes (KDIGO) Anaemia Work Group clinical practice guideline for anemia in chronic kidney disease. Kidney Int. Suppl. (2012) 2 279-335 82. John R., Webb M., Young A., Stevens P.E.. Unreferred chronic kidney disease: A longitudinal study. Am. J. Kidney Dis. (2004) 43 825-835. DOI: 10.1053/j.ajkd.2003.12.046 83. Modjadji S., Alberts M., Mamabolo R.. Folate and iron status of South African non-pregnant rural women of childbearing age, before and after fortification of foods. S. Afr. J. Clin. Nutr. (2007) 20 89-93. DOI: 10.1080/16070658.2007.11734132 84. 84. NDOH Standard Treatment Guidelines and Essential Medicines List for South Africa. Hospital Level, Adults4th ed.NDOHPretoria, South Africa2015. Standard Treatment Guidelines and Essential Medicines List for South Africa. Hospital Level, Adults (2015) 85. Vishnu P., Aboulafia D.M.. Haematological manifestations of human immune deficiency virus infection. Br. J. Haematol. (2015) 171 695-709. DOI: 10.1111/bjh.13783 86. Paltiel O., Falutz J., Veilleux M., Rosenblatt D.S., Gordon K.. Clinical correlates of subnormal vitamin B12 levels in patients infected with the human immunodeficiency virus. Am. J. Hematol. (1995) 49 318-322. DOI: 10.1002/ajh.2830490410 87. Hepburn M.J., Dyal K., Runser L.A., Barfield R.L., Hepburn L.M., Fraser S.L.. Low serum vitamin B12 levels in an outpatient HIV-infected population. Int. J. STD AIDS (2004) 15 127-133. DOI: 10.1258/095646204322764334 88. Metz J.. Haematological implications of folate food fortification. S. Afr. Med. J. (2013) 103 978-981. DOI: 10.7196/SAMJ.7022 89. Flores-Guerrero J.L., Minović I., Groothof D., Gruppen E.G., Riphagen I.J., Kootstra-Ros J., Kobold A.M., Hak E., Navis G., Gansevoort R.T.. Association of plasma concentration of vitamin B12 with all-cause mortality in the general population in the Netherlands. JAMA Netw. Open (2020) 3 e1919274. DOI: 10.1001/jamanetworkopen.2019.19274 90. Wolffenbuttel B.H.R., Heiner-Fokkema M.R., Green R., Gans R.O.B.. Relationship between serum B12 concentrations and mortality: Experience in NHANES. BMC Med. (2020) 18. DOI: 10.1186/s12916-020-01771-y 91. Labadarios D.. National Food Consumption SurveyFortification Baseline (NFCS-FB): South Africa, 2005 (2007) 92. Greer J.P., Arber D.A., Glader B., List A.F., Means R.T.J., Paraskevas F., Rodgers G.M.. Wintrobe’s Clinical Hematology (2014) 93. Kasper D., Fauci A., Hauser S., Longo D., Jameson J., Loscalzo J.. Harrison’s Principles of Internal Medicine, 19e (2015) 94. Essential Drugs Programme. Primary Healthcare Standard Treatment Guideline and Essential Medicine List (2020) 95. Petrosyan I., Blaison G., Andrès E., Federici L.. Anaemia in the elderly: An aetiologic profile of a prospective cohort of 95 hospitalised patients. Eur. J. Intern. Med. (2012) 23 524-528. DOI: 10.1016/j.ejim.2012.03.013 96. Sedgwick P.. Cross sectional studies: Advantages and disadvantages. BMJ (2014) 348 g2276. DOI: 10.1136/bmj.g2276
--- title: YTHDF1 Attenuates TBI-Induced Brain-Gut Axis Dysfunction in Mice authors: - Peizan Huang - Min Liu - Jing Zhang - Xiang Zhong - Chunlong Zhong journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966860 doi: 10.3390/ijms24044240 license: CC BY 4.0 --- # YTHDF1 Attenuates TBI-Induced Brain-Gut Axis Dysfunction in Mice ## Abstract The brain-gut axis (BGA) is a significant bidirectional communication pathway between the brain and gut. Traumatic brain injury (TBI) induced neurotoxicity and neuroinflammation can affect gut functions through BGA. N6-methyladenosine (m6A), as the most popular posttranscriptional modification of eukaryotic mRNA, has recently been identified as playing important roles in both the brain and gut. However, whether m6A RNA methylation modification is involved in TBI-induced BGA dysfunction is not clear. Here, we showed that YTHDF1 knockout reduced histopathological lesions and decreased the levels of apoptosis, inflammation, and oedema proteins in brain and gut tissues in mice after TBI. We also found that YTHDF1 knockout improved fungal mycobiome abundance and probiotic (particularly Akkermansia) colonization in mice at 3 days post-CCI. Then, we identified the differentially expressed genes (DEGs) in the cortex between YTHDF1-knockout and WT mice. These genes were primarily enriched in the regulation of neurotransmitter-related neuronal signalling pathways, inflammatory signalling pathways, and apoptotic signalling pathways. This study reveals that the ITGA6-mediated cell adhesion molecule signalling pathway may be the key feature of m6A regulation in TBI-induced BGA dysfunction. Our results suggest that YTHDF1 knockout could attenuate TBI-induced BGA dysfunction. ## 1. Introduction Traumatic brain injury (TBI) is a major public health concern, with up to 69 million TBIs occurring worldwide each year [1,2]. Due to the poor understanding of TBI’s heterogeneity and complexity, there has been limited clinical success, leading to a huge social and economic burden. TBI exerts profound effects on the gut’s functions [3]. The brain–gut axis (BGA), a bidirectional communication network connecting the central nervous system (CNS) and enteric nervous system (ENS), is the key to CNS and gastrointestinal homeostasis and regulates diverse functions including gut barrier functions, intestinal motility, and neurobehaviours [4]. TBI-induced neurotoxicity and neuroinflammation can affect gut functions through BGA [5,6]. However, the mechanisms of BGA dysfunction induced by TBI are still elusive, especially in posttranscriptional regulation. N6-methyladenosine (m6A), the most prevalent mRNA posttranscriptional modification in eukaryotes, requires various regulatory proteins encoded by writing genes (writers), erasing genes (erasers), and reading genes (readers) [7]. An increasing amount of research has revealed that m6A modification can influence almost all aspects of RNA metabolism, including RNA transcription, splicing, nuclear output, translation, decay, and RNA-protein interactions [8,9,10,11]. Substantial lines of study have shown that m6A exquisitely regulates various spatial and temporal physiological processes, including gametogenesis, embryogenesis, cell fate determination, sex determination, DNA damage response, circadian rhythm, heat shock response, pluripotency, and neuronal functions [10,12,13]. It has recently been identified that m6A plays significant roles in both the brain and gut. m6A modification is abundant in the CNS, modulates the activation of various nerve conduction pathways, and plays an important role in the development, differentiation, and regeneration of neurons [14]. Meanwhile, it shows a significant impact on the communication between the gut microbiome and the host [15,16]. Thus, m6A modification may also play an important role in BGA. As one of the m6A reader proteins, YTHDF1 promotes mRNA translation. The knockout of YTHDF1 does not alter the m6A/A ratio of total mRNA but impacts on the association of its target mRNA with ribosome [17]. YTHDF1 regulates the subcellular distribution and translation status of the m6A-modified mRNA [17]. YTHDF1 plays an important role in both the brain and gut. In the brain, YTHDF1 recognizes and binds the mRNAs of m6A-modified glutamate ionotropic receptor NMDA type subunit 1 and 2A (GRIN1, GRIN2A), glutamate ionotropic receptor AMPA 1 (GRIA1), calcium/calmodulin-dependent protein kinase (CaM kinase) II alpha and II beta (CAMK2A, CAMK2B) genes to promote translation, while YTHDF1 deletion causes the ectopic translation of related proteins, resulting in memory loss [18]; moreover, the upregulation of YTHDF1 promotes cancer stem cell properties in glioblastoma cells [19]. In the gut, YTHDF1 recognizes the target TNF receptor associated factor 6 (TRAF6) transcript to modulate the gut immune response to bacterial infection, by the unique interaction mechanism between the P/Q/N-rich domain and host factor DEAD (Asp-Glu-Ala-Asp) box polypeptide 60 (DDX60) death domain [20]. YTHDF1-mediated exportin 1 (XPO1) activates the NF-κB pathway and then induces an increased expression of IL-8 in gut cells, resulting in the development and aggravation of celiac disease [21]. YTHDF1 exhibits the highest diagnostic value for chronic obstructive pulmonary disease (COAD) [22]. However, the roles of YTHDF1 in the TBI-induced brain-gut axis dysfunction are unclear; therefore, the functions of YTHDF1 in TBI need to be investigated. To confirm the involvement of m6A RNA methylation in TBI-induced BGA dysfunction, we performed controlled cortical impact (CCI) on YTHDF1-knockout mice and C57BL/6J mice, and compared their differences in the brain defect area; surviving neuron count; the levels of apoptosis, inflammation, and oedema proteins in brain tissues; the ratio of villus height to crypt depth (V/C); the levels of apoptosis and inflammation proteins in gut tissues; and the composition of the faecal microbiome. Then, using RNA-seq, we identified the differentially expressed genes (DEGs) in the cortex and further explored the possible pathways and key genes of YTHDF1 regulating TBI-induced BGA dysfunction. ## 2.1. YTHDF1-Knockout Decreases the Cortical Tissue Losses while Increasing the Neuronal Cell Survivals and the Colon Tissue V/C Ratio after CCI Three days after CCI, HE revealed that there was no significant brain tissue loss in either the WT + sham group or the YTHDF1-knockout + sham group, while marked brain tissue losses were observed in the CCI groups compared with the sham groups (Figure 1B). Notably, compared with the YTHDF1-knockout + CCI group, there was a significant loss of cortical tissue in the WT + CCI group (Figure 1B,C) ($p \leq 0.05$). In addition, compared with the sham groups, the YTHDF1-knockout + CCI group and WT + CCI group displayed a significant decrease in the total neuron count in the perilesional zone to trauma, with a greater decrease in the WT + CCI group. However, neuronal cell survivals were not significantly different between the YTHDF1-knockout + sham group and the WT + sham group (Figure 1D,E). In the colon tissue at 3 days after CCI, the villus height decreased while the crypt depth increased, and the V/C ratios of the YTHDF1-knockout + CCI group and WT + CCI group significantly decreased compared with that of the sham groups. Notably, the V/C ratio of the WT + CCI group decreased more markedly than that compared with the YTHDF1-knockout + CCI group. However, there was no significant difference between the YTHDF1-knockout + sham group and the WT + sham group (Figure 2). ## 2.2. YTHDF1-Knockout Decreases Pro-Apoptosis, Pro-Inflammation and Pro-Oedema Protein Levels and Increases the Anti-Apoptosis and Anti-Oedema Protein Levels following CCI At 3 days after CCI, Western blots revealed that the apoptosis proteins BCL2-associated X protein (BAX), caspase3, and Cleaved caspase3 were notably decreased, while B-cell CLL/lymphoma 2 (BCL2) was markedly increased in the YTHDF1-knockout group compared with the WT group in injured cortical tissues (Figure 3A). Meanwhile, compared with the WT group, the inflammatory protein CD68 significantly decreased, forkhead box P3 (FOXP3) markedly increased (Figure 3B), and the oedema protein aquaporin 4 (AQP4) notably decreased, while claudin5 markedly increased (Figure 3C) in the YTHDF1-knockout group. In colon tissues, the pro-apoptosis proteins, such as BAX, Caspase3, and Cleaved caspase3 in the YTHDF1-knockout group, displayed a significant decrease, while the anti-apoptosis protein BCL2 exhibited a notable increase compared with the WT group at 3 days post-CCI (Figure 4A). In addition, the pro-inflammation protein CD68 reduced markedly and the anti-inflammation protein FOXP3 was significantly enhanced in the YTHDF1-knockout group compared with the WT group (Figure 4B). ## 2.3. YTHDF1-Knockout Increases the Abundance of Fungal Mycobiome, and Alters Mycobiome Structure and Mycobiome Colonization after CCI The raw data from the Illumina (San Diego, CA, USA) sequencing were processed using the Quantitative Insights Into Microbial Ecology (QIIME2) pipeline, and their relative abundance was calculated and grouped according to mouse origin (Figure 5). The results revealed the different profiles of fungal communities in the faecal mycobiomes of YTHDF1-knockout and WT mice at 3 days after CCI. The majority of fungal communities were in the Firmicutes phylum, which dominated the faecal mycobiome of mice, contributing approximately > $50.7\%$ of the total identified fungi. The second major phylum in the make-up of the YTHDF1-knockout mouse faecal mycobiome was Bacteroidetes (accounting for $24.8\%$), while in WT mice, the faecal mycobiome was Proteobacteria (accounting for $24.8\%$). The third major phylum in the composition of the YTHDF1-knockout mouse faecal mycobiome was Verrucomicrobia (accounting for $15.7\%$), while in the WT mouse faecal mycobiome it was Bacteroidetes (accounting for $19.0\%$). At the order level, the results indicated many intriguing composition differences in the faecal mycobiome between YTHDF1-knockout and WT mice at 3 days post-CCI. The most abundant fungi in the YTHDF1-knockout group were Eubacteriales ($26.5\%$), Bacteroidales ($22.9\%$), Lactobacillales ($16.0\%$), Verrucomicrobiales ($15.7\%$), and Erysipelotrichales ($7\%$), while the WT group mycobiomes were dominated by Lactobacillales ($26.6\%$), Eubacteriales ($23.5\%$), Desulfovibrionales ($12.9\%$), Bacteroidales ($9.0\%$), and Chitinophagales ($9.0\%$). For alpha diversity, the ACE and Chao1 indices show the abundance of the fungal mycobiome, while Shannon’s and Simpson’s indices indicate the diversity of the fungal mycobiome. The results demonstrated that the ACE and Chao1 indices of fungal microbiomes in YTHDF1-knockout mice were markedly higher than that in WT mice at 3 days after CCI ($p \leq 0.05$) (Figure 6A,B). However, the Shannon and Simpson indices of fungal microbiomes in YTHDF1-knockout and WT mice were similar (Figure 6C,D). The beta diversity revealing the community distance between samples was evaluated by weighted and unweighted UniFrac distance (Figure 7A,B), which exhibited a significant difference ($p \leq 0.05$) in mycobiome profiles between the YTHDF1-knockout and WT groups at 3 days following CCI. The result of the clustering dendrogram of mouse faecal bacteria was consistent with the above conclusion (Figure 7C). The differential taxa between YTHDF1-knockout and WT mice were analysed using the linear discriminant analysis effect size (LEfSe) method. The results in Figure 8A demonstrate the enriched fungal taxa in each group that had a >two-fold change and $p \leq 0.05$ (Kruskal-Wallis test). The faecal mycobiome of YTHDF1-knockout mice greatly diverged from that of WT mice at 3 days post-CCI (Figure 8B), revealing dissimilar mycobiome colonization. The fungi Akkermansia, Verrucomicrobia, Muribaculum, Erysipelotrichia, and Dubosiella were enriched in the YTHDF1-knockout group, while the fungi from various genera, such as Proteobacteria, Kineothrix, Desulfovibrio, Chitinophagia, and Sediminibacterium, were enriched in the WT mice. ## 2.4. YTHDF1-Knockout Affects Gene Expression in the Mouse Cerebral Cortex Post-TBI To further investigate the influence of YTHDF1-knockout on gene expression, using the RNA sequencing data, we measured the levels of mRNA changes in mice cerebral cortex after TBI. Figure 9A reveals the principal component analyses of the YTHDF1-knockout and WT groups. It shows that the confidence ellipses of the samples among the YTHDF1-knockout and WT groups are separated from each other, suggesting that the gene expression patterns are similar within the same treatment group, but markedly different between the YTHDF1-knockout and WT groups (Figure 9A). The volcano plot (Figure 9B) reveals the significantly upregulated and downregulated mRNA in the YTHDF1-knockout group. We marked the top five most significantly upregulated genes (Cdh15, Col23a1, Fcrls, Tox, and Slc17a6) and top five downregulated genes (Kdm5d, Ddx3y, Eif2s3y, Uty, and Lyrm7) in the volcano plot. The heatmap reveals the relative expression levels of the three YTHDF1-knockout samples, and the three WT samples in the same group have similar presentation (Figure 9C). It shows that a total of 154 mRNAs had increased expression and 57 mRNAs decreased expression ($p \leq 0.05$, log2FC > 1) in YTHDF1-knockout mouse cerebral cortex after TBI (Figure 9D). The top 20 differentially changed mRNAs are displayed in Table 1. Gene Ontology (GO) analyses revealed that the YTHDF1-knockout was associated with three parts of biological information—biological process: localization, aromatic amino acid family metabolic process, response to peptide hormone; cellular component: membrane, MKS complex, cytoplasmic microtubule; and molecular function: ionotropic glutamate receptor activity, glutamate receptor activity, GDP-dissociation inhibitor activity (Figure 10A). The Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses showed the significantly changed genes in the YTHDF1-knockout mice. The most significant KEGG pathways related to these genes included neuroactive ligand-receptor interaction, glutamatergic synapse, axon guidance, and cAMP signalling pathway (Figure 10B). ## 3. Discussion An increasing number of studies have focused on the role of RNA m6A methylation in the development of various neurological diseases, such as Parkinson’s disease [23], Alzheimer’s disease [24,25], multiple sclerosis [26], tumours [27,28,29], epilepsy [30], and neuropsychiatric disorders [31]. However, to date, there is limited research focusing on the role of RNA m6A methylation in TBI and BGA. We observed that YTHDF1-knockout could reduce the brain defect area, rescue neuron cells, and downregulate the levels of apoptosis, inflammation, and oedema proteins in the brain tissues of mice with CCI treatment. Meanwhile, YTHDF1-knockout increased the ratio of V/C and downregulated the levels of apoptosis and inflammation proteins in the gut tissues of mice post-CCI. Then, we identified that the DEGs in the cortex between YTHDF1-knockout and WT mice were primarily enriched in the regulation of neurotransmitter-related neuronal signalling pathways, inflammatory signalling pathways, and apoptotic signalling pathways. Our results revealed that the ITGA6-mediated cell adhesion molecules signalling pathway might be the key feature of m6A regulation in TBI-induced BGA dysfunction. Then, using RNA-Seq, we identified the DEGs in the cortex between YTHDF1-knockout and WT mice and investigated the possible m6A-mediated signalling pathways which were involved in TBI-induced BGA dysfunction. After TBI, neuronal cells release a large number of pro-inflammatory cytokines, such as TNF-α, IL-1, and IL-6, leading to gut inflammation and damage to the gut structure. Our results showed that the deletion of YTHDF1 could mitigate structural lesions in both brain and gut tissues post-TBI. At the protein level, apoptosis is potentiated [32,33,34], marked by upregulated caspase3 [35,36,37,38], Cleaved caspase-3 [39] and BAX, as well as downregulated BCL2 [40,41]. Meanwhile, inflammation is enhanced [42], such as increased CD68 cells [42,43,44] and decreased Foxp3-mediated regulatory T cells (Tregs) [45,46,47]. In addition, brain oedema is marked by enhanced AQP4 expression [48] and attenuated claudin5 expression [49,50]. Our results indicated that the lack of YTHDF1 might block apoptosis, inflammation, and oedema following TBI, thereby attenuating TBI-induced BGA dysfunction in mice. GO and KEGG functional analyses showed that the DEGs in the mouse cortex post-TBI, in the YTHDF1-knockout group compared with the WT group, were primarily enriched in the regulation of neurotransmitter-related neuronal signalling pathways (neuroactive ligand-receptor interaction, glutamatergic synapse, axon guidance, and cAMP signalling pathway); response to inflammation (cell adhesion molecule signalling pathway); and the regulation of apoptotic process (calcium signalling pathway). These signalling pathways are involved in the pathophysiological processes post-TBI [51,52,53,54,55,56]. Among them, cell adhesion molecules play important roles in brain development, and maintaining synaptic structure, function, and plasticity [57,58]. An in vivo study demonstrated that TBI-induced intercellular adhesion molecule-1 (ICAM-1) regulated neuroinflammation and cell death through oxidative stress, vascular endothelial growth factor (VEGF), and matrix metalloproteinase (MMP) pathways. The deletion of ICAM-1 exhibited a better outcome in alleviating neuroinflammation and cell death, as shown by markers such as Cleaved-caspase-3, IL-1β, NF-kB, and TNF-α. In the present study, we identified that integrin a6 (ITGA6) is one of the DEGs enriched in the cell adhesion molecule signalling pathway. ITGA6 is commonly used as a glioblastoma stem-like cell (GSC) marker [59]. ITGA6 inhibition can weaken the radioresistance of mesenchymal GSCs, and it decreases proliferation and stemness in proneural GSCs [60]. In addition, ITGA6 also participates in the tumour development and progression of lower-grade gliomas [61]. However, the impact of ITGA6 on TBI has not been reported to date. Our results provide a new perspective to further investigate the pathophysiological roles of ITGA6. Furthermore, m6A can affect mRNA stability [62], while the m6A reader YTHDF1 mainly promotes mRNA translation efficiency [17]. A clinical study revealed that in bladder cancer cells, m6A is highly enriched in the ITGA6 transcripts, and increased m6A methylation of the ITGA6 mRNA 3′UTR promotes the translation of ITGA6 mRNA by binding YTHDF1 [63]. The deletion of YTHDF1 weakens the ability of RNA-binding proteins to identify m6A, thereby impeding mRNA translation and affecting downstream biological functions. Thus, YTHDF1 knockout likely downregulates the expression of ITGA6 via the cell adhesion molecule signalling pathway, thereby alleviating neuroinflammation and cell apoptosis and ultimately attenuating TBI-induced BGA dysfunction. However, the detailed mechanisms need further investigation. It is well known that the role of the gut microbiome in BGA exhibits a complex bidirectional communication system that includes neuroimmunoendocrine mediators and network pathways between the gut mucosa, ENS, and CNS [64]. The gut microbiome is a rich and complex ecosystem composed of bacteria, fungi, archaea, protists, viruses, and (sometimes) helminths [65]. While gut bacteria are essential for microbiome-BGA [66], mycobiome equilibrium is also critical for microbiome stability [67]. Mycobiome interactions may participate in mycobiome-BGA communication via immune- and nonimmune-regulated crosstalk systems, similar to those in the microbiome-BGA [68]. Increasing evidence indicates that TBI causes alteration of the gut microbiome by disrupting BGA [6]; meanwhile, host m6A modification affects the gut microbiome by inducing gut inflammatory responses [69,70]. M6A could act as a molecule to be involved in the interaction of the host and microbiome along with noncoding RNAs, chromatin remodelling, and histone modifications [15]. The knockout of YTHDF1 also significantly decreased gastric cancer cell proliferation and tumorigenesis in vivo [71]. Despite a relatively small number of gut fungi, they profoundly affect nutrition, metabolism, and immunity in the gut. Not only do gut fungi shape the functions of the gut, but they also affect the physiological functions of other crucial extraintestinal organs, such as the liver, lung, and brain [72]. For example, fungi are implicated in the inflammatory immune disorder of inflammatory bowel disease (IBD) [73,74,75], while mucosa-associated fungi (MAF) reinforced gut epithelial function and protected mice against gut injury and bacterial infection [76]. Meanwhile, fungi are able to synthesize and release neurotransmitters, which increases locomotor activity and aggressive behaviour and decreases anxiety reactions. Conversely, neuromediators may have an impact on gut fungi [65]. Therefore, we examined the alterations of fungal microbiome after TBI. In the current study, the results showed that the diversity of the mycobiome varied dramatically between YTHDF1-knockout and WT mice at 3 days following CCI. In alpha diversity, the ACE and Chao1 indices of the YTHDF1-knockout mice were markedly higher than those in WT mice, while Shannon’s and Simpson’s indices were similar. This indicated that YTHDF1 deficiency could promote the abundance of the fungal mycobiome but not the diversity of the fungal mycobiome after CCI. The beta diversity analysis showed the specific mycobiome structure detected from the group samples. The weighted and unweighted UniFrac distances indicated that the faecal mycobiomes of YTHDF1-knockout and WT mice had distinct community structures. The various characteristics of the mycobiome can be identified from family to species levels. The results of this study showed that the compositions of the mycobiome differed between the YTHDF1-knockout group and the WT group. Akkermansia exhibited a marked enrichment in YTHDF1-knockout mice. Akkermansia muciniphilia (A. muciniphila) belongs to Akkermansia, and it is regarded as a probiotic [77]. The abundance of A. muciniphila is positively associated with mucus layer thickness and it can protect gut barrier integrity in humans and animals [78,79,80,81]. Moreover, the colonization of A. muciniphila enhances the development of host innate and adaptive immune systems with anti-inflammatory effects [82]. The expressions of Foxp3 and retinoic orphan receptor gamma T (RORγt) in colonic tissue were both positively associated with A. muciniphila colonization, and A. muciniphila administration markedly promoted colonic RORγt+ Treg cell responses to ameliorate colitis [83]. A. muciniphila could promote 5-HT levels in the colons of mice via its outer membrane protein Amuc_1100 and TLR2 signalling pathway, thus improving gastrointestinal diseases and metabolic disorders [84]. Additionally, a mouse experiment showed that A. muciniphila and its extracellular vesicles could enhance the 5-HT levels in the colon and hippocampus [85]. As the key mediator of the development and function of the ENS and CNS, 5-HT may play an important role in microbiome-BGA communication. 5-HT could improve certain gut bacterial growth and influence the gut microbiota via the host immune system, which likely affects the colonization and interaction of gut bacteria [84]. A correlation study indicated a high correlation between m6A genes and the TLR2 gene [86]. A. muciniphila could influence specific m6A modifications in mono-associated mice [87]. In addition, YTHDF1 promotes the translation of m6A-modified mRNA [12,15]. YTHDF1-knockout weakens the ability of RNA-binding proteins to identify m6A, thereby impeding mRNA translation and affecting downstream biological functions. Thus, it is reasonable to deduce that the deletion of YTHDF1 may increase A. muciniphila colonization to enhance mouse anti-inflammatory effects, by promoting the expression of Foxp3 and promoting 5-HT levels, ultimately attenuating TBI-induced BGA dysfunction in mice. Our study provides new therapeutic targets for BGA dysfunction after TBI. Attempts to develop agonists or inhibitors of these new molecular targets may offer a potential strategy to attenuate TBI-induced BGA dysfunction. Further studies are necessary to explore the mechanisms responsible for the participation of these genes in BGA dysfunction following TBI. This study harboured several limitations. First, we only elucidated DEGs correlated with TBI between the YTHDF1-knockout and WT mouse cortices, and whether the DEGs in the colon are consistent with those in the cortex post-TBI needs to be explored. Second, we only illustrated several possible pathways by which YTHDF1 regulates BGA dysfunction following TBI, and there may be more pathways that require further investigation. Third, we only described the DEGs in the cortex and the alteration of faecal mycobiomes between the YTHDF1-knockout and WT mice after TBI, and the mechanisms of YTHDF1 mediating TBI-induced BGA dysfunction remain to be explored in future studies. Fourth, our study only includes male mice; the data of female animals need to be further researched. In summary, this study revealed that YTHDF1 deficiency could attenuate TBI-induced BGA dysfunction, afforded a comprehensive analysis of the DEGs related to TBI, and further confirmed the hub genes associated with TBI progression. ## 4.1. Animals and Grouping YTHDF1-knockout and C57BL/6J (WT) male mice, aged 8–12 weeks and weighing 24 ± 3 g, were provided by the Animal Center, Nanjing Agricultural University (Jiangsu, China). The YTHDF1-knockout mice were generated based on CRISPR/Cas9 [18]. sgRNA expression plasmids were generated by annealing and cloning oligos that were designed to target exon 4 of YTHDF1 into the BsaI sites of pUC57-sgRNA (Addgene 51132). m YTHDF1-E4-1 T7 gRNA up: TAGGATAGTAACTGGACAGGTA m YTHDF1-E4-1 gRNA down: AAACTACCTGTCCAGTTACTAT m YTHDF1-E4-2 T7 gRNA up: TAGGCACCATGGTCCACTGCAG m YTHDF1-E4-2 gRNA down: AAACCTGCAGTGGACCATGGTG. The in vitro transcription and microinjection of CRISPR/Cas9 was performed as follows [88]: Briefly, the Cas9 expression construct pST1374-Cas9-N-NLS-Flag-linker-D10A (Addgene 51130) was linearized with Age I and transcribed using the mMACHINE™ T7 Ultra Kit (Ambion, AM1345). Cas9 mRNA was purified by an RNeasy Mini Kit (Qiagen, 74104). pUC57-sgRNA expression vectors were linearized by Dra I and transcribed using the MEGAshortscript Kit (Ambion, AM1354). sgRNAs were purified using the MEGAclear Kit (Ambion, AM1908). A mixture of Cas9 mRNA (20 ng/μL) and two sgRNAs (5 ng/μL each) was injected into the cytoplasm and the male pronucleus of zygotes obtained by the mating of CBF1. Injected zygotes were transferred into pseudopregnant CD1 female mice. Founder mice used for experiments were backcrossed to C57BL/6J for at least five generations. m YTHDF1-E4 C9 For: CACCTGAGTTCAGATCATTAC m YTHDF1-E4 C9 Rev: GCTCCAGACTGTTCATCC. Amplicon length: 650 bp. Applicable to genotyping founders and targeted embryonic stem cell (ESC). The mice were housed in a standardized SPF animal laboratory under a 12 h light-dark cycle at a constant temperature (24 °C) and humidity ($50\%$) and allowed free access to food and water. They were randomly assigned to 2 groups after 1 week of acclimation. [ 1] C57BL/6J + sham group (WT + sham, $$n = 3$$); [2] C57BL/6J + CCI group (WT + CCI, $$n = 9$$); [3] YTHDF1-knockout + sham group (YTHDF1-KO + sham, $$n = 3$$); [4] YTHDF1-knockout + CCI group (YTHDF1-KO + CCI, $$n = 9$$). There were no significant differences among the 4 groups in terms of general health, reactivity, locomotor activity, and neurological reflexes. ## 4.2. CCI Procedure Sodium pentobarbital (65 mg/kg) was injected intraperitoneally to anaesthetize the mice before CCI injury, and surgery began when pedal reflexes were absent. A heating pad was used to maintain the core body temperature at 37 °C during surgery. The heads of mice were fixed in a stereotaxic frame, and a 4 mm diameter craniotomy was performed at 2.0 mm lateral to the midline over the right hemisphere and 2.0 mm posterior to bregma. A 3.0 mm rounded metal tip attached to the Pin-Point™ CCI device (Model PCI3000, Hatteras Instruments Inc., Cary, NC, USA) was angled vertically towards the brain surface. A severe injury was performed with 3.0 m/s speed, 2.0 mm depth, and 180 ms procedural duration. After the operation, the mice were removed from the stereotaxic holder, and the wound was lightly sutured. The sham groups underwent the same anaesthesia and surgical procedures but without CCI injury. All mice were placed in heated cages to maintain their body temperature after surgery, and the mice were not returned to their original cages until they were fully awake. ## 4.3. Sample Collection In each of the 4 groups, 3 mice were used for haematoxylin-eosin staining (H&E) at 3 days post-CCI. The animals were euthanized by intraperitoneally injecting sodium pentobarbital (65 mg/kg), then perfused transcardially with phosphate-buffered saline followed by 50 mL of $4\%$ paraformaldehyde. The brains were quickly dissected from the mouse body and fixed in $4\%$ paraformaldehyde at 4 °C for 48 h. Coronal sections, which were obtained using a vibratome (Leica VT 1000S, Wetzlar, Germany), should contain the entire hippocampus (–0 mm, –3.5 mm relative to bregma). Serial coronal sections (30 μm thick) for H&E staining ($$n = 3$$ per group) were cut by a cryostat (Leica CM 1950). At the same time, the colon tissue was excised and fixed in a $4\%$ paraformaldehyde solution, dehydrated using ethanol and xylene, and embedded in paraffin. Moreover, 5 mm thick sections were cut for H&E staining. In addition, the injured cerebral cortex (respective $$n = 3$$) and colon tissue (respective $$n = 3$$) were quickly removed, weighed, frozen in liquid nitrogen, and then stored at −80 °C for Western blotting, and the cerebral cortex on the injured side (respective $$n = 3$$) was used for RNA-Seq. Furthermore, the faecal samples (respective $$n = 3$$) were quickly collected, weighed, frozen in liquid nitrogen, and then stored at −80 °C for SMRT sequencing. ## 4.4. HE Brain and colon sections were rinsed with dH2O, stained in haematoxylin for 6 min, and then decolorized in acid alcohol for 1 s. Afterwards, the sections were rinsed with dH2O for 3 s and counterstained in eosin for 15 s before being immersed in LiCO3. Next, the sections were rinsed with dH2O and dehydrated with $95\%$ ethyl alcohol for 2–3 min and $100\%$ ethyl alcohol for 2–3 min. Then, the sections were cleared with xylene for 2–5 min, mounted with DePeX (Thermo Fisher Scientific Inc., Waltham, MA, USA) in a fume hood, and visualized using an inverted microscope at 100× magnification (Nikon, Tokyo, Japan). Digital images were captured using a SPOT microscope camera (Diagnostic Instruments, Sterling Heights, MI, USA). ## 4.5. Western Blots The total proteins from the brain and colon samples 3 days after CCI were extracted with ice-cold RIPA lysis buffer containing phosphatase inhibitors and protease inhibitors (Beyotime, Shanghai, China), and the supernatants were collected after the lysates were centrifuged at 4 °C. The supernatants were mixed with 5× dual colour protein loading buffer (Fudebio, Hangzhou, China) and boiled at 100 °C for 5 min. An equal amount of protein (30 µg/lane) was separated on an SDS-PAGE gel. The proteins were transferred onto a PVDF membrane. The membranes were developed with Femto ECL reagent (Fudebio, Hangzhou, China) after blocking and incubating with the primary antibodies and the secondary antibodies. The relative protein expression levels were analysed with GAPDH or β-tubulin as an internal control using Gel-Pro Analyser software (Media Cybernetics, Rockville, MD, USA). The primary antibody information is listed in Table 2. ## 4.6. PCR Amplification and SMRT Sequencing Total DNA was extracted from the faecal samples using the E.Z.N.A.® Soil DNA Kit (Omega Biotek, Norcross, GA, USA) according to the manufacturer’s protocols. The V1-V9 region of the bacterial 16S ribosomal RNA gene was amplified by PCR (95 °C for 2 min, followed by 27 cycles at 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 60 s, with a final extension at 72 °C for 5 min) using primers 27F 5′-AGRGTTYGATYMTGGCTCAG-3′ and 1492R 5′-RGYTACCTTGTTACGACTT-3′, where the barcode is an eight-base sequence unique to each sample. PCRs were performed in triplicate in a 20 μL mixture containing 4 μL of 5× FastPfu Buffer, 2 μL of 2.5 mM dNTPs, 0.8 μL of each primer (5 μM), 0.4 μL of FastPfu Polymerase, and 10 ng of template DNA. Amplicons were extracted from $2\%$ agarose gels and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) following the manufacturer’s instructions. SMRTbell libraries were prepared from the amplified DNA by blunt ligation according to the manufacturer’s instructions (Pacific Biosciences). Purified SMRTbell libraries from the Zymo and HMP mock communities were sequenced on dedicated PacBio Sequel II 8 M cells using Sequencing Kit 2.0 chemistry. Purified SMRTbell libraries from the pooled and barcoded samples were sequenced on a single PacBio Sequel II cell. ## 4.7. RNA-Seq Total RNA from the cerebral cortices of WT and YTHDF1-knockout mice 3 days after CCI (each $$n = 3$$) was isolated using TRIzol reagent (Invitrogen, Waltham, MA, USA) following the manufacturer’s protocol. Afterwards, the quantity and quality of RNA were assessed using a NanoDrop ND-1000 (NanoDrop). RNA integrity was assessed using denaturing agarose gel electrophoresis. Next, mRNA was extracted using NEBNextR Poly (A) mRNA Magnetic Isolation Module (New England Biolabs, Hertfordshire, UK) according to the manufacturer’s procedure. Then, RNA libraries were prepared using a KAPA Stranded RNA-Seq Library Prep Kit (Illumina) following the manufacturer’s protocol. Finally, libraries were sequenced using Illumina HiSeq 4000 platforms. ## 4.8. Statistical Analyses All data are presented as the mean ± standard error (SE). Student’s t tests were used to test the difference between two groups. A one-way ANOVA with Tukey’s multiple comparison test was used to evaluate the differences among multiple groups. The significance level was set to $p \leq 0.05.$ All analyses were performed using SPSS version 25.0 (IBM, New York, NY, USA). GO and KEGG analyses of peaks and differentially expressed peaks were performed using R based on the hypergeometric distribution. Bar charts show the most abundant phyla and species (>$1\%$). Relative abundances were detected between sequencing technologies using paired Student’s t tests. The significant differences in taxa between the YTHDF1-knockout and WT groups at 3 days following CCI was compared with LEfSe, which uses a nonparametric factorial Kruskal–Wallis with a subsequent unpaired Wilcoxon test. An LDA higher than 4 and a p value lower than 0.05 were considered significant. Alpha diversity (ACE, Chao, Simpson, and Shannon indices) was compared between samples from the YTHDF1-knockout and WT mice by a t test with two dependent means and considering a p value < 0.05 as significant. Principle coordinate analysis (PCoA) plots and a clustering dendrogram were generated to visualize the beta diversity of the faecal mycobiome of the YTHDF1-knockout and WT groups. ## References 1. Jaganathan K.S., Sullivan K.A.. Traumatic Brain Injury Rehabilitation: An Exercise Immunology Perspective. Exerc. Immunol. Rev. (2022) 28 90-97. PMID: 35452396 2. Dewan M.C., Rattani A., Gupta S., Baticulon R.E., Hung Y.-C., Punchak M., Agrawal A., Adeleye A.O., Shrime M.G., Rubiano A.M.. Estimating the global incidence of traumatic brain injury. J. Neurosurg. (2018) 130 1080-1097. DOI: 10.3171/2017.10.JNS17352 3. Mercado N.M., Zhang G., Ying Z., Gómez-Pinilla F.. Traumatic brain injury alters the gut-derived serotonergic system and associated peripheral organs. Biochim. Biophys Acta Mol. Basis Dis. (2022) 1868 166491. DOI: 10.1016/j.bbadis.2022.166491 4. Hanscom M., Loane D.J., Shea-Donohue T.. Brain-gut axis dysfunction in the pathogenesis of traumatic brain injury. J. Clin. Investig. (2021) 131 e143777. DOI: 10.1172/JCI143777 5. Urban R.J., Pyles R.B., Stewart C.J., Ajami N., Randolph K.M., Durham W.J., Danesi C.P., Dillon E.L., Summons J.R., Singh C.K.. Altered Fecal Microbiome Years after Traumatic Brain Injury. J. Neurotrauma (2020) 37 1037-1051. DOI: 10.1089/neu.2019.6688 6. Kharrazian D.. Traumatic Brain Injury and the Effect on the Brain-Gut Axis. Altern. Ther. Health Med. (2015) 21 28-32 7. Wang J., Sha Y., Sun T.. m. Front. Oncol. (2021) 11 611660. DOI: 10.3389/fonc.2021.611660 8. Roignant J.-Y., Soller M.. m. Trends Genet. (2017) 33 380-390. DOI: 10.1016/j.tig.2017.04.003 9. Coker H., Wei G., Brockdorff N.. m6A modification of non-coding RNA and the control of mammalian gene expression. Biochim. Biophys Acta Gene Regul. Mech. (2019) 1862 310-318. DOI: 10.1016/j.bbagrm.2018.12.002 10. Sun T., Wu R., Ming L.. The role of m6A RNA methylation in cancer. Biomed. Pharmacother. (2019) 112 108613. DOI: 10.1016/j.biopha.2019.108613 11. Zhang Y., Geng X., Li Q., Xu J., Tan Y., Xiao M., Song J., Liu F., Fang C., Wang H.. m6A modification in RNA: Biogenesis, functions and roles in gliomas. J. Exp. Clin. Cancer Res. (2020) 39 192. DOI: 10.1186/s13046-020-01706-8 12. Huang P., Liu M., Zhang J., Zhong X., Zhong C.. The Potential Role of m. Antioxidants (2022) 11. DOI: 10.3390/antiox11081521 13. Ma S., Chen C., Ji X., Liu J., Zhou Q., Wang G., Yuan W., Kan Q., Sun Z.. The interplay between m6A RNA methylation and noncoding RNA in cancer. J. Hematol. Oncol. (2019) 12 121. DOI: 10.1186/s13045-019-0805-7 14. Zhang N., Ding C., Zuo Y., Peng Y., Zuo L.. N6-methyladenosine and Neurological Diseases. Mol. Neurobiol. (2022) 59 1925-1937. DOI: 10.1007/s12035-022-02739-0 15. Zhuo R., Xu M., Wang X., Zhou B., Wu X., Leone V., Chang E.B., Zhong X.. The regulatory role of N. Wiley Interdiscip Rev. RNA (2022) 13 e1725. DOI: 10.1002/wrna.1725 16. Wu J., Zhao Y., Wang X., Kong L., Johnston L.J., Lu L., Ma X.. Dietary nutrients shape gut microbes and intestinal mucosa via epigenetic modifications. Crit. Rev. Food Sci. Nutr. (2022) 62 783-797. DOI: 10.1080/10408398.2020.1828813 17. Wang X., Zhao B.S., Roundtree I.A., Lu Z., Han D., Ma H., Weng X., Chen K., Shi H., He C.. N. Cell (2015) 161 1388-1399. DOI: 10.1016/j.cell.2015.05.014 18. Shi H., Zhang X., Weng Y.-L., Lu Z., Liu Y., Lu Z., Li J., Hao P., Zhang Y., Zhang F.. m. Nature (2018) 563 249-253. DOI: 10.1038/s41586-018-0666-1 19. Yarmishyn A.A., Yang Y.-P., Lu K.-H., Chen Y.-C., Chien Y., Chou S.-J., Tsai P.-H., Ma H.-I., Chien C.-S., Chen M.-T.. Musashi-1 promotes cancer stem cell properties of glioblastoma cells via upregulation of YTHDF1. Cancer Cell Int. (2020) 20 597. DOI: 10.1186/s12935-020-01696-9 20. Zong X., Xiao X., Shen B., Jiang Q., Wang H., Lu Z., Wang F., Jin M., Min J., Wang F.. The. Nucleic Acids Res. (2021) 49 5537-5552. DOI: 10.1093/nar/gkab343 21. Olazagoitia-Garmendia A., Zhang L., Mera P., Godbout J.K., Sebastian-DelaCruz M., Garcia-Santisteban I., Mendoza L.M., Huerta A., Irastorza I., Bhagat G.. Gluten-induced RNA methylation changes regulate intestinal inflammation via allele-specific XPO1 translation in epithelial cells. Gut (2022) 71 68-76. DOI: 10.1136/gutjnl-2020-322566 22. Xu D., Shao J., Song H., Wang J.. The YTH Domain Family of N6-Methyladenosine “Readers” in the Diagnosis and Prognosis of Colonic Adenocarcinoma. Biomed. Res. Int. (2020) 2020 9502560. DOI: 10.1155/2020/9502560 23. Quan W., Li J., Liu L., Zhang Q., Qin Y., Pei X., Chen J.. Influence of N6-Methyladenosine Modification Gene HNRNPC on Cell Phenotype in Parkinson’s Disease. Parkinsons Dis. (2021) 2021 9919129. DOI: 10.1155/2021/9919129 24. Xu C., Huang H., Zhang M., Zhang P., Li Z., Liu X., Fang M.. Methyltransferase-Like 3 Rescues the Amyloid-beta protein-Induced Reduction of Activity-Regulated Cytoskeleton Associated Protein Expression via YTHDF1-Dependent N6-Methyladenosine Modification. Front. Aging Neurosci. (2022) 14 890134. DOI: 10.3389/fnagi.2022.890134 25. Feng Z., Zhou F., Tan M., Wang T., Chen Y., Xu W., Li B., Wang X., Deng X., He M.-L.. Targeting m6A modification inhibits herpes virus 1 infection. Genes Dis. (2022) 9 1114-1128. DOI: 10.1016/j.gendis.2021.02.004 26. Ye F., Wang T., Wu X., Liang J., Li J., Sheng W.. J. Transl. Med. (2021) 19 316. DOI: 10.1186/s12967-021-02981-5 27. Li B., Wang F., Wang N., Hou K., Du J.. Identification of Implications of Angiogenesis and m6A Modification on Immunosuppression and Therapeutic Sensitivity in Low-Grade Glioma by Network Computational Analysis of Subtypes and Signatures. Front. Immunol. (2022) 13 871564. DOI: 10.3389/fimmu.2022.871564 28. Miao Y.-Q., Chen W., Zhou J., Shen Q., Sun Y., Li T., Wang S.-C.. N(6)-adenosine-methyltransferase-14 promotes glioma tumorigenesis by repressing argininosuccinate synthase 1 expression in an m6A-dependent manner. Bioengineered (2022) 13 1858-1871. DOI: 10.1080/21655979.2021.2018386 29. Zhang G., Zheng P., Lv Y., Shi Z., Shi F.. m6A Regulatory Gene-Mediated Methylation Modification in Glioma Survival Prediction. Front. Genet. (2022) 13 873764. DOI: 10.3389/fgene.2022.873764 30. Wu Y., Yang Z., Su S., Xu X., Li Y., Li X., Gao Y., Sun D., Wan S., Pen M.. Differential epitranscriptome and proteome modulation in the brain of neonatal mice exposed to isoflurane or sevoflurane. Cell Biol. Toxicol. (2022). DOI: 10.1007/s10565-022-09701-9 31. Wang X.-L., Wei X., Yuan J.-J., Mao Y.-Y., Wang Z.-Y., Xing N., Gu H.-W., Lin C.-H., Wang W.-T., Zhang W.. Downregulation of Fat Mass and Obesity-Related Protein in the Anterior Cingulate Cortex Participates in Anxiety- and Depression-Like Behaviors Induced by Neuropathic Pain. Front. Cell Neurosci. (2022) 16 884296. DOI: 10.3389/fncel.2022.884296 32. Kunz A., Dirnagl U., Mergenthaler P.. Acute pathophysiological processes after ischaemic and traumatic brain injury. Best Pract. Res. Clin. Anaesthesiol. (2010) 24 495-509. DOI: 10.1016/j.bpa.2010.10.001 33. Monier B., Suzanne M.. Orchestration of Force Generation and Nuclear Collapse in Apoptotic Cells. Int. J. Mol. Sci. (2021) 22. DOI: 10.3390/ijms221910257 34. Unnisa A., Greig N.H., Kamal M.A.. Inhibition of Caspase 3 and Caspase 9 Mediated Apoptosis: A Multimodal Therapeutic Target in Traumatic Brain Injury. Curr. Neuropharmacol. (2022). DOI: 10.2174/1570159X20666220327222921 35. Clark R.S.B., Kochanek P., Watkins S., Chen M., Dixon C.E., Seidberg N.A., Melick J., Loeffert J.E., Nathaniel P.D., Jin K.L.. Caspase-3 mediated neuronal death after traumatic brain injury in rats. J. Neurochem. (2000) 74 740-753. DOI: 10.1046/j.1471-4159.2000.740740.x 36. Abrahamson E.E., Ikonomovic M.D., Ciallella J.R., Hope C.E., Paljug W.R., Isanski B.A., Flood D.G., Clark R.S., DeKosky S.T.. Caspase inhibition therapy abolishes brain trauma-induced increases in Abeta peptide: Implications for clinical outcome. Exp. Neurol. (2006) 197 437-450. DOI: 10.1016/j.expneurol.2005.10.011 37. Shi Y., Cui W., Wang Q., Zhou J., Wu X., Wang J., Zhang S., Hu Q., Han L., Du Y.. MicroRNA-124/Death-Associated Protein Kinase 1 Signaling Regulates Neuronal Apoptosis in Traumatic Brain Injury via Phosphorylating NR2B. Front. Cell Neurosci. (2022) 16 892197. DOI: 10.3389/fncel.2022.892197 38. Lorente L., Martín M.M., Argueso M., Ramos L., Solé-Violán J., Riaño-Ruiz M., Jiménez A., Borreguero-León J.M.. Serum caspase-3 levels and mortality are associated in patients with severe traumatic brain injury. BMC Neurol. (2015) 15. DOI: 10.1186/s12883-015-0485-z 39. He J., Mao J., Hou L., Jin S., Wang X., Ding Z., Jin Z., Guo H., Dai R.. Minocycline attenuates neuronal apoptosis and improves motor function after traumatic brain injury in rats. Exp. Anim. (2021) 70 563-569. DOI: 10.1538/expanim.21-0028 40. Raghupathi R., Strauss K.I., Zhang C., Krajewski S., Reed J.C., McIntosh T.K.. Temporal alterations in cellular Bax:Bcl-2 ratio following traumatic brain injury in the rat. J. Neurotrauma (2003) 20 421-435. DOI: 10.1089/089771503765355504 41. Kim D.-H., Ko I.-G., Kim B.-K., Kim T.-W., Kim S.-E., Shin M.-S., Kim C.-J., Kim H., Kim K.-M., Baek S.-S.. Treadmill exercise inhibits traumatic brain injury-induced hippocampal apoptosis. Physiol. Behav. (2010) 101 660-665. DOI: 10.1016/j.physbeh.2010.09.021 42. Harting M.T., Jimenez F., Adams S.D., Mercer D.W., Cox C.S.. Acute, regional inflammatory response after traumatic brain injury: Implications for cellular therapy. Surgery (2008) 144 803-813. DOI: 10.1016/j.surg.2008.05.017 43. Neri M., Frati A., Turillazzi E., Cantatore S., Cipolloni L., Di Paolo M., Frati P., La Russa R., Maiese A., Scopetti M.. Immunohistochemical Evaluation of Aquaporin-4 and its Correlation with CD68, IBA-1, HIF-1alpha, GFAP, and CD15 Expressions in Fatal Traumatic Brain Injury. Int. J. Mol. Sci. (2018) 19. DOI: 10.3390/ijms19113544 44. Härtig W., Michalski D., Seeger G., Voigt C., Donat C.K., Dulin J., Kacza J., Meixensberger J., Arendt T., Schuhmann M.U.. Impact of 5-lipoxygenase inhibitors on the spatiotemporal distribution of inflammatory cells and neuronal COX-2 expression following experimental traumatic brain injury in rats. Brain Res. (2013) 1498 69-84. DOI: 10.1016/j.brainres.2012.12.022 45. Xie D., Miao W., Xu F., Yuan C., Li S., Wang C., Junagade A., Hu X.. IL-33/ST2 Axis Protects Against Traumatic Brain Injury Through Enhancing the Function of Regulatory T Cells. Front. Immunol. (2022) 13 860772. DOI: 10.3389/fimmu.2022.860772 46. Krämer T.J., Hack N., Brühl T.J., Menzel L., Hummel R., Griemert E.-V., Klein M., Thal S.C., Bopp T., Schäfer M.K.E.. Depletion of regulatory T cells increases T cell brain infiltration, reactive astrogliosis, and interferon-gamma gene expression in acute experimental traumatic brain injury. J. Neuroinflammat. (2019) 16 163. DOI: 10.1186/s12974-019-1550-0 47. Yu Y., Cao F., Ran Q., Sun X.. Regulatory T cells exhibit neuroprotective effect in a mouse model of traumatic brain injury. Mol. Med. Rep. (2016) 14 5556-5566. DOI: 10.3892/mmr.2016.5954 48. Dadgostar E., Rahimi S., Nikmanzar S., Nazemi S., Taheri M.N., Alibolandi Z., Aschner M., Mirzaei H., Tamtaji O.R.. Aquaporin 4 in Traumatic Brain Injury: From Molecular Pathways to Therapeutic Target. Neurochem. Res. (2022) 47 860-871. DOI: 10.1007/s11064-021-03512-w 49. Jiang Q., Wei D., He X., Gan C., Long X., Zhang H.. Phillyrin Prevents Neuroinflammation-Induced Blood-Brain Barrier Damage Following Traumatic Brain Injury via Altering Microglial Polarization. Front. Pharmacol. (2021) 12 719823. DOI: 10.3389/fphar.2021.719823 50. Main B.S., Villapol S., Sloley S.S., Barton D.J., Parsadanian M., Agbaegbu C., Stefos K., McCann M.S., Washington P.M., Rodriguez O.C.. Apolipoprotein E4 impairs spontaneous blood brain barrier repair following traumatic brain injury. Mol. Neurodegener. (2018) 13 17. DOI: 10.1186/s13024-018-0249-5 51. Yang Z., Li X., Luo W., Wu Y., Tang T., Wang Y.. The Involvement of Long Non-coding RNA and Messenger RNA Based Molecular Networks and Pathways in the Subacute Phase of Traumatic Brain Injury in Adult Mice. Front. Neuroinform. (2022) 16 794342. DOI: 10.3389/fninf.2022.794342 52. Elliott J.E., De Luche S.E., Churchill M.J., Moore C., Cohen A.S., Meshul C.K., Lim M.M.. Dietary therapy restores glutamatergic input to orexin/hypocretin neurons after traumatic brain injury in mice. Sleep (2018) 41 zsx212. DOI: 10.1093/sleep/zsx212 53. Chandran R., Sharma A., Bhomia M., Balakathiresan N.S., Knollmann-Ritschel B.E., Maheshwari R.K.. Differential expression of microRNAs in the brains of mice subjected to increasing grade of mild traumatic brain injury. Brain Inj. (2017) 31 106-119. DOI: 10.1080/02699052.2016.1213420 54. Atkins C.M., Oliva A.A., Alonso O.F., Pearse D.D., Bramlett H.M., Dietrich W.D.. Modulation of the cAMP signaling pathway after traumatic brain injury. Exp. Neurol. (2007) 208 145-158. DOI: 10.1016/j.expneurol.2007.08.011 55. Bhowmick S., Malat A., Caruso D., Ponery N., D’Mello V., Finn C., Abdul-Muneer P.M.. Intercellular Adhesion Molecule-1-Induced Posttraumatic Brain Injury Neuropathology in the Prefrontal Cortex and Hippocampus Leads to Sensorimotor Function Deficits and Psychological Stress. eNeuro (2021) 8. DOI: 10.1523/ENEURO.0242-21.2021 56. Fröhlich A., Heuvel F.O., Rehman R., Krishnamurthy S.S., Li S., Li Z., Bayer D., Conquest A., Hagenston A.M., Ludolph A.. Neuronal nuclear calcium signaling suppression of microglial reactivity is mediated by osteoprotegerin after traumatic brain injury. J. Neuroinflammat. (2022) 19 279. DOI: 10.1186/s12974-022-02634-4 57. Yamagata M., Sanes J., Weiner J.. Synaptic adhesion molecules. Curr. Opin. Cell Biol. (2003) 15 621-632. DOI: 10.1016/S0955-0674(03)00107-8 58. Duncan B.W., Murphy K.E., Maness P.F.. Molecular Mechanisms of L1 and NCAM Adhesion Molecules in Synaptic Pruning, Plasticity, and Stabilization. Front. Cell Dev. Biol. (2021) 9 625340. DOI: 10.3389/fcell.2021.625340 59. Erhart F., Blauensteiner B., Zirkovits G., Printz D., Soukup K., Klingenbrunner S., Fischhuber K., Reitermaier R., Halfmann A., Lötsch D.. Gliomasphere marker combinatorics: Multidimensional flow cytometry detects CD44+/CD133+/ITGA6+/CD36+ signature. J. Cell Mol. Med. (2019) 23 281-292. DOI: 10.1111/jcmm.13927 60. Stanzani E., Pedrosa L., Bourmeau G., Anezo O., Noguera-Castells A., Esteve-Codina A., Passoni L., Matteoli M., de la Iglesia N., Seano G.. Dual Role of Integrin Alpha-6 in Glioblastoma: Supporting Stemness in Proneural Stem-Like Cells While Inducing Radioresistance in Mesenchymal Stem-Like Cells. Cancers (2021) 13. DOI: 10.3390/cancers13123055 61. Lin T., Cheng H., Liu D., Wen L., Kang J., Xu L., Shan C., Chen Z., Li H., Lai M.. A Novel Six Autophagy-Related Genes Signature Associated with Outcomes and Immune Microenvironment in Lower-Grade Glioma. Front. Genet. (2021) 12 698284. DOI: 10.3389/fgene.2021.698284 62. An Y., Duan H.. The role of m6A RNA methylation in cancer metabolism. Mol. Cancer (2022) 21 14. DOI: 10.1186/s12943-022-01500-4 63. Jin H., Ying X., Que B., Wang X., Chao Y., Zhang H., Yuan Z., Qi D., Lin S., Min W.. N. EBioMedicine (2019) 47 195-207. DOI: 10.1016/j.ebiom.2019.07.068 64. Foster J.A., Baker G.B., Dursun S.M.. The Relationship Between the Gut Microbiome-Immune System-Brain Axis and Major Depressive Disorder. Front. Neurol. (2021) 12 721126. DOI: 10.3389/fneur.2021.721126 65. Enaud R., Vandenborght L.-E., Coron N., Bazin T., Prevel R., Schaeverbeke T., Berger P., Fayon M., Lamireau T., Delhaes L.. The Mycobiome: A Neglected Component in the Microbiota-Gut-Brain Axis. Microorganisms (2018) 6. DOI: 10.3390/microorganisms6010022 66. Gabanyi I., Lepousez G., Wheeler R., Vieites-Prado A., Nissant A., Wagner S., Moigneu C., Dulauroy S., Hicham S., Polomack B.. Bacterial sensing via neuronal Nod2 regulates appetite and body temperature. Science (2022) 376 eabj3986. DOI: 10.1126/science.abj3986 67. Qiu X., Zhang F., Yang X., Wu N., Jiang W., Li X., Li X., Liu Y.. Changes in the composition of intestinal fungi and their role in mice with dextran sulfate sodium-induced colitis. Sci. Rep. (2015) 5 10416. DOI: 10.1038/srep10416 68. Carabotti M., Scirocco A., Maselli M.A., Severi C.. The gut-brain axis: Interactions between enteric microbiota, central and enteric nervous systems. Ann. Gastroenterol. (2015) 28 203-209. PMID: 25830558 69. Lu M., Zhang Z., Xue M., Zhao B.S., Harder O., Li A., Liang X., Gao T.Z., Xu Y., Zhou J.. Nat. Microbiol. (2020) 5 584-598. DOI: 10.1038/s41564-019-0653-9 70. Zong X., Zhao J., Wang H., Lu Z., Wang F., Du H., Wang Y.. Mettl3 Deficiency Sustains Long-Chain Fatty Acid Absorption through Suppressing Traf6-Dependent Inflammation Response. J. Immunol. (2019) 202 567-578. DOI: 10.4049/jimmunol.1801151 71. Pi J., Wang W., Ji M., Wang X., Wei X., Jin J., Liu T., Qiang J., Qi Z., Li F.. YTHDF1 Promotes Gastric Carcinogenesis by Controlling Translation of FZD7. Cancer Res. (2021) 81 2651-2665. DOI: 10.1158/0008-5472.CAN-20-0066 72. Li X.V., Leonardi I., Iliev I.D.. Gut Mycobiota in Immunity and Inflammatory Disease. Immunity (2019) 50 1365-1379. DOI: 10.1016/j.immuni.2019.05.023 73. Sokol H., Leducq V., Aschard H., Pham H.P., Jegou S., Landman C., Cohen D., Liguori G., Bourrier A., Nion-Larmurier I.. Fungal microbiota dysbiosis in IBD. Gut (2017) 66 1039-1048. DOI: 10.1136/gutjnl-2015-310746 74. Ott S.J., Kühbacher T., Musfeldt M., Rosenstiel P., Hellmig S., Rehman A., Drews O., Weichert W., Timmis K.N., Schreiber S.. Fungi and inflammatory bowel diseases: Alterations of composition and diversity. Scand. J. Gastroenterol. (2008) 43 831-841. DOI: 10.1080/00365520801935434 75. Wheeler M.L., Limon J.J., Bar A.S., Leal C.A., Gargus M., Tang J., Brown J., Funari V.A., Wang H.L., Crother T.R.. Immunological Consequences of Intestinal Fungal Dysbiosis. Cell Host Microbe. (2016) 19 865-873. DOI: 10.1016/j.chom.2016.05.003 76. Leonardi I., Gao I.H., Lin W.-Y., Allen M., Li X.V., Fiers W.D., De Celie M.B., Putzel G.G., Yantiss R.K., Johncilla M.. Mucosal fungi promote gut barrier function and social behavior via Type 17 immunity. Cell (2022) 185 831-846.e14. DOI: 10.1016/j.cell.2022.01.017 77. Zhai Q., Feng S., Arjan N., Chen W.. A next generation probiotic, Akkermansia muciniphila. Crit. Rev. Food Sci. Nutr. (2019) 59 3227-3236. DOI: 10.1080/10408398.2018.1517725 78. Collado M.C., Derrien M., Isolauri E., de Vos W.M., Salminen S.. Intestinal integrity and Akkermansia muciniphila, a mucin-degrading member of the intestinal microbiota present in infants, adults, and the elderly. Appl. Environ. Microbiol. (2007) 73 7767-7770. DOI: 10.1128/AEM.01477-07 79. Zhong Y., Teixeira C., Marungruang N., Sae-Lim W., Tareke E., Andersson R., Fåk F., Nyman M.. Barley malt increases hindgut and portal butyric acid, modulates gene expression of gut tight junction proteins and Toll-like receptors in rats fed high-fat diets, but high advanced glycation end-products partially attenuate the effects. Food Funct. (2015) 6 3165-3176. DOI: 10.1039/C5FO00150A 80. Reunanen J., Kainulainen V., Huuskonen L., Ottman N., Belzer C., Huhtinen H., de Vos W.M., Satokari R.. Akkermansia muciniphila Adheres to Enterocytes and Strengthens the Integrity of the Epithelial Cell Layer. Appl. Environ. Microbiol. (2015) 81 3655-3662. DOI: 10.1128/AEM.04050-14 81. Liu M.-J., Yang J.-Y., Yan Z.-H., Hu S., Li J.-Q., Xu Z.-X., Jian Y.-P.. Recent findings in Akkermansia muciniphila-regulated metabolism and its role in intestinal diseases. Clin. Nutr. (2022) 41 2333-2344. DOI: 10.1016/j.clnu.2022.08.029 82. Derosa L., Routy B., Thomas A.M., Iebba V., Zalcman G., Friard S., Mazieres J., Audigier-Valette C., Moro-Sibilot D., Goldwasser F.. Intestinal Akkermansia muciniphila predicts clinical response to PD-1 blockade in patients with advanced non-small-cell lung cancer. Nat. Med. (2022) 28 315-324. DOI: 10.1038/s41591-021-01655-5 83. Liu Y., Yang M., Tang L., Wang F., Huang S., Liu S., Lei Y., Wang S., Xie Z., Wang W.. TLR4 regulates RORgammat(+) regulatory T-cell responses and susceptibility to colon inflammation through interaction with Akkermansia muciniphila. Microbiome (2022) 10 98. DOI: 10.1186/s40168-022-01296-x 84. Wang J., Xu W., Wang R., Cheng R., Tang Z., Zhang M.. The outer membrane protein Amuc_1100 of Akkermansia muciniphila promotes intestinal 5-HT biosynthesis and extracellular availability through TLR2 signalling. Food Funct. (2021) 12 3597-3610. DOI: 10.1039/D1FO00115A 85. Yaghoubfar R., Behrouzi A., Ashrafian F., Shahryari A., Moradi H.R., Choopani S., Hadifar S., Vaziri F., Nojoumi S.A., Fateh A.. Modulation of serotonin signaling/metabolism by Akkermansia muciniphila and its extracellular vesicles through the gut-brain axis in mice. Sci. Rep. (2020) 10 22119. DOI: 10.1038/s41598-020-79171-8 86. Mo B.-W., Li X.-M., Li S.-M., Xiao B., Yang J., Li H.-M.. m6A echoes with DNA methylation: Coordinated DNA methylation and gene expression data analysis identified critical m6A genes associated with asthma. Gene (2022) 828 146457. DOI: 10.1016/j.gene.2022.146457 87. Jabs S., Biton A., Bécavin C., Nahori M.-A., Ghozlane A., Pagliuso A., Spanò G., Guérineau V., Touboul D., Gianetto Q.G.. Impact of the gut microbiota on the m. Nat. Commun. (2020) 11 1344. DOI: 10.1038/s41467-020-15126-x 88. Shen B., Zhang W., Zhang J., Zhou J., Wang J., Chen L., Wang L., Hodgkins A., Iyer V., Huang X.. Efficient genome modification by CRISPR-Cas9 nickase with minimal off-target effects. Nat. Methods (2014) 11 399-402. DOI: 10.1038/nmeth.2857
--- title: 'In the Beginning Was the Bud: Phytochemicals from Olive (Olea europaea L.) Vegetative Buds and Their Biological Properties' authors: - Marijana Popović - Franko Burčul - Maja Veršić Bratinčević - Nikolina Režić Mužinić - Danijela Skroza - Roberta Frleta Matas - Marija Nazlić - Tonka Ninčević Runjić - Maja Jukić Špika - Ana Bego - Valerija Dunkić - Elda Vitanović journal: Metabolites year: 2023 pmcid: PMC9966879 doi: 10.3390/metabo13020237 license: CC BY 4.0 --- # In the Beginning Was the Bud: Phytochemicals from Olive (Olea europaea L.) Vegetative Buds and Their Biological Properties ## Abstract Even though Olea europaea L. is one of the most important and well-studied crops in the world, embryonic parts of the plants remain largely understudied. In this study, comprehensive phytochemical profiling of olive vegetative buds of two Croatian cultivars, Lastovka and Oblica, was performed with an analysis of essential oils and methanol extracts as well as biological activities (antioxidant, antimicrobial, and cytotoxic activities). A total of 113 different volatiles were identified in essential oils with hydrocarbons accounting for up to $60.30\%$ and (Z)-3-heptadecene being the most abundant compound. Oleacein, oleuropein, and 3-hydroxytyrosol had the highest concentrations of all phenolics in the bud extracts. Other major compounds belong to the chemical classes of sugars, fatty acids, and triterpenoid acids. Antioxidant, antimicrobial, and cytotoxic activities were determined for both cultivars. Apart from antioxidant activity, essential oils had a weak overall biological effect. The extract from cultivar Lastovka showed much better antioxidant activity than both isolates with both methods (with an oxygen radical absorbance capacity value of 1835.42 μM TE/g and DPPH IC50 of 0.274 mg/mL), as well as antimicrobial activity with the best results against Listeria monocytogenes. The human breast adenocarcinoma MDA-MB-231 cell line showed the best response for cultivar Lastovka bud extract (IC50 = 150 μg/mL) among three human cancer cell lines tested. These results demonstrate great chemical and biological potential that is hidden in olive buds and the need to increase research in the area of embryonic parts of plants. ## 1. Introduction The olive tree, Olea europaea L., is one of the most important crops in the Mediterranean since ancient times [1]. Different parts of the olive tree, including predominantly leaves but also fruits, oil, seeds, and bark, have been used to treat various conditions for centuries [2]. Traditionally, it was used in different forms to heal inflammation, diabetes, the gut, hypertension, asthma, diarrhea, and several other conditions [3]. Nowadays, a great deal of research is performed, especially for the fruits, leaves, and olive oil, resulting in in vitro and in vivo evidence of antioxidant, anti-inflammatory, immunomodulatory, antimicrobial, antiviral, antihypertensive, anticancer, antihyperglycemic, gastroprotective, and several other biological activities [4,5,6]. Olives, as well as olive oil and extra virgin olive oil (EVOO), which are one of the most important olive products, are rich in fatty acids, triacylglycerols, tocopherols, sterols, and phenolic compounds [4,5,7,8]. Olive leaves are also rich in phenolic compounds as well as in polyalcohols and triterpenoids [9] jointly contributing to the beneficial effect that they exert on human health. Given the number of beneficial components in different parts of the olive tree, it is no surprise that the concept of the Mediterranean diet, based on the regular consumption of EVOO and other olive derivatives (amongst other foods), is universally recognized by medical professionals, given that it provides extended health benefits and a protective dietary pattern for disease prevention and health maintenance [10,11]. Olive leaf extract is widely used in phytotherapy for the treatment of various conditions and is generally safe even at high doses [12]. Leaf extracts have several bioactive compounds (predominantly oleuropein (OLE) and hydroxytyrosol) that show positive effects on the parameters related to diabetes [13,14], lipid regulation, hypertension, and protection of the cardiovascular system [15,16]. Recent studies show the beneficial effect of olive leaf extract on healing herpes simplex virus labialis [17] as well as on improving the clinical status of COVID-19 patients [18]. The fruit extract has also recently been explored in experimental animals, with positive effects on hepatic lipid accumulation, chronic fatigue syndrome, and antioxidant capacity [19,20,21]. There was a popular concept in France in the middle of the 20th century named gemmotherapy, with the belief that using extracts derived from meristematic tissues would be more beneficial for human therapy than adult plant parts, since they contain specific bioactive compounds that are later subjected to metabolic transformations [22]. Although some studies on the health benefits of bud extracts have been conducted [23,24,25], the entire area is largely unexplored. Olive trees and olive derivatives are the subjects of the research area of many laboratories and scientists around the world; however, research on phytochemicals from olive buds as well as their biological properties is rare. Only a few studies have reported the chemical composition of olive buds but were mainly oriented on their phenolic compounds and scarce biological activities [26,27]. The aim of this study was to obtain phytochemical profiles of olive vegetative bud essential oil (EO) and methanol extract from two Croatian olive cultivars (cvs.), Lastovka and Oblica, as well as to investigate some of their biological properties (antioxidant, antimicrobial, and cytotoxic activities). To the best of our knowledge, this is the first report of the chemical profile of essential oils from olive vegetative buds and their biological activities. ## 2.1. Chemicals Standards of volatile and phenolic compounds were commercially obtained as follows: (Z)-3-hexen-1-ol (Toronto research chemicals Inc., Toronto, CA, USA); linalool, nonanal, phenylethyl alcohol, decanal, eugenol, 1,3,5-trimethoxybenzene, 3-hydroxytyrosol, caffeic acid, vanillin, trans-p-coumaric acid, trans-o-coumaric acid, apigenin-7-glucoside, oleuropein, pinoresinol, quercetin, luteolin, apigenin, diosmetin tyrosol, rutin, luteolin-7-glucoside, oleacein, and ligstroside from Sigma-Aldrich (St. Louis, MO, USA); verbascoside (HWI group, Rülzheim, Germany); and oleuroside (Phytolab, Vestenbergsgreuth, Germany). Alkane standard solutions C8-C20 and C10, C20-C40, and derivatization reagent N,O-Bis(trimethylsilyl)trifluoroacet-amide (BSTFA) were purchased from Sigma-Aldrich (St. Louis, MO, USA); pentane, diethyl-ether, acetonitrile, and methanol were from VWR (Radnor, PA, USA); and anhydrous sodium sulfate was from Kemika (Zagreb, Croatia). ## 2.2. Plant Material Olive vegetative buds were sampled in the experimental olive orchard in Kaštel Stari (43°55′ N; 16°35′ E) belonging to the Institute for Adriatic Crops and Karst Reclamation. The buds were collected from two Croatian olive cultivars, Lastovka and Oblica, in April 2020. Part of the samples was stored at −80 °C until further analysis, while the other part was left to dry at room temperature. ## 2.3.1. Essential Oil Distillation Fresh buds were air-dried at room temperature for 15 days. Afterward, three replicate samples of dried buds (100 g) were simultaneously hydrodistilled in a Clevenger apparatus for 150 min. Pentane and diethyl-ether (1:3) were used as a trap for the essential oil [28]. After the distillation, essential oil samples were dried over anhydrous sodium sulfate, the solvent was evaporated under a stream of nitrogen, and the samples were stored in dark glass vials at 4 °C until analysis. ## 2.3.2. Methanolic Extraction The plant material was freeze-dried (FreeZone 2.5, Labconco, Kansas City, MO, USA) and a sample of the dry buds was ground to a coarse powder using a stainless-steel mill (A 11 Analytical mill, IKA, Staufen, Germany). The modified procedure described by Marinova et al. [ 29] was used to extract phenols. Briefly, 0.25 g of powdered tissue was extracted with 10 mL of methanol/water (80:20, by volume) for 20 min with an ultrasonic bath (Sonorex Digitec DT 100H, Bandelin, Berlin, Germany). An aliquot was centrifuged for 5 min at 14,000 RPM/21,255 RCF (Beckman Instruments J2-21, Palo Alto, CA, USA). ## 2.4. Methanolic Extract Derivatization Prepared methanolic extracts (1 mL) were evaporated in a centrifugal evaporator (RC10-22, Jouan, Herblain, France) at room temperature until completely dry. A derivatizing agent (50 μL of BSTFA) was added to the dried extracts for 20 min at 20 °C prior to the analysis [30]. Commercial phenolic standards (1 mg) were also derivatized by the addition of 50 μL of BSTFA derivatizing agent for 20 min at 20 °C. ## 2.5.1. GC-MS Conditions for Essential Oil Analysis Essential oils were diluted in hexane (v/v, 1:1000) and analyzed by gas chromatography-mass spectrometry (GC-MS) with gas chromatograph model 8890 GC (Agilent Inc., Santa Clara, CA, USA), equipped with automatic liquid injector model 7693A and tandem mass spectrometer (MS/MS) model 7000D GC/TQ (Agilent Inc., Santa Clara, CA, USA). The samples were separated on a nonpolar HP-5MS UI column (30 m length, inner diameter of 0.25 mm, and stationary phase layer thickness of 0.25 µm, Agilent Inc., Santa Clara, CA, USA). Helium was used as the carrier gas, and the flow rate was set to 1 mL/min. The inlet temperature was set at 250 °C, and the volume of the injected sample was 1 μL, at a split ratio of 1:50. An initial column temperature of 60 °C was set for the first 3 min and then increased to 246 °C at a rate of 3 °C/min and maintained for 25 min. Mass spectrometer conditions were set as follows: ionization energy of 70 eV, ion source temperature of 230 °C, and a scanning range of 40–350 m/z. The individual peaks were identified by comparison of their retention indices with the series of n-hydrocarbons (C8–C40), along with computer matching of mass spectra with commercial databases (Wiley 9N08 & NIST 2017) as well as by comparison with literature data [31]. The percentages in Table 1 and Table S1 were calculated as the mean value of component percentages on the HP-5MS UI column. All analyses were performed in triplicate. ## 2.5.2. GC-MS Conditions for Derivatized Extracts’ Analysis Derivatized extracts were analyzed by gas chromatography-mass spectrometry (GC-MS) with Shimadzu (Kyoto, Japan) Nexis GC-2030 gas chromatograph coupled with Shimadzu QP2020 NX mass detector. Helium was used as the carrier gas with a flow rate of 2.46 mL/min. The samples were separated on nonpolar column SH-5MS (30 m length, inner diameter of 0.25 mm, and stationary phase layer thickness of 0.25 µm, Shimadzu, Kyoto, Japan). The inlet temperature was set at 280 °C, and the volume of the injected sample was 1 μL, at a split ratio of 1:10. The initial column temperature of 120 °C was set for the first 3 min, increased to 292 °C at a rate of 5 °C/min, then increased to 320 °C at a rate of 30 °C/min, and maintained for 17 min. The measurement was performed with a scanning range of 35–750 m/z and with an electron impact ionization energy of 70 eV [30,32,33]. The identification of compounds in derivatized extracts was performed by comparing their trimethylsilyl (TMS) derivative mass spectra and GC retention indices relative to series of n-hydrocarbons, by computer matching with commercial libraries (Wiley 12 & NIST 2020) and those of derivatized phenolic commercial standards, and by comparison with literature data. Sample extracts were injected and analyzed in triplicate. ## 2.5.3. HPLC Conditions for Phenolics Determination The separation, identification, and quantification of 19 standards of phenolic compounds were performed using a Shimadzu Nexera LC-40 HPLC system (Shimadzu, Kyoto, Japan), equipped with a UV-VIS detector. The column used for phenolic separation was a C18 reversed-phase chromatography column (250 mm length, 4.6 mm width, and particle size 5 μm; Phenomenex, Torrance, CA, USA). Sample elution was performed at a flow rate of 1 mL/min and the temperature was set to 35 °C. The mobile phase A was ultra-pure water/$85\%$ o-phosphoric acid (v/v 99.8:0.2), and mobile phase B was methanol/acetonitrile (v/v 1:1), all HPLC grade. The chromatographic conditions were optimized in our laboratory, with a total run time of 55 min using a gradient elution as follows: initially $4\%$ B; 25 min $20\%$ B; 40 min $50\%$ B; 45 min $40\%$ B; 50 min $0\%$ B; 52 min $4\%$ B; and 55 min $4\%$ B. In order to create calibration curves for the quantification of the tested phenolics, six concentration levels were prepared and injected into HPLC in triplicate using an autosampler. Calibration curves’ ranges were 0.5 mg/L–50 mg/L for tyrosol, caffeic acid, vanilin, trans-p-coumaric acid, rutin, verbascoside, luteolin-7-glucoside, trans-o-coumaric acid, apigenin-7-glucoside, oleuroside, ligstroside, pinoresinol, quercetin, luteolin, apigenin, and diosmetin and 25–250 mg/L for 3-hydroxytyrosol, oleacein, and oleuropein. Olive bud extracts were filtered through 0.45 µm polyvinylidene difluoride (PVDF) membrane filters prior to HPLC analysis. All samples were injected in triplicate in a volume of 10 µL, and the results were expressed as mg/kg of olive bud extract. ## 2.6.1. Oxygen Radical Absorbance Capacity Assay (ORAC) The assay was performed on a Tecan Infinite 200 PRO spectrophotometer (TecanTrading AG, Männedorf, Switzerland), using 96-well black polystyrene microtiter plates (Porvair Sciences, Leatherhead, UK), according to a method described by Nazlić et al. [ 34]. Each reaction contained 180 µL of fluorescein (1 µM), 70 µL of 2,2′-Azobis (2-methyl-propionamidine) dihydrochloride (AAPH, Acros Organics) (300 mM), and 30 µL of plant extract or reference standard Trolox (6.25–50 µM) (Sigma–Aldrich). All experimental solutions were prepared in a phosphate buffer (0.075 mM, pH 7.0). Essential oils were diluted in acetone with a starting concentration of 31.8 µg/mL for Lastovka and 35.42 µg/mL for Oblica and then further diluted in phosphate buffer for the experiment by 40× and 80×. The extract was prepared in $70\%$ methanol (1 mg/mL) and was further diluted with phosphate buffer to 40 µg/mL. The measurements were performed in triplicate. The ORAC values were expressed as µmol of Trolox equivalents (TE) per gram of isolate (EOs or phenolic compounds). ## 2.6.2. Measurement of the DPPH Radical Scavenging Activity The antioxidant capacity of the extracts was assessed by the DPPH method previously described by Nazlić et al. [ 34]. The assay was performed on a Tecan Infinite 200 PRO spectrophotometer (Tecan-Trading AG, Männedorf, Switzerland) using 96-well microtiter plates for the reaction of reduction of alcoholic DPPH (2,2-diphenyl-1-picrylhydrazyl) solution (Sigma–Aldrich) in the presence of a hydrogen-donating antioxidant. Plant extracts were prepared as follows: essential oils diluted in $70\%$ acetone with a starting concentration of 35.42 µg/mL for Oblica EO and 31.8 µg/mL for Lastovka EO and bud extracts diluted in $70\%$ methanol with a starting concentration of 1 mg/mL. The first step was pipetting 100 μL of methanol (Kemika, Zagreb, Croatia) and 200 μL standard and/or sample into each well. Serial dilutions of the standard and samples were prepared (starting with the mentioned concentrations for the samples) by pipetting 100 μL from the first row with a multichannel pipette into the wells in the second row and so on to the last row, where 100 μL of the solution was ejected after mixing. In the first column, in 96-well plates, a blank sample was always added ($70\%$ methanol), and in the second column, Trolox standard of 200 μM concentration was added. After the last step of adding 100 µL of a methanolic solution of DPPH (200 µM) to each well, the reaction started, and initial absorbance was measured immediately at 517 nm. After 30 min of incubation, the absorbance was measured again, and the percentage of DPPH inhibition was calculated according to the following formula by Yen and Duh [35]:% inhibition = ((AC[0] − AA(t))/AC[0]) × 100, where AC[0] is the absorbance of the control at $t = 0$ min and AA(t) is the absorbance of the antioxidant at $t = 30$ min. All measurements were performed in triplicate. ## 2.7. Antimicrobial Activity Evaporated bud extracts were dissolved in $4\%$ DMSO at a concentration of 16 mg/mL in order to determine the minimal inhibitory concentration (MIC) by microdilution-method experiments. Mueller–Hinton broth (MHB) was added in a 1:1 ratio to the diluted extracts and 100 µL of the mixture was subjected to the first wells of the 96-well microtiter plate. Two-fold dilutions were performed in the following adjacent wells (4–0.06 mg/mL). To prepare the inoculum, bacterial cultures were grown in MHB for 24 h. The inoculum size was prepared according to the growth curves of the bacteria in the log phase (1 × 105 colony-forming units (CFU)/mL)). After the addition of 50 µL of the inoculum into each well, each plate was shaken on a microtiter plate shaker for 1 min at 600 rpm (Plate Shaker-Thermostat PST-60 HL, Biosan, Riga, Latvia). Along with the samples, $4\%$ DMSO used in sample preparation was tested as well as a positive control (50 µL of inoculum and 50 µL of broth media), a negative control (50 µL of broth media and 50 µL of essential oil/extract), and a blank (100 µL of broth media). After 24 h of incubation at 37 °C, 20 µL of the indicator of bacterial metabolic activity, 2-p-iodophenyl-3-p-nitrophenyl-5-phenyl tetrazolium chloride (INT, 2 mg/mL), was added. Plates were then shaken in the plate shaker and incubated for 1 h at 37 °C. MIC values were determined visually as the lowest concentration of the extract at which suppression of bacterial growth by the reduction of INT to red formazan was not recorded [36]. The minimal bactericide concentration (MBC) of olive bud essential oils and extracts was determined as the lowest concentration at which no microbial growth was detected. Briefly, MBC is measured by reculturing 10 uL of broth from the wells in which the MIC was determined and from the wells with higher concentrations of the extract on the Mueller–Hinton agar (MHA) plates [37]. After 24 h of incubation, a reduction in bacterial growth ($99.9\%$) was observed, and the lowest number of bacterial colonies represents the MBC. Essential oils and extracts were tested against foodborne pathogen bacteria including two Gram-negative (Escherichia coli ATCC 25922 and Salmonella enteritidis ATCC 13076) and four Gram-positive (Enterococcus faecalis ATCC 29212, Listeria monocytogenes ATCC 7644, Staphylococcus aureus ATCC 25923, and Bacillus cereus ATCC 14579) strains. ## 2.8. Cytotoxic Activity In order to determine the cytotoxic activity of olive bud essential oils and extracts, a cell viability assay (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, MTT) was performed on three cell lines: human breast adenocarcinoma (MDA-MB-231), human breast metastatic adenocarcinoma (MCF-7), and human ovarian carcinoma (OVCAR-3) cell line (LGC Standards) [38,39]. MDA-MB-231, MCF-7, and OVCAR-3 cell lines were incubated overnight in 96-well plates at a density of 9000 cells/well for MDA-MB-231 and MCF-7 and 6000 cells/well for OVCAR-3 followed by incubation with test extracts at concentrations in the range of 1–200 µg/mL for 24 h, 48 h, and 72 h (in triplicate). Afterward, cells were incubated with 0.5 g MTT/L at 37 °C for 2 h; the medium was removed, and $10\%$ dimethylsulfoxide (DMSO) was added for another 10 min at 37 °C. The indicator of metabolically active cells, formazan, was formed and measured at 570 nm using a microplate reader (BioSan, Riga, Latvia). The half maximal inhibitory concentration (IC50) value is a quantitative measure that indicates how much of a particular inhibitory substance is needed to inhibit, in vitro, a given biological process or component by $50\%$. We performed its calculation with Microsoft Excel 2016 with data normalization by the measurements of untreated controls. To determine the differences between tested concentrations, analysis of variance (one-way ANOVA) was performed using Past 3.X software (version 3.14, University of Oslo, Oslo, Norway), with the significance level at $p \leq 0.05.$ ## 3. Results As far as authors know, this is the first report of volatile compounds identified in olive bud essential oils by gas chromatography-mass spectrometry (GC-MS). Croatian cultivars Lastovka and Oblica were thoroughly characterized and the results are presented in Table 1 and Table S1 and Figure S1. A total of 113 volatiles from 18 different compound classes were identified in the EOs, with a total of $92.08\%$ (cv. Lastovka) and $88.59\%$ (cv. Oblica) identified compounds. The main component of cv. Lastovka was (Z)-3-heptadecene ($16.25\%$), followed by 1,3,5-trimethoxybenzene ($12.60\%$) and tricosane ($7.08\%$), while in cv. Oblica, (Z)-3-heptadecene ($8.13\%$) was also the most abundant compound followed by nonacosane ($7.93\%$) and heneicosane ($6.02\%$). The most abundant class of compounds in both cvs. were hydrocarbons: saturated hydrocarbons ($33.48\%$ and $35.36\%$, respectively), followed by unsaturated hydrocarbons ($22.71\%$ and $13.14\%$, respectively) and aromatic hydrocarbons ($4.11\%$ and $6.19\%$, respectively), yielding overall $60.3\%$ of hydrocarbons in cv. Lastovka and $54.69\%$ in cv. Oblica. Other highly represented compounds belong to the classes of aldehydes ($9.69\%$ and $11.29\%$, respectively), heterocyclic compounds (1.70–$3.96\%$), and alcohols (1.69–$3.23\%$). Other identified compounds belong to chemical classes of ketones, esters, organic acids, phenols, terpenes (mono-, sesqui-, and tri-), terpene alcohols (mono-, sesqui-, and di-), and furans. To deepen the knowledge of phenolic compounds from olive vegetative bud extract from Croatian domestic cvs. Lastovka and Oblica, we have analyzed their methanol extracts with high-performance liquid chromatography (Table 2, Figure S2). To further explore the phytochemical profile and expand the knowledge of metabolites from olive vegetative buds, we have evaporated methanol and performed trimethylsilyl (TMS) derivatization of extracts. By doing so, 42 compounds have been identified, most of which belong to the chemical class of sugars (13 compounds), followed by phenolic (6 compounds), fatty acids (5 compounds), and triterpenoid acids (4 compounds) (Figure 1, Table S2). Along with the chemical characterization of olive vegetative buds, their biological properties were screened as well. Both EOs and extracts of cvs. Lastovka and Oblica were subjected to antioxidant, antimicrobial, and cytotoxic analyses. Antioxidant potential was evaluated with two different methods, ORAC and DPPH. Both methods showed superior antioxidant potential for bud extracts in comparison to the EOs. From the data in Table 3, it can also be concluded that all cv. Lastovka isolates have higher antioxidative potential than cv. Oblica, measured by both methods. To test antimicrobial activity against foodborne pathogens, the minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) were determined for EOs and for extracts on Gram-positive and Gram-negative bacteria (Table 4). The tested EOs did not show antimicrobial activity or inhibited bacterial growth at a concentration of 4 mg/mL. On the other hand, the bud extracts effectively inhibited the bacterial growth of almost all tested bacteria at a concentration of 4 mg/mL. Moreover, an MIC of 2 mg/mL against L. monocytogenes was observed for both extracts. The cytotoxic activity of olive bud EOs and extracts from cvs. Lastovka and Oblica were tested against three human carcinoma cell lines: breast adenocarcinoma (MDA-MB-231), breast metastatic adenocarcinoma (MCF-7), and ovarian carcinoma (OVCAR-3). The results of the percentage of metabolically active cells and p-values for tested concentrations after 24 h, 48 h, and 72 h of incubation are shown in Figure 2, Figure S3, and Table S3. Generally, EOs showed very weak activity, and none of the samples reached the IC50 value regardless of incubation time or cell line tested. Methanol extracts of cv. Lastovka showed the best results, especially in the case of the MDA-MB-231 cell line reaching the IC50 value of 150 μg/mL for both 48 h and 72 h incubation times. Methanol extracts from cv. Oblica reached ca. $75\%$ inhibition at concentration of 200 μg/mL after 72 h for the same cell line. Additionally, both cv. methanol extracts exhibited the same $75\%$ inhibition at 200 μg/mL activity for the MCF-7 cancer cell line as well. No IC50 was reached for the OVCAR-3 cell line. ## 4. Discussion In recent years, interest in natural products and herbal medicine has greatly increased, but the embryonic parts of plants are still largely unexplored. Phytochemicals from olive vegetative buds were scarcely investigated. This is the first report of volatile compounds from olive bud EOs as well as their biological activities. Essential oils are widely used for different applications because of their antibacterial, antifungal, and insecticidal properties, most often in perfumery, in makeup and sanitary products, as food preservers and additives, in agriculture, in pharmacy as natural remedies, and in dentistry [40]. A great diversity of volatile compounds from olive vegetative bud EOs was found in Croatian cvs. Lastovka and Oblica. As already mentioned, we identified 108 volatile compounds from cv. Lastovka and 110 compounds from cv. Oblica; altogether, 113 different volatile compounds from 18 different compound classes were found in EOs from olive vegetative buds (Table S1, Figure S1). Hydrocarbons (aliphatic and aromatic) were the most abundant class of molecules in both EOs. Saturated and unsaturated hydrocarbons often serve as the key signal for chemical mimicry, acting as female mating signals and attracting male insects, which makes them an important part of the pollination mechanism [41,42]. One of the most abundant compounds with the largest differences between cultivars was 1,3,5-trimethoxybenzene, an anisole derivative derived from phenylpropanoid metabolic pathways, with a yield of $12.60\%$ in cv. Lastovka and $2.12\%$ cv. Oblica. Until now, it was mostly found in rose floral scent (family Rosaceae) [43] but also found in Eugenia confuse leaf EO from the Myrtaceae family [44]. Other abundant classes of molecules (in amount > $2\%$) are aldehydes, alcohols, ketones, fatty acids, and monoterpene alcohols. Aldehydes and ketones are associated with sweet and sometimes pungent odors and have many other biological properties. Aldehydes in EOs are often related to antibacterial properties as well as immunomodulatory properties [45], while ketones from EOs should be used with great caution since they can have neurotoxic effects [46]. Monoterpene alcohols have similar characteristics as aldehydes but are generally more potent compounds and can also act as insecticides and repellents against pests [47]. Jurešić Grubešić et al. [ 48] studied volatile compounds present in olive leaf EO of cv. Oblica during a 6-month period (from December to May). A comparison of EOs from buds and leaves both sampled in April did not result in a great degree of similarity in volatile profiles. The most represented class of compounds sampled in April in leaf EO was ketones: β-ionone ($20.48\%$), α-ionone ($18.56\%$), and (E)-β-damascenone ($5.02\%$). In our study, β-ionone ($0.43\%$) and β-damascenone ($0.1\%$) were found in much smaller amounts in cvs. Oblica and Lastovka ($0.29\%$ and $0.04\%$, respectively). Popović et al. [ 28] investigated the volatiles of olive leaf EO from cvs. Oblica and Lastovka during a three-month period (from August to October), also stating that the group of the most abundant compounds in all months was ketones, namely dihydrodehydro-β-ionone for cv. Oblica ($22.53\%$) and (E)-β-damascenone ($15\%$) for cv. Lastovka, followed by the class of sesquiterpenes and, only afterward, the class of aldehydes. Studies of phytochemicals from bud extract are scarce and are mostly performed on particular phenolic compounds [49,50]. Phenolic compounds have a role during fruit development, with OLE being one of the most abundant compounds in olive fruit and also in the bud. Malik et al. [ 26] studied the OLE level in the transition from vegetative bud to mature black fruit and revealed that the highest OLE concentrations were in the vegetative bud stage, and the results were later confirmed by Taamalli et al. [ 27]. Our results confirm high OLE levels in vegetative bud extract; however, the compound with the highest concentration in buds for cv. Lastovka was oleacein. When comparing OLE levels in our study (2.05 and 2.69 mg/g fresh weight, FW, for cv. Lastovka and Oblica, respectively) with those of Malik et al. [ 26] (with 58.36 mg/g FW), it may be concluded that the difference in content could come from sample preparation before storage, since we did not freeze buds in liquid nitrogen prior to storage at −80°C. Cecchi et al. [ 51] also showed that the absence of liquid nitrogen treatment prior to unripe olive fruit storage results in a loss of OLE as much as $68\%$, but since the matrixes of the olive bud and fruit are different, a similar study should be performed for olive buds. The main category of olive fruit phenols are secoiridoids, including oleacein (3,4-DHPEA-EDA), which also represents one of the most abundant phenol compounds of extra virgin olive oil. In addition to oleacein, other secoiridoids such as oleuropein aglycone (3,4-DHPEA-EA) and oleocanthal (p-HPEA-EDA) are also present in large concentrations [52,53]. Other phenolic compounds found in higher amounts were phenolic alcohol 3-hydroxytyrosol, terpene glycoside oleuroside, glycosylated flavonoid rutin, and phenylpropanoid glycoside verbascoside. Phenolic compounds derived from olive fruit, leaves, and oil are largely responsible for their beneficial health effects [54]. Furthermore, we derivatized bud extract and performed GC-MS analysis to broaden the knowledge and the spectrum of existing compounds (Figure 1, Table S2). Dastkar et al. studied the differential expression of genes in buds of ON- vs. OFF-crop trees and found differences in the expression of genes related to carbohydrate metabolism as well as in genes involved in the secondary metabolism pathway—precisely, genes involved in phenolic biosynthesis [55]. Most of the identified compounds belong to the class of sugars (13 from 42 identified). Several phenolic compounds have also been identified, followed by fatty acids, triterpenoid acids, and organic acids. Fatty acids from olive fruit and EVOO, especially monounsaturated fatty acids (MUFAs), are widely known for their health benefits, especially on the cardiovascular system. Oleic acid is a ω-9 fatty acid, one of the most abundant MUFAs in EVOO, and is often associated with beneficial anti-inflammatory effects and improvement of immune system function [56]. Linoleic acid is a ω-6 essential fatty acid and cannot be synthesized in humans. There are a lot of controversies regarding the health implications of linoleic acid, but if consumed moderately and mostly used as a replacement for solid fats, it could be beneficial for the improvement of cardiovascular risk as well as long-term glycemic control and insulin resistance [57,58]. Pentacyclic triterpenes (including triterpenoid acids and alcohols) from olive extracts were previously studied and showed various biological benefits, such as immunomodulatory, anti-inflammatory, antioxidant, anticancer, antiviral, and antimicrobial activity [59]. As indicated by the results obtained using DPPH and ORAC assays (Table 3), all tested extracts exhibited antioxidant activity. By comparing the results obtained from both methods, the best ability to neutralize free radicals was shown by extracts of cv. Lastovka, both for EOs and phenolics. However, as expected, phenolic extracts of both cultivars yielded better results than bud EO extracts. The DPPH IC50 value for cv. Lastovka bud phenolic extracts and EO were 0.274 mg/mL and 30.51 mg/mL, respectively. ORAC values, expressed as μmol of Trolox equivalents (TE) per gram of extract, were much higher for cv. Lastovka phenolic bud extracts (1835.42) than EO extract (139.95). There is a lack of information about the antioxidant activity of olive buds in contrast to other olive tree parts such as olive fruits, olive leaves, and olive oil [60]. To the best of our knowledge, this is the first study that investigated the antioxidant activity of EOs and phenolic extracts from olive buds from two domestic olive cultivars, cvs. Lastovka and Oblica, using two methods, DPPH and ORAC. Rekik et al. [ 60] measured the antioxidant activity of olive flower extracts using a DPPH radical scavenging assay, which ranged from 5.24 to 11.37 µg/mL, and concluded that the antioxidant activity increases with the development stage of the flower. Kouka et al. [ 61] tested four olive blossom polyphenolic extracts using the same method and obtained IC50 values from 40.5 to 73.25 µg/mL. Their results can be related to a certain extent to the results from our study, where the obtained values were weaker, 274 µg/mL for cv. Lastovka and 385 µg/mL for cv. Oblica. It is reported that hydroxytyrosol, due to its structure, has beneficial antioxidant, antimicrobial, anti-inflammatory, and anticancer properties [62]. The higher concentration of 3-hydroxytyrosol in cv. Lastovka extracts could be responsible for the better antioxidant activity compared to cv. Oblica extracts. Jurišić Grubešić et al. [ 48] studied the antioxidant capacity of EOs of cv. Oblica leaves using DPPH and ORAC methods, with values for the DPPH method ranging from 23.58 to 130.71 mg/mL and for ORAC from 4.43 to 73.12 µmol TE. When compared to our results of DPPH measurements, IC50 (cv. Lastovka 30.51 and cv. Oblica 55.36 µmol TE) results were in favor of olive bud EO, that is, bud EO showed a better antioxidant capacity than leaf EO. The same results were obtained using the ORAC method, where only one of six measurements (conducted in February; 73.12 µmol TE) gave a better result compared to the results obtained for cv. Oblica (43.11 µmol TE). The ORAC method was used for the assessment of monitoring quenching free peroxyl radicals. Šimat et al. [ 63] tested the antioxidant activity of olive leaf extracts from six Mediterranean olive cultivars using the ORAC method, among others, and the antioxidant activity was in favor of cv. Oblica leaf extracts, since the reported values were higher for Oblica than the Lastovka cultivar. The antioxidant activity of parts of the olive tree is related to different groups of bioactive components such as fatty acids, triterpenic acids, polyphenols, and phytosterols as well as their synergistic effect, since it is not possible to predict the total antioxidant potential of the samples from the antioxidant activity of individual compounds [64,65]. Our results showed a higher presence of compounds such as tripenoids, acids and alcohols (betulinic acid, ursolic acid, maslinic acid, and erythrodiol), secoiridoids (oleacein, oleuropein, and ligstroside), β-sitosterol, and 3-hydroxytyrosol in the extracts of cv. Lastovka, which could explain the better antioxidant activity in both phenolic extract and EO when compared to cv. Oblica. Bud extracts are overall more abundant with compounds that possess higher antioxidant properties than Eos, which was confirmed by the results of ORAC and DPPH. Bud EOs are the richest in aliphatic hydrocarbons that do not possess such activities. The total antioxidant potential of the extract is mostly due to their combined effects (synergistic, antagonistic, and additive) and not only from the antioxidant activity of individual compounds [65]. Foodborne pathogens can have a great impact on human health and cause a large number of diseases [66]. We tested the antimicrobial activity of EOs and extracts against selected foodborne pathogens, Gram-positive and Gram-negative bacteria, and the results were expressed as minimal inhibitory concentration (MIC) and minimal bactericidal concentration (MBC) (Table 4). At the highest concentration of EOs (4 mg/mL), the MIC was not recorded for any of the tested bacteria, so the MBC was not tested. The bud extracts showed much better results. The MIC was determined for all bacterial strains, and the lowest MIC (2 mg/mL) was for L. monocytogenes. As far as the authors know, there are no previous data for the antimicrobial activity of olive bud extracts except for several reports where the antimicrobial activity of extracts from table olives, olive leaves, olive oils, and olive mill wastewater was tested. Guo et al. [ 67] tested olive oil polyphenol extract (OOPE) on L. monocytogenes and showed that the bacterial colony did not grow at an OOPE concentration of 1.25 mg/mL. Olive oil polyphenol extract affected the intracellular adenosine 5′-triphosphate (ATP) concentration level and cell membrane potential and led to a reduction in bacterial protein and DNA levels and a change in cell morphology. The extract from the cv. Lastovka variety showed slightly better results for S. enteridis, with an equal MIC and MBC concentration of 4 mg/mL, whereas for cv. Oblica, the MIC was reached at 4 mg/mL but the MBC could not be determined. A previous study by Liu et al. [ 68] on the antimicrobial activity of olive leaf extract (OLE) on S. enteridis, L. monocytogenes, and E. coli showed that, at a concentration of 62.5 mg/mL, the growth of S. enteritidis and L. monocytogenes was completely inhibited. The antimicrobial activity against the Gram-positive bacteria E. faecalis was also tested for the bud extracts of cultivars, and the MIC as well as the MBC was obtained at a concentration of 4 mg/mL. A previous study on olive leaf extract against E. faecalis inhibited bacterial growth at a concentration of 0.60 mg/mL, while the MBC could not be determined [69]. In the same study, an extract of table olives was also tested against E. faecalis, and the MIC was achieved at a concentration of 5 mg/mL, while the MBC could not be determined. The growth of S. aureurs, B. cereus, and E. coli was also affected by olive bud extracts of both cultivars, with MIC values of 4 mg/mL and the same MBC values for E.coli and S. aureurs (only for cv. Oblica extract). Šimat et al. [ 63] studied the antibacterial activity of olive leaf extract against the same bacterial strains. Olive leaf extract showed inhibitory activity against S. aureurs (MIC and MBC = 2 mg/mL) and B. cereus (MIC and MBC = 4 mg/mL for cv. Lastovka; MIC and MBC = 8 mg/mL for cv. Oblica), while no antimicrobial activity was detected against E. coli. Phenolic compounds are a large and diverse class of compounds with different effects on microorganisms. Their structure is related to antibacterial activity, which can be mediated by different mechanisms [70]. In addition to individual phenolic compounds, a group of phenols can also interact and have a synergistic, additive, or antagonistic effect [36]. Other than phenolics, triterpenoid acids also act against different microorganisms [71]. The extract of industrial olive oil waste is rich in oleanolic and maslinic acid. Blanco-Cabra et al. [ 72] tested both compounds along with their derivatives against several bacterial strains including S. aureus, E. faecalis, and E. coli. The MIC was determined for maslinic acid in very low concentrations for S. aureus and E. faecalis (15 µg/mL for both strains), while there was no activity for E. coli. Similar results were obtained for oleanolic acid, but MIC values were slightly higher for S. aureus and E. faecalis (30 µg/mL for both strains), and there was no activity for E. coli. Oleanolic and maslinic acid derivatives showed even better results on tested bacterial strains and great antibiofilm activity for S. aureus. Overall, the chemical profile correlates with antimicrobial activity; extracts rich in phenolics and triterpenoid acids had better results than EOs that were high in hydrocarbons. The Mediterranean diet is known to be linked with lower incidences of major illnesses such as cancer and cardiovascular disease. Frequent consumption of olives and olive oil, rich in antioxidant compounds that possess chemoprotective effects, is one of the different nutritional habits of the population from the Mediterranean basin [73]. Several authors have studied the effect of olive, EVOO, and olive leaf extract on different cancer cell lines, but there are no studies performed on extracts from olive bud. EOs and extracts of cvs. Lastovka and Oblica were tested for cytotoxic activity against human breast adenocarcinoma MDA-MB-231, human breast metastatic adenocarcinoma MCF7, and human ovarian carcinoma OVCAR-3 cell lines (Figure 2 and Figure S3 and Table S3). Essential oil had no activity for any of the tested cell lines (Figure S3). Methanol extracts showed better results than EOs for both MDA-MB-231 and MCF-7 cell lines. The OVCAR-3 cell line was equally resistant to both isolates from both cultivars for the tested concentrations. Methanol extracts of cv. Lastovka showed somewhat better results, especially for the MDA-MB-231 cell line, where IC50 was reached after 48 and 72 h of incubation (150 μg/mL for both measurements). Methanol extracts from cv. Oblica were less potent but reached $75\%$ inhibition at a concentration of 200 μg/mL after 72 h for the same cell line as well as for the MCF-7 cell line. Benot-Dominguez et al. [ 74] studied the effect of olive leaf extract on different cancer cell lines, including MDA-MB-231 and OVCAR-3, as well as on nontumoral cells and found that OLE specifically inhibits MDA-MB-231 and OVCAR-3 cell viability (IC50 = 200 μg/mL) without affecting nontumoral cells. Oleuropein was a major component of OLE extracts ($87\%$ of the total components) and was able to induce a cytotoxic effect on both cell lines. The results show that OLE has multiple effects on cancer cell lines: it promotes cell cycle arrest, promotes apoptosis, selectively increases ROS production and alters the protein levels of oxidative stress pathway-related proteins, and compromises mitochondrial function. Elamin et al. [ 75] studied the cytotoxic effects of OLE, a major phenol from olive oil, on human breast adenocarcinoma MDA-MB-231 and human breast metastatic adenocarcinoma MCF7 cell lines. The authors determined specific cytotoxicity against breast cancer cells, for MDA-MB-231 with LC50 = 200 µM and for MCF7 with LC50 = 150 µM. Other than OLE, cytotoxic effects on MCF-7 and MDA-MB-231 cell lines were also found significant for other phenolic compounds, such as verbascoside [76] and hydroxytyrosol [77,78]. Pentacyclic triterpenes are also potent cytotoxic molecules and are known to have a cytotoxic effect on all the cancer cell lines used in this study [79]. Overall, the results for the tested biological activities of olive buds show a moderate but promising result. Since phenolics from other olive extracts seem to have the largest biological effect on metabolic disorders and the cardiovascular system, further research can be focused in that direction to examine whether this combination of phenolics and triterpenes could result in a similar or better effect. Since gemmotherapy is a scarcely investigated area, these results point to the potential hidden in embryological tissues and the need for a comprehensive search in this area for novel, promising natural compounds and/or their synergistic effect. ## 5. Conclusions Regardless of the fact that Olea europaea L. is one of the world’s most important and studied crops originating from the Mediterranean basin, phytochemicals from olive vegetative buds are scarcely investigated at present. A comprehensive chemical analysis of olive vegetative bud from two Croatian cultivars revealed numerous compounds in EO and methanol extract that could be of great use for nutraceutical or biotechnological application. Olive buds are rich in phenolics, especially oleacin, oleuropein, and hydroxytyrosol. Previous studies on olive leaf extracts, which are widely used in phytotherapy for the treatment of various conditions such as lipid regulation, hypertension, and cardiac system protection, point to oleuropein and hydroxytyrosol as the main constituents. The results of preformed biological activities (antioxidant activity, antimicrobial activity, and cytotoxic activity) show moderate biological potential of olive bud extract and the need for further investigation in different biological systems. Research on embryonic plant parts could lead to the discovery of novel compounds as well as their nutraceutical or biotechnological application potential. Since the entire field is understudied, a further increase in research in this area is needed. ## References 1. Langgut D., Cheddadi R., Carrión J.S., Cavanagh M., Colombaroli D., Eastwood W.J., Greenberg R., Litt T., Mercuri A.M., Miebach A.. The Origin and Spread of Olive Cltivation in the Mediterranean Basin: The Fossil Pollen Evidence. Holocene (2019) 29 902-922. DOI: 10.1177/0959683619826654 2. Wren R.C., Holmes E.M.. Potter’s Cyclopaedia of Botanical Drugs and Preparations (1915) 111 3. Hashmi M.A., Khan A., Hanif M., Farooq U., Perveen S.. Traditional Uses, Phytochemistry, and Pharmacology of. Evid. Based Complement. Altern. Med. (2015) 2015 541591. DOI: 10.1155/2015/541591 4. la Lastra A.d.C., Barranco D.M., Motilva V., Herrerias M.J.. Mediterrranean Diet and Health Biological Importance of Olive Oil. Curr. Pharm. Des. (2001) 7 933-950. DOI: 10.2174/1381612013397654 5. Rocha J., Borges N., Pinho O.. Table olives and health: A review. J. Nutr. Sci. (2020) 9 e57. DOI: 10.1017/jns.2020.50 6. Vogel P., Kasper Machado I., Garavaglia J., Zani V.T., de Souza D., Morelo Dal Bosco S.. Polyphenols Benefits of Olive Leaf (. Nutr. Hosp. (2014) 31 1427-1433. PMID: 25726243 7. Jukić Špika M., Žanetić M., Kraljić K., Soldo B., Ljubenkov I., Politeo O., Škevin D.. Differentiation Between Unfiltered and Filtered. J. Food Sci. (2019) 84 877-885. DOI: 10.1111/1750-3841.14494 8. Boskou D., Blekas G., Tsimidou M.. Phenolic Compounds in Olive Oil and Olives. Curr. Top. Nutraceutical Res. (2005) 3 125-136 9. Guinda Á., Castellano J.M., Santos-Lozano J.M., Delgado-Hervás T., Gutiérrez-Adánez P., Rada M.. Determination of Major Bioactive Compounds from Olive Leaf. LWT Food Sci. Technol. (2015) 64 431-438. DOI: 10.1016/j.lwt.2015.05.001 10. Gorzynik-Debicka M., Przychodzen P., Cappello F., Kuban-Jankowska A., Marino Gammazza A., Knap N., Wozniak M., Gorska-Ponikowska M.. Potential Health Benefits of Olive Oil and Plant Polyphenols. Int. J. Mol. Sci. (2018) 19. DOI: 10.3390/ijms19030686 11. Sofi F., Abbate R., Gensini G.F., Casini A.. Accruing Evidence on Benefits of Adherence to the Mediterranean Diet on Health: An Updated Systematic Review and Meta-analysis. Am. J. Clin. Nutr. (2010) 92 1189-1196. DOI: 10.3945/ajcn.2010.29673 12. Hassen I., Casabianca H., Hosni K.. Biological activities of the natural antioxidant oleuropein: Exceeding the expectation–A mini-review. J. Funct. Foods (2015) 18 926-940. DOI: 10.1016/j.jff.2014.09.001 13. Wainstein J., Ganz T., Boaz M., Bar Dayan Y., Dolev E., Kerem Z., Madar Z.. Olive Leaf Extract as a Hypoglycemic Agent in Both Human Diabetic Subjects and in Rats. J. Med. Food (2012) 15 605-610. DOI: 10.1089/jmf.2011.0243 14. Abunab H., Dator W.L., Hawamdeh S.. Effect of Olive Leaf Extract on Glucose Levels in Diabetes-Induced Rats: A Systematic Review and Meta-Analysis. J. Diabetes (2017) 9 947-957. DOI: 10.1111/1753-0407.12508 15. Lockyer S., Rowland I., Spencer J.P.E., Yaqoob P., Stonehouse W.. Impact of Phenolic-Rich Olive Leaf Extract on Blood Pressure, Plasma Lipids and Inflammatory Markers: A Randomised Controlled Trial. Eur. J. Nutr. (2017) 56 1421-1432. DOI: 10.1007/s00394-016-1188-y 16. Acar-Tek N., Ağagündüz D.. Olive leaf (. Ann. Nutr. Metab. (2020) 76 10-15. DOI: 10.1159/000505508 17. Toulabi T., Delfan B., Rashidipour M., Yarahmadi S., Ravanshad F., Javanbakht A., Almasian M.. The Efficacy of Olive Leaf Extract on Healing Herpes Simplex Virus Labialis: A Randomized Double-Blind Study. Explore (2022) 18 287-292. DOI: 10.1016/j.explore.2021.01.003 18. Ahmadpour E., Toulabi T., Yadegarinia D., Yarahmadi S., Mohammadi R., Keyvanfar A.. Efficacy of Olive Leaves Extract on the Outcomes of Hospitalized COVID-19 Patients: A Randomized, Triple-Blinded Clinical Trial. Explore (2022) 19. Yeh Y.T., Cho Y.Y., Hsieh S.C., Chiang A.N.. Chinese Olive Extract Ameliorates Hepatic Lipid Accumulation in Vitro and in Vivo by Regulating Lipid Metabolism. Sci. Rep. (2018) 8 1057. DOI: 10.1038/s41598-018-19553-1 20. Gupta A., Vij G., Chopra K.. Possible Role of Oxidative Stress and Immunological Activation in Mouse Model of Chronic Fatigue Syndrome and its Attenuation by Olive Extract. J. Neuroimmunol. (2010) 226 3-7. DOI: 10.1016/j.jneuroim.2010.05.021 21. Wang M., Zhang S., Zhong R., Wan F., Chen L., Liu L., Yi B., Zhang H.. Olive Fruit Extracts Supplement Improve Antioxidant Capacity via Altering Colonic Microbiota Composition in Mice. Front. Nutr. (2021) 8 645099. DOI: 10.3389/fnut.2021.645099 22. Donno D., Beccaro G.L., Cerutti A.K., Mellano M.G., Bounous G., Rao A.V., Leticia G.R.. Bud Extracts as New Phytochemical Source for Herbal Preparations—Quality Control and Standardization by Analytical Fingerprint. Phytochemicals—Isolation, Characterisation and Role in Human Health (2015) 187-218 23. Hoefler C., Fleurentin J., Mortier F., Pelt J.M., Guillemain J.. Comparative choleretic and hepatoprotective properties of young sprouts and total plant extracts of rosmarinus officinalis in rats. J. Ethnopharmacol. (1987) 19 133-143. DOI: 10.1016/0378-8741(87)90037-7 24. Brigo B.. L’uomo, la Fitoterapia, la Gemmoterapia. 211 Sindromi Cliniche Trattate con Fitocomplessi e Gemmoderivati (2009) 768 25. Kovalikova Z., Lnenicka J., Andrys R.. The Influence of Locality on Phenolic Profile and Antioxidant Capacity of Bud Extracts. Foods (2021) 10. DOI: 10.3390/foods10071608 26. Malik N.S.A., Bradford J.M.. Changes in Oleuropein Levels During Differentiation and Development of Floral Buds in ‘. Sci. Hortic. (2006) 110 274-278. DOI: 10.1016/j.scienta.2006.07.016 27. Taamalli A., Abaza L., Arráez Román D., Segura Carretero A., Fernández Gutiérrez A., Zarrouk M., Ben Youssef Nabil B.. Characterisation of Phenolic Compounds by HPLC–TOF/IT/MS in Buds and Open Flowers of ‘. Phytochem. Anal. (2013) 24 504-512. DOI: 10.1002/pca.2450 28. Popović M., Jukić Špika M., Veršić Bratinčević M., Ninčević T., Matešković A., Mandušić M., Rošin J., Nazlić M., Dunkić V., Vitanović E.. Essential Oil Volatile Fingerprint Differentiates Croatian cv.. Molecules (2021) 26. DOI: 10.3390/molecules26123533 29. Marinova D., Ribarova F., Atanassova M.. Total Phenolics and Total Flavonoids in Bulgarian Fruits and Vegetables. J. Chem. Technol. Metall. (2005) 40 255-260 30. Zarate E., Boyle V., Rupprecht U., Green S., Villas-Boas S.G., Baker P., Pinu F.R.. Fully Automated Trimethylsilyl (TMS) Derivatisation Protocol for Metabolite Profiling by GC-MS. Metabolites (2017) 7. DOI: 10.3390/metabo7010001 31. Adams R.P.. Identification of Essential Oil Components by Gas Chromatography/Quadrupole Mass Spectrometry (2007) 32. Lisec J., Schauer N., Kopka J., Willmitzer L., Fernie A.R.. Gas Chromatography Mass Spectrometry-Based Metabolite Profiling in Plants. Nat. Protoc. (2006) 1 387-396. DOI: 10.1038/nprot.2006.59 33. Frleta R., Popović M., Smital T., Šimat V.. Comparison of Growth and Chemical Profile of Diatom. Mar. Drugs (2022) 20. DOI: 10.3390/md20110697 34. Nazlić M., Kremer D., Grubešić R.J., Soldo B., Vuko E., Stabentheiner E., Ballian D., Bogunić F., Dunkić V.. Endemic. Plants (2020) 9. DOI: 10.3390/plants9121646 35. Yen G.C., Duh P.D.. Scavenging Effect of Methanolic Extracts of Peanut Hulls on Free-Radical and Active-Oxygen Species. J. Agric. Food Chem. (1994) 42 629-632. DOI: 10.1021/jf00039a005 36. Skroza D., Šimat V., Smole Možina S., Katalinić V., Boban N., Generalić Mekinić I.. Interactions of Resveratrol with Other Phenolics and Activity Against Food-Borne Pathogens. Food Sci. Nutr. (2019) 7 2312-2318. DOI: 10.1002/fsn3.1073 37. Čagalj M., Skroza D., Razola-Díaz M.d.C., Verardo V., Bassi D., Frleta R., Generalić Mekinić I., Tabanelli G., Šimat V.. Variations in the Composition, Antioxidant and Antimicrobial Activities of. Mar. Drugs (2022) 20. DOI: 10.3390/md20010064 38. Srisawat T., Chumkaew P., Heed-Chim W., Sukpondma Y., Kanokwiroon K.. Phytochemical Screening and Cytotoxicity of Crude Extracts of. Trop. J. Pharm. Res. (2013) 12 71-76. DOI: 10.4314/tjpr.v12i1.12 39. Popović M., Maravić A., Čikeš Čulić V., Đulović A., Burčul F., Blažević I.. Biological Effects of Glucosinolate Degradation Products from Horseradish: A Horse That Wins the Race. Biomolecules (2020) 10. DOI: 10.3390/biom10020343 40. Bakkali F.B., Averbeck S., Averbeck D., Idaomar M.. Biological Effects of Essential Oils—A Review. Food Chem. Toxicol. (2008) 46 446-475. DOI: 10.1016/j.fct.2007.09.106 41. Schiestl F.P., Cozzolino S.. Evolution of Sexual Mimicry in the Orchid Subtribe Orchidinae: The Role of Preadaptations in the Attraction of Male Bees as Pollinators. BMC Evol. Biol. (2008) 8. DOI: 10.1186/1471-2148-8-27 42. Son H.-D., Yun S.A., Kim S.-C., Im H.-T.. Hydrocarbon Patterns in. Korean J. Plant Taxon. (2020) 50 148-153. DOI: 10.11110/kjpt.2020.50.2.148 43. Shi S., Zhang Z.. Genetic and Biochemical Aspects of Floral Scents in Roses. Int. J. Mol. Sci. (2022) 23. DOI: 10.3390/ijms23148014 44. Setzer W.N., Noletto J.A., Vincent M.A.. 1,3,5-Trimethoxybenzene and 2,4,6-Trimethoxystyrene are the Major Components in the Leaf Oil of. Nat. Prod. Commun. (2006) 1 43-45. DOI: 10.1177/1934578X0600100108 45. Valdivieso-Ugarte M., Gomez-Llorente C., Plaza-Díaz J., Gil Á.. Antimicrobial, Antioxidant, and Immunomodulatory Properties of Essential Oils: A Systematic Review. Nutrients (2019) 11. DOI: 10.3390/nu11112786 46. Tisserand R., Young R., Tisserand R., Young R.. Essential Oil Composition. Essential Oil Safety (2014) 5-22 47. Reis S.L., Mantello A.G., Macedo J.M., Gelfuso E.A., Da Silva C.P., Fachin A.L., Cardoso A.M., Beleboni R.O.. Typical Monoterpenes as Insecticides and Repellents Against Stored Grain Pests. Molecules (2016) 21. DOI: 10.3390/molecules21030258 48. Jurišić Grubešić R., Nazlić M., Miletić T., Vuko E., Vuletić N., Ljubenkov I., Dunkić V.. Antioxidant Capacity of Free Volatile Compounds From. Antioxidants (2021) 10. DOI: 10.3390/antiox10111832 49. Bendini A., Cerretani L., Carrasco-Pancorbo A., Gómez-Caravaca A.M., Segura-Carretero A., Fernández-Gutiérrez A., Lercker G.. Phenolic Molecules in Virgin Olive Oils: A Survey of Their Sensory Properties, Health Effects, Antioxidant aActivity and Analytical Methods. An Overview of the Last Decade. Molecules (2007) 12 1679-1719. DOI: 10.3390/12081679 50. Silva S., Gomes L., Leitão F., Coelho A.V., Boas L.V.. Phenolic Compounds and Antioxidant Activity of. Food Sci. Technol. Int. (2006) 12 385-395. DOI: 10.1177/1082013206070166 51. Cecchi L., Guerrini L., Bellumori M., Balli D., Xie P., Parenti A., Mulinacci N.. Optimization of the Production Process of Dried Unripe Olives (. LWT (2020) 129 109569. DOI: 10.1016/j.lwt.2020.109569 52. Naruszewicz M., Czerwinska E.M., Kiss K.A.. Oleacein. Translation From Mediterranean Diet to Potential Antiatherosclerotic Drug. Curr. Pharm. Des. (2015) 21 1205-1212. DOI: 10.2174/1381612820666141007141137 53. Alagna F., Mariotti R., Panara F., Caporali S., Urbani S., Veneziani G., Esposto S., Taticchi A., Rosati A., Rao R.. Olive Phenolic Compounds: Metabolic and Transcriptional Profiling During Fruit Development. BMC Plant Biol. (2012) 12. DOI: 10.1186/1471-2229-12-162 54. Romani A., Ieri F., Urciuoli S., Noce A., Marrone G., Nediani C., Bernini R.. Health Effects of Phenolic Compounds Found in Extra-Virgin Olive Oil, By-Products, and Leaf of. Nutrients (2019) 11. DOI: 10.3390/nu11081776 55. Dastkar E., Soleimani A., Jafary H., de Dios Alche J., Bahari A., Zeinalabedini M., Salami S.A.. Differential Expression of Genes in Olive Leaves and Buds of ON-versus OFF-Crop Trees. Sci. Rep. (2020) 10 15762. DOI: 10.1038/s41598-020-72895-7 56. Sales-Campos H., Souza P.R., Peghini B.C., da Silva J.S., Cardoso C.R.. An Overview of the Modulatory Effects of Oleic Acid in Health and Disease. Mini Rev. Med. Chem. (2013) 13 201-210. PMID: 23278117 57. Jandacek R.J.. Linoleic acid: A Nutritional Quandary. Healthcare (2017) 5. DOI: 10.3390/healthcare5020025 58. Marangoni F., Agostoni C., Borghi C., Catapano A.L., Cena H., Ghiselli A., La Vecchia C., Lercker G., Manzato E., Pirillo A.. Dietary Linoleic Acid and Human Health: Focus on Cardiovascular and Cardiometabolic Effects. Atherosclerosis (2020) 292 90-98. DOI: 10.1016/j.atherosclerosis.2019.11.018 59. Rufino-Palomares E.E., Pérez-Jiménez A., García-Salguero L., Mokhtari K., Reyes-Zurita F.J., Peragón-Sánchez J., Lupiáñez J.A.. Nutraceutical Role of Polyphenols and Triterpenes Present in the Extracts of Fruits and Leaves of. Molecules (2022) 27. DOI: 10.3390/molecules27072341 60. Rekik O., Ben Mansour A., Bouaziz M.. Evaluation of Phenolic Composition and Antioxidant Activity Changes in Olive Flowers During Development Using HPLC/DAD and LC-MS/MS. Electrophoresis (2018) 39 1663-1672. DOI: 10.1002/elps.201700200 61. Kouka P., Tekos F., Valta K., Mavros P., Veskoukis A.S., Angelis A., Skaltsounis A.-L., Kouretas D.. Olive Tree Blossom Polyphenolic Extracts Exert Antioxidant and Antimutagenic Activities in Vitro and in Various Cell Lines. Oncol. Rep. (2019) 42 2814-2825. DOI: 10.3892/or.2019.7386 62. Martínez-Zamora L., Peñalver R., Ros G., Nieto G.. Olive Tree Derivatives and Hydroxytyrosol: Their Potential Effects on Human Health and Its Use as Functional Ingredient in Meat. Foods (2021) 10. DOI: 10.3390/foods10112611 63. Šimat V., Skroza D., Tabanelli G., Čagalj M., Pasini F., Gómez-Caravaca A.M., Fernández-Fernández C., Sterniša M., Smole Možina S., Ozogul Y.. Antioxidant and Antimicrobial Activity of Hydroethanolic Leaf Extracts from Six Mediterranean Olive Cultivars. Antioxidants (2022) 11. DOI: 10.3390/antiox11091656 64. Khwaldia K., Attour N., Matthes J., Beck L., Schmid M.. Olive Byproducts and Their Bioactive Compounds as a Valuable Source for Food Packaging Applications. Compr. Rev. Food Sci. Food Saf. (2022) 21 1218-1253. DOI: 10.1111/1541-4337.12882 65. Skroza D., Šimat V., Vrdoljak L., Jolić N., Skelin A., Čagalj M., Frleta R., Generalić Mekinić I.. Investigation of Antioxidant Synergisms and Antagonisms among Phenolic Acids in the Model Matrices Using FRAP and ORAC Methods. Antioxidants (2022) 11. DOI: 10.3390/antiox11091784 66. Bintsis T.. Foodborne Pathogens. AIMS Microbiol. (2017) 3 529-563. DOI: 10.3934/microbiol.2017.3.529 67. Guo L., Sun Q., Gong S., Bi X., Jiang W., Xue W., Fei P.. Antimicrobial Activity and Action Approach of the Olive Oil Polyphenol Extract Against. Front. Microbiol. (2019) 10 1586. DOI: 10.3389/fmicb.2019.01586 68. Liu Y., McKeever L.C., Malik N.S.A.. Assessment of the Antimicrobial Activity of Olive Leaf Extract Against Foodborne Bacterial Pathogens. Front. Microbiol. (2017) 8 113. DOI: 10.3389/fmicb.2017.00113 69. Karygianni L., Cecere M., Skaltsounis A.L., Argyropoulou A., Hellwig E., Aligiannis N., Wittmer A., Al-Ahmad A.. High-Level Antimicrobial Efficacy of Representative Mediterranean Natural Plant Extracts Against Oral Microorganisms. Biomed. Res. Int. (2014) 2014 839019. DOI: 10.1155/2014/839019 70. Xie Y., Yang W., Tang F., Chen X., Ren L.. Antibacterial Activities of Flavonoids: Structure-Activity Relationship and Mechanism. Curr. Med. Chem. (2015) 22 132-149. DOI: 10.2174/0929867321666140916113443 71. Ndjateu F.S.T., Tsafack R.B.N., Nganou B.K., Awouafack M.D., Wabo H.K., Tene M., Tane P., Eloff J.N.. Antimicrobial and Antioxidant Activities of Extracts and Ten Compounds from Three Cameroonian medicinal plants:. S. Afr. J. Bot. (2014) 91 37-42. DOI: 10.1016/j.sajb.2013.11.009 72. Blanco-Cabra N., Vega-Granados K., Moya-Andérico L., Vukomanovic M., Parra A., Álvarez de Cienfuegos L., Torrents E.. Novel Oleanolic and Maslinic Acid Derivatives as a Promising Treatment Against Bacterial Biofilm in Nosocomial Infections: An in Vitro and in Vivo Study. ACS Infect. Dis. (2019) 5 1581-1589. DOI: 10.1021/acsinfecdis.9b00125 73. Owen R.W., Haubner R., Würtele G., Hull E., Spiegelhalder B., Bartsch H.. Olives and Olive Oil in Cancer Prevention. Eur. J. Cancer Prev. (2004) 13 319-326. DOI: 10.1097/01.cej.0000130221.19480.7e 74. Benot-Dominguez R., Tupone M.G., Castelli V., d’Angelo M., Benedetti E., Quintiliani M., Cinque B., Forte I.M., Cifone M.G., Ippoliti R.. Olive Leaf Extract Impairs Mitochondria by Pro-Oxidant Activity in MDA-MB-231 and OVCAR-3 Cancer Cells. Biomed. Pharmacother. (2021) 134 111139. DOI: 10.1016/j.biopha.2020.111139 75. Elamin M.H., Daghestani M.H., Omer S.A., Elobeid M.A., Virk P., Al-Olayan E.M., Hassan Z.K., Mohammed O.B., Aboussekhra A.. Olive Oil Oleuropein has Anti-Breast Cancer Properties with Higher Efficiency on ER-Negative Cells. Food Chem. Toxicol. (2013) 53 310-316. DOI: 10.1016/j.fct.2012.12.009 76. Şenol H., Tulay P., Ergören M., Hanoğlu A., Çalış İ., Mocan G.. Cytotoxic Effects of Verbascoside on MCF-7 and MDA-MB-231. Turk. J. Pharm. Sci. (2021) 18 637-644. DOI: 10.4274/tjps.galenos.2021.36599 77. Calahorra J., Martínez-Lara E., Granadino-Roldán J.M., Martí J.M., Cañuelo A., Blanco S., Oliver F.J., Siles E.. Crosstalk Between Hydroxytyrosol, a Major Olive Oil Phenol, and HIF-1 in MCF-7 Breast Cancer Cells. Sci. Rep. (2020) 10 6361. DOI: 10.1038/s41598-020-63417-6 78. Sánchez-Quesada C., Gutiérrez-Santiago F., Rodríguez-García C., Gaforio J.J.. S Synergistic Effect of Squalene and Hydroxytyrosol on Highly Invasive MDA-MB-231 Breast Cancer Cells. Nutrients (2022) 14. DOI: 10.3390/nu14020255 79. Chudzik M., Korzonek-Szlacheta I., Król W.. Triterpenes as Potentially Cytotoxic Compounds. Molecules (2015) 20 1610-1625. DOI: 10.3390/molecules20011610
--- title: 'Genetic Predisposition to a Higher Whole Body Water Mass May Increase the Risk of Atrial Fibrillation: A Mendelian Randomization Study' authors: - Qi Zhu - Qiyu Chen - Ying Tian - Jing Zhang - Rui Ran - Shiyu Shu journal: Journal of Cardiovascular Development and Disease year: 2023 pmcid: PMC9966889 doi: 10.3390/jcdd10020076 license: CC BY 4.0 --- # Genetic Predisposition to a Higher Whole Body Water Mass May Increase the Risk of Atrial Fibrillation: A Mendelian Randomization Study ## Abstract Background: Observational studies have found an association between increased whole body water mass (BWM) and atrial fibrillation (AF). However, the causality has yet to be confirmed. To provide feasible protective measures on disease development, we performed Mendelian randomization (MR) design to estimate the potential causal relationship between increased BWM and AF. Methods: We implemented a two-sample MR study to assess whether increased BWM causally influences AF incidence. For exposure, 61 well-powered genetic instruments extracted from UK Biobank ($$n = 331$$,315) were used as the proxies of BWM. Summary genetic data of AF were obtained from FinnGen (Ncase = 22,068; Ncontrol = 116,926). Inverse-variance weighted (IVW), MR-Egger and weighted median methods were selected to infer causality, complemented with a series of sensitivity analyses. MR-Pleiotropy Residual Sum and Outlier (MR-PRESSO) and Radial MR were employed to identify outliers. Furthermore, risk factor analyses were performed to investigate the potential mechanisms between increased BWM and AF. Results: Genetic predisposition to increased BWM was demonstrated to be significantly associated with AF in the IVW model (OR = 2.23; $95\%$ CI = 1.47–3.09; $$p \leq 1.60$$ × 10−7), and the result was consistent in other MR approaches. There was no heterogeneity or pleiotropy detected in sensitivity analysis. MR-PRESSO identified no outliers with potential pleiotropy after excluding outliers by Radial MR. Furthermore, our risk factor analyses supported a positive causal effect of genetic predicted increased BWM on edematous diseases. Conclusions: MR estimates showed that a higher BWM could increase the risk of AF. Pathological edema is an important intermediate link mediating this causal relationship. ## 1. Introduction Atrial fibrillation (AF), a prevalent arrhythmia that poses a significant threat to the global public health burden, has increased the risk of cardiogenic embolic stroke [1], dementia [2] and death [3] due to irregular discharge activity. Unfortunately, with aging populations, the number of people with AF is predicted to reach 17.9 million in Europe by 2060, raising public health concerns [4]. Various risk factors have been found in previous studies: age, obesity, hypertension, diabetes, coronary heart disease, rheumatism and heart valve disease [5,6]. However, the prevention of AF is still insufficient [7]. Considering the severe threat that AF represents to human health, it is particularly imperative to uncover other modifiable risk factors contributing to the development of AF, which may facilitate effective protective strategies and reduce disabling complications of the disorder [8]. An increasing number of researches have demonstrated that owing to the convenience of measurements, anthropometric indicators (such as fat mass, fat-free mass, waist circumference, and waist-to-hip ratio) showed a tendency to be valuable predictors of AF [9,10,11]. Herein, we suspect whether other alternations of body composition could be an indicator of AF. Body water mass (BWM) can be easily obtained through bio-impedance, which correlates with several health problems (such as sleep apnea) owing to the increased amount [12]. However, there are lacking studies evaluating whether increased BWM fuels the risk of AF episodes. Recently some researches have validated that overhydration status might be responsible for the risk of AF, indicating a positive impact of increased BWM on AF [13,14]. In contrast, the decreased BWM appears to be an independent predictor for the incidence of AF in another observational study [15]. The inconsistent results of these studies might be due to the inability to avoid the influence of high correlation among different body composition characteristics in these observational studies. Additionally, owing to the interference of residual confounding and potential reverse causality, the causal relationship between increased BWM and AF remains unclear. Therefore, this study aims to decipher the causal effect of increased BWM on the risk of AF using a method free from the abovementioned drawbacks. Mendelian randomization (MR) is a method of assessing potential causality between exposures and outcomes of interest. This approach utilizes single nucleotide polymorphisms (SNPs) as the confusion-free proxies for exposure, which effectively minimizes confounding bias and reverse causality in conventional designs [16]. Since this genetic variation is randomly allocated at fertilization before the onset of disease, such analysis mimics a randomized controlled trial (RCT) with randomly assigned SNPs in the offspring [17]. Herein, without substantial staffing resources and time-consuming subsequent tasks, MR analysis plays a crucial role in causal inference in the lack of RCTs [18]. Using the MR method, an increasing number of risk factors [19,20] for AF and the relationship between AF and other diseases [21,22] have been reported before. However, the causal effect of increased BWM on AF has not been confirmed yet. In this context, based on publicly available genome-wide association studies (GWAS) data from a large European population, we performed a two-sample MR approach to elucidate the potential causal relationship between increased BWM and AF. Furthermore, since edematous diseases such as chronic kidney disease (CKD) [23,24], type 2 diabetes (T2D) [25,26], hypertension [27,28] and heart failure [29] are well established risk factors for AF, a similar MR method was conducted in our analysis to decipher the potential mediating mechanisms on the pathway from a higher BWM to AF. ## 2. Methods We performed a two-sample MR method to explore the potential causal relationship between increased BWM and AF. Ethical approval of our study was not required due to the absence of individual-level data. The study applied the two-sample MR package (version 0.5.6) and RadialMR package (version 1.0) of the R program (version 4.2.1) to conduct all statistical analyses. All details of datasets employed in our study are displayed in Table 1. The current MR design is shown in Figure 1. ## 2.1. Data Sources The UK Biobank (UKB) is a large-scale and detailed prospective cohort study that explores in-depth genetic and health information in European populations to promote human healthcare and provide new insights into the prevention and treatment strategies of various chronic diseases (https://www.ukbiobank.ac.uk/, accessed on 1 October 2022). Between 2006 and 2010, over 500,000 participants aged 40–69 years were recruited to complete a range of baseline measurements [30]. The genome-wide genotyping was conducted on 488,377 participants with the UK Biobank Lung Exome Variant Evaluation (UK BiLEVE) and UK Biobank Axiom arrays. Approximately 90 million variants were imputed with the IMPUTE4 program by using the Haplotype Reference Consortium (HRC) and the merged UK10K and 1000 Genomes phase 3 reference panels [31]. BWM-related GWAS data were obtained from Neale Lab (http://www.nealelab.is/, accessed on 1 October 2022). Neale Lab implemented a GWAS analysis among thousands of human characteristics in 331,315 unrelated European populations using the data from UKB. The BWM of participants was assessed by impedance technique (measured in kg). Data were accurate to 0.1 kg. Summary-level GWAS data for AF were gained from the FinnGen biobank with up to 138,994 participants (22,068 cases with AF and 116,926 control samples without AF). According to FinnGen (https://risteys.finngen.fi, accessed on 1 October 2022), the definition of AF is “A disorder characterized by an electrocardiographic finding of a supraventricular arrhythmia characterized by the replacement of consistent p waves by rapid oscillations or fibrillatory waves that vary in size, shape and timing and are accompanied by an irregular ventricular response.” This disease was determined by reviewing the medical documentation based on ICD-10 criteria. Detailed information for outcome data was provided in Supplementary Table S1. ## 2.2. Selection of Genetic Instruments A rigorous filtering procedure was carried out to control the quality of our analysis strictly. Only individuals of European descent were enrolled in the study to minimize potential confounding bias associated with descent. Ideally, well-powered instrumental variables (IVs) should fulfill three assumptions (Figure 2): (i) IVs are strongly related to exposure of interest ($p \leq 5$ × 10–8); (ii) IVs are independent of possible confounders; (iii) IVs should not be related to the outcome through any alternative pathway other than via a hypothesized one [32]. In accordance with the core assumptions of the MR study, we conducted a collection of filtering steps as following (Figure 1): (a) We selected valid SNPs closely associated with BWM as IVs from a summary-level GWAS dataset with a stringent threshold ($p \leq 5$ × 10–8, IV Assumption i, Figure 2). ( b) To guarantee the IVs chosen for BWM are independent of each other, we set strict requirements (LD threshold of r2 < 0.001) to minimize the effect of linkage disequilibrium (LD). ( c) We extracted the SNPs from the dataset of outcomes (AF). If IVs for BWM were not available in the outcome data, we then searched online (https://snipa.helmholtz-muenchen.de/snipa3/index.php, accessed on 2 October 2022) for proxy SNPs (LD threshold of r2 > 0.80). For those absent in the outcome without suitable proxies, we discarded them. ( d) We excluded the palindromic and incompatible SNPs after harmonization, a process that enables the SNPs-exposure and SNPs-outcome to correspond to the same allele. In this step, we also removed SNPs related to the outcome dataset (IV Assumption iii, Figure 2). ( e) To ensure the reliability of our study, Radial regression of MR (Radial MR) [33] was then performed to identify and exclude outliers with possible pleiotropy as an alternative approach to MR-pleiotropy residual sum and outlier (MR-PRESSO). F-statistics were performed to test the strength of chosen SNPs. Typically, F > 10 may indicate sufficient strength in causal Inference [34]. The flowchart of SNPs filtering is shown in Figure 2, and Supplementary Table S2 displays the final list of complex traits of selected SNPs. ## 2.3. MR Estimates To demonstrate the causal relationship between increased BWM and AF, we conducted three primary methods in a two-sample MR analysis, namely inverse variance weighted method (IVW), weight median, and MR-Egger regression. Different methods were implemented to settle the effect of variant heterogeneity and pleiotropy due to their different fundamental assumptions. We used the IVW model as the primary estimate, a traditional approach to combine the Wald ratio of multiple SNPs and obtain a pooled causal effect by conducting a meta-analysis. The weighted median estimator provides a robust result, requiring that at least half of the SNPs used in the analysis are valid [35]. Notably, MR-Egger regression has less statistical power than IVW but provides a broader confidence interval since this method allows all IVs for horizontal pleiotropic effect [36]. In addition, we also applied other MR estimates, including simple mode and weighted mode, as complementary tools to explore the causality. The consistent results of estimates among all MR models provide credible evidence for our analysis. $p \leq 0.05$ was set as significance. If there are any inconsistent results among different MR models, we then tighten the p value threshold [37]. The statistical power of our MR estimates was calculated with a significance value of 0.05 based on Brion et al. [ 38] (https://shiny.cnsgenomics.com/mRnd/, accessed on 2 October 2022). Generally, adequate power of $80\%$ or more was suggested. ## 2.4. Sensitivity Analysis Sensitivity analysis has been a crucial device for detecting potential heterogeneity and pleiotropy in MR studies. When IVs associated with exposure (BWM) act directly on the outcome (AF) via various pathways other than the one of interest, this indicates the presence of horizontal pleiotropy. Herein, the MR-Egger intercept test ($p \leq 0.05$ was considered to be absent of pleiotropy), MR-Pleiotropy Residual Sum and Outlier methods (MR-PRESSO), Radial MR and funnel plot were applied to explore the underlying pleiotropy in the MR estimates. MR-PRESSO causes regression in the estimates of SNP-exposure on the estimates of SNP-outcome, aiming to detect outlier SNPs and yield calibrated causality [39]. Cochran Q-test was also conducted to appraise the heterogeneity of our MR results. Specifically, there was a presence of underlying heterogeneity in our MR analysis when $p \leq 0.05$ in the Cochran Q-test [40]. In addition, we used the leave-one-out (LOO) analysis to evaluate whether any single SNP determined causality by removing SNPs by turns and recomputing the result. To check whether the core assumption of MR analysis was violated by residual confounds (IV Assumption ii, Figure 2), we also screened PhenoScanner for SNPs related to any-well accepted risk factors of AF (www.phenoscanner.medschl.cam.ac.uk, accessed on 3 October 2022) with the threshold of 1 × 10−5, including BMI, coronary artery heart disease (CHD), hyperlipidemia, smoking, alcohol use, and coffee intake. If any IVs associated with potential confounders were noted, we then discarded them manually and recomputed MR analysis to verify that the result was consistent. ## 2.5. Risk Factors To reveal the potential mediating mechanisms genetically linking increased BWM and AF, we further performed the IVW estimate to illustrate the causal relationship between increased BWM and several common edematous diseases, including CKD, T2D, heart failure, and hypertension. Genome-wide summary data for heart failure and hypertension were obtained from the FinnGen dataset. The endpoint definition and detailed characteristics of these two disorders can be found online (https://risteys.finngen.fi, accessed on 1 October 2022). GWAS data for CKD were extracted from a meta-analysis of up to 133,413 participants performed by Pattarot et al. Disease is determined under the guidelines of the National Kidney Foundation [41]. For diabetes, we obtained genetic information for combined sexes from Angli Xue et al. for causal inference [42]. Details of GWAS data are shown in Table 1. Using BWM as exposure while the potential risk factors described above as outcomes, we implemented MR analysis to uncover the intermediate pathways between increased BWM and AF. Estimates from the IVW method were employed as the main results with a statistical significance of $p \leq 0.05.$ ## 3. Results In the present study, we selected 61 valid SNPs as genetic instruments for predicted BWM after a range of vigorous filtering procedures (Supplementary Table S2). There were 6 SNPs lost in the process of analysis, but no appropriate proxy was noted. Radical MR estimator identified 316 outliers (Figure 3), and we removed them manually. The F-statistics for all chosen SNPs ranged from 12 to 91, indicating no weak IVs in this MR analysis. ## 3.1. Estimation of Causal Effect of BWM on AF The figure displays the results of the two-sample MR analysis for a credible relationship between increased BWM and AF episodes (Figure 4). The IVW model provided robust evidence that a higher BWM can significantly increase the risk of AF (OR = 2.233; $95\%$ CI = 1.654–3.016; $$p \leq 1.60$$ × 10−7). Likewise, the consistent estimates were demonstrated using MR-Egger regression (OR = 2.221; $95\%$ CI = 1.064–4.637; $$p \leq 0.04$$) and weighted median method (OR = 2.145; $95\%$ CI = 1.380–3.334; $$p \leq 6.95$$ × 10−4). Furthermore, a positive causality between genetic liability for increased BWM and AF was also detected using other approaches ($p \leq 0.05$ in both Simple mode and Weighted mode methods), strengthening the reliability of our instrumental-variable analysis. To eliminate the potential confounding effect of risk factors on causal inference, we performed SNPs search in Phenoscanner. Specifically, 15 SNPs related to confounders were detected. 13 SNPs (rs111640872, rs1286138, rs2281175, rs2815753, rs545608, rs57636386, rs66922415, rs7298201, rs73052033, rs836519, rs9826759, rs9861443 and rs9951619) genetically predicted BMI; 2 SNPs (rs28391281 and rs28929474) were associated with CHD. Similar results were obtained from repeated MR analysis (OR = 2.134; $95\%$ CI = 1.473–3.090; $$p \leq 6.03$$ × 10−5) after excluding these SNPs, and there was no evidence to show any existence of heterogeneity and pleiotropy. This suggested no violations from residual confounds in our analysis and the essential assumptions of MR study were met. We implemented a range of sensitivity analyses to validate the robustness of MR analysis, including MR-Egger intercept test, MR-PRESSO, funnel plot and LOO analysis. Cochran’s Q-test did not identify any heterogeneity in IVW method ($Q = 19.64$; $$p \leq 0.99$$). The intercept term obtained from MR-Egger intercept test showed no evidence of possible horizontal pleiotropy (intercept = 9.09 × 10−5; $$p \leq 0.99$$). MR-PRESSO and Radial MR also found no valid proof of any outlier SNPs with potential pleiotropic effects. A Scatter plot of the current analysis can be found in Supplementary Figure S1. The Funnel plot (Supplementary Figure S2) was symmetrical, suggesting that no estimate was violated. Moreover, the leave-one-out analysis (Supplementary Figure S3) showed a steady estimate when throwing a single SNP one by one, indicating that the pooled IVW result was not driven by one single SNP. Power analysis revealed that our study provided sufficient power ($100\%$) (Supplementary Table S3) to infer the impact of increased BWM on AF in the context of massive sample size [138,994] and the threshold of 0.05. ## 3.2. Risk Factor Analysis To further explore the potential intermediating factors connecting increased BWM to AF, we conducted similar MR analyses of edematous conditions that a higher BWM may influence. Table 2 shows the results of the risk factor analysis for the association between increased BWM and edematous diseases. IVW model provided credible evidence of a causal association between genetically determined increased BWM and CKD (OR = 1.432; $95\%$ CI = 1.231–1.667; $$p \leq 3.48$$ × 10−6), T2D (OR = 1.339; $95\%$ CI = 1.166–1.537; $$p \leq 3.40$$ × 10−5), heart failure (OR = 1.555; $95\%$ CI = 1.371–1.763; $$p \leq 5.95$$ × 10−12) and hypertension (OR = 1.119; $95\%$ CI = 1.002–1.249; $$p \leq 0.046$$). MR-Egger intercept analysis demonstrated no presence of potential pleiotropy for those disorders, complemented with a sufficient statistical power of all results (Supplementary Table S3). Overall, based on the above analyses, our study illustrated edematous diseases such as CKD, T2D, heart failure, and hypertension might be the pivotal mediators of the causal effect of increased BWM and a higher risk of AF. ## 4. Discussion Using large-scale GWAS data from UKB and FinnGen biobank, we applied a two-sample MR analysis to comprehensively investigate whether genetic predisposition toward BWM increases the incidence of AF. We found credible evidence to verify that the genetic liability for a higher BWM might be responsible for AF episodes. Specifically, per SD (Std.dev = 8 kg) increase in BWM was associated with a 2.233-fold higher risk of AF according to the MR method. Furthermore, to identify the potential mediating mechanisms on the pathway from increased BWM to AF, we demonstrated that genetically increased BWM was associated with several common edematous disorders, including CKD, T2D, heart failure, and hypertension according to our risk factor analyses. There is still a higher prevalence of AF in both developed and developing countries [3], and our study brings new elements to the prevention and management of AF to decrease the worldwide burden of AF and its sequelae effectively. Atrial rhythm disturbances are a common phenomenon across the world [43], with AF being the most common type of arrhythmia. With the incidence of AF increasing over the years, it is crucial to explore the various risk factors of AF. Previous studies have demonstrated the significant contribution of cytoskeletal protein variants and calcium ions in the development of AF from a molecular biological perspective [44,45,46]. And in the present research, we focus on the role of a higher BWM in the onset of AF. However, the contrary results have been observed in clinical studies exploring the relationship between BWM and AF. In 5 hospitals in Finland, Kaartinen et al. performed a prospective study of 69 individuals in the condition of end-stage renal disease. This research provided compelling evidence that fluid overload played a critical role in AF with insertable cardiac to identify AF and body composition monitor for overhydration status detection [13]. Recently, Anaszewicz et al. implemented a study of 120 patients hospitalized for AF and 240 patients clinically diagnosed with other cardiovascular disorders. A significant association between increased BWM and increased risk of AF was elucidated by applying anthropometric examination [14]. A higher BWM may lead to maladaptive alterations in cardiac structure, such as left ventricular hypertrophy and fibrosis, facilitating the development of AF [47]. Paradoxically, according to an observational study performed by Anna et al., hospitalized heart failure patients in a dehydrated state tended to be more prone to AF [15]. The contradictory findings may be explained by the inherent faults of traditional observational studies, such as confounding factors and reverse causality in causality inferences. Moreover, there is the possibility of reverse causality even in prospective studies. Compared to the comparatively impractical large-scale prospective clinical trials that require long-term observation, our MR study revealed a positive causal relationship between increased BWM and AF in a time-conserving and low-cost manner, which was consistent with the conclusions derived from some previous studies. Our results extend the literature on the issue of increased BWM in AF events in a new manner. The main finding of our research is that the relationship between a higher BWM and AF reported in previous observational studies is essentially causal. In the MR frame, the causality was validated by the consistent magnitude and direction estimates from five MR models (IVW, Weighted median, MR-Egger, Simple mode, and Weighted mode method), complemented with a series of sensitivity analyses as validation of primary MR results. Specifically, an increase of each SD in BWM predicted a 2.233-fold increase in the risk of AF based on the two-sample MR analysis, suggesting that we can achieve early prevention and intervention of disease by monitoring BWM. In addition, even excluding SNPs strongly correlated with confounding factors, the causal relationship remains, illustrating that increased BWM dose acts as an independent predictor of AF. Moreover, further risk factors analysis showed a robust association between genetic liability for a higher BWM and various diseases. This likewise provides new insights into the early identification of patients at higher risk. The present analysis provided convincing evidence for a causal effect of increased BWM on AF, and various intermediary factors might mediate this relationship. The multiple potential mechanisms might be attributed to the followings (Supplementary Figure S4). First, an increasing trend of BWM may be accompanied by chronic volume overload (CVO), leading to atrial dilatation and myocardial fibrosis [48]. And it has been reported that reentrant and focal activation, commonly considered to be triggered by microreentry and unusual automaticity in CVO-induced atrial dilatation, should be responsible for the incidence of AF. The secondary slowing conduction also acts as a key player in facilitating reentry during AF [48]. Additionally, based on the results of risk factor analysis, our study deciphered that genetic liability for increased BWM was fueled by edematous diseases (CKD, T2D, heart failure, hypertension). Hence, we assume that patients with a higher BWM are more likely to develop edematous diseases, increasing their vulnerability to AF. A recent study showed that dysfunction of pericytes or endothelial cells arising from tissue edema might be a potential trigger for the onset of AF in patients exposed to COVID-19 infection [49], which supported that pathologic edema might play an influential part in the development of AF. Furthermore, as the well-recognized risk factor for AF, obesity may have a crucial mediating impact on the causal relationship between increased BWM and AF [50]. Since obese individuals generally have a higher BWM than those with average weight [51], increased BWM tends to be tied to obesity condition. The growth in body weight is paralleled by tissue fibrosis, infiltration of fat cells into the adjacent myocardium, shortened action potential interval, and epicardial fat-associated inflammatory hallmarks [52] contributing to AF susceptibility among people. Additionally, AF episodes might be fueled by the deleterious effects on profibrillatory hemodynamics, musculoskeletal and adipose tissue caused by overweight [53]. Moreover, it is noteworthy that obesity also fosters the development of diseases (such as diabetes and hypertension) and enlarges the size of the atria, facilitating AF episodes [52]. Given the growing prevalence of AF, it is imperative to recognize the modifiable risk factors linked to AF, which may help us implement targeted interventions to decrease the burden of the disease. As BWM can be obtained conveniently from impedance, the utilization of this modifiable risk factor as a potential independent indicator of AF may be efficient guidance to disease screening and therapy. According to the results of our MR analysis, there is a causal relationship between genetic predisposition toward increased BWM and AF, supported by further sensitivity analyses and power statistics. Our research presents a new paradigm for the screening, targeted therapy strategies and sequelae prevention of AF. Specifically, from the public health perspective, more concern should be given to patients with a higher BWM, and to the fact that monitoring and management of fluid overload in this population may be a vital strategy for AF prevention. Recently a meta-analysis showed that diuretics significantly decreased the incidence of new and recurrent AF [54], which is consistent with our findings. According to Solun et al., diuretics were beneficial in preventing cardiovascular complications (such as AF and stroke) in patients with hypertension [55]. Furthermore, the risk of AF in individuals who reported a history of heart failure was effectively reduced by using diuretics [54]. All these studies indicated that effective fluid management was essential to prevent the development of AF. Further studies are required to evaluate the efficiency of different diuretics response to AF and its long-term sequelae such as stroke. During the therapeutic phase of the disease, the treatment modalities of AF should also consider a new pillar targeting fluid management, which may be an effective element of the multi-modality therapeutic intervention of AF. Moreover, a study has shown that increased physical activity positively reduces the incidence of AF, while obesity and edema status are generally accompanied by sedentary activity [52]. Consequently, lifestyle interventions, including losing weight and increased physical exercise [8] should be taken into account while setting a new management paradigm of AF for the patients with obesity or edema. To the best of our knowledge, it is the first analysis to systematically appraise the casual association between genetically determined increased BWM and AF by using the MR method. Our research has several strengths. First, the analytical method employed in this study allows for mimicking RCTs in an observational environment. Secondly, using reliable genetic instruments (F statistic > 10) as the proxy for BWM in a robust MR framework (power statistics = $100\%$), our analysis minimized the bias of reverse causality and underlying confounders commonly seen in observational studies. Finally, we deployed a range of sensitivity analyses to detect the robustness of the results and minimize the interference of horizontal pleiotropy. The present analysis revealed a significant causality between genetically increased BWM and AF, providing a new insight into the refinement of intervention strategies. ## 5. Limitation Notably, there are still some things that could be improved in our analysis. First, our MR analysis comprised only European participants to maintain a consistent genetic background. Therefore, the outcomes may not present as applicable to other races. Second, although we excluded a series of common confounders in our sensitivity analysis, some unidentified potential confounders that were not removed might still exist, limiting the capability of avoiding the confounding bias of each SNP. Third, summary-level data were employed in the current study for causal inference. Hence, we could not exclude the probability that the causal effect of increased BWM on AF might be non-linear. Finally, our research on the potential mechanisms mediating this causal effect is only an incomplete exploration. Therefore, more researches are warranted to shed light on the detailed biological mechanisms underlying this causal relationship in future research. ## 6. Conclusions In summary, our MR study demonstrates that genetic liability for a higher BWM may increase the risk of AF, in which pathological edema is a crucial potential mediator of this causal relationship. Further studies are needed to explore the possible non-linear association, evaluate the efficiency of different diuretics for a more optimized treatment option and illustrate various potential mechanisms. ## References 1. Pistoia F., Sacco S., Tiseo C., Degan D., Ornello R., Carolei A.. The Epidemiology of Atrial Fibrillation and Stroke. Cardiol. Clin. (2016) 34 255-268. DOI: 10.1016/j.ccl.2015.12.002 2. Bunch T.J.. Atrial Fibrillation and Dementia. Circulation (2020) 142 618-620. DOI: 10.1161/CIRCULATIONAHA.120.045866 3. Chugh S.S., Roth G.A., Gillum R.F., Mensah G.A.. Global burden of atrial fibrillation in developed and developing nations. Glob. Heart. (2014) 9 113-119. DOI: 10.1016/j.gheart.2014.01.004 4. Schnabel R.B., Yin X., Gona P., Larson M.G., Beiser A.S., McManus D.D., Newton-Cheh C., Lubitz S.A., Magnani J.W., Ellinor P.T.. 50 year trends in atrial fibrillation prevalence, incidence, risk factors, and mortality in the Framingham Heart Study: A cohort study. Lancet (2015) 386 154-162. DOI: 10.1016/S0140-6736(14)61774-8 5. Rahman F., Kwan G.F., Benjamin E.J.. Global epidemiology of atrial fibrillation. Nat. Rev. Cardiol. (2014) 11 639-654. DOI: 10.1038/nrcardio.2014.118 6. Kallistratos M.S., Poulimenos L.E., Manolis A.J.. Atrial fibrillation and arterial hypertension. Pharmacol. Res. (2018) 128 322-326. DOI: 10.1016/j.phrs.2017.10.007 7. Benjamin E.J., Chen P.S., Bild D.E., Mascette A.M., Albert C.M., Alonso A., Calkins H., Connolly S.J., Curtis A.B., Darbar D.. Prevention of atrial fibrillation: Report from a national heart, lung, and blood institute workshop. Circulation (2009) 119 606-618. DOI: 10.1161/CIRCULATIONAHA.108.825380 8. Fitzgerald J.L., Middeldorp M.E., Gallagher C., Sanders P.. Lifestyle Modification and Atrial Fibrillation: Critical Care for Successful Ablation. J. Clin. Med. (2022) 11. DOI: 10.3390/jcm11092660 9. Aronis K.N., Wang N., Phillips C.L., Benjamin E.J., Marcus G.M., Newman A.B., Rodondi N., Satterfield S., Harris T.B., Magnani J.W.. Associations of obesity and body fat distribution with incident atrial fibrillation in the biracial health aging and body composition cohort of older adults. Am. Heart J. (2015) 170 498-505.e2. DOI: 10.1016/j.ahj.2015.06.007 10. Tedrow U.B., Conen D., Ridker P.M., Cook N.R., Koplan B.A., Manson J.E., Buring J.E., Albert C.M.. The long- and short-term impact of elevated body mass index on the risk of new atrial fibrillation the WHS (women’s health study). J. Am. Coll. Cardiol. (2010) 55 2319-2327. DOI: 10.1016/j.jacc.2010.02.029 11. Tikkanen E., Gustafsson S., Knowles J.W., Perez M., Burgess S., Ingelsson E.. Body composition and atrial fibrillation: A Mendelian randomization study. Eur. Heart J. (2019) 40 1277-1282. DOI: 10.1093/eurheartj/ehz003 12. Huang H.C., Walters G., Talaulikar G., Figurski D., Carroll A., Hurwitz M., Karpe K., Singer R.. Sleep apnea prevalence in chronic kidney disease-association with total body water and symptoms. BMC Nephrol. (2017) 18. DOI: 10.1186/s12882-017-0544-3 13. Kaartinen K., Rautavaara J., Aitkoski A., Anttonen O., Ahvonen J., Vilpakka M., Koistinen J., Vääräniemi K., Miettinen M., Ylitalo A.. Fluid overload is an independent predictor of atrial fibrillation in end-stage renal disease: A prospective study using insertable cardiac and body composition monitors. Clin. Nephrol. (2020) 94 127-134. DOI: 10.5414/CN109997 14. Anaszewicz M., Banaś W., Wawrzeńczyk A., Budzyński J.. Body Composition in Patients with Atrial Fibrillation. Acta Cardiol. Sin. (2019) 35 484-492. DOI: 10.6515/ACS.201909_35(5).20190323A 15. Chuda A., Kaszkowiak M., Banach M., Maciejewski M., Bielecka-Dabrowa A.. The Relationship of Dehydration and Body Mass Index With the Occurrence of Atrial Fibrillation in Heart Failure Patients. Front. Cardiovasc. Med. (2021) 8 668653. DOI: 10.3389/fcvm.2021.668653 16. Haycock P.C., Burgess S., Wade K.H., Bowden J., Relton C., Davey Smith G.. Best (but oft-forgotten) practices: The design, analysis, and interpretation of Mendelian randomization studies. Am. J. Clin. Nutr. (2016) 103 965-978. DOI: 10.3945/ajcn.115.118216 17. Davey Smith G., Hemani G.. Mendelian randomization: Genetic anchors for causal inference in epidemiological studies. Hum. Mol. Genet. (2014) 23 R89-R98. DOI: 10.1093/hmg/ddu328 18. Sekula P., Del Greco M.F., Pattaro C., Köttgen A.. Mendelian Randomization as an Approach to Assess Causality Using Observational Data. J. Am. Soc. Nephrol. (2016) 27 3253-3265. DOI: 10.1681/ASN.2016010098 19. Chen W., Cai X., Yan H., Pan Y.. Causal Effect of Obstructive Sleep Apnea on Atrial Fibrillation: A Mendelian Randomization Study. J. Am. Heart. Assoc. (2021) 10 e022560. DOI: 10.1161/JAHA.121.022560 20. Larsson S.C., Lee W.H., Burgess S., Allara E.. Plasma Cortisol and Risk of Atrial Fibrillation: A Mendelian Randomization Study. J. Clin. Endocrinol. Metab. (2021) 106 e2521-e2526. DOI: 10.1210/clinem/dgab219 21. Park S., Lee S., Kim Y., Lee Y., Kang M.W., Kim K., Kim Y.C., Han S.S., Lee H., Lee J.P.. Atrial fibrillation and kidney function: A bidirectional Mendelian randomization study. Eur. Heart J. (2021) 42 2816-2823. DOI: 10.1093/eurheartj/ehab291 22. Liao L.Z., Wen X.Y., Zhang S.Z., Li W.D., Zhuang X.D.. Hypertension and Atrial Fibrillation: A Study on Epidemiology and Mendelian Randomization Causality. Front. Cardiovasc. Med. (2021) 8 644405. DOI: 10.3389/fcvm.2021.644405 23. Alonso A., Lopez F.L., Matsushita K., Loehr L.R., Agarwal S.K., Chen L.Y., Soliman E.Z., Astor B.C., Coresh J.. Chronic kidney disease is associated with the incidence of atrial fibrillation: The Atherosclerosis Risk in Communities (ARIC) study. Circulation (2011) 123 2946-2953. DOI: 10.1161/CIRCULATIONAHA.111.020982 24. Qiu H., Ji C., Liu W., Wu Y., Lu Z., Lin Q., Xue Z., Liu X., Wu H., Jiang W.. Chronic Kidney Disease Increases Atrial Fibrillation Inducibility: Involvement of Inflammation, Atrial Fibrosis, and Connexins. Front. Physiol. (2018) 9 1726. DOI: 10.3389/fphys.2018.01726 25. Echouffo-Tcheugui J.B., Shrader P., Thomas L., Gersh B.J., Kowey P.R., Mahaffey K.W., Singer D.E., Hylek E.M., Go A.S., Peterson E.D.. Care Patterns and Outcomes in Atrial Fibrillation Patients With and Without Diabetes: ORBIT-AF Registry. J. Am. Coll. Cardiol. (2017) 70 1325-1335. DOI: 10.1016/j.jacc.2017.07.755 26. Wang A., Green J.B., Halperin J.L., Piccini J.P.. Atrial Fibrillation and Diabetes Mellitus: JACC Review Topic of the Week. J. Am. Coll. Cardiol. (2019) 74 1107-1115. DOI: 10.1016/j.jacc.2019.07.020 27. Conen D., Tedrow U.B., Koplan B.A., Glynn R.J., Buring J.E., Albert C.M.. Influence of systolic and diastolic blood pressure on the risk of incident atrial fibrillation in women. Circulation (2009) 119 2146-2152. DOI: 10.1161/CIRCULATIONAHA.108.830042 28. Grundvold I., Skretteberg P.T., Liestøl K., Erikssen G., Kjeldsen S.E., Arnesen H., Erikssen J., Bodegard J.. Upper normal blood pressures predict incident atrial fibrillation in healthy middle-aged men: A 35-year follow-up study. Hypertension (2012) 59 198-204. DOI: 10.1161/HYPERTENSIONAHA.111.179713 29. Benjamin E.J., Levy D., Vaziri S.M., D’Agostino R.B., Belanger A.J., Wolf P.A.. Independent risk factors for atrial fibrillation in a population-based cohort. The Framingham Heart Study. JAMA (1994) 271 840-844. DOI: 10.1001/jama.1994.03510350050036 30. Sudlow C., Gallacher J., Allen N., Beral V., Burton P., Danesh J., Downey P., Elliott P., Green J., Landray M.. UK biobank: An open access resource for identifying the causes of a wide range of complex diseases of middle and old age. PLoS Med. (2015) 12. DOI: 10.1371/journal.pmed.1001779 31. Bycroft C., Freeman C., Petkova D., Band G., Elliott L.T., Sharp K., Motyer A., Vukcevic D., Delaneau O., O’Connell J.. The UK Biobank resource with deep phenotyping and genomic data. Nature (2018) 562 203-209. DOI: 10.1038/s41586-018-0579-z 32. Boef A.G., Dekkers O.M., le Cessie S.. Mendelian randomization studies: A review of the approaches used and the quality of reporting. Int. J. Epidemiol. (2015) 44 496-511. DOI: 10.1093/ije/dyv071 33. Bowden J., Spiller W., Del Greco M.F., Sheehan N., Thompson J., Minelli C., Davey Smith G.. Improving the visualization, interpretation and analysis of two-sample summary data Mendelian randomization via the Radial plot and Radial regression. Int. J. Epidemiol. (2018) 47 1264-1278. DOI: 10.1093/ije/dyy101 34. Burgess S., Thompson S.G.. Avoiding bias from weak instruments in Mendelian randomization studies. Int. J. Epidemiol. (2011) 40 755-764. DOI: 10.1093/ije/dyr036 35. Bowden J., Davey Smith G., Haycock P.C., Burgess S.. Consistent Estimation in Mendelian Randomization with Some Invalid Instruments Using a Weighted Median Estimator. Genet. Epidemiol. (2016) 40 304-314. DOI: 10.1002/gepi.21965 36. Bowden J., Davey Smith G., Burgess S.. Mendelian randomization with invalid instruments: Effect estimation and bias detection through Egger regression. Int. J. Epidemiol. (2015) 44 512-525. DOI: 10.1093/ije/dyv080 37. Ong J.S., MacGregor S.. Implementing MR-PRESSO and GCTA-GSMR for pleiotropy assessment in Mendelian randomization studies from a practitioner’s perspective. Genet. Epidemiol. (2019) 43 609-616. DOI: 10.1002/gepi.22207 38. Brion M.J., Shakhbazov K., Visscher P.M.. Calculating statistical power in Mendelian randomization studies. Int. J. Epidemiol. (2013) 42 1497-1501. DOI: 10.1093/ije/dyt179 39. Verbanck M., Chen C.Y., Neale B., Do R.. Detection of widespread horizontal pleiotropy in causal relationships inferred from Mendelian randomization between complex traits and diseases. Nat. Genet. (2018) 50 693-698. DOI: 10.1038/s41588-018-0099-7 40. Burgess S., Bowden J., Fall T., Ingelsson E., Thompson S.G.. Sensitivity Analyses for Robust Causal Inference from Mendelian Randomization Analyses with Multiple Genetic Variants. Epidemiology (2017) 28 30-42. DOI: 10.1097/EDE.0000000000000559 41. Pattaro C., Teumer A., Gorski M., Chu A.Y., Li M., Mijatovic V., Garnaas M., Tin A., Sorice R., Li Y.. Genetic associations at 53 loci highlight cell types and biological pathways relevant for kidney function. Nat. Commun. (2016) 7 10023. DOI: 10.1038/ncomms10023 42. Xue A., Wu Y., Zhu Z., Zhang F., Kemper K.E., Zheng Z., Yengo L., Lloyd-Jones L.R., Sidorenko J., Wu Y.. Genome-wide association analyses identify 143 risk variants and putative regulatory mechanisms for type 2 diabetes. Nat. Commun. (2018) 9 2941. DOI: 10.1038/s41467-018-04951-w 43. Henckell A.K., Gusetu G., Rosu R., Ciobanu D.M., Istratoaie S., Muresan L., Lazea C., Pop D., Cismaru G., Bârsu C.. Low Atrial Rhythm in a Large Cohort of Children from Transylvania, Romania. Life (2022) 12. DOI: 10.3390/life12111895 44. van Wijk S.W., Su W., Wijdeveld L.F.J.M., Ramos K.S., Brundel B.J.J.M.. Cytoskeletal Protein Variants Driving Atrial Fibrillation: Potential Mechanisms of Action. Cells (2022) 11. DOI: 10.3390/cells11030416 45. Reinhardt F., Beneke K., Pavlidou N.G., Conradi L., Reichenspurner H., Hove-Madsen L., Molina C.E.. Abnormal Calcium Handling in Atrial Fibrillation Is Linked to Changes in Cyclic AMP Dependent Signaling. Cells (2021) 10. DOI: 10.3390/cells10113042 46. Sánchez J., Trenor B., Saiz J., Dössel O., Loewe A.. Fibrotic Remodeling during Persistent Atrial Fibrillation: In Silico Investigation of the Role of Calcium for Human Atrial Myofibroblast Electrophysiology. Cells (2021) 10. DOI: 10.3390/cells10112852 47. Cabrera C., Brunelli S.M., Rosenbaum D., Anum E., Ramakrishnan K., Jensen D.E., Stålhammar N.O., Stefánsson B.V.. A retrospective, longitudinal study estimating the association between interdialytic weight gain and cardiovascular events and death in hemodialysis patients. BMC Nephrol. (2015) 16. DOI: 10.1186/s12882-015-0110-9 48. Hirose M., Takeishi Y., Miyamoto T., Kubota I., Laurita K.R., Chiba S.. Mechanism for atrial tachyarrhythmia in chronic volume overload-induced dilated atria. J. Cardiovasc. Electrophysiol. (2005) 16 760-769. DOI: 10.1046/j.1540-8167.2005.40331.x 49. Stone E., Kiat H., McLachlan C.S.. Atrial fibrillation in COVID-19, A review of possible mechanisms. FASEB J. (2020) 34 11347-11354. DOI: 10.1096/fj.202001613 50. Lavie C.J., Pandey A., Lau D.H., Alpert M.A., Sanders P.. Obesity and Atrial Fibrillation Prevalence, Pathogenesis, and Prognosis: Effects of Weight Loss and Exercise. J. Am. Coll. Cardiol. (2017) 70 2022-2035. DOI: 10.1016/j.jacc.2017.09.002 51. Sartorio A., Malavolti M., Agosti F., Marinone P.G., Caiti O., Battistini N., Bedogni G.. Body water distribution in severe obesity and its assessment from eight-polar bioelectrical impedance analysis. Eur. J. Clin. Nutr. (2005) 59 155-160. DOI: 10.1038/sj.ejcn.1602049 52. Chung M.K., Eckhardt L.L., Chen L.Y., Ahmed H.M., Gopinathannair R., Joglar J.A., Noseworthy P.A., Pack Q.R., Sanders P., Trulock K.M.. Lifestyle and Risk Factor Modification for Reduction of Atrial Fibrillation: A Scientific Statement From the American Heart Association. Circulation (2020) 141 e750-e772. DOI: 10.1161/CIR.0000000000000748 53. Nattel S.. Atrial Fibrillation and Body Composition: Is it Fat or Lean That Ultimately Determines the Risk?. J. Am. Coll. Cardiol. (2017) 69 2498-2501. DOI: 10.1016/j.jacc.2017.03.566 54. Neefs J., van den Berg N.W., Limpens J., Berger W.R., Boekholdt S.M., Sanders P., de Groot J.R.. Aldosterone Pathway Blockade to Prevent Atrial Fibrillation: A Systematic Review and Meta-Analysis. Int. J. Cardiol. (2017) 231 155-161. DOI: 10.1016/j.ijcard.2016.12.029 55. Solun B., Marcoviciu D., Dicker D.. Does treatment of hypertension decrease the incidence of atrial fibrillation and cardioembolic stroke?. Eur. J. Intern Med. (2009) 20 125-131. DOI: 10.1016/j.ejim.2008.07.001
--- title: The Relationship between Dietary Flavonols Intake and Metabolic Syndrome in Polish Adults authors: - Joanna Popiolek-Kalisz journal: Nutrients year: 2023 pmcid: PMC9966903 doi: 10.3390/nu15040854 license: CC BY 4.0 --- # The Relationship between Dietary Flavonols Intake and Metabolic Syndrome in Polish Adults ## Abstract Metabolic syndrome (MetS) is a cluster of metabolic disorders primarily caused by central obesity, which results in chronic inflammation leading to hypertension, diabetes and atherogenic dyslipidemia. Inflammation underlying MetS could be the target for dietary flavonols as they present antioxidative properties. The aim of this paper was to analyze the differences in habitual intake of selected flavonols (quercetin, kaempferol, isorhamnetin and myricetin) between MetS patients and healthy participants, and its relationship with MetS advancement. Ninety participants were enrolled in this study. The one-year flavonol intake was assessed with a dedicated food frequency questionnaire. The patients with MetS consumed significantly less quercetin ($$p \leq 0.01$$), kaempferol ($$p \leq 0.04$$), isorhamnetin ($p \leq 0.001$), total flavonols ($$p \leq 0.01$$), tomatoes ($$p \leq 0.001$$) and wine ($$p \leq 0.01$$) daily. Further analysis revealed a moderate inverse correlation between quercetin ($$p \leq 0.001$$), kaempferol ($$p \leq 0.01$$), isorhamnetin ($p \leq 0.001$), total flavonols ($$p \leq 0.001$$) and tomato consumption ($$p \leq 0.004$$) and MetS stage. The analysis of laboratory parameters showed that dietary intake of flavonols was not correlated with lipid profile, glucose level or renal function. On the basis of this observation, a potential protective effect of dietary flavonols, mainly from tomatoes, against MetS could be suggested. However, when referring to MetS components, flavonols probably mainly impact central obesity and blood pressure, without a significant impact on conventional lipid-profile parameters and glucose level. ## 1. Introduction Metabolic syndrome (MetS) is becoming a global problem that is associated with the progressive change in the lifestyle of modern societies. MetS is a cluster of metabolic disorders mainly caused by central obesity [1]. They include insulin resistance, atherogenic dyslipidemia, central obesity and elevated blood pressure. MetS has become an important healthcare problem. It is estimated that MetS prevalence in the US reached $35\%$ in 2012, which was a $10\%$ increase compared to a $25\%$ prevalence in 1994 [2]. The situation in Europe is better, but still very serious, as MetS prevalence was about $24\%$ in 2015 [3]. The problem is very important, because MetS leads to an increased risk of diabetes (if not already present) and cardiovascular disease (CVD). CVD and diabetes are the leading causes of death globally [4]. MetS pathogenesis is very complex; however, it generally initiates from excess visceral fat tissue, which via adipokine production leads to insulin resistance, chronic low-grade inflammation and neurohormonal activation. Secondarily, these processes result in endothelium dysfunction and glucose and lipid metabolism abnormalities, among other problems. The fat tissue distribution is an important factor as visceral adipose tissue is more vulnerable to macrophage infiltration and thus to inflammation development [5] and free fatty acid release [6]. According to the new IDF definition, MetS diagnosis requires the presence of central obesity, which is defined as a waist circumference of 94 cm or more in men or 80 cm or more in women in Poland, or a BMI of 30 kg/m2 or higher, accompanied by two out of four additional criteria: [1] elevated blood triglycerides (TG) 150 mg/dL or greater (or hypertriglyceridemia treatment), [2] reduced high-density lipoprotein cholesterol (HDL) less than 40 mg/dL in men or less than 50 mg/dL in women (or hypercholesterolemia treatment), [3] elevated fasting glucose of 100 mg/dL or greater (or diabetes treatment), [4] blood pressure values of systolic (SBP) 130 mmHg or higher and/or diastolic (DBP) 85 mmHg or higher (or hypertension treatment) [7]. Although insulin resistance is the main mechanism of MetS, it is not directly captured in the diagnostic criteria as insulin level measurement (essential for this purpose) is cumbersome in everyday clinical practice. However, insulin resistance is represented by the waist circumference criterion, because they correlate [8]. Excessive fat tissue located in the abdomen area, which is the foundation of MetS, is the result of interactions among lifestyle factors and genetic predispositions [9]. The main lifestyle contributors include improper dietary patterns (mainly high caloric intake) and lack of physical activity. In connection with the mechanisms described above, MetS prevention and treatment are based on lifestyle changes, including dietary modifications. They involve, e.g., increased consumption of fruit and vegetables, which are good sources of antioxidants and fiber. There have not been established any recommendations for antioxidative agent supplementation or details of their dietary intake in terms of MetS prevention [10]. Flavonols are a group of flavonoids distinguished by their chemical structure, including a 3-hydroxyflavone backbone, which are known for their antioxidative properties. They are present mainly in fruits, vegetables, and tea. The most important flavonols are quercetin and kaempferol, followed by less-prevalent compounds such as myricetin, isorhamnetin, morin, galangin, fisetin, kaempferide, azaleatin, natsudaidain, pachypodol and rhamnazin. The major contributors to everyday dietary flavonol intake are onions, tea, and apples [11]. Other flavonol-rich products are kale, lettuce, tomatoes, broccoli, grapes, berries, and red wine [12,13]. Studies of CVD patients showed that the main dietary contributors to flavonol intake are blueberries and apples among the fruits; onions and tomatoes among the vegetables; and tea (black and green), coffee and wine among the beverages [14]. Low-grade inflammation underlying MetS could be the target for flavonols as they present antioxidative properties. Existing findings suggest a positive impact of selected flavonols, mainly quercetin, on MetS single components; however, there are not many studies investigating the impact of single flavonols intake on MetS as a set of disorders. The results of the already-conducted studies are not consistent. What is more, most of the studies focus only on quercetin, while there is a shortage of studies investigating other flavonols’ impact on metabolic parameters. There has not been any study conducted summarizing all the components of MetS together. The objective of this study was to analyze the differences in habitual dietary intake of flavonols (quercetin, kaempferol, isorhamnetin and myricetin) between patients with and without MetS and the relationship between their habitual intake and MetS advancement. Additionally, the relationship between dietary flavonols intake, their main dietary sources consumption and metabolic parameters (glucose level and lipid profile) in MetS patients was also investigated. The main research question was: Is there a relationship between flavonols intake and MetS? The hypotheses were: (i) Does flavonols intake differ between MetS and healthy participants? ( ii) Is flavonols intake related to MetS advancement? ( iii) Is flavonols intake related to the main laboratory parameters in MetS patients? This is the first study that addresses the intake of all four main flavonols, not only quercetin, on MetS. Moreover, in the course of this study, the relationship between dietary flavonols intake and MetS as a holistic set of disorders was analyzed. As explained above, the components of MetS are usually related to each other; thus, analysis of only single ones separated from others could potentially not provide all the information. The foregoing studies were focused on single components of MetS, but there has not been any analysis conducted which acknowledged MetS as a set of disorders together. This is the first study to present such an attitude. ## 2. Materials and Methods Ninety participants (53 women and 37 men) were enrolled in this study between April and December 2022. The inclusion criteria were: [1] age 18–85 years, [2] written consent and [3] mental condition that enabled a one-year retrospective dietary interview. The exclusion criteria were: [1] age <18 or >85 years, [2] lack of written consent, [3] abnormal mental condition, [4] pregnancy and [5] special diet due to health reasons. The food-frequency questionnaire dedicated to one-year specific flavonol intake assessment was administered to the participants [15]. The questionnaire gathered information about the mean consumption of 140 flavonol sources during the preceding year. The full questionnaire is available as supplementary material [15]. The selected flavonols were the four most widespread in food sources according to the USDA database [16]. The suggested portions of the products were based on typical servings in everyday life (e.g., one piece, a glass) and described for the participants by a suggested serving (e.g., a piece, a glass) and a weight in grams. The participants were asked to provide a frequency of selected product consumption (never or almost never, once a month, few times a month with a number of times per month given by the responder, once a week, few times a week with a number of times per week given by the responder, once a day, few times every day with a number of times per day given by the responder). The amounts of quercetin, kaempferol, isorhamnetin, and myricetin in each product were based on the data available in the USDA database [16]. On the basis of this information, the mean daily consumption of each product and flavonol was calculated for each participant. Total flavonol intake was calculated by adding the values of quercetin, kaempferol, isorhamnetin, and myricetin. The daily intake of each compound was expressed relative to body mass. The patient’s weight was measured with 0.05 kg accuracy by a trained professional. The patient was permitted to wear only underwear for this measurement. The information about the mean daily intake of flavonol sources was also derived from the above-described questionnaire [15]. The fasting glucose, lipid profile and creatinine level were assessed in venous blood. The patients were not allowed to eat for 12 h before the test. The blood samples were gathered by a trained nurse. The samples for glucose tests were gathered with dedicated probes (EDTA + sodium fluoride) and then measured using the enzyme (hexokinase) method with a Cobas Pro (Roche Diagnostics, Mannheim, Germany) analyzer. The lipid profile test samples were gathered with dedicated probes (heparinized) and then performed using colorimetric enzyme assays with a Cobas Pro (Roche Diagnostics, Mannheim, Germany) analyzer. Creatinine samples were gathered with dedicated probes (heparinized) and then measured using a colorimetric test based on the Jaffe method with a Cobas Pro (Roche Diagnostics, Mannheim, Germany) analyzer. ## 2.1. Ethical Concerns The study was approved by the local Bioethics Committee of the Medical University of Lublin (consent no. KE-$\frac{0254}{9}$/$\frac{01}{2022}$). The study was conducted in line with the directives of the Declaration of Helsinki on Ethical Principles for Medical Research. All participants signed a written consent agreement. ## 2.2. Statistical Analysis Statistical analyses were performed with the RStudio software v. 4.2.0. The normality of the distribution of each parameter was checked with the Shapiro–Wilk test. The variables were presented as means (±SD). The MetS diagnosis was based on fulfilling the above-mentioned criteria (1 basic and 2 additional criteria). The comparison of flavonol intake between patients with and without MetS was performed with Mann–Whitney tests. A p value below 0.05 was considered significant. Then, the MetS stage was expressed as the number of fulfilled criteria of MetS. The one-way ANOVA test was used to compare the mean flavonol intakes between these groups. Then the linear correlation between flavonol intake and MetS progression was investigated with the Pearson correlation test. The cut-off points used for the correlation coefficient were the same as above: <0.20 as low, 0.20–0.49 as moderate and ≥0.50 as high correlation. A p value below 0.05 was considered significant. The two-way ANOVA was used to compare the mean flavonol intakes between the subgroups regarding BMI impact. A p value below 0.05 was considered significant. Pearson correlation was also used to analyze the linear association between selected flavonol mean daily intake and laboratory parameters, and between selected product mean daily intake and laboratory parameters. The cut-off points used for the correlation coefficient were as follows: <0.20 as low, 0.20–0.49 as moderate and ≥0.50 as high correlation. A p value below 0.05 was considered significant. ## 3.1. The Characteristics of the Participant Group The final group included 89 Europid patients (55 women and 37 men). One participant with MetS was excluded from the analysis due to blood sample hemolysis. A total of 32 participants met the MetS diagnosis criteria and 57 participants were not diagnosed with MetS. The mean age was 45.8 ± 21.9 years. The patients were non-smokers. The mean daily intakes for each flavonol were as follows: 0.63 ± 0.39 mg/kg for quercetin, 0.22 ± 0.13 mg/kg for kaempferol, 0.06 ± 0.06 for isorhamnetin, 0.08 ± 0.05 for myricetin and 0.98 ± 0.57 mg/kg for total flavonols. The mean body mass was 71.39 ± 14.49 kg and the mean BMI was 25.34 ± 4.98 kg/m2. The MetS group was older (67.29 ± 9.12 years) than healthy controls (33.56 ± 17.35 years) and included a smaller percentage of women ($46\%$) than in healthy controls ($67\%$). ## 3.2. Flavonols Intake in Participants with and without MetS The subgroup analysis between the patients diagnosed with MetS and those without MetS revealed significant differences in total flavonol ($$p \leq 0.001$$), quercetin ($$p \leq 0.01$$), kaempferol ($$p \leq 0.04$$) and isorhamnetin ($p \leq 0.001$) intakes. Among flavonol sources, patients without MetS eat significantly more tomatoes than patients with MetS (0.94 ± 0.76 portions/day vs. 0.58 ± 0.72 portions/day, $$p \leq 0.001$$) and drink more wine (0.15 ± 0.32 portions/day vs. 0.08 ± 0.21 portions/day, $$p \leq 0.01$$). The detailed results are presented in Table 1. The box-plots showing the differences in flavonols intake between patients with and without MetS diagnosis are presented in Figure 1. When participants were divided into subgroups by BMI—normal (<25kg/m2), overweight (25–29.99 kg/m2), and obese (≥30 kg/m2)—significant differences were present for all flavonols: quercetin ($p \leq 0.001$), kaempferol ($$p \leq 0.004$$), isorhamnetin ($$p \leq 0.001$$), myricetin ($$p \leq 0.01$$) and total flavonols ($p \leq 0.001$). The detailed results are presented in the box-plot in Figure 2. ## 3.3. Flavonols Intake and MetS Advancement The comparison of flavonol intake between the subgroups meeting consecutive numbers of MetS criteria showed that total flavonol ($$p \leq 0.003$$), quercetin ($$p \leq 0.005$$), kaempferol ($$p \leq 0.03$$) and isorhamnetin ($p \leq 0.001$) intakes differed significantly between these subgroups. The detailed results are presented in Table 2. The further analysis revealed linear characteristics of these relationships (R: −0.31; $95\%$ CI: −0.486 to −0.108; $$p \leq 0.003$$ for total flavonols; R: −0.30; $95\%$ CI: −0.476 to −0.095; $$p \leq 0.001$$ for quercetin; R: −0.23; $95\%$ CI: −0.421 to −0.026; $$p \leq 0.01$$ for kaempferol; R: −0.40; $95\%$ CI: −0.559 to −0.206; $p \leq 0.001$ for isorhamnetin). Among flavonol sources, MetS advancement was inversely correlated with tomato consumption (R: −0.30, $95\%$ CI: −0.483 to −0.103; $$p \leq 0.004$$). Detailed results are presented in Table 3. The analysis regarding flavonol intakes and BMI impact showed significant differences between subgroups in terms of each flavonol: quercetin, kaempferol, isorhamnetin, myricetin and total flavonols ($p \leq 0.001$ for each). When participants were divided into subgroups by BMI—normal (< 25kg/m2), overweight (25–29.99 kg/m2) and obese (≥ 30 kg/m2)—the correlation was present only in the overweight subgroup for isorhamnetin intake (R: −0.38; $95\%$ CI: −0.623 to −0.012; $$p \leq 0.04$$) and tomato consumption (R:−0.47; $95\%$ CI: −0.712 to −0.122; $$p \leq 0.01$$). ## 3.4. The Analysis of Laboratory Parameters in MetS Patients The analysis of the relationship between flavonol intake and laboratory metabolic parameters (glucose, TC, HDL, LDL, TG, creatinine) revealed that total and selected flavonol intake was not correlated with any of them. The detailed results are presented in Table 4. The subgroup analysis did not show any significant correlation in men and women or in BMI-stratified subgroups. Mean kaempferol intake was highly inversely correlated with glucose level among men; however, this relationship was still not significant (R: −0.58; $95\%$ CI: −0.876 to 0.027; $$p \leq 0.06$$). ## 3.5. The Analysis of the Flavonol Source Consumption in MetS Patients The results showed that onion and tomato were the main contributors to flavonol intake among vegetables, blueberries and apples among fruit, tea and coffee among non-alcoholic beverages and wine as an alcoholic drink. The analysis of the relationship between these flavonol sources’ intake and laboratory metabolic parameters (glucose, TC, HDL, LDL, TG, creatinine) did not reveal any significant correlation. The detailed results are presented in Table 5. The subgroup analysis in MetS patients regarding BMI (overweight and obese) showed a significant correlation in the overweight subgroup for tomato consumption and TC (R = −0.68; $95\%$ CI: −897 to −0.214; $$p \leq 0.01$$), TG (R = −0.64; $95\%$ CI: −0.882 to −0.144; $$p \leq 0.02$$) and LDL levels (R = −0.58; $95\%$ CI: −0.858 to −0.047; $$p \leq 0.04$$). ## 4. Discussion MetS is the set of disorders such as central obesity, elevated blood pressure, elevated glucose level and atherogenic dyslipidemia which is defined as elevated TG and decreased HDL levels. All of them are the main risk factors of CVD. This is why MetS prevention and treatment are crucial for CVD prevention as well. This study is the first to analyze MetS as a set of disorders, not only separate conditions, in terms of dietary flavonol intake. The analysis of the flavonols intake between patients with and without diagnosed MetS revealed significant differences in quercetin, kaempferol, isorhamnetin and total flavonol consumption. The MetS patients were characterized by lower quercetin, kaempferol, isorhamnetin and total flavonol intakes compared to healthy ones. Thus, a preliminary conclusion can be provided that dietary flavonol consumption could play a potentially protective role against MetS development. Then, to analyze the relationship between flavonols intake and MetS progression, a comparison between the subgroups meeting the consecutive numbers of MetS criteria was performed. There are no defined stages of MetS as its criteria are complex, which is why this simplified form of MetS advancement was used for the purpose of this study. This detailed investigation revealed significant differences in quercetin, kaempferol, isorhamnetin and total flavonol intake also between the subgroups meeting different numbers of MetS criteria. Further analysis showed the linear nature of this relationship, as quercetin, kaempferol, isorhamnetin and total flavonol habitual intake were moderately inversely correlated with MetS stage. The details of this trend were the subject of further investigation in the course of this study. Among flavonols sources, there were significant differences in tomato and wine consumption between patients with and without MetS. Tomatoes are good sources of flavonols, mainly kaempferol and quercetin (from 15 μg/mL in juice to 70 μg/g in puree) [17,18]. It is worth noting that they also contain other bioactive ingredients such as lycopene, which is also proven to present a protective potential against MetS [19]. Apart from the mentioned study focusing on lycopene from tomatoes, no other human study is available that links tomato consumption with MetS incidence. It is also worth noting that antioxidant bioavailability in different foods could be impacted by cultivation practice, meal preparation techniques or storage. Most of the phenolic compounds are located in tomato skin [17]. Quercetin and kaempferol content in tomatoes decreases in the course of peeling, dicing and heat treatment [20,21]. On the other hand, studies show that lycopene bioavailability from tomatoes increases after thermal treatment [22]. Storage of tomato-derived industry-processed products did not change the content of quercetin [23]. In the present study, wine consumption is inversely associated with MetS occurrence, which is in line with other observational studies regarding moderate wine consumption [24,25]. The mechanisms of this phenomenon could include wine polyphenols’ impact on the gut microbiota [26]. On the other hand, the direct positive impact of wine polyphenols on MetS laboratory components was not confirmed [27], which is also in line with the present study. Nonetheless, it is worth noting that wine consumption could be part of a more complex lifestyle pattern that could play an additional role in MetS prevention [28]. What is more, heat treatment increases flavonol content in wine pomace [29]. As already mentioned, MetS is a set of disorders. Referring to BMI values, significant differences were observed regarding all flavonol intake between participants with normal body mass, overweight and obese. Central obesity prevalence, which is the primary cause of MetS, has already been linked to lower flavonol consumption [30]. Excessive fat adipose tissue presence, especially in the abdominal region, leads to chronic inflammation, which is one of the causes of insulin resistance and endothelium dysfunction. They result in other MetS components: hypertension, diabetes and atherogenic dyslipidemia development. A positive impact of quercetin supplementation on blood pressure levels has been observed [31,32]. As one of the MetS criteria is elevated blood pressure level, this group of patients could possibly benefit from quercetin supplementation [33,34]. The relationship between habitual flavonol consumption and blood pressure in men was also confirmed in observational studies [14]. This is why these parameters were not the subject of investigation in the course of this study. In terms of glucose metabolism, this study showed that flavonol consumption was not correlated with the fasting glucose level. On the other hand, interventional studies in diabetic patients showed that supplementation of quercetin and myricetin could decrease glucose levels [35,36,37]. The difference might have been caused by the fact that none of these studies was conducted in MetS patients, i.e., with other comorbidities apart from diabetes. Mean kaempferol intake was highly inversely correlated with glucose level among men; however, this relationship was still not significant. There is not any data available from human studies that investigated this particular topic. Other parameters related to glucose metabolism, such as insulin or glycated hemoglobin, were not analyzed because they are not acknowledged in the MetS criteria at the moment. Flavonol intake was not correlated with lipid parameters as well. This observation is in line with results by other authors, as among MetS patients, quercetin supplementation did not alter total cholesterol (TC), triglycerides and the LDL/HDL, TC/HDL and triglycerides/HDL ratio [33]. Nonetheless, in the same study, it significantly decreased plasma concentrations of oxidized LDL, which is known for its atherogenic influence [33]. In the present study, only basic lipid parameters available in everyday clinical practice were measured; however, more detailed observation regarding other parameters such as oxidized LDL would be helpful. In interventional studies, the combination of quercetin and kaempferol also did not improve the lipid profile in healthy men; however, the applied doses of 20.2 mg and 3.4 mg, respectively, were relatively low [38] compared even to habitual intake, which was 40.0 mg/day for quercetin and 14.0 mg/day for kaempferol in the present study. What is more, the participants of this study were non-smokers, while the significant positive impact of quercetin supplementation in other studies was observed only in smokers [35]. Renal function is an essential aspect of the clinical assessment of MetS patients; however, it is not acknowledged as a MetS criterion. The present analysis showed that flavonol consumption was also not correlated with creatinine levels. The results from animal model studies suggest that quercetin has a protective role against diabetic nephropathy, as well as in human clinical trials [39,40]. The differences might be caused by the quercetin dosage, as in the interventional environment high daily doses, such as 500 mg and 1000 mg, were applied, while the observed mean habitual intake was 41.3 mg/day. Among other flavonol sources, tomato intake was moderately inversely correlated with TC and LDL levels; however, this relationship was not significant. Nonetheless, tomato consumption was highly inversely correlated with TC, TG and LDL levels in the overweight subgroup. This could suggest a potential beneficial role of tomato consumption in overweight MetS patients. This observation is in line with the studies suggesting a beneficial role of tomato juice supplementation in MetS [41]. In the mentioned study, tomato juice supplementation did not significantly decrease TC levels; however, it significantly decreased LDL levels and elevated HDL levels. The differences might have been caused by different dosages; however, it is unclear as the authors described that the patients were told to consume tomato juice four times per week, but they did not detail the portion capacity to evaluate the mean daily flavonoid intake. This study has its limitations. The flavonol intake was based on questionnaire responses, so it shares all the limitations of this type of study. The cross-sectional type of the study does not allow the establishment of a cause/result relationship, so these observations should be confirmed in an interventional study. Even though the questionnaire used was proved suitable for this type of study and the study aimed to investigate long-term habits, additional blood tests for flavonol serum levels could support these observations. Moreover, the questionnaire was based on the USDA database; thus, it could not analyze the year-long consumption of each possible variety of every available fruit or vegetable. What is more, laboratory data were available only in MetS patients, so future laboratory analysis including healthy participants would be helpful. The healthy controls were volunteers, and women were more interested in study participation; thus the majority of the group was women. Moreover, healthy participants showed more interest in the study participation; thus, the sizes of the MetS and control groups were not equal, which disabled age or sex matching. Nonetheless, this is the first study to analyze the relationship between dietary flavonols as single compounds and the complexity of MetS. This approach is highly valuable, as it provides a direction for future interventional studies regarding the preventive role of selected flavonols in MetS. This study aimed to analyze general clinical trends; however, a detailed investigation of the mechanisms responsible for these observations is needed, as it might be beyond the antioxidative properties. ## 5. Conclusions This study showed that healthy participants consume more flavonols than those with MetS. Moreover, higher habitual flavonol intake was inversely associated with MetS progression. On the basis of this observation, a potential protective effect of dietary flavonol intake against MetS could be suggested. However, when referring to MetS components, habitual intake of selected flavonols was related mainly to central obesity and blood pressure, without a significant correlation with conventional lipid profile parameters or glucose levels. Further investigation regarding additional parameters of lipid and glucose metabolism could provide additional information upon this topic. Among flavonols sources, the patients without MetS ate significantly more tomatoes than MetS patients, which consumption was also inversely correlated with MetS stage. This could suggest a potential role of tomato consumption in MetS prevention. ## References 1. Matsuzawa Y., Funahashi T., Nakamura T.. The Concept of Metabolic Syndrome: Contribution of Visceral Fat Accumulation and Its Molecular Mechanism. J. Atheroscler. Thromb. (2011) 18 629. DOI: 10.5551/jat.7922 2. Moore J.X., Chaudhary N., Akinyemiju T.. Metabolic Syndrome Prevalence by Race/Ethnicity and Sex in the United States, National Health and Nutrition Examination Survey, 1988–2012. Prev. Chronic Dis. (2017) 14 160287. DOI: 10.5888/pcd14.160287 3. Scuteri A., Laurent S., Cucca F., Cockcroft J., Cunha P.G., Mañas L.R., Raso F.U.M., Muiesan M.L., Ryliškyte L., Rietzschel E.. Metabolic Syndrome across Europe: Different Clusters of Risk Factors. Eur. J. Prev. Cardiol. (2015) 22 486-491. DOI: 10.1177/2047487314525529 4. Roth G.A., Abate D., Abate K.H., Abay S.M., Abbafati C., Abbasi N., Abbastabar H., Abd-Allah F., Abdela J., Abdelalim A.. Global, Regional, and National Age-Sex-Specific Mortality for 282 Causes of Death in 195 Countries and Territories, 1980–2017: A Systematic Analysis for the Global Burden of Disease Study 2017. Lancet (2018) 392 1736-1788. DOI: 10.1016/S0140-6736(18)32203-7 5. Blaszczak A.M., Jalilvand A., Liu J., Wright V.P., Suzo A., Needleman B., Noria S., Lafuse W., Hsueh W.A., Bradley D.. Human Visceral Adipose Tissue Macrophages Are Not Adequately Defined by Standard Methods of Characterization. J. Diabetes Res. (2019) 2019 8124563. DOI: 10.1155/2019/8124563 6. Patel P., Abate N.. Body Fat Distribution and Insulin Resistance. Nutrients (2013) 5 2019-2027. DOI: 10.3390/nu5062019 7. The IDF Consensus Worldwide Definition of the Metabolic Syndrome. Obe. Metab. (2005) 2 47-49. DOI: 10.14341/2071-8713-4854 8. Ross R., Neeland I.J., Yamashita S., Shai I., Seidell J., Magni P., Santos R.D., Arsenault B., Cuevas A., Hu F.B.. Waist Circumference as a Vital Sign in Clinical Practice: A Consensus Statement from the IAS and ICCR Working Group on Visceral Obesity. Nat. Rev. Endocrinol. (2020) 16 177-189. DOI: 10.1038/s41574-019-0310-7 9. Fathi Dizaji B.. The Investigations of Genetic Determinants of the Metabolic Syndrome. Diabetes Metab. Syndr. Clin. Res. Rev. (2018) 12 783-789. DOI: 10.1016/j.dsx.2018.04.009 10. Visseren F.L.J., Mach F., Smulders Y.M., Carballo D., Koskinas K.C., Bäck M., Benetos A., Biffi A., Boavida J.-M., Capodanno D.. 2021 ESC Guidelines on Cardiovascular Disease Prevention in Clinical Practice. Eur. Heart J. (2021) 42 3227-3337. DOI: 10.1093/eurheartj/ehab484 11. Cassidy A., O’Reilly É.J., Kay C., Sampson L., Franz M., Forman J., Curhan G., Rimm E.B.. Habitual Intake of Flavonoid Subclasses and Incident Hypertension in Adults. Am. J. Clin. Nutr. (2011) 93 338-347. DOI: 10.3945/ajcn.110.006783 12. Zamora-Ros R., Knaze V., Luján-Barroso L., Slimani N., Romieu I., Fedirko V., Santucci de Magistris M., Ericson U., Amiano P., Trichopoulou A.. Estimated Dietary Intakes of Flavonols, Flavanones and Flavones in the European Prospective Investigation into Cancer and Nutrition (EPIC) 24 Hour Dietary Recall Cohort. Br. J. Nutr. (2011) 106 1915-1925. DOI: 10.1017/S000711451100239X 13. Panche A.N., Diwan A.D., Chandra S.R.. Flavonoids: An Overview. J. Nutr. Sci. (2016) 5 e47. DOI: 10.1017/jns.2016.41 14. Popiolek-Kalisz J., Blaszczak P., Fornal E.. Dietary Isorhamnetin Intake Is Associated with Lower Blood Pressure in Coronary Artery Disease Patients. Nutrients (2022) 14. DOI: 10.3390/nu14214586 15. Popiolek-Kalisz J., Fornal E.. Dietary Isorhamnetin Intake Is Inversely Associated with Coronary Artery Disease Occurrence in Polish Adults. Int. J. Environ. Res. Public Health (2022) 19. DOI: 10.3390/ijerph191912546 16. Bhagwat S., Haytowitz D.B., Holden J.M.. USDA Database for the Flavonoid Content of Selected Foods: Release 3 (2011) 17. Stewart A.J., Bozonnet S., Mullen W., Jenkins G.I., Lean M.E.J., Crozier A.. Occurrence of Flavonols in Tomatoes and Tomato-Based Products. J. Agric. Food Chem. (2000) 48 2663-2669. DOI: 10.1021/jf000070p 18. Slimestad R., Fossen T., Verheul M.J.. The Flavonoids of Tomatoes. J. Agric. Food Chem. (2008) 56 2436-2441. DOI: 10.1021/jf073434n 19. Li N., Wu X., Zhuang W., Xia L., Chen Y., Wu C., Rao Z., Du L., Zhao R., Yi M.. Tomato and Lycopene and Multiple Health Outcomes: Umbrella Review. Food Chem. (2021) 343 128396. DOI: 10.1016/j.foodchem.2020.128396 20. Vallverdú-Queralt A., Medina-Remón A., Andres-Lacueva C., Lamuela-Raventos R.M.. Changes in Phenolic Profile and Antioxidant Activity during Production of Diced Tomatoes. Food Chem. (2011) 126 1700-1707. DOI: 10.1016/j.foodchem.2010.12.061 21. Kamiloglu S., Demirci M., Selen S., Toydemir G., Boyacioglu D., Capanoglu E.. Home Processing of Tomatoes (. J. Sci. Food Agric. (2014) 94 2225-2233. DOI: 10.1002/jsfa.6546 22. Agarwal A., Shen H., Agarwal S., Rao A.V.. Lycopene Content of Tomato Products: Its Stability, Bioavailability and in Vivo Antioxidant Properties. J. Med. Food (2001) 4 9-15. DOI: 10.1089/10966200152053668 23. Koh E., Charoenprasert S., Mitchell A.E.. Effects of Industrial Tomato Paste Processing on Ascorbic Acid, Flavonoids and Carotenoids and Their Stability over One-Year Storage. J. Sci. Food Agric. (2012) 92 23-28. DOI: 10.1002/jsfa.4580 24. Janssen I., Powell L.H., Wildman R.P.. Moderate Wine Consumption Inhibits the Development of the Metabolic Syndrome: The Study of Women’s Health Across the Nation (SWAN). J. Wine Res. (2011) 22 113-117. DOI: 10.1080/09571264.2011.603217 25. Tresserra-Rimbau A., Medina-Remón A., Lamuela-Raventós R.M., Bulló M., Salas-Salvadó J., Corella D., Fitó M., Gea A., Gómez-Gracia E., Lapetra J.. Moderate Red Wine Consumption Is Associated with a Lower Prevalence of the Metabolic Syndrome in the PREDIMED Population. Br. J. Nutr. (2015) 113 S121-S130. DOI: 10.1017/S0007114514003262 26. Moreno-Indias I., Sánchez-Alcoholado L., Pérez-Martínez P., Andrés-Lacueva C., Cardona F., Tinahones F., Queipo-Ortuño M.I.. Red Wine Polyphenols Modulate Fecal Microbiota and Reduce Markers of the Metabolic Syndrome in Obese Patients. Food Funct. (2016) 7 1775-1787. DOI: 10.1039/C5FO00886G 27. Woerdeman J., Van Poelgeest E., Ket J.C.F., Eringa E.C., Serné E.H., Smulders Y.M.. Do Grape Polyphenols Improve Metabolic Syndrome Components? A Systematic Review. Eur. J. Clin. Nutr. (2017) 71 1381-1392. DOI: 10.1038/ejcn.2016.227 28. Rosell M., de Faire U., Hellénius M.L.. Low Prevalence of the Metabolic Syndrome in Wine Drinkers—Is It the Alcohol Beverage or the Lifestyle?. Eur. J. Clin. Nutr. (2003) 57 227-234. DOI: 10.1038/sj.ejcn.1601548 29. Del Pino-García R., González-SanJosé M.L., Rivero-Pérez M.D., García-Lomillo J., Muñiz P.. The Effects of Heat Treatment on the Phenolic Composition and Antioxidant Capacity of Red Wine Pomace Seasonings. Food Chem. (2017) 221 1723-1732. DOI: 10.1016/j.foodchem.2016.10.113 30. Popiolek-Kalisz J.. The Impact of Dietary Flavonols on Central Obesity Parameters in Polish Adults. Nutrients (2022) 14. DOI: 10.3390/nu14235051 31. Edwards R.L., Lyon T., Litwin S.E., Rabovsky A., Symons J.D., Jalili T.. Quercetin Reduces Blood Pressure in Hypertensive Subjects. J. Nutr. (2007) 137 2405-2411. DOI: 10.1093/jn/137.11.2405 32. Popiolek-Kalisz J., Fornal E.. The Effects of Quercetin Supplementation on Blood Pressure—Meta-Analysis. Curr. Probl. Cardiol. (2022) 47 101350. DOI: 10.1016/j.cpcardiol.2022.101350 33. Egert S., Bosy-Westphal A., Seiberl J., Kürbitz C., Settler U., Plachta-Danielzik S., Wagner A.E., Frank J., Schrezenmeir J., Rimbach G.. Quercetin Reduces Systolic Blood Pressure and Plasma Oxidised Low-Density Lipoprotein Concentrations in Overweight Subjects with a High-Cardiovascular Disease Risk Phenotype: A Double-Blinded, Placebo-Controlled Cross-over Study. Br. J. Nutr. (2009) 102 1065-1074. DOI: 10.1017/S0007114509359127 34. Popiolek-Kalisz J., Fornal E.. The Impact of Flavonols on Cardiovascular Risk. Nutrients (2022) 14. DOI: 10.3390/nu14091973 35. Lee K.-H., Park E., Lee H.-J., Kim M.-O., Cha Y.-J., Kim J.-M., Lee H., Shin M.-J.. Effects of Daily Quercetin-Rich Supplementation on Cardiometabolic Risks in Male Smokers. Nutr. Res. Pract. (2011) 5 28. DOI: 10.4162/nrp.2011.5.1.28 36. Abidov M., Ramazanov A., Jimenez Del Rio M., Chkhikvishvili I.. Effect of Blueberin on Fasting Glucose, C-Reactive Protein and Plasma Aminotransferases, in Female Volunteers with Diabetes Type 2: Double-Blind, Placebo Controlled Clinical Study. Georgian Med. News (2006) 141 66-72 37. Ahrens M.J., Thompson D.L.. Effect of Emulin on Blood Glucose in Type 2 Diabetics. J. Med. Food (2013) 16 211-215. DOI: 10.1089/jmf.2012.0069 38. Suomela J.P., Ahotupa M., Yang B., Vasankari T., Kallio H.. Absorption of Flavonols Derived from Sea Buckthorn (. J. Agric. Food Chem. (2006) 54 7364-7369. DOI: 10.1021/jf061889r 39. Hu T., Yue J., Tang Q., Cheng K.-W., Chen F., Peng M., Zhou Q., Wang M.. The Effect of Quercetin on Diabetic Nephropathy (DN): A Systematic Review and Meta-Analysis of Animal Studies. Food Funct. (2022) 13 4789-4803. DOI: 10.1039/D1FO03958J 40. Knab A.M., Shanely R.A., Henson D.A., Jin F., Heinz S.A., Austin M.D., Nieman D.C.. Influence of Quercetin Supplementation on Disease Risk Factors in Community-Dwelling Adults. J. Am. Diet. Assoc. (2011) 111 542-549. DOI: 10.1016/j.jada.2011.01.013 41. Tsitsimpikou C., Tsarouhas K., Kioukia-Fougia N., Skondra C., Fragkiadaki P., Papalexis P., Stamatopoulos P., Kaplanis I., Hayes A.W., Tsatsakis A.. Dietary Supplementation with Tomato-Juice in Patients with Metabolic Syndrome: A Suggestion to Alleviate Detrimental Clinical Factors. Food Chem. Toxicol. (2014) 74 9-13. DOI: 10.1016/j.fct.2014.08.014
--- title: Sustained Release of Antifungal and Antibacterial Agents from Novel Hybrid Degradable Nanofibers for the Treatment of Polymicrobial Osteomyelitis authors: - Yung-Heng Hsu - Yi-Hsun Yu - Ying-Chao Chou - Chia-Jung Lu - Yu-Ting Lin - Steve Wen-Neng Ueng - Shih-Jung Liu journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966905 doi: 10.3390/ijms24043254 license: CC BY 4.0 --- # Sustained Release of Antifungal and Antibacterial Agents from Novel Hybrid Degradable Nanofibers for the Treatment of Polymicrobial Osteomyelitis ## Abstract This study aimed to develop a drug delivery system with hybrid biodegradable antifungal and antibacterial agents incorporated into poly lactic-co-glycolic acid (PLGA) nanofibers, facilitating an extended release of fluconazole, vancomycin, and ceftazidime to treat polymicrobial osteomyelitis. The nanofibers were assessed using scanning electron microscopy, tensile testing, water contact angle analysis, differential scanning calorimetry, and Fourier-transform infrared spectroscopy. The in vitro release of the antimicrobial agents was assessed using an elution method and a high-performance liquid chromatography assay. The in vivo elution pattern of nanofibrous mats was assessed using a rat femoral model. The experimental results demonstrated that the antimicrobial agent-loaded nanofibers released high levels of fluconazole, vancomycin, and ceftazidime for 30 and 56 days in vitro and in vivo, respectively. Histological assays revealed no notable tissue inflammation. Therefore, hybrid biodegradable PLGA nanofibers with a sustainable release of antifungal and antibacterial agents may be employed for the treatment of polymicrobial osteomyelitis. ## 1. Introduction Despite advances in surgery, chronic osteomyelitis treatment remains a great challenge, often associated with a significant financial burden on healthcare systems. Osteomyelitis is an acute or chronic inflammatory process affecting the bone and its structure, secondary to the infection with pyogenic organisms, including bacteria, fungi, and mycobacteria [1,2]. Certain fungal and bacterial infections can be identified by the formation of biofilms that enhance antifungal and antibiotic drug resistance [3]. Bacterial and fungal coinfection was found in a small group of patients with osteomyelitis [4,5]. Candida albicans and Staphylococcus aureus are the nosocomial pathogens frequently responsible for severe morbidity and mortality, even with appropriate treatment. Coinfection is associated with mixed biofilms, more severe clinical manifestations, and enhanced drug resistance [6,7,8]. The mixed biofilm provides microbes with a stable environment that allows them to tolerate high antimicrobial concentrations, frequently resulting in therapeutic failure [6,7]. Antifungal and antibiotic drug concentrations a hundred-fold to a thousand-fold higher than the minimum inhibitory concentration (MIC) are generally necessary to treat coinfections [9,10]. The consensus treatment for bacterial and fungal osteomyelitis often involves both surgery and administration of antimicrobial agents [11,12]. Antibiotic-loaded polymethyl methacrylate (PMMA) (also known as bone cement) beads or spacers are well accepted for the treatment of bacterial osteomyelitis to provide a highly sustained local antibiotic concentration [13,14,15]. However, the effect of their antifungal-loading on osteomyelitis treatment is not well defined because of inconsistent drug release. Although successful eradication of fungal osteomyelitis or periprosthetic joint infections has been reported, [16] there are conflicting results [17,18]. High local concentrations of antifungal and antibiotic agents following radical debridement surgery can be a novel approach for treating complex polymicrobial infections. However, to the best of our knowledge, no study has addressed the simultaneous delivery of local, sustained, multiple antimicrobial agents for the treatment of fungal and bacterial-associated coinfections. In this study, we developed hybrid biodegradable antifungal and antibacterial poly lactic-co-glycolic acid (PLGA) nanofibers via electrospinning that provide extended release of fluconazole, vancomycin, and ceftazidime. Fluconazole is an antifungal agent used against several fungal infections, including candidiasis, blastomycosis, coccidioidomycosis, etc. [ 16]. Vancomycin is an antibiotic and the treatment of choice for methicillin-resistant S. aureus osteomyelitis [19]. Ceftazidime is an antibiotic used to treat several bacterial infections, including joint infections, meningitis, pneumonia, sepsis, etc. [ 20]. Electrospinning is a versatile method to produce nanofibers or nanofiber mats from diverse polymers or polymer blends. In this process, high voltage electricity (5 to 50 kV) is applied to both a liquid solution and a collector, allowing the solution to extrude from a needle to form a jet. Once the solvent has evaporated, the jet solidifies and deposits fibers on the collector [21]. Electrospinning can be used to prepare fibers with diameters ranging from tens to hundreds of nanometers, sometimes up to a few micrometers [22]. Owing to their small diameter, large surface-to-volume ratio, and 3D networks, which mimic native extracellular matrices, nanofibers or nanofibrous mats can be used for various applications [23,24,25,26]. Distinct materials have been adopted for delivering drugs and biomolecules. PLGA is a biodegradable polymer that has received approval from pharmaceutical authorities as a therapeutic vehicle because of its extraordinary biocompatibility and biodegradability [27,28]. Our aim was to manufacture and assess electrospun antimicrobial-loaded degradable PLGA nanofibers and nanofibrous mats, and determine their in vitro and in vivo antimicrobial agent discharge pattern as a possible treatment for osteomyelitis. ## 2.1. Characterization of Electrospun Nanofibers Nanofibers incorporated with biodegradable hybrid antimicrobial agents were successfully manufactured via electrospinning. Figure 1 shows the SEM images and size distributions of the electrospun nanofibers. The measured diameter range for pure PLGA nanofibers was 1.45 ± 0.15 μm. The calculated fiber diameter range was 1.05 ± 0.11 μm and 120.0 ± 37.1 nm for fluconazole-loaded and vancomycin/ceftazidime-incorporated nanofibers, respectively. The nanofiber diameters decreased with the addition of antimicrobial agents. The increase in drug concentration reduced the PLGA percentage in the solution, which became less viscous and was more easily extended by the electric force during the spinning process. Hence, the size of the electrospun nanofibers decreased. The wetting angles of spun mats decreased with the increasing percentage of pharmaceuticals, that is, 125.6°, 107.1°, and 117.8°, for the pristine PLGA nanofibrous mats, fluconazole-, and vancomycin/ceftazidime-embedded mats, respectively (Figure 2). The fluconazole-embedded nanofibers exhibited greater hydrophilicity than the vancomycin/ceftazidime-loaded nanofibers, mainly because fluconazole is more hydrophilic than the antibiotics. Figure 3 illustrates the stress–strain curves of the electrospun nanofibers. The ultimate tensile strength of the spun nanofibers decreased with the incorporation of antimicrobial agents. The incorporation of pharmaceuticals decreased the percentage of polymers in the nanofibers, which reduced the resistance of the fibers to external tensile loads. Thus, the measured tensile properties were compromised. FTIR spectroscopy was performed to confirm the presence of antimicrobial agents in the electrospun nanofibrous mats. In the drug-loaded nanofibers, the absorbance at 970 cm−1 was promoted by the C–H bonds of vancomycin and ceftazidime [29,30]. The new vibration at 1616 cm−1 was caused by the –NH bond of the added pharmaceuticals [29,30,31]. Additionally, the C=C peaks at 1500 cm−1 were significantly developed for pharmaceutical-loaded nanofibers with respect to the peak of the pristine PLGA nanofibrous mats. The FTIR assay results confirmed the successful incorporation of antimicrobial agents into the electrospun hybrid nanofibers (Figure 4). The thermal behaviors of pristine PLGA, vancomycin/ceftazidime-loaded PLGA and fluconazole-loaded PLGA were assessed (Figure 5a,b). The peaks of vancomycin and ceftazidime disappeared after incorporation into the PLGA matrix (Figure 5a), while a new peak at 295 °C was noted [32,33]. Meanwhile, the endothermic peak of fluconazole at 140 °C diminished after blending with PLGA (Figure 5b) [34]. These results demonstrated the successful embedding of antimicrobial agents into the PLGA nanofibers. Table 1 lists the average entrapment efficiency of the nanofiber membranes containing vancomycin/ceftazidime or fluconazole. While the entrapment efficiencies of ceftazidime and fluconazole were high, the efficiency of vancomycin only reached $62\%$. This might be due to the fact that vancomycin could not be completely dissolved in HFIP during electrospinning. Drug precipitation was noted at the outlet of the spinning needle. The drug entrapped in spun nanofibers decreased accordingly. ## 2.2. In Vitro and In Vivo Release Patterns of Antimicrobial Agents The release characteristics of fluconazole, vancomycin, and ceftazidime from nanofibrous mats are shown in Figure 6. Triphasic liberation curves showed a high release on day 1, accompanied by a slow discharge for several days. The second set of peaks was noted at day 15 for vancomycin and ceftazidime, and day 23 for fluconazole; thereafter, the release diminished gradually. The relatively small errors in the curves suggested that the antimicrobial agents were uniformly distributed in the electrospun PLGA mats. Overall, the hybrid antimicrobial agent-embedded nanofibers provided sustained release of pharmaceuticals in vitro for more than 30 days. Moreover, the in vivo discharge patterns indicated that the drug-incorporated nanofibrous mats discharged high levels of fluconazole, vancomycin, and ceftazidime (above the MIC) for up to 56 days in vivo (Figure 7). ## 2.3. Histological Analysis Figure 8 shows the histological images on postoperative days 1, 7, 14 and 28. The hematoxylin and eosin-stained specimens showed notable mononuclear cell infiltrates of lymphocytes, plasma cells, and eosinophils in the muscle tissues surrounding the nanofibrous mats at day 7. Since then, the number of polymorphonuclear leukocytes diminished progressively with time up to day 28 post-operation. ## 3. Discussion Over the past three decades, the pathogenesis of osteomyelitis has been clarified, along with the identification of factors responsible for infection. In addition to surgical treatment, many antimicrobial agents have been adopted for the treatment of osteomyelitis. Bacterial and fungal coinfection has been reported in upper respiratory tract infections and cultures from distinct medical devices, including dentures, implants, endotracheal tubes, and, most commonly, catheters [35]. Although this has rarely been reported in osteomyelitis [4,5], the treatment of orthopedic coinfections remains a great challenge [36]. The major difficulty in the treatment of osteomyelitis or other co-infections is biofilm formation. Three-dimensional bacterial and fungal biofilms protect the microbial community from antimicrobial agent damage, increasing chronic infections [6]. This dominant feature contributes to therapeutic failure. In particular, mixed biofilms generated by different species, such as Candida albicans and Staphylococcus, lead to an increase in antimicrobial agent resistance by a hundred-fold of MIC [9,10]. As morbidity and mortality increase, the successful treatment of fungal and bacterial osteomyelitis has become more challenging. Lerch et al. reported that the successful treatment of S. aureus and C. albicans coinfection requires serial and radical debridement and the use of a higher dose of fluconazole (800 mg/day or 12 mg/kg body weight) than a regular dose of 400 mg/day for an extended period. However, the role of antimicrobial agent-loaded bone cement beads was not clearly described in their study [4]. The therapeutic milestones for fungal or bacterial osteomyelitis have been reported to be radical debridement and adequate systemic/local antimicrobial agents [13,14,15,36,37,38]. Recently, many advanced antibiotic delivery methods have been developed [39,40], including our previous study presenting drug-eluting degradable PLGA beads that provide a sustained release of antifungal and antibacterial agents [12,41]. Advances in nanotechnology have led to encouraging treatments for osteomyelitis. In this study, degradable PLGA nanofibers were exploited for the sustained discharge of antimicrobial agents into the target area for infection control. Due to their elevated surface-area-to-volume ratio, nanofibers offer an advantageous vehicle for the transport of water-insoluble or poorly soluble pharmaceuticals. The three-dimensional network structure of nanofibers also mimics the architecture of the extracellular matrix of natural tissues, allowing enhanced cell functionality after the incorporation of drugs and multiple factors [42]. Our previous study examined the toxicity of electrospun nanofibers using a 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay of human fibroblast proliferation [43]. The electrospun nanofibers showed no signs of cytotoxicity. Additionally, the antimicrobial agent release from the nanofibers was determined with a disk diffusion method. After the electrospinning process, the bioactivities of released antibiotics remained high, ranging from $40\%$ to $100\%$ [44]. Additionally, nanofibrous mats also possess the benefits of conventional solid dosage forms, including easy processing, excellent drug stability, and simple packaging/shipping [45,46]. The biodegradable PLGA/vancomycin/ceftazidime/fluconazole nanofibers developed in this study could release high concentrations of antimicrobial agents for over 30 days in vitro, which provides advantages in terms of orthopedic infection control. This study is the first to develop biodegradable hybrid antibiotic/antifungal nanofibers using an electrospinning technique that concurrently offers the sustained discharge of elevated local concentrations of vancomycin, ceftazidime, and fluconazole. Several factors may affect the release of pharmaceuticals from PLGA matrix systems [47], including the molecular weight and hydrophilicity of incorporated drugs, the rate of aqueous medium infiltration into the matrix, and the rate of matrix erosion. In general, the release curves can be partitioned into three stages: a primary blast, diffusion-governed discharge, and degradation-controlled release [48]. After the spinning procedure, most drugs are distributed in the nanofiber volume. Drugs on the fiber surface may lead to a primary blast release in the first few days. Afterwards, drug release is mainly governed by drug diffusion and polymer degradation. Successive drug release peaks were observed at days 14, 15, and 25 for vancomycin, ceftazidime, and fluconazole, respectively. Subsequently, a persistent steady discharge of antibiotics and antifungals was observed. Due to the different characteristics of embedded vancomycin, ceftazidime, and fluconazole, the nanofibers may present distinct antimicrobial agent-release profiles. In addition, a lower amount of vancomycin was released during both the in vitro and in vivo elution processes. Due to a relatively low solubility in HFIP, vancomycin precipitation was observed during the spinning process causing an accumulation of the precipitate at the exit of the needle and a lower conveyance to the nanofiber matrix. The drug release decreased accordingly. The advantage of local drug delivery is high, sustained, local levels of pharmaceuticals without high systemic doses, thus, minimizing systemic toxicity [49]. Intravenous administration of antimicrobial agents leads to drug concentrations that are above MIC for susceptible microorganisms, but the levels reached are lower and the period above MIC is limited [50]. Roy et al. [ 51] found that both intra-articular and intravenous administration of vancomycin reached therapeutic concentrations in the synovial fluid of the knee, but intra-articular delivery resulted in peak concentrations several orders of magnitude higher, and also led to therapeutic serum levels. The half-life of intra-articular-delivered vancomycin was slightly over 3 h, with concentrations persisting above therapeutic level 24 h after injection in both the joint and serum. Traditionally, osteomyelitis is treated with 4–6 weeks of parenteral antimicrobial agents after debridement procedure [1,13,52], and the antimicrobial concentration required to eradicate microorganisms inside biofilms can be a hundred to a thousand-fold higher than the MIC. The results of our in vivo experiments showed that the released concentrations of vancomycin, ceftazidime, and fluconazole (approximately 50–100, 1000, and 1000 μg/mL, respectively) from the nanofibers was much higher than their corresponding MICs (1.0, 2.0, and 0.5 μg/mL, respectively [12]) for more than 8 weeks (Figure 7). This provides advantages for sustained local therapeutic concentrations of antimicrobial agents in osteomyelitis treatment and prophylaxis [48,53]. Additionally, the histological assay suggested significant mononuclear cell infiltrates of lymphocytes, plasma cells, and eosinophils in muscle tissues surrounding the membrane post-implantation. Nevertheless, the number of polymorphonuclear leukocytes progressively diminished after 4 weeks (Figure 8). Despite the proven effectiveness of the hybrid antimicrobial agent-incorporated nanofibrous mats, there were limitations in this study. First, restriction lies in the limited type of animals used. Second, a non-infected animal model was used in this study. It is unknown whether the hybrid drug-eluting PLGA nanofibers will perform differently in an infected model. Finally, the relationship between these findings and osteomyelitis in humans is unclear and requires further investigation. ## 4.1. Manufacturing of Hybrid Drug-Loaded Nanofibers PLGA was used in nanofiber fabrication (LA:GA = 75:25, with a molecular weight of 76,000–115,000 Da; Sigma-Aldrich, St. Louis, MO, USA). Vancomycin hydrochloride, ceftazidime hydrate, fluconazole, and hexafluoro-2-propanol (HFIP) (Sigma-Aldrich) were used as the antimicrobial agents and solvent, respectively. Two-layer hybrid nanofibers were prepared using a lab-scale electrospinning apparatus. To fabricate the antifungal-loaded nanofibers, 1120 mg PLGA and 560 mg fluconazole were blended with 6 mL HFIP and subsequently spun into nanofibrous mats using a syringe/needle (internal diameter of 0.42 mm) at a temperature of 25 °C and relative humidity of $65\%$. The mixture transport speed was 0.5 mL/h. The voltage employed was 18,000 V, and the distance between the needle and the collector was 150 mm. This was followed by electrospinning of the PLGA/vancomycin/ceftazidime nanofibers. PLGA/vancomycin/ceftazidime (1120 mg/280 mg/280 mg) was blended with 6 mL of HFIP using the same spinning parameters as those for PLGA/fluconazole nanofibers. After electrospinning, hybrid antifungal and antibacterial nanofibers were obtained, with a thickness of approximately 0.18 mm. For comparison, pure PLGA nanofibers were also prepared by mixing PLGA (1120 mg) with HFIP (6 mL) and electrospinning into nanofibrous mats. ## 4.2. Assessment of Electrospun Nanofibers The electrospun nanofibrous mats were evaluated using a Joel JSM-7500F scanning electron microscope (SEM; Tokyo, Japan). One hundred randomly selected nanofibers from the SEM micro-image were used to determine the distribution of the nanofiber size. A general-purpose goniometer (First Ten Angstroms, Newark, CA, USA) ($$n = 3$$) was used to evaluate the wetting angle of the nanofibrous mats. The tensile properties of the nanofibers were evaluated using a Lloyd tensiometer (Ametek, Berwyn, PA, USA). Nanofibrous samples with dimensions of 10 mm × 50 mm were clamped between two grips with a distance of 3 cm between the grips. The sample was stretched by the top grip at a rate of 60 mm/min for a distance of 10 cm ($$n = 3$$) before the grip returned to its starting point. The thermal properties of pristine PLGA, vancomycin/ceftazidime-loaded PLGA, and fluconazole-loaded PLGA were identified using TA-DSC25 differential scanning calorimeter (New Castle, DE, USA). The heating rate was maintained at 10 °C/min over a 30–350 °C scan range. The spectra of the electrospun antimicrobial agent-incorporated mats were obtained using a Thermo Fisher Nicolet iS5 Fourier-transform infrared (FTIR) spectrometer (Waltham, MA, USA). The nanofibrous samples were pressed into KBr discs and evaluated under the absorption mode. The resolution was 4 cm−1 with 32 scans, with a 400–4000 cm−1 range. ## 4.3. Entrapment Efficiency To determine the entrapment efficiency (EE%) of added drugs, the nanofibrous mats with vancomycin/ceftazidime or fluconazole were weighed accurately in triplicate and extracted in HFIP employing a magnetic stirrer at 20 rpm. The extract solution was then centrifuged at 10,000 rpm for 10 min at ambient temperature. The drug concentrations in the eluents were characterized using a Hitachi L-2200R high-performance liquid chromatograph (HPLC; Tokyo, Japan) ($$n = 3$$). The EE % was calculated by the following equation [54]:[1]EE(%)=WmWa×$100\%$ where EE is the entrapment efficiency, Wm is the amount of drug measured in the nanofibrous mat, and Wa is the drug added in the electrospun fibers. ## 4.4. In Vitro Pharmaceutical Discharge The discharge pattern of antimicrobial agents from drug-incorporated nanofibrous mats was assessed using an in vitro elution scheme. Nanofibrous mats with a dimension of 10 mm × 10 mm (~5 mg) were submerged in a phosphate-buffered solution of 1 mL at 37 °C. The solution was then collected and evaluated after 24 h of isothermal incubation. The buffer was refreshed every 24 h, and the process was repeated for 30 days. The concentrations of fluconazole, vancomycin, and ceftazidime in the collected solutions were determined using the HPLC assay. ## 4.5. In Vivo Investigations All animal-correlated processes received institutional approval (CGU109-005), and all animals were cared for under the supervision of a licensed veterinarian, in line with the ARRIVE guidelines and the regulations of the Department of Health and Welfare, Taiwan. Male Sprague Dawley rats (7-week-old, ~250 g each) were enrolled in the tests. The rats were kept in independent cages with temperature and light control and had ad libitum access to standard mouse chow and sterilized drinking water. The rats first received general anesthesia through isoflurane inhalation provided by a vaporizer in a polymethyl methacrylate box. Anesthesia was maintained through inhalation of isoflurane via a mask. After sedation, the right thigh of each rat was depilated, cleaned using soap, and disinfected with $70\%$ ethanol prior to surgery. The right femoral shaft was explored using an anterolateral approach under aseptic conditions. Membranes (1 cm × 2 cm) were cut from the electrospun fluconazole/vancomycin/ceftazidime-loaded nanofibers (Figure 9a) and surgically enveloped around the right femoral shaft (Figure 9b). The wound was then sealed layer-by-layer. In vivo pharmaceutical levels were assessed by collecting tissues around the nanofibers at 1, 3, 7, 14, 28, 42 and 56 days post-operation. In vivo concentrations of vancomycin, ceftazidime, and fluconazole in the collected tissue specimens were assayed using HPLC ($$n = 3$$). In addition, a tissue biopsy was performed for histological analysis at 1, 7, 14 and 28 days post-implantation. ## 5. Conclusions We prepared degradable antimicrobial agents incorporated in PLGA nanofibers and assessed their release behaviors. The in vitro drug discharge was assessed using HPLC, whereas a rat bone model was used for the evaluation of in vivo drug elution. The results showed that all nanofibrous mats discharged effective levels of vancomycin, ceftazidime, and fluconazole in vitro for 30 days. Animal tests also showed that the nanofibers released effective concentrations of antimicrobial agents for over 56 days after surgery. Histological assays did not reveal any notable tissue inflammation. Therefore, degradable nanofibers with sustained release of antimicrobial agents may be a potential treatment for polymicrobial osteomyelitis. ## References 1. Pincher B., Fenton C., Jeyapalan R., Barlow G., Sharma H.K.. A systematic review of the single-stage treatment of chronic osteomyelitis. J. Orthop. Surg. Res. (2019) 14 393. DOI: 10.1186/s13018-019-1388-2 2. Gamaletsou M.N., Walsh T.J., Sipsas N.V.. Epidemiology of fungal osteomyelitis. Curr. Fungal. Infect. Rep. (2014) 8 262-270. DOI: 10.1007/s12281-014-0200-3 3. Desai J.V., Mitchell A.P., Andes D.R.. Fungal biofilms, drug resistance, and recurrent infection. Cold Spring Harb. Perspect. Med. (2014) 4 a019729. DOI: 10.1101/cshperspect.a019729 4. Sanz-Rodriguez C., Hernandez-Surmann F., Bueno A.G., Goizueta C., Noguerado A.. Candida and bacterial mandibular osteomyelitis in an AIDS patient. Eur. J. Clin. Microbiol. Infect. Dis. (1998) 17 531-532. DOI: 10.1007/bf01691142 5. Lerch K., Kalteis T., Schubert T., Lehn N., Grifka J.. Prosthetic joint infections with osteomyelitis due to. Mycoses (2003) 46 462-466. DOI: 10.1046/j.0933-7407.2003.00928.x 6. Scheunemann G., Fortes B.N., Lincopan N., Ishida K.. Caspofungin inhibits mixed biofilms of. Microbiol. Spectr. (2021) 9 e0074421. DOI: 10.1128/Spectrum.00744-21 7. Kong E.F., Tsui C., Kucharíková S., Andes D., Van Dijck P., Jabra-Rizk M.A.. Commensal protection of. mBio (2016) 7 e01365-16. DOI: 10.1128/mBio.01365-16 8. Eichelberger K.R., Cassat J.E.. Metabolic adaptations during. Front. Immunol. (2021) 12 797550. DOI: 10.3389/fimmu.2021.797550 9. Høiby N., Bjarnsholt T., Givskov M., Molin S., Ciofu O.. Antibiotic resistance of bacterial biofilms. Int. J. Antimicrob. Agents (2010) 35 322-332. DOI: 10.1016/j.ijantimicag.2009.12.011 10. Kuhn D.M., George T., Chandra J., Mukherjee P.K., Ghannoum M.A.. Antifungal susceptibility of Candida biofilms: Unique efficacy of amphotericin B lipid formulations and echinocandins. Antimicrob. Agents Chemother. (2002) 46 1773-1780. DOI: 10.1128/AAC.46.6.1773-1780.2002 11. Barakat A., Schilling W.H.K., Sharma S., Guryel E., Freeman R.. Chronic osteomyelitis: A review on current concepts and trends in treatment. Orthop. Trauma (2019) 33 181-187. DOI: 10.1016/j.mporth.2019.03.005 12. Hsu Y.H., Chen H.Y., Chen J.C., Yu Y.H., Chou Y.C., Ueng S.W.N., Liu S.J.. Resorbable beads provide extended release of antifungal medication: In vitro and in vivo analyses. Pharmaceutics (2019) 11. DOI: 10.3390/pharmaceutics11110550 13. Mader J.T., Calhoun J., Cobos J.. In vitro evaluation of antibiotic diffusion from antibiotic-impregnated biodegradable beads and polymethylmethacrylate beads. Antimicrob. Agents Chemother. (1997) 41 415-418. DOI: 10.1128/AAC.41.2.415 14. Kontogiannopoulos K.N., Assimopoulou A.N., Tsivintzelis I., Panayiotou C., Papageorgiou V.P.. Electrospun fiber mats containing shikonin and derivatives with potential biomedical applications. Int. J. Pharm. (2011) 409 216-228. DOI: 10.1016/j.ijpharm.2011.02.004 15. Hsu Y.H., Hu C.C., Hsieh P.H., Shih H.N., Ueng S.W.N., Chang Y.. Vancomycin and ceftazidime in bone cement as a potentially effective treatment for knee periprosthetic joint infection. J. Bone Joint Surg. Am. (2017) 99 223-231. DOI: 10.2106/JBJS.16.00290 16. Wu M.H., Hsu K.Y.. Candidal arthritis in revision knee arthroplasty successfully treated with sequential parenteral-oral fluconazole and amphotericin B-loaded cement spacer. Knee Surg. Sports Traumatol. Arthrosc. (2011) 19 273-276. DOI: 10.1007/s00167-010-1211-4 17. Sealy P.I., Nguyen C., Tucci M., Benghuzzi H., Cleary J.D.. Delivery of antifungal agents using bioactive and non-bioactive bone cements. Ann. Pharmacother. (2009) 43 1606-1615. DOI: 10.1345/aph.1M143 18. Goss B., Lutton C., Weinrauch P., Jabur M., Gillett G., Crawford R.. Elution and mechanical properties of antifungal bone cement. J. Arthroplast. (2007) 22 902-908. DOI: 10.1016/j.arth.2006.09.013 19. Liu S.J., Ueng S.W.N., Lin S.S., Chan E.C.. In vivo release of vancomycin from biodegradable beads. J. Biomed. Mater. Res. (2002) 63 807-813. DOI: 10.1002/jbm.10406 20. Hsu Y.H., Chen D.W., Li M.J., Yu Y.H., Chou Y.C., Liu S.J.. Sustained delivery of analgesic and antimicrobial agents to the knee joint by direct injections of electrosprayed multipharmaceutical-loaded nano/microparticles. Polymers (2018) 10. DOI: 10.3390/polym10080890 21. Islam M.S., Ang B.C., Andriyana A., Afifi A.M.. A review on fabrication of nanofibers via electrospinning and their applications. SN Appl. Sci. (2019) 1 1248. DOI: 10.1007/s42452-019-1288-4 22. Fadil F., Affandi N.D.N., Misnon M.I., Bonnia N.N., Harun A.M., Alam M.K.. Review on electrospun nanofiber-applied products. Polymers (2021) 13. DOI: 10.3390/polym13132087 23. Torres-Martinez E.J., Bravo J.M.C., Medina A.S., Gonzalez G.L.P., Gomez L.J.V.. A summary of electrospun nanofibers as drug delivery system: Drugs loaded and biopolymers used as matrices. Curr. Drug Deliv. (2018) 15 1360-1374. DOI: 10.2174/1567201815666180723114326 24. Hrib J., Sirc J., Hobzova R., Hampejsova Z., Bosakova Z., Munzarova M., Michalek J.. Nanofibers for drug delivery—Incorporation and release of model molecules, influence of molecular weight and polymer structure. Beilstein J. Nanotechnol. (2015) 6 1939-1945. DOI: 10.3762/bjnano.6.198 25. Cleeton C., Keirouz A., Chen X., Radacsi N.. Electrospun nanofibers for drug delivery and biosensing. ACS Biomater. Sci. Eng. (2019) 5 4183-4205. DOI: 10.1021/acsbiomaterials.9b00853 26. Madhukiran D.R., Jha A., Kumar M., Ajmal G., Bonde G.V., Mishra B.. Electrospun nanofiber-based drug delivery platform: Advances in diabetic foot ulcer management. Expert Opin. Drug Deliv. (2021) 18 25-42. DOI: 10.1080/17425247.2021.1823966 27. Ghitman J., Biru E.I., Stan R., Iovu H.. Review of hybrid PLGA nanoparticles: Future of smart drug delivery and theranostics medicine. Mater. Des. (2020) 193 108805. DOI: 10.1016/j.matdes.2020.108805 28. Chiu H.I., Samad N.A., Fang L., Lim V.. Cytotoxicity of targeted PLGA nanoparticles: A systematic review. RSC Adv. (2021) 11 9433-9449. DOI: 10.1039/D1RA00074H 29. Hagbani T.A., Yadav H., Moin A., Lila A.S.A., Mehmood K., Alshammari F., Khan S., Khafagy E.S., Hussain T., Rizvi S.M.D.. Enhancement of vancomycin potential against pathogenic bacterial strains via gold nano-formulations: A nano-antibiotic approach. Materials (2022) 15. DOI: 10.3390/ma15031108 30. Huang Y., Zhang Y., Yan Z., Liao S.. Assay of ceftazidime and cefepime based on fluorescence quenching of carbon quantum dots. Luminescence (2015) 30 1133-1138. DOI: 10.1002/bio.2871 31. Atay O., Selck F.. Quantitative determination of fluconazole by infrared spectrophotometry. Anal. Lett. (1996) 29 2163-2176. DOI: 10.1080/00032719608002239 32. Zhou J., Fang T., Wen J., Shao Z., Dong J.. Silk coating on poly(ε-caprolactone) microspheres for the delayed release of vancomycin. J. Microencapsul. (2011) 28 99-107. DOI: 10.3109/02652048.2010.534824 33. Safdari M., Shakiba E., Kiaie S.H., Fattahi A.. Preparation and characterization of ceftazidime loaded electrospun silk fibroin/gelatin mat for wound dressing. Fiber Polym. (2016) 17 744-750. DOI: 10.1007/s12221-016-5822-3 34. Alkhamis K.A., Obaidat A.A., Nuseirat A.F.. Solid-state characterization of fluconazole. Pharm. Dev. Technol. (2002) 7 491-503. DOI: 10.1081/PDT-120015052 35. Shirtliff M.E., Peters B.M., Jabra-Rizk M.A.. Cross-kingdom interactions:. FEMS Microbiol Lett (2009) 299 1-8. DOI: 10.1111/j.1574-6968.2009.01668.x 36. Jorge L.S., Chueire A.G., Rossit A.R.B.. Osteomyelitis: A current challenge. Braz. J. Infect. Dis. (2010) 14 310-315. DOI: 10.1016/S1413-8670(10)70063-5 37. Hatzenbuehler J., Pulling T.J.. Diagnosis and management of osteomyelitis. Am. Fam. Physician (2011) 84 1027-1033. PMID: 22046943 38. Blyth C.C., Gomes L., Sorrell T.C., da Cruz M., Sud A., Chen S.C.. Skull-base osteomyelitis: Fungal vs. bacterial infection. Clin. Microbiol. Infect. (2011) 17 306-311. DOI: 10.1111/j.1469-0691.2010.03231.x 39. Stewart R.L., Cox J.T., Volgas D., Stannard J., Duffy L., Waites K.B., Chu T.M.. The use of a biodegradable, load-bearing scaffold as a carrier for antibiotics in an infected open fracture model. J. Orthop. Trauma (2010) 24 587-591. DOI: 10.1097/BOT.0b013e3181ed1349 40. Khan A.A., Alanazi A.M., Jabeen M., Khan S., Malik A.. Additive potential of combination therapy against cryptococcosis employing a novel amphotericin B and fluconazole loaded dual delivery system. Eur. J. Pharm. Sci. (2018) 119 171-178. DOI: 10.1016/j.ejps.2018.04.015 41. Hsu Y.H., Chen D.W.C., Tai C.D., Chou Y.C., Liu S.J., Ueng S.W.N., Chan E.C.. Biodegradable drug-eluting nanofiber-enveloped implants for sustained release of high bactericidal concentrations of vancomycin and ceftazidime: In vitro and in vivo studies. Int. J. Nanomed. (2014) 9 4347-4355. DOI: 10.2147/IJN.S66526 42. Chen Y.P., Liu H.Y., Liu Y.W., Lee T.Y., Liu S.J.. Determination of electrospinning parameters’ strength in Poly(D,L)-lactide-co-glycolide micro/nanofibers diameter tailoring. J. Nanomater. (2019) 2019 2626085. DOI: 10.1155/2019/2626085 43. Liu S.J., Kau Y.C., Chou C.Y., Chen J.K., Wu R.C., Yeh W.L.. Electrospun PLGA/collagen nanofibrous membrane as early-stage wound dressing. J. Membr. Sci. (2010) 355 53-59. DOI: 10.1016/j.memsci.2010.03.012 44. Hsu Y.H., Lin C.T., Yu Y.H., Chou Y.C., Liu S.J., Chan E.C.. Dual delivery of active antibactericidal agents and bone morphogenetic protein at sustainable high concentrations using biodegradable sheath-core-structured drug-eluting nanofibers. Int. J. Nanomed. (2016) 11 3927-3937. DOI: 10.2147/IJN.S107250 45. Luraghi A., Peri F., Moroni L.. Electrospinning for drug delivery applications: A review. J. Control. Release (2021) 334 463-484. DOI: 10.1016/j.jconrel.2021.03.033 46. Martínez-Pérez C.A.. Electrospinning: A promising technique for drug delivery systems. Rev. Adv. Mater. Sci. (2020) 59 441-454. DOI: 10.1515/rams-2020-0041 47. Varma M.V., Kaushal A.M., Garg A., Garg S.. Factors Affecting mechanism and kinetics of drug release from matrix-based oral controlled drug delivery systems. Am. J. Drug Deliv. (2004) 2 43-57. DOI: 10.2165/00137696-200402010-00003 48. Liu K.S., Kao C.W., Tseng Y.Y., Chen S.K., Lin Y.T., Lu C.J., Liu S.J.. Assessment of antimicrobial agents, analgesics, and epidermal growth factors-embedded anti-adhesive poly (lactic-co-glycolic acid) nanofibrous membranes: In vitro and in vivo studies. Int. J. Nanomed. (2021) 16 4471-4480. DOI: 10.2147/IJN.S318083 49. Eccleston D.S., Horrigan M.C., Ellis S.G.. Rationale for local drug delivery. Semin. Interv. Cardiol. (1996) 1 8-16. PMID: 9552480 50. McLaren A.C.. Alternative materials to acrylic bone cement for delivery of depot antibiotics in orthopaedic infections. Clin. Orthop. Relat. Res. (2004) 427 101-106. DOI: 10.1097/01.blo.0000143554.56897.26 51. Roy M.E., Peppers M.P., Whiteside L.A., LaZear R.M.. Vancomycin concentration in synovial fluid: Direct injection into the knee vs. intravenous infusion. J. Arthroplast. (2014) 29 564-568. DOI: 10.1016/j.arth.2013.08.017 52. Haidar R., Der Boghossian A., Atiyeh B.. Duration of post-surgical antibiotics in chronic osteomyelitis: Empiric or evidence-based?. Int. J. Infect. Dis. (2010) 14 e752-e758. DOI: 10.1016/j.ijid.2010.01.005 53. Lou Aie Banerjee P., Drusano G.L., Shayegani M., Miller M.H.. Interaction between fluconazole and amphotericin B in mice with systemic infection due to fluconazole-susceptible or -resistant strains of. Antimicrob. Agents Chemother. (1999) 43 2841-2847. DOI: 10.1128/AAC.43.12.2841 54. Ye P., Wei S., Luo C., Wang Q., Li A., Wei F.. Long-term effect against Methicillin-resistant. Biomolecules (2020) 10. DOI: 10.3390/biom10030362
--- title: 'A Critical Appraisal of the Diagnostic and Prognostic Utility of the Anti-Inflammatory Marker IL-37 in a Clinical Setting: A Case Study of Patients with Diabetes Type 2' authors: - Zvonimir Bosnić - František Babič - Viera Anderková - Mario Štefanić - Thomas Wittlinger - Ljiljana Trtica Majnarić journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9966907 doi: 10.3390/ijerph20043695 license: CC BY 4.0 --- # A Critical Appraisal of the Diagnostic and Prognostic Utility of the Anti-Inflammatory Marker IL-37 in a Clinical Setting: A Case Study of Patients with Diabetes Type 2 ## Abstract Background: The role of the cytokine interleukin-37 (IL-37) has been recognized in reversing inflammation-mediated metabolic costs. The aim was to evaluate the clinical utility of this cytokine as a diagnostic and prognostic marker in patients with type 2 diabetes (T2D). Methods: We included 170 older (median: 66 years) individuals with T2D (females: 95) and classified as primary care attenders to assess the association of factors that describe patients with plasma IL-37 levels (expressed as quartiles) using multinomial regression models. We determined the diagnostic ability of IL-37 cut-offs to identify diabetes-related complications or patient subgroups by using Receiver Operating Characteristic analysis (c-statistics). Results: Frailty status was shown to have a suppressive effect on IL-37 circulating levels and a major modifying effect on associations of metabolic and inflammatory factors with IL-37, including the effects of treatments. Situations in which IL-37 reached a clinically significant discriminating ability included the model of IL-37 and C-Reactive Protein in differentiating among diabetic patients with low–normal/high BMI ((<25/≥25 kg/m2), and the model of IL-37 and Thyroid Stimulating Hormone in discriminating between women with/without metabolic syndrome. Conclusions: The study has revealed limitations in using classical approaches in determining the diagnostic and prognostic utility of the cytokine IL-37 in patients with T2D and lain a foundation for new methodology approaches. ## 1. Introduction Interleukin-37 (IL-37) is a member of the IL-1 cytokine family, otherwise known for its role in inflammation promotion [1]. This cytokine, formerly known as family member 7 of the IL-1 cytokine family, was characterized by computational cloning as having a role as a negative regulator of IL-18 which, in synergism with IL-1β, acts as the critical proinflammatory cytokine of this cytokine family [2]. IL-37 is a part of the mechanisms of self-control developed during evolution to limit the harmful effects of excessive inflammation [3,4]. The broad suppressor activities of IL-37 in both innate and adaptive immunity arise from its dual modes of action, which include the effects induced by the binding of this cytokine to the cell surface IL-18 decoy receptors and the effects that are exerted by its binding to the nuclear DNA (Deoxyribonucleic acid), where it impedes the transcription of proinflammatory genes [5,6]. In addition, complex mechanisms are involved in its activation since IL-37 transcripts are released in the cytosol in the precursor forms, which may account for additional variability in this cytokine’s abundance and specific actions [7,8]. Besides its anti-inflammatory activity, IL-37 acts to restore cell metabolic homeostasis during inflammation and reverse chronic inflammation’s metabolic costs [9,10]. The results of experimental studies cannot be directly translated into real-life situations. In human diseases, in different conditions that occur under the same disease label, the prevailing proinflammatory or anti-inflammatory effect of IL-37 may relate to either IL-37 genetic variations or other specific patient-related or disease-related contexts. Regarding IL-37 genetic variations, differences in the expression of several existing IL-37 isoforms, either as mRNA (messenger Ribonucleic acid) variants or variations in single nucleotide polymorphisms, were found to influence the abundance of the circulating IL-37 and disease pathways [11,12,13,14,15,16]. However, the level of knowledge of the clinical consequences of these variations is still low [7,8,17,18]. It is worth noting that cytokine regulation, and in particular regarding IL-37, as a negative regulator of inflammation is generally more complex in multifactorial age-related conditions (such as type 2 diabetes (T2D), where patients with the same diagnostic label share multiple pathophysiology pathways and exhibit heterogeneous clinical expression), than in autoimmune and immunologically mediated diseases, which are characterized by high-impact genetic and environmental exposures, and which are more homogeneous in phenotypic expression [19,20]. That differences in IL-37 gene polymorphisms in combination with behavioral or other specific patient-related factors may differently direct phenotypic variability was demonstrated in a recent epidemiologic study, where authors searched for IL-37 gene polymorphisms in individuals with and without hypercholesterolemia [21]. Due to its functioning at the crossroad of inflammatory and metabolic pathways, IL-37 has attracted growing interest in cardio-metabolic conditions for its translational perspective. The theoretical framework is based on the evidence indicating that obesity and its related disorders, metabolic syndrome (MS) (glucose-related metabolism impairment associated with abdominal obesity and hypertension), T2D, and cardiovascular disease (CVD) are those factors that contribute the most to inflammation associated with accelerated aging [22]. Increased activity of the innate immune system as well as increased production of metabolic intermediates and reactive oxygen species due to the glucose-related metabolism impairment generate inflammatory signals [23,24]. The proinflammatory microenvironment turns the balance from the predomination of anti-inflammatory regulatory T cells (Treg) to the development of the proinflammatory Th1/Th17 immune pathway by metabolic reprogramming and epigenetic alterations, for which IL-37 acts as a reversing factor [9,25,26]. The therapeutic potential of IL-37 has been demonstrated in experimental models [27,28,29]. However, clinical studies in this setting are insufficient to provide the specific biological properties of IL-37 across a range of clinical conditions [30,31,32]. Before the clinical utility of IL-37 becomes possible in cardio-metabolic conditions, there is a need to establish the reference range for this cytokine in healthy adults and patient subgroups [33]. The problem with these conditions is that they are characterized by high complexity, as already shown for patients with T2D. That means that multiple comorbidities, which change in scope and intensity can influence the pathophysiology pathways and inflammatory responses, depending on the dynamics of disease progression and patient age [34]. It implies the need to search for new research approaches when validating the prognostic value of circulating IL-37 in cardio-metabolic conditions [35]. Accordingly, this study aimed to identify the range values of plasma levels of IL-37 in older primary care (PC) patients diagnosed with T2D, to critically evaluate their diagnostic and prognostic relevance, and to identify limitations of using classical approaches in determining the diagnostic utility of this cytokine in older patients with T2D. ## 2.1. Participants and Study Design Participants were patients diagnosed with T2D, aged 50 years or more (median 66), and attendees in a PC setting. The study was conducted in 2020 and lasted for four months. Four PC practices were included, and doctors agreed to participate in the study. The Expert and Ethics Council of the Health Centre, where these PC practices were located, approved the study (ID: 1433-$\frac{1}{020}$). Only patients who were able to come to the doctor personally, but not those dependent on the care of others, were recruited. They were selected consecutively at their visits, except for those who met the exclusion criteria. For the exclusion criteria, we excluded acute conditions, malignant diseases in the active treatment phase, noticeable or diagnosed cognitive impairments, and individuals with an amputated lower limb, transplanted kidneys, or those on chronic renal replacement therapy. The calculation of the sample size, based on the significance level of 0.05 and power of 0.8, indicated a minimum sample size of 180 individuals (GPower, 3.1.9.4.). The final number of participants included in the study was 170 (Male: Female, 75:95), as we adjusted the sample size to the size of the cytokine diagnostic kits. ## 2.2. Data Collection Participants were described by a total of 62 variables, including sociodemographic characteristics, anthropometric measures, comorbidities, medications, frailty, nutritional status, markers of inflammation, and laboratory tests indicating metabolic status and renal function (Table S1a,b—Supplementary Materials). Some data were taken from patient health records, such as comorbidities and medication prescription information. During their visits, patients undertook anthropometric measurements and an assessment of their frailty and nutritional status. They were referred to the county hospital’s biochemical laboratory for venipuncture and laboratory testing. To assess frailty, we used Fried’s phenotype model, which requires a walking exercise and hand grip strength assessment by the dynamometer and involves information on the level of activity, weight loss, and subjective feeling of exhaustion [36]. Positive 1-2 criteria out of a maximum of five indicate pre-frailty, and three or more positive criteria indicate frailty. Nutritional status was assessed by using the 18-item Mini Nutritional Assessment—Short Form (MNA-SF) test [37]. This test can discriminate among good nutritional status, risk of malnutrition, and malnutrition. The level of sarcopenia (muscle loss) was determined by measuring mid-arm circumference (mac) [38]. Body mass index (BMI, kg/m2) was used to determine general obesity, and waist circumference (wc) was used to determine the abdominal type of obesity. According to the guidelines, for older patients with T2D and characterized with multimorbidity, the threshold of HbA1c (Glycated hemoglobin) of 8.0 mmol/l or even 9.0 mmol/L can be considered a reasonable hyperglycemia control [39]. To determine whether patients had MS, we used the modified criteria of the National Cholesterol Education Program, Adult Treatment Panel III (NCP ATP III) definition [40]. The NCP ATP III definition of MS is: ≥2 of the following: wc ≥ 102 cm (88 cm for F), Triglycerides (TG) ≥ 1.7, HDL (High Density Lipoprotein)-cholesterol < 1.0 (1.2 for F), and diagnosis of hypertension. Among laboratory tests, we also performed Thyroid-Stimulating Hormone (TSH) screening to detect patients with latent hypothyroidism—the common condition in patients with T2D, which can contribute to metabolic disorders [41]. Renal function decline, a common disorder in patients with T2D due to the development of diabetic nephropathy, was assessed from serum creatinine levels and information on patient age and gender by using the Modification of Diet in Renal Disease (MDRD) equation and online calculator [42]. An estimated glomerular filtration rate (eGFR) ≤ 60 mL/min was considered to indicate decreased renal function, corresponding with stages 3 and 4 of chronic renal insufficiency [43]. Diagnoses of other major diabetic complications (such as CVD, including coronary artery disease (CAD), chronic heart disease (CHD), cerebrovascular disease, and periphery artery disease), as well as of diabetic retinopathy were recorded in patient health records if the specialist examinations confirmed these diagnoses. Since cardiac imaging is necessary to evaluate CHD, and this data was not systematically recorded in patient health records, information was missing on the grades of CHD progression [44]. For prescribed medications, we used data on the prescription rates of the main groups of antidiabetic and antihypertensive drugs, on the hypolipidemic drugs “statins”, and on non-steroidal anti-inflammatory drugs (NSAID), all of which may have influenced metabolic parameters or the level of inflammation [45,46]. As a marker of chronic inflammation, we used the neutrophil-to-lymphocyte ratio (NLR), which can be estimated from the complete differential blood count [47]. We also used classical markers of inflammation and nutritional status, C-reactive protein (CRP), number of erythrocytes, hemoglobin, and hematocrit [48,49]. For IL-37 analysis, a part of the blood sample was separated, placed into heparinized tubes, and centrifuged for plasma. These specimens were transported in the transporter refrigerator to the Laboratory for Clinical Immunology and Allergology Diagnostics of the University Hospital Centre of Osijek, the administrative center of the eastern part of Croatia. For the determination of IL-37, the quantitative sandwich immunoassay technique was used (IL-37 Human Uncoated ELISA kit, Invitrogen, ThermoFisher Scientific, SAD). ## 2.3. Statistical Analysis The data description was provided as mean ± SD (standard deviation) or the median and interquartile range (IQR) for numerical data and as absolute and relative frequencies for categorical data. Discrimination among patients by levels of IL-37 was expressed as quartiles of IL-37. The collinearity and multicollinearity were investigated for numerical attributes using correlation analysis (Spearman’s correlation coefficient) and Variance Inflation Factor (VIF). A value of VIF between 1 and 5 indicates a moderate correlation between the given predictor variable and other predictor variables in the model, but this is often not severe enough to require attention. A value greater than 5 indicates a potentially severe correlation between the given predictor variable and other predictor variables in the model. In this case, the coefficient estimates and p-values in the regression outputs are likely unreliable. The methods used to assess differences in other examined variables among quartiles of IL-37 were the Chi-square (χ2) test or the Fisher’s exact test for categorical variables, and the one-way analysis of variance (ANOVA) or the Kruskal–Wallis rank–sum test for numerical variables, depending on the type of distribution (standard or not). The Games-–Howell post-hoc test was used to compare differences between the quartile pairs for numerical variables that showed significant differences. The significance level of $p \leq 0.05$ was considered statistically significant in all cases. Associations of variables with quartiles of IL-37 were assessed by using the multinomial logistic regression (MLR) model from R statistics. The Akaike Information Criterion (AIC) measured the model’s predictive performance quality. To estimate the optimal cut-off values of IL-37 and other inflammatory markers for patient subgroups, we used the function “cutpointr” from R statistics. This function works in a way to take the sum of sensitivity and specificity to maximize the metric (separation) function. The Receiver Operating Characteristics Curve (ROC) was used to test the predictive power of the identified cut-off values [50]. We used the ROC and the Area Under Curve (AUC) for evaluation, also known as the c-statistics [51]. The AUC metric varies between 0.5 and 1.00 (an excellent value). A value of AUC above 0.80 indicates a good classifier. ## 3. Results Participants were mostly 50–75 years old, and women participated more than men (Table S1a,b—Supplementary Materials). They were primarily overweight/obese (Figure 1). Most of the patients also had MS, as confirmed by the fact that a large proportion of them expressed the abdominal (visceral) type of obesity (wc ≥ 88 cm for females and ≥102 for males), and the majority were diagnosed with hypertension. There were patients with short-term (0–5 years) and long-term (>10 years) hypertension and T2D duration. Most of the patients were in a good nutritional state, and only a small proportion were at risk of malnutrition. No individuals had severe sarcopenia, as indicated by the fact that no one had mac ≤ 22 cm. Only about a fifth of patients were frail ($\frac{31}{170}$). In most patients, HbA1c was <$8.5\%$, indicating well-controlled hyperglycemia. Concerning diabetic complications, about one-third of patients were diagnosed with CAD. There was a high proportion of those diagnosed with CHD, although information on severity grades was missing. About two-thirds of participants had decreased renal function but, in most cases, renal function decline was of a mild (eGFR 90–60 mL/min) or a moderate degree (eGFR < 60–45 mL/min). When compared to the epidemiologic data, a high proportion of patients had diabetic retinopathy (>$30\%$) [46]. Of non-cardiovascular comorbidities, the descriptive data indicated a high burden of musculoskeletal diseases and anxiety disorders. A high proportion of patients had been prescribed antihypertensive drugs of the ACE-INH/ARB (angiotensin-converting enzyme inhibitors/angiotensin receptor blockers) group and hypolipidemic statin drugs, indicating that family physicians in the study area have made efforts to adhere to the guidelines [39]. In addition, they were more prone to prescribe the old-fashioned antidiabetic drug, metformin, as the first-line therapy, while newly recommended cardio- and renal-protective drugs, GLP1ra (Glucagon-like peptide-1 receptor agonists) and SGLT2inh (Sodium-Glucose Transport Protein 2 inhibitors), were prescribed at low rates [39,46]. The values of IL-37 for most of the patients oscillated between 3.40 and 38.0 pg/mL, being skewed around the median value (Table 1 and Figure 2). The variability of IL-37 was higher in the upper part of the range values, reaching up to 258.80 pg/mL (excluding one extreme value of 1788.4 pg/mL), than in the lower part of the range values, where the minimum value was 0.14 pg/mL. Two other conventional markers of inflammation, NLR and CRP, also showed low variability and distributions being skewed around the median values. It can be seen from correlation analyses (Table 2) that measures of the body’s shape and nutritional status (BMI, wc, and mac) are strongly correlated with each other. However, only two variables, total cholesterol, and LDL (Low Density Lipoprotein) cholesterol, indicating whether there is hypercholesterolemia or not, showed multicollinearity (Cholesterol: 9.33, LDL: 8.70), but these variables were not selected in predictive models (see Table 3). We assessed which variables significantly changed among quartiles of IL-37 using the following cut-off values: between quartiles 1 and 2–3.40, between quartiles 2 and 3–13.40, and between quartiles 3 and 4–38.00. There were several variables that showed significant variations. These included erythrocytes, hemoglobin, and hematocrit, indicating chronic inflammation, and variables HbA1c (a borderline significance) and LDL-cholesterol, indicating metabolic disorders (Table S2—Supplementary Materials). All these variables can also be considered as signs of blood viscosity [52]. Of comorbid disorders, only CAD and gastrointestinal diseases were shown to be significant (Table S3—Supplementary Materials). In addition, the new generation of oral antidiabetic drugs, including DPP4inh (Dipeptidyl peptidase-4 inhibitors), SGLT2inh, and GLP1ra taken as a group, have shown significant variations. Variables indicating the proportion of diabetic patients with CVD, including CAD and CHD, showed a tendency to increase according to the increasing quartile rank of IL-37. On the contrary, the variables hemoglobin, hematocrit, and HbA1c, and variables indicating the proportion of patients to whom new-fashioned oral antidiabetic drugs were prescribed, showed a decreasing tendency. It is worth mentioning here that chronic inflammation associated with metabolic conditions such as T2D may cause anemia (known as anemia of chronic disease). Anemia acts to decrease blood viscosity, while, on the contrary, the burden of inflammatory mediators and metabolic substances, such as glucose and lipids, tend to increase it. The net effect on blood viscosity of these two opposite tendencies may, in turn, affect the level of inflammation in a variable way. The group of variables—erythrocytes, hemoglobin, hematocrit, HbA1c, and LDL-cholesterol—all showing significant changes across quartiles of IL-37, but in a non-linear manner, are likely to reflect this scenario. The degrees of renal function decline did not show variations according to increasing levels (expressed as quartiles) of IL-37 (Tables S2 and S3—Supplementary Materials). As demonstrated by the regression models, many clinical characteristics of diabetic patients are associated with higher-ranked quartiles of IL-37 compared with the basic one showing both negative and positive associations (Table 3). In all models, the frailty index modulates associations of other variables with IL-37, negatively influencing IL-37 levels. Figure 3 and Figure 4 show how IL-37 changes in dependence on two variables with opposite effects on IL-37 levels. The differences in both figures were statistically evaluated, i.e., if the respective data samples were normally distributed (Shapiro–Wilk test), the difference was investigated by the ANOVA test. Otherwise, we applied the Kruskal–Wallis test. The following were observed from Figure 3: in data sample BMI < 25 kg/m2 (no significant difference), BMI 25–30 kg/m2 (no significant difference), BMI > 30 kg/m2 (no significant difference), between BMI < 25 kg/m2 and BMI 25–30 kg/m2 (no significant difference), between BMI 25–30 kg/m2 and BMI > 30 kg/m2 (no significant difference). However, the tendency of elevated IL-37 in diabetic patients with CVD can be seen if they are obese (BMI > 30 kg/m2). Data from Figure 4 showed: in data sample BMI < 25 kg/m2 (no significant difference), BMI 25–30 kg/m2 (no significant difference), BMI > 30 kg/m2 (no significant difference), between BMI < 25 kg/m2 and BMI 25–30 kg/m2 (no significant difference), between BMI 25–30 kg/m2 and BMI > 30 kg/m2 (no significant difference). Similarly, as in the previous case, the tendency of IL-37 to increase can be seen in frail diabetic patients if they are obese, compared to those who are not. The clinical utility of IL-37 cut-off values in identifying subgroups of diabetic patients, such as those with CAD, CHD, or other complications, does not seem promising (Table S4—Supplementary Materials). Only in a few cases were their use likely to be feasible (showing accuracy > $80\%$). These situations involve negative prediction (exclusion) of patients with low renal function (eGFR < 45 mL/min) or CAD, as well as positive prediction (recognition) of those with diabetic retinopathy or high-level comorbidity (>3 comorbid disorders). The levels of IL-37 that are higher than 38.2 pg/mL (from the upper quartile upward) are likely to indicate, with a high level of confidence, older diabetic patients with low BMI (<25 kg/m2). Interestingly, the only model that showed significant discriminative ability of determined cut-off values of IL-37 (expressed as AUC) was the one where IL-37 was combined with TSH in discriminating between women with and without MS (Table S5—Supplementary Materials) (Figure 5). The combination with TSH also improved the diagnostic accuracy of the basic NLR model. ROC analysis also indicates that models involving CRP are better than NLR in discriminating among diabetic patients according to BMI categories (Figure 6). ## 4. Discussion In this sample of older diabetic patients, we have recorded significant variations in plasma levels of IL-37, ranging from 0.14 pg/mL to 258.80 pg/mL (except for one extreme value). In most participants, however, these values were at low levels, ranging from 3.40 to 38.00 pg/mL. According to the available evidence, these range values belong to the lower reference range for healthy adults [33]. There was a gentle rise in the 3rd, compared to the 2nd quartile, and a steep uprise in the upper quartile, so the distribution curve acquired the non-linear “J” shape. The low values of IL-37 in most patients may have been as a result of low inflammatory stimuli, intrinsically low anti-inflammatory response to inflammatory stimuli, or homeostasis breakdown due to organ damage. The fact that supports the first option is that most patients were metabolically well controlled, which may have attenuated the metabolic and inflammatory challenges. This statement is confirmed by HbA1c interquartile values ranging from $5.2\%$ to $7.7\%$, which is within the recommended target values for older diabetic patients [39]. As indicated by the regression models, antidiabetic medications significantly associated with different levels of IL37 were metformin and DPP4 inh. These medications are known to have direct anti-inflammatory effects beyond the effect on lowering hyperglycemia, which can be due to their mechanisms of action [53]. Adverse associations of these medications with plasma IL-37 levels can be viewed in the light of evidence indicating obesity as a proinflammatory state and the fact that these medications are usually prescribed to obese diabetic patients as being weight-neutral [54]. The benefit of these medications may depend, as shown in this study (Figure 3), on whether obese diabetic patients have or do not have CVD. In the case of CVD, treating obese diabetic patients with these medications may suppress the protective homeostatic anti-inflammatory response. The need for personalization of diabetic therapy is especially emphasized when considering recent evidence that some IL-37 gene polymorphisms may be associated with low IL-37 production and increased susceptibility to T2D [21]. The contradictory results that hypolipidemic statin medications did not show associations with plasma IL-37 levels, despite their anti-inflammatory mechanism of action, also support this idea. More recently, statins’ alternating proinflammatory and anti-inflammatory effects have been observed, which are proposed to be clinical context-dependent or associated with gene polymorphisms [21,55]. To conclude, knowing more about the CV risk profile of patients in the sample is necessary to realize whether low IL-37 is prognostically beneficial or detrimental. To make the range of IL-37 that we obtained in this study more reliable, we compared these values with the results of some prognostic studies from a similar setting. For example, even in the highest stage of hypertension, which is associated with overt atherosclerotic disease, plasma IL-37 levels reached values that were not higher than 50 pg/mL, which is consistent with the lower part of the upper IL-37 quartile in this study [30]. In patients with CHD, the plasma threshold of IL-37, shown to be prognostically negative, was 100 pg/mL, ranging within the limit of about 170 pg/mL [31]. These values apply to this study’s upper quartile range values of IL-37. In acute coronary syndrome, as an acute state of serious homeostasis breakdown, plasma IL-37 levels were shown to be lower than in the control and mostly below 40 pg/mL, which is consistent with the Q3 range values in this study [32]. To conclude, plasma IL-37 levels in these described pathologic conditions are consistent with Q3 and Q4 values of this study, where there was a noticeable rise in IL-37 away from the low values in Q1 and Q2 (above the cut-off of 3.40 pg/mL). However, this comparison still needs to answer whether patients in Q3 and Q4 are at increased CV risk. This uncertainty arises from the fact that the range values of the studies mentioned above are comparable with the reference range values for healthy adult controls, as reported in the recent meta-analysis [33]. Resolving this issue would require criteria reconsideration for subject selection in the control groups. The large variability of circulating levels of IL-37, showing the skewed distribution we found in this study, is consistent with the view of T2D as a complex disease [35]. In order words, this implicates the existence of multiple patient subgroups. The complexity is especially emphasized in diabetic patients of older age for their increased risk for “harmful “geriatric conditions, such as sarcopenia, malnutrition, and frailty, that are known to change the course and clinical expression of T2D [56]. Of the utmost importance is recognizing frailty, because of its proven influence on negative outcomes and its modifying effect on treatments [57,58]. Frailty is a progressive disorder characterized by muscle loss, low activity, and disturbed homeostatic reserves in multiple organs and systems [59]. The closely related disorders, T2D, MS, and CVD, are all known to strongly associate with frailty [20,60,61,62]. In this study, about half of the patients were diagnosed with CVD, and a little fewer than half had pre-frailty/frailty. More than three-quarters of male and almost all female patients met the criteria for MS. The results of this study reflect these associations. First, factors that were found to associate with IL-37 indicated metabolic disorders, chronic inflammation, and comorbidities of CVD. The second, pre-frailty/frailty status was highlighted as having a pivotal modifying role in these associations. Notably, a negative direction of associations of pre-frailty/frailty with IL-37, which was stably maintained across the regression models, is suggestive of the suppressive effect of pre-frailty/frailty on circulating levels of IL-37. Some clarifications come from the recent study, where it was found that plasma levels of IL-37 are lower in healthy elderly individuals than in middle-aged and young ones, despite elevations in proinflammatory markers [63]. These results suggest that IL-37 responsiveness to proinflammatory stimuli declines with age. Pre-frailty/frailty, rather than age per se, may have a role as the primary negative regulator of IL-37 [64], as our studies indicate that, at least in older diabetic patients, pre-frailty/frailty rates overcome those in the general population. Moreover, associations between pre-frailty/frailty and IL-37 were shown to be non-linear, that is, discontinued among quartiles of IL-37. These results fit the evidence indicating that frailty status can change in severity (pre-frailty vs. frailty) with age and the number of comorbidities [65]. Lastly, these changes express as two metabolic phenotypes of frailty—one associated with obesity (myosteatosis) and another one associated with advanced stages of target organ damage, where frailty is characterized by both muscle and weight loss, and diffusely impaired physiological reserves (sarcopenic or cachectic frailty) [66,67,68]. Within this context, we can also consider the close associations between CVD and frailty. Evidence suggests that CVD can already appear in the early phases of T2D development, which is usually associated with obesity [69]. More often, and in more severe forms, CVD appears after a more extended period of T2D duration, which often co-exists with renal function decline; all these together increase the risk of sarcopenic frailty [34,70,71]. The role of non-cardiovascular comorbidities, which usually co-exist with T2D and CVD, in contributing to inflammation and the development of frailty, should also not be neglected. Literature evidence, together with results from this study, highlight the importance of musculoskeletal and gastrointestinal disorders in these conditions [72,73]. These results have pointed out one more fact. Abdominal obesity, presented with the variable wc, and its related disorders of MS, is, on the one hand, highly correlated with BMI (a sign of general obesity) and, on the other hand, with mac (a sign of muscle mass). In this regard, evidence indicated that associations of MS with frailty and mortality in the population of older adults remain stable across categories of BMI, including both obese and non-obese individuals with MS [74,75]. For example, a non-obese form of MS often accompanies chronic kidney disease (and is consistent with muscle mass loss and sarcopenic frailty) [76]. Taken together, if frailty is present in individuals with MS, this acts suppressively to IL-37; otherwise, obesity-related MS may regulate IL-37 through the intensity of inflammatory challenges [77]. Thus, as shown in Figure 4, to accurately validate circulating levels of IL-37 as a diagnostic tool in T2D patients, there is the need to recognize several MS phenotypes, relating them to the degree of frailty, sarcopenia, and categories of BMI. When cut-offs of plasma IL-37 levels were considered with respect to their practical utility in identifying some specific subgroups of older diabetic patients, the best results were obtained in terms of negative prediction (exclusion) (NPV) of patients with low renal function (GFR < 45 mL/min) and those who have experienced CAD, and for positive prediction (recognition) (PPV) of those with low body mass (BMI < 25 kg/m2), those with diabetic retinopathy, and those who are characterized with comorbidities (>3). The problem with the practical usage of these results is in the fact that the estimated cut-off values belong to different parts of the IL-37 distribution curve (different quartiles) and that there is an overlap between the subgroups of diabetic patients when they are defined using one single label. For example, the subgroup with low body mass is suggestive of sarcopenic frailty. However, this can only be conclusively determined if we have insights into a more comprehensively defined profile of patients in this subgroup. Because of the complexity of older patients with T2D, using the circulating IL-37 levels as a diagnostic tool in patients with T2D will require different methodological approaches to those usually used. For example, in our recently published paper, we have suggested a profile-based approach [78]. In that study, we used the cytokine IL-17A, in addition to IL-37, as two major players of chronic inflammation and target organ damage in cardio-metabolic conditions, with the former having a pro-inflammatory role and the latter one acting as its suppressor. We have realized that these two cytokines are differently regulated, as indicated by their non-linear correlations, which may depend on the degree to which inflammation, tissue repair process, and fibrosis, take place in target organs. Many factors are likely to dictate the rates of these processes, which at least include patients’ age and gender, the time of T2D onset, being obese or not at the time of T2D onset, T2D and hypertension duration, and comorbidity-related frailty. Thus, neither IL-37 alone, or combinations of IL-17A and IL-37, can be used to predict T2D progression and bad outcomes, without knowing the wider contexts in which cytokine patterns are embedded. Moreover, variable clustering (based on their “natural” tendency to associate together), could be a way to indicate reliable patient subgroups (phenotypes) which differ from each other in risks for bad outcomes. As we have learned from previous experience, inflammatory markers are only moderate predictors in the setting of CVD, even if used in combination with each other or with other, organ-specific, and pathway-specific biomarkers [79,80]. In this study, the classical marker of inflammation, CRP, taken alone or in combination with other markers of inflammation, including IL-37, was indicated to be feasible in separating older diabetic patients according to whether they are overweight/obese or not. However, this finding is of little practical value because it remains unknown whether patients with low–normal BMI are relatively healthy diabetics or diabetics with sarcopenic frailty. Exciting results also indicate that TSH, used alone or in combination with markers of inflammation, is a potential biomarker for identifying older female diabetic patients with MS. According to both our results (Table 3, the regression model “C”) and the literature evidence, the frailty status is able to modify correlations between TSH and IL-37. In our recently published review paper, we elaborated the role of TSH—a sign of the activated hypothalamus-pituitary-thyroid (HPT) axis—in conditions associated with inflammation and impaired glucose-related metabolism, such as aging and cardio-metabolic conditions [81]. In brief, a hypothyroid state in older individuals, indicated by increased TSH, is viewed as the system’s homeostatic reaction which takes place when IL-37, as a major tissue homeostatic mechanism under inflammatory conditions, fails to restore the metabolic costs of inflammation. This explains the negative correlation between TSH and IL-37 in the regression model. In this respect, in numerous observational studies, obesity and MS have been shown to strongly correlate with TSH. In the case of the system’s homeostasis failure, such as in fully developed frailty, the HPT axis also tends to be disrupted. Thus, TSH can be used as a biomarker of older diabetic women with MS if they are separated according to the frailty status (as non-frail, pre-frail, frail). It is worth mentioning in this part of the discussion that it is still unclear how many entities of MS female diabetic patients may involve, considering variations in factors such as BMI, stages of renal function decline, age, and the presence of CVD. Obviously, there is the need to change the research approaches in complex conditions such as T2D. We are now witnessing the new wave in research on T2D, which has become possible due to technological advances in molecular biology and methods for data analysis. The focus is on identifying metabolic phenotypes or pathways that can optimally support the personalized management of diabetic patients [82,83]. In this elaborative study, we have revealed limitations of using classical data analysis approaches to determine the diagnostic utility of the cytokine IL-37 in older patients with T2D, and we have laid a foundation for introducing new methodology approaches. ## 5. Conclusions The results show that the regulation of circulating levels of IL-37 in older diabetic individuals is highly complex, mostly due to the high heterogeneity of these patients. Frailty was shown to have a suppressive effect on IL-37 circulating levels and a modifying role in associations of metabolic and inflammatory factors with IL-37, including the effect of treatments. Situations in which IL-37 attained clinically significant discriminating ability included the model of IL-37 and C-Reactive Protein in differentiating diabetic patients with low/normal–high BMI (<25/≥25 kg/m2), and the model of IL-37 and Thyroid Stimulating Hormone in discriminating between women with/without metabolic syndrome. ## References 1. Dinarello C.A.. Overview of the IL-1 family in innate inflammation and acquired immunity. Immunol. Rev.* (2018.0) 281 8-27. DOI: 10.1111/imr.12621 2. Dinarello C.A., Nold-Petry C., Nold M., Fujita M., Li S., Kim S., Bufler P.. Suppression of innate inflammation and immunity by interleukin-37. Eur. J. Immunol. (2016.0) 46 1067-1081. DOI: 10.1002/eji.201545828 3. Molgora M., Barajon I., Mantovani A., Garlanda C.. Regulatory Role of IL-1R8 in Immunity and Disease. Front. Immunol. (2016.0) 7 149. DOI: 10.3389/fimmu.2016.00149 4. Jia H., Liu J., Han B.. Reviews of Interleukin-37: Functions, Receptors, and Roles in Diseases. BioMed Res. Int. (2018.0) 2018 3058640. DOI: 10.1155/2018/3058640 5. Wang L., Quan Y., Yue Y., Heng X., Che F.. Interleukin-37: A crucial cytokine with multiple roles in disease and potentially clinical therapy. Oncol. Lett. (2018.0) 15 4711-4719. DOI: 10.3892/ol.2018.7982 6. Yan X., Xie B., Wu G., Hu J., Wang D., Cai X., Li J.. Interleukin-37: The Effect of Anti-Inflammatory Response in Human Coronary Artery Endothelial Cells. Mediat. Inflamm. (2019.0) 2019 2650590. DOI: 10.1155/2019/2650590 7. Wang P., Wang H., Li C., Zhang X., Xiu X., Teng P., Wang Z.. Dysregulation of microRNA-657 influences inflammatory response via targeting interleukin-37 in gestational diabetes mellitus. J. Cell. Physiol. (2019.0) 234 7141-7148. DOI: 10.1002/jcp.27468 8. Zhao M., Li Y., Guo C., Wang L., Chu H., Zhu F., Li Y., Wang X., Wang Q., Zhao W.. IL-37 isoform D downregulates pro-inflammatory cytokines expression in a Smad3-dependent manner. Cell Death Dis. (2018.0) 9 582. DOI: 10.1038/s41419-018-0664-0 9. Cavalli G., Tengesdal I.W., Gresnigt M., Nemkov T., Arts R., Domínguez-Andrés J., Molteni R., Stefanoni D., Cantoni E., Cassina L.. The anti-inflammatory cytokine interleukin-37 is an inhibitor of trained immunity. Cell Rep. (2021.0) 35 108955. DOI: 10.1016/j.celrep.2021.108955 10. Cavalli G., Justice J.N., Boyle K.E., D’Alessandro A., Eisenmesser E.Z., Herrera J.J., Hansen K.C., Nemkov T., Stienstra R., Garlanda C.. Interleukin 37 reverses the metabolic cost of inflammation, increases oxidative respiration, and improves exercise tolerance. Proc. Natl. Acad. Sci. USA (2017.0) 114 2313-2318. DOI: 10.1073/pnas.1619011114 11. Bufler P., Gamboni-Robertson F., Azam T., Kim S.H., Dinarello C.A.. Interleukin-1 homologues IL-1F7b and IL-18 contain functional mRNA instability elements within the coding region responsive to lipopolysaccharide. Biochem. J. (2004.0) 381 503-510. DOI: 10.1042/BJ20040217 12. McNamee E.N., Masterson J.C., Jedlicka P., McManus M., Grenz A., Collins C.B., Nold M.F., Nold-Petry C., Bufler P., Dinarello C.A.. Interleukin 37 expression protects mice from colitis. Proc. Natl. Acad. Sci. USA (2011.0) 108 16711-16716. DOI: 10.1073/pnas.1111982108 13. Zeng H., Zhou K., Ye Z.. Biology of interleukin-37 and its role in autoimmune diseases (Review). Exp. Ther. Med. (2022.0) 24 495. DOI: 10.3892/etm.2022.11422 14. Zhang L., Zhang J., Gao P.. The potential of interleukin-37 as an effective therapeutic agent in asthma. Respir. Res. (2017.0) 18 192. DOI: 10.1186/s12931-017-0675-x 15. Li Y., Wang Y., Liu Y., Wang Y., Zuo X., Li Y., Lu X.. The possible role of the novel cytokines il-35 and il-37 in inflammatory bowel disease. Mediat. Inflamm. (2014.0) 2014 136329. DOI: 10.1155/2014/136329 16. Kim M.S., Baek A.R., Lee J.H., Jang A.S., Kim D.J., Chin S.S., Park S.W.. IL-37 Attenuates Lung Fibrosis by Inducing Autophagy and Regulating TGF-β1 Production in Mice. J. Immunol. (2019.0) 203 2265-2275. DOI: 10.4049/jimmunol.1801515 17. Yan J., Zhang Y., Cheng S., Kang B., Peng J., Zhang X., Yuan M., Chu W., Zhang W., Shen J.. Common genetic heterogeneity of human interleukin-37 leads to functional variance. Cell. Mol. Immunol. (2017.0) 14 783-791. DOI: 10.1038/cmi.2016.48 18. Yin D., Naji D.H., Xia Y., Li S., Bai Y., Jiang G., Zhao Y., Wang X., Huang Y., Liu Y.. Genomic Variant in IL-37 Confers A Significant Risk of Coronary Artery Disease. Sci. Rep. (2017.0) 7 42175. DOI: 10.1038/srep42175 19. Ramos-Lopez O., Milagro F.I., Riezu-Boj J.I., Martinez J.A.. Epigenetic signatures underlying inflammation: An interplay of nutrition, physical activity, metabolic diseases, and environmental factors for personalized nutrition. Inflamm. Res. (2021.0) 70 29-49. DOI: 10.1007/s00011-020-01425-y 20. Roden M., Shulman G.I.. The integrative biology of type 2 diabetes. Nature (2019.0) 576 51-60. DOI: 10.1038/s41586-019-1797-8 21. López-Bautista F., Posadas-Sánchez R., Vázquez-Vázquez C., Fragoso J.M., Rodríguez-Pérez J.M., Vargas-Alarcón G.. IL-37 Gene and Cholesterol Metabolism: Association of Polymorphisms with the Presence of Hypercholesterolemia and Cardiovascular Risk Factors. The GEA Mexican Study. Biomolecules (2020.0) 10. DOI: 10.3390/biom10101409 22. Franceschi C., Garagnani P., Parini P., Giuliani C., Santoro A.. Inflammaging: A new immune-metabolic viewpoint for age-related diseases. Nat. Rev. Endocrinol. (2018.0) 14 576-590. DOI: 10.1038/s41574-018-0059-4 23. Russo S., Kwiatkowski M., Govorukhina N., Bischoff R., Melgert B.N.. Meta-Inflammation and Metabolic Reprogramming of Macrophages in Diabetes and Obesity: The Importance of Metabolites. Front. Immunol. (2021.0) 12 746151. DOI: 10.3389/fimmu.2021.746151 24. Wang X., Antony V., Wang Y., Wu G., Liang G.. Pattern recognition receptor-mediated inflammation in diabetic vascular complications. Med. Res. Rev. (2020.0) 40 2466-2484. DOI: 10.1002/med.21711 25. Zhang S., Gang X., Yang S., Cui M., Sun L., Li Z., Wang G.. The Alterations in and the Role of the Th17/Treg Balance in Metabolic Diseases. Front. Immunol. (2021.0) 12 678355. DOI: 10.3389/fimmu.2021.678355 26. Loftus R.M., Finlay D.K.. Immunometabolism: Cellular Metabolism Turns Immune Regulator. J. Biol. Chem. (2016.0) 291 1-10. DOI: 10.1074/jbc.R115.693903 27. Ballak D.B., Li S., Cavalli G., Stahl J.L., Tengesdal I.W., van Diepen J.A., Klück V., Swartzwelter B., Azam T., Tack C.J.. Interleukin-37 treatment of mice with metabolic syndrome improves insulin sensitivity and reduces pro-inflammatory cytokine production in adipose tissue. J. Biol. Chem. (2018.0) 293 14224-14236. DOI: 10.1074/jbc.RA118.003698 28. Zhang X., Zhu Y., Zhou Y., Fei B.. Interleukin 37 (IL-37) Reduces High Glucose-Induced Inflammation, Oxidative Stress, and Apoptosis of Podocytes by Inhibiting the STAT3-Cyclophilin A (CypA) Signaling Pathway. Med. Sci. Monit. (2020.0) 26 e922979. DOI: 10.12659/MSM.922979 29. Liu J., Lin J., He S., Wu C., Wang B., Liu J., Duan Y., Liu T., Shan S., Yang K.. Transgenic Overexpression of IL-37 Protects Against Atherosclerosis and Strengthens Plaque Stability. Cell Physiol. Biochem. (2018.0) 45 1034-1050. DOI: 10.1159/000487344 30. Ye J., Wang Y., Wang Z., Lin Y., Liu L., Zhou Q., Wang M., Xu Y., Ye D., Zhang J.. Circulating IL-37 levels are elevated in patients with hypertension. Exp. Ther. Med. (2021.0) 21 558. DOI: 10.3892/etm.2021.9990 31. Shou X., Lin J., Xie C., Wang Y., Sun C.. Plasma IL-37 Elevated in Patients with Chronic Heart Failure and Predicted Major Adverse Cardiac Events: A 1-Year Follow-Up Study. Dis. Mark. (2017.0) 2017 9134079. DOI: 10.1155/2017/9134079 32. Liu C., Cui Y., Zhang D., Tian X., Zhang H.. Analysis of Serum Interleukin-37 Level and Prognosis in Patients with ACS. Comput. Math. Methods Med. (2021.0) 2021 3755458. DOI: 10.1155/2021/3755458 33. Santarelli D.M., Vincent F.B., Rudloff I., Nold-Petry C.A., Nold M.F., Russo M.A.. Circulating Interleukin-37 Levels in Healthy Adult Humans—Establishing a Reference Range. Front. Immunol. (2021.0) 12 708425. DOI: 10.3389/fimmu.2021.708425 34. Zoungas S., Woodward M., Li Q., Cooper M.E., Hamet P., Harrap S., Heller S., Marre M., Patel A., Poulter N.. ADVANCE Collaborative group. Impact of age, age at diagnosis and duration of diabetes on the risk of macrovascular and microvascular complications and death in type 2 diabetes. Diabetologia (2014.0) 57 2465-2474. DOI: 10.1007/s00125-014-3369-7 35. McCarthy M.I.. Painting a new picture of personalised medicine for diabetes [published correction appears in Diabetologia. Diabetologia (2017.0) 60 793-799. DOI: 10.1007/s00125-017-4210-x 36. Fried L.P., Tangen C.M., Walston J.. Cardiovascular health study collaborative research group. Frailty in older adults: Evidence for a phenotype. J. Gerontol. A Biol. Sci. Med. Sci. (2001.0) 56 M146-M156. DOI: 10.1093/gerona/56.3.M146 37. Kalan U., Arik F., Isik A.T., Soysal P.. Nutritional profiles of older adults according the Mini-Nutritional Assessment. Aging Clin. Exp. Res. (2020.0) 32 673-680. DOI: 10.1007/s40520-019-01235-1 38. Landi F., Russo A., Liperoti R., Pahor M., Tosato M., Capoluongo E., Bernabei R., Onder G.. Midarm muscle circumference, physical performance and mortality: Results from the aging and longevity study in the Sirente geographic area ilSIRENTE study. Clin. Nutr. (2010.0) 29 441-447. DOI: 10.1016/j.clnu.2009.12.006 39. Cosentino F., Grant P.J., Aboyans V., Bailey C.J., Ceriello A., Delgado V., Federici M., Filippatos G., Grobbee D.E., Hansen T.B.. 2019 ESC Guidelines on diabetes, pre-diabetes, and cardiovascular diseases developed in collaboration with the EASD. Eur. Heart J. (2020.0) 41 255-323. DOI: 10.1093/eurheartj/ehz486 40. Huang P.L.. A comprehensive definition for metabolic syndrome. Dis. Model. Mech. (2009.0) 2 231-237. DOI: 10.1242/dmm.001180 41. Han C., He X., Xia X., Li Y., Shi X., Shan Z., Teng W.. Subclinical Hypothyroidism and Type 2 Diabetes: A Systematic Review and Meta-Analysis. PLoS ONE (2015.0) 10. DOI: 10.1371/journal.pone.0135233 42. National Kidney Fundation. eGFR Calculator 43. Levey A.S.. A decade after the KDOQI CDK guidelines. Am. J. Kidney Dis. Off. J. Natl. Kidney Found. (2012.0) 60 683-685. DOI: 10.1053/j.ajkd.2012.08.019 44. McDonagh T.A., Metra M., Adamo M., Gardner R.S., Baumbach A., Böhm M., Burri H., Butler J., Čelutkienė J., Chioncel O.. ESC Scientific Document Group. 2021 ESC Guidelines for the diagnosis and treatment of acute and chronic heart failure. Eur. Heart J. (2021.0) 42 3599-3726. DOI: 10.1093/eurheartj/ehab368 45. Visseren F.L.J., Mach F., Smulders Y.M., Carballo D., Koskinas K.C., Bäck M., Benetos A., Biffi A., Boavida J.-M., Capodanno D.. ESC National Cardiac Societies; ESC Scientific Document Group. 2021 ESC Guidelines on cardiovascular disease prevention in clinical practice. Eur. Heart J. (2021.0) 42 3227-3337. DOI: 10.1093/eurheartj/ehab484 46. Buse J.B., Wexler D.J., Tsapas A., Rossing P., Mingrone G., Mathieu C., D’Alessio D.A., Davies M.J.. 2019 update to: Management of hyperglycaemia in type 2 diabetes, 2018. A consensus report by the American Diabetes Association (ADA) and the European Association for the Study of Diabetes (EASD). Diabetologia (2020.0) 63 221-228. DOI: 10.1007/s00125-019-05039-w 47. Angkananard T., Anothaisintawee T., McEvoy M., Attia J., Thakkinstian A.. Neutrophil lymphocyte ratio and cardiovascular disease risk: A systematic review and meta-analysis. BioMed Res. Int. (2018.0) 2018 2703518. DOI: 10.1155/2018/2703518 48. Weiss G., Ganz T., Goodnough L.T.. Anemia of inflammation. Blood (2019.0) 133 40-50. DOI: 10.1182/blood-2018-06-856500 49. Smidowicz A., Regula J.. Effect of nutritional status and dietary patterns on human serum C-reactive protein and interleukin-6 concentrations. Adv. Nutr. (2015.0) 6 738-747. DOI: 10.3945/an.115.009415 50. Hajian-Tilaki K.. Receiver Operating Characteristic (ROC) curve analysis for medical diagnostic test evaluation. Caspian J. Intern. Med. (2013.0) 4 627-635. PMID: 24009950 51. Carter J.V., Pan J., Rai S.N., Galandiuk S.. ROC-ing along: Evaluation and interpretation of receiver operating characteristic curves. Surgery (2016.0) 159 1638-1645. DOI: 10.1016/j.surg.2015.12.029 52. Sun J., Han K., Xu M., Li L., Qian J., Li L., Li X.. Blood Viscosity in Subjects With Type 2 Diabetes Mellitus: Roles of Hyperglycemia and Elevated Plasma Fibrinogen. Front. Physiol. (2022.0) 25 827428. DOI: 10.3389/fphys.2022.827428 53. Tomovic K., Lazarevic J., Kocic G., Deljanin-Ilic M., Anderluh M., Smelcerovic A.. Mechanisms and pathways of anti-inflammatory activity of DPP-4 inhibitors in cardiovascular and renal protection. Med. Res. Rev. (2019.0) 39 404-422. DOI: 10.1002/med.21513 54. Nauck M.A., Meininger G., Sheng D., Terranella L., Stein P.P.. Efficacy and safety of the dipeptidyl peptidase-4 inhibitor, sitagliptin, compared with the sulfonylurea, glipizide, in patients with type 2 diabetes inadequately controlled on metformin alone: A randomized, double-blind, non-inferiority trial. Diabetes Obes. Metab. (2007.0) 9 194-205. DOI: 10.1111/j.1463-1326.2006.00704.x 55. Koushki K., Shahbaz S.K., Mashayekhi K., Sadeghi M., Zayeri Z.D., Taba M.Y., Banach M., Al-Rasadi K., Johnston T.P., Sahebkar A.. Anti-inflammatory Action of Statins in Cardiovascular Disease: The Role of Inflammasome and Toll-Like Receptor Pathways. Clin. Rev. Allergy Immunol. (2021.0) 60 175-199. DOI: 10.1007/s12016-020-08791-9 56. Bellary S., Kyrou I., Brown J.E., Bailey C.J.. Type 2 diabetes mellitus in older adults: Clinical considerations and management. Nature reviews. Endocrinology (2021.0) 17 534-548. PMID: 34172940 57. Pothier K., Gana W., Bailly N., Fougère B.. Associations Between Frailty and Inflammation, Physical, and Psycho-Social Health in Older Adults: A Systematic Review. Front. Psychol. (2022.0) 13 805501. DOI: 10.3389/fpsyg.2022.805501 58. Strain W.D., Down S., Brown P., Puttanna A., Sinclair A.. Diabetes and Frailty: An Expert Consensus Statement on the Management of Older Adults with Type 2 Diabetes. Diabetes Ther. (2021.0) 12 1227-1247. DOI: 10.1007/s13300-021-01035-9 59. Fried L.P., Tangen C.M., Walston J., Newman A.B., Hirsch C., Gottdiener J.. Cardiovascular Health Study Collaborative Research Group. Frailty in older adults: Evidence for a phenotype. J. Gerontol. A Biol. Sci. Med. Sci. (2001.0) 56 M146-M156. DOI: 10.1093/gerona/56.3.M146 60. Nwaneri C., Cooper H., Bowen-Jones D.. Mortality in type 2 diabetes mellitus: Magnitude of the evidence from a systematic review and meta-analysis. Br. J. Diabetes Vasc. Dis. (2013.0) 13 192-207. DOI: 10.1177/1474651413495703 61. Kleipool E.E.F., Hoogendijk E.O., Trappenburg M.C., Handoko M.L., Huisman M., Peters M.J., Muller M.. Frailty in older adults with cardiovascular disease: Cause, effect or both?. Aging Dis. (2018.0) 9 489-497. DOI: 10.14336/AD.2017.1125 62. Ekram A.R.M.S., Espinoza S., Ernst M., Beilin L., Stocks N.P., Ryan J., Woods R.. The Association Between Metabolic Syndrome and Frailty in Healthy Community-Dwelling Older Adults. Innov. Aging (2021.0) 5 531-532. DOI: 10.1093/geroni/igab046.2049 63. Brunt V.E., Ikoba A.P., Ziemba B.P., Ballak D.B., Hoischen A., Dinarello C.A., Ehringer M.A., Seals D.R.. Circulating interleukin-37 declines with aging in healthy humans: Relations to healthspan indicators and IL37 gene SNPs. GeroScience (2022.0) 1-20. DOI: 10.1007/s11357-022-00587-3 64. O’Caoimh R., Galluzzo L., Rodriguez-Laso A., Van Der Heyden J., Ranhoff A.H., Lamprini-Koula M., Ciutan M., López-Samaniego L., Carcaillon-Bentata L., Kennelly S.. Prevalence of frailty at population level in European ADVANTAGE Joint Action Member States: A systematic review and meta-analysis. Ann. Ist. Super. Sanit. (2018.0) 54 226-238 65. Vetrano D.L., Palmer K., Marengoni A., Marzetti E., Lattanzio F., Roller-Wirnsberger R., Samaniego L.L., Rodríguez-Mañas L., Bernabei R., Onder G.. Joint Action ADVANTAGE WP4 Group. Frailty and Multimorbidity: A Systematic Review and Meta-analysis. J. Gerontol. Biol. Sci. Med. Sci. (2019.0) 74 659-666. DOI: 10.1093/gerona/gly110 66. Yuan L., Chang M., Wang J.. Abdominal obesity, body mass index and the risk of frailty in community-dwelling older adults: A systematic review and meta-analysis. Age Ageing (2021.0) 50 1118-1128. DOI: 10.1093/ageing/afab039 67. Amblàs-Novellas J., A Murray S., Oller R., Torné A., Martori J.C., Moine S., Latorre-Vallbona N., Espaulella J., Santaeugènia S.J., Gómez-Batiste X.. Frailty degree and illness trajectories in older people towards the end-of-life: A prospective observational study. BMJ Open (2021.0) 11 e042645. DOI: 10.1136/bmjopen-2020-042645 68. Abdelhafiz A.H., Emmerton D., Sinclair A.J.. Impact of frailty metabolic phenotypes on the management of older people with type 2 diabetes mellitus. Geriatr. Gerontol. Int. (2021.0) 1 614-622. DOI: 10.1111/ggi.14214 69. Wu J.D., Liang D.L., Xie Y.. Prediabetes and risk of heart failure: The link grows stronger. Cardiovasc. Diabetol. (2021.0) 20 112. DOI: 10.1186/s12933-021-01302-w 70. Wu P.Y., Chao C.T., Chan D.C., Huang J.W., Hung K.Y.. Contributors, risk associates, and complications of frailty in patients with chronic kidney disease: A scoping review. Ther. Adv. Chronic Dis. (2019.0) 5 382. DOI: 10.1177/2040622319880382 71. Bekić S., Babič F., Pavlišková V., Paralič J., Wittlinger T., Majnarić L.T.. Clusters of Physical Frailty and Cognitive Impairment and Their Associated Comorbidities in Older Primary Care Patients. Healthcare (2021.0) 9. DOI: 10.3390/healthcare9070891 72. Tembo M.C., Mohebbi M., Holloway-Kew K.L., Gaston J., Sui S.X., Brennan-Olsen S.L., Williams L.J., Kotowicz M.A., Pasco J.A.. The contribution of musculoskeletal factors to physical frailty: A cross-sectional study. BMC Musculoskelet Disord. (2021.0) 22. DOI: 10.1186/s12891-021-04795-4 73. Li C., Li Y., Wang N., Ge Z., Shi Z., Wang J., Ding B., Bi Y., Wang Y., Hong Z.. Intestinal Permeability Associated with the Loss of Skeletal Muscle Strength in Middle-Aged and Older Adults in Rural Area of Beijing, China. Healthcare (2022.0) 10. DOI: 10.3390/healthcare10061100 74. Afonso C., Sousa-Santos A.R., Santos A., Borges N., Padrão P., Moreira P., Amaral T.F.. Frailty status is related to general and abdominal obesity in older adults. Nutr. Res. (2021.0) 85 21-30. DOI: 10.1016/j.nutres.2020.10.009 75. de Hollander E.L., Bemelmans W.J., Boshuizen H.C., Friedrich N., Wallaschofski H., Guallar-Castillón P., Walter S., Zillikens M.C., Rosengren A., Lissner L.. The association between waist circumference and risk of mortality considering body mass index in 65- to 74-year-olds: A meta-analysis of 29 cohorts involving more than 58 000 elderly persons. Int. J. Epidemiol. (2022.0) 41 805-817. DOI: 10.1093/ije/dys008 76. Ciardullo S., Ballabeni C., Trevisan R., Perseghin G.. Metabolic Syndrome, and Not Obesity, Is Associated with Chronic Kidney Disease. Am. J. Nephrol. (2021.0) 52 666-672. DOI: 10.1159/000518111 77. Rehman K., Akash M.S.. Mechanisms of inflammatory responses and development of insulin resistance: How are they interlinked?. J. Biomed. Sci. (2016.0) 23 87. DOI: 10.1186/s12929-016-0303-y 78. Bosnic Z., Yildirim P., Babič F., Šahinović I., Wittlinger T., Martinović I., Majnaric L.T.. Clustering Inflammatory Markers with Sociodemographic and Clinical Characteristics of Patients with Diabetes Type 2 Can Support Family Physicians’ Clinical Reasoning by Reducing Patients’ Complexity. Healthcare (2021.0) 9. DOI: 10.3390/healthcare9121687 79. Kritharides L.. Inflammatory markers and outcomes in cardiovascular disease. PLoS Med. (2009.0) 6. DOI: 10.1371/journal.pmed.1000147 80. Zethelius B., Berglund L., Sundström J., Ingelsson E., Basu S., Larsson A., Venge P., Arnlöv J.. Use of multiple biomarkers to improve the prediction of death from cardiovascular causes. N. Engl. J. Med. (2008.0) 358 2107-2116. DOI: 10.1056/NEJMoa0707064 81. Majnarić L.T., Bosnić Z., Štefanić M., Wittlinger T.. Cross-Talk between the Cytokine IL-37 and Thyroid Hormones in Modulating Chronic Inflammation Associated with Target Organ Damage in Age-Related Metabolic and Vascular Conditions. Int. J. Mol. Sci. (2022.0) 23. DOI: 10.3390/ijms23126456 82. Li L., Cheng W.Y., Glicksberg B.S., Gottesman O., Tamler R., Chen R., Bottinger E.P., Dudley J.T.. Identification of type 2 diabetes subgroups through topological analysis of patient similarity. Sci. Transl. Med. (2015.0) 7 311ra174. DOI: 10.1126/scitranslmed.aaa9364 83. Wesolowska-Andersen A., Brorsson C.A., Bizzotto R., Mari A., Tura A., Koivula R., Mahajan A., Vinuela A., Tajes J.F., Sharma S.. IMI DIRECT Consortium. Four groups of type 2 diabetes contribute to the etiological and clinical heterogeneity in newly diagnosed individuals: An IMI DIRECT study. Cell reports. Medicine (2022.0) 3 100477. PMID: 35106505
--- title: Kidney Function, Male Gender, and Aneurysm Diameter Are Predictors of Acute Kidney Injury in Patients with Abdominal Aortic Aneurysms Treated Endovascularly authors: - Bartłomiej Antoń - Sławomir Nazarewski - Jolanta Małyszko journal: Toxins year: 2023 pmcid: PMC9966909 doi: 10.3390/toxins15020130 license: CC BY 4.0 --- # Kidney Function, Male Gender, and Aneurysm Diameter Are Predictors of Acute Kidney Injury in Patients with Abdominal Aortic Aneurysms Treated Endovascularly ## Abstract Abdominal aortic aneurysm (AAA) is a degenerative disease of the aortic wall with potentially fatal complications. The widespread adoption of endovascular aneurysm repair (EVAR), which is less invasive and equally (if not more) effective for abdominal aortic aneurysms (AAA), is due to the obvious advantages of the procedure compared to the traditional open repair. As the popularity of endovascular procedures grows, related complications become more evident, with kidney damage being one of them. Although acute kidney injury following EVAR is relatively common, its true incidence is still uncertain. The purpose of this study was to assess the incidence of acute kidney injury among patients treated with endovascular repair of ruptured AAA. In addition, we aimed to determine the predictors of PC-AKI in patients with abdominal aortic aneurysm treated with EVAR. Patients and Methods: We retrospectively analyzed a prospective registry of abdominal aortic aneurysm of 247 patients operated endovascularly at a single center between 2015 and 2021. Due to a lack of clinical data, data of 192 patients were reviewed for postcontrast acute kidney injury. Additional comorbidities were included in this study: hypertension, diabetes mellitus, atrial fibrillation, chronic coronary syndrome, COPD, and chronic kidney disease. Follow-up examinations were performed before the procedure and 48 h after contrast administration. Results: The group of 36 patients developed PC-AKI, which is $19\%$ of the entire study population. Hypertension, diabetes, chronic kidney disease, male gender, and incidence of PC-AKI were more prevalent in patients with higher aortic aneurysm diameter ≥67 mm. In multiple regression analyses, independent predictors of PC-AKI were serum creatinine, chronic kidney disease, male gender, and aortic aneurysm diameter ≥67 mm. Conclusions: One of the major complications after EVAR is acute kidney injury, which is linked to higher death and morbidity rates. Independent risk factors for postcontrast acute kidney injury were chronic kidney disease, male gender, and aortic diameter. Only aortic diameter could be modifiable risk factor, and earlier surgery could be considered to yield better outcomes. More research is critically needed to determine how AKI affects long-term outcomes and to look at preventive options. ## 1. Introduction In clinical medicine, contrast agents are frequently used to enhance the visibility of interior organs and structures in X-ray-based imaging procedures including computed tomography (CT) and radiography. High contrast density is achieved by iodine contrast media, which contain one or two tri-iodobenzene rings. Iodine-based contrast agents are often categorized into iso-, low-, and high-osmolar substances based on their osmolarity. In recent years, the use of contrast agents has increased, particularly for several diagnostic imaging procedures such as angiography [1]. From minor symptoms, such as itching, to major and even fatal reactions, such anaphylactic reactions, radiographic contrast media can cause a variety of adverse reactions. Kidney failure is known to be one of the main side consequences of administering contrast media. When contrast medium is administered intravenously or intraarterially, a serious response known as contrast-induced nephropathy (CIN) can occur. CIN is defined as an acute renal failure with no other cause that results in a serum creatinine rise of at least $25\%$ (or 44 mol/L or 0.5 mg/dL) following the administration of contrast agents [2]. The criteria used for postcontrast acute kidney injury (PC-AKI) research, developed by the Kidney Disease Improving Global Outcome (KDIGO) initiative, are more accurately derived than those utilized by the CIN definition. KDIGO defines PC-AKI as acute kidney injury caused secondary to contrast administration. The definition of AKI provides one of the following: an increase in SCr of ≥0.3 mg/dL (≥26.5 µmol/L) within 48 h; an increase in SCr of ≥1.5 times baseline that is known or presumed to have occurred within the previous 7 days; or an increase in urine volume of 0.5 mL/kg/h for 6 h. AKI can be divided into three stages. Stage one requires one of the following: a reduction in urine production to <0.5 mL/kg/h during a 6 h block, increase in SCr by ≥0.3 mg/dL (≥26.5 µmol/L) within 48 h, or increase in SCr to ≥1.5–1.9 times baseline, which is known or presumed to have occurred within the prior 7 days. Stage two demands one of the following: serum creatinine increase ≥ 2.0–2.9 times baseline or reduction in urine production to <0.5 mL/kg/h during two 6 h blocks. Stage three finally is defined by one of the following: serum creatinine increase ≥3.0 times baseline, serum creatinine increase to >4.0 mg/dL (353 µmol/L), initiation of renal replacement therapy, reduction in urine production to <0.3 mL/kg/h during more than 24 h, or anuria for more than 12 h [3]. It has been demonstrated that factors such as advanced age, diabetes mellitus, pre-existing renal impairment, intravascular volume depletion after surgery, congestive heart failure, or concurrent use of other nephrotoxic medicines enhance the chance of developing PC-AKI [4,5,6,7,8]. Additionally, sarcopenia—musculoskeletal loss was associated with AKI occurrence after AAA repair—is known to be an independent mortality factor following AAA therapy. According to studies [9,10] PC-AKI is the third most common reason for hospital-acquired acute renal failure and is linked to a high mortality rate. Due to this, a number of methods (such as pre and posthydration and the injection of N-acetylcysteine) have been shown to be somewhat useful in preventing CIN [11], but they are not always practical, particularly when treating patients who have suffered serious injuries. Additionally, there are other potential causes of the creatinine rise in individuals who have had serious injuries (e.g., contrast-induced, haemorrhagic shock, blood transfusions, advanced age). Abdominal aortic aneurysm (AAA) is a degenerative disease of the aortic wall with potentially fatal complications. The widespread adoption of endovascular aneurysm repair (EVAR) for AAA is due to the obvious advantages of the procedure compared to the traditional open repair. However, these advantages have to be weighed against the increased risk of renal dysfunction with EVAR [12,13,14]. Acute kidney injury (AKI) is a severe complication after infrarenal abdominal aortic aneurysm repair [15]. Although acute kidney injury following EVAR is relatively common, its true incidence is still uncertain. The purpose of this study was to assess the incidence of acute kidney injury among patients treated with endovascular repair of ruptured AAA. In addition, we aimed to determine the predictors of PC-AKI in patients with abdominal aortic aneurysm treated with EVAR. ## 2. Results The data of 192 patients were analyzed. The median age of those patients was 73 years. The proportion of female patients was $24\%$. The group of 36 patients developed PC-AKI, which is $19\%$ of the entire study population. Mean iodine contrast volume was 149.6 mL in the whole population, 179.3 mL in the PC-AKI population, and 142.1 mL in the non-PC-AKI population. Mean aortic diameter for all study participants was 57.2 mm. Aortic diameter in the population with PC-AKI was 66.9 mm, and in the population without PC-AKI, it was 55.7 mm. Chronic kidney disease pre-existed in 67 of all study participants: 24 patients with PC-AKI and 43 patients without PC-AKI. Table 1 shows that in the male population, pre-existing chronic kidney disease and blood urea nitrogen concentration were significantly higher in the PC-AKI group ($p \leq 0.001$). In Figure 1, receiver operating curves for aortic aneurysm diameter in predicting postcontrast-induced kidney injury are presented for all patients and for CKD patients. In Table 2, we compared patients in relation to the aortic aneurysm diameter (cut off ≥67 mm). Hypertension, diabetes, chronic kidney disease, male gender, and incidence of PC-AKI were more prevalent in patients with higher aortic aneurysm diameter ≥67 mm. In addition, in multiple regression analyses independent predictors of PC-AKI were serum creatinine, chronic kidney disease, male gender, and aortic aneurysm diameter ≥67 mm. The correlations between eGFR and iodine contrast volume, as well as with aortic aneurysm diameter, determined using Pearson’s correlation, are presented as rank correlation (r) in Figure 2. In Table 3, we assess that PC-AKI is more common in the group with aortic diameter greater than the cut-off point (67 mm) (OR 1.364) with $$p \leq 0.01.$$ ## 3. Discussion Nowadays, due to its clear advantages in reducing perioperative morbidity and mortality, EVAR has become the standard procedure for AAA repair [12,13,14]. This study supports previous findings that EVAR has a significant impact on renal function. Together with other published studies that utilized defined criteria to identify AKI, it can be said that $15\%$ to $20\%$ of EVAR patients develop AKI in the immediate postoperative interval. [ 15,16,17,18,19] Saratzis et al. divided a variety of AKIs after EVAR, which are: microembolization, suprarenal fixation, accessory renal artery blockage, and CIN, as well as the related inflammatory and ischemic reactions [19]. This study corresponds with the analysis of Lee et al. that the main cause of post-EVAR AKI is related to contrast due to association of contrast dose and AKI frequency [20]. In contrast-induced nephropathy, the administration of contrast results in an increase in vasoconstrictive forces, a reduction in local prostaglandin and nitric oxide-mediated vasodilatation, a direct toxic effect of oxygen free radicals on renal tubular cells, an increase in oxygen consumption, an increase in intratubular pressure due to contrast-induced diuresis, an increase in urinary viscosity, and tubular obstruction. As with other studies [15] this analysis confirms that AKI is more common in people with decreased renal function prior to EVAR (baseline). Those who developed AKI in this group had a substantially lower eGFR at baseline. This study does not include perioperative hydration as a prophylaxis of PC-AKI, although proper patient preparation prior to EVAR utilizing hydration with sodium bicarbonates and N-acetylcysteine seems to offer better postoperation outcomes [15]. Due to a lack of data supporting the usage of N-acetylcysteine or hydration with sodium bicarbonates [21], more research is diligently needed to evaluate additional strategies as well as the best strategy to provide hydration to this group of patients. Compared to data analyzed by Krasznai et al. [ 22] where the incidence of PC-AKI amongst the population with eGFR < 30 mL/min/1.73 m² is more than $50\%$, in this study $64\%$ of patients with eGFR < 30 mL/min/1.73 m² developed AKI. It is well recognized that contrast media can result in renal vasoconstriction, hypoxia damage, and acute tubular necrosis. [ 23,24] Contrast volume between patients who developed PC-AKI (179.3cc) versus patients who did not develop AKI (142.1cc) was much less dispersed compared with Mun et al. [ 25] in whose study it was 249.17cc in PC-AKI patients and 179.43cc in non-PC-AKI patients. Although the Dutch Randomized Endovascular Aneurysm Management (DREAM) trial demonstrated that post Hosmer-Lemeshow operative changes in SCr were comparable with no statistical differences in the incidence of AKI (as per their definition) or need for dialysis, [26] this study significantly demonstrates that adequate patient preparation prior to the procedure as well as management perioperatively aid in the prevention of PC-AKI. Other authors, such as Rastogi et al. [ 27], estimated that the aneurysmal neck was an important factor in the development of PC-AKI, and the wider it is, the higher the incidence of PC-AKI. This study shows that the aneurysm diameter itself was also connected with the rate of PC-AKI development, and in patients with an aneurysm diameter higher than 67 mm it was $30\%$, and in those with diameter lower than 67 mm it was $8.1\%$. In addition, in the recent paper by Arbănași et al. [ 28] the diameter of the abdominal aorta at different levels has better accuracy and a higher predictive role of rupture than the maximal diameter of AAA. ## 4. Conclusions One of the major complications after EVAR is acute kidney injury, which is linked to higher death and morbidity rates. This study showed that pre-existing chronic kidney disease is an independent risk factor for postcontrast acute kidney injury. Additionally, male sex and aortic diameter are other independent factors in development of PC-AKI. Only aortic diameter could be a modifiable risk factor, and earlier surgery could be considered to yield better outcomes. More research is critically needed to determine how AKI affects long-term outcomes and to look at preventive options. ## 5. Materials and Methods This research is a retrospective analysis based on a prospective registry of abdominal aortic aneurysm patients that was conducted at a single center between 2015 and 2021. A group of 247 patients with AAA was operated endovascularly. Due to a lack of clinical data, 55 participants were omitted from this study. In total, data from 192 patients were reviewed for postcontrast acute kidney injury. Additional comorbidities were included in this study: hypertension, diabetes mellitus, atrial fibrillation, chronic coronary syndrome, COPD, and chronic kidney disease. Follow-up examinations were performed before the procedure and 48 h after contrast administration. ## Statistical Analysis Normality of the distribution was assessed with Shapiro–Wilk test. The characteristics of the studied population are presented as means with standard deviations (SD) for normally distributed continuous, and as the number of cases and percentage (for categorical variables). Statistical significance of differences between two groups were determined using the χ2 and Mann–Whitney U tests when appropriate. The correlations between psychical variables and eGFR were determined using Pearson’s correlation. Data are presented as rank correlation (r). The threshold of statistical significance for all tests was set at $p \leq 0.05.$ All analyses were performed using MS Excel (Microsoft, 2020, version 16.40, Redmond, WA, USA) and XL Stat (Addinsoft, 2020, version 2020.03.01, New York, NY, USA). For the purpose of this study, PC-AKI was defined as an absolute increase of serum creatinine ≥0.3 mg/dL or a relative increase ≥$150\%$ from baseline value within the first 48–72 h of intervention. ## References 1. Boozari M., Hosseinzadeh H.. Preventing contrast-induced nephropathy (CIN) with herbal medicines: A review. Phytother. Res. (2021) 35 1130-1146. DOI: 10.1002/ptr.6880 2. Stacul F., van der Molen A.J., Reimer P., Webb J.A.W., Thomsen H.S., Morcos S.K., Almén T., Aspelin P., Bellin M.-F., Clement O.. On behalf of the Contrast Media Safety Committee of the European Society of Urogenital Radiology. Contrast induced nephropathy: Updated ESUR Contrast Media Safety Committee guidelines. Eur. Radiol. (2011) 21 2527-2541. DOI: 10.1007/s00330-011-2225-0 3. KDIGO Clinical Practice Guideline for Acute Kidney Injury. Summary of Recommendation Statements. Kidney Int. Suppl. (2012) 2 8-12. DOI: 10.1038/kisup.2012.7 4. Chong E., Poh K.K., Liang S., Tan H.C.. Risk factors and clinical outcomes for contrast-induced nephropathy after percutaneous coronary intervention in patients with normal serum creatinine. Ann. Acad. Med. Singap. (2010) 39 374-380. DOI: 10.47102/annals-acadmedsg.V39N5p374 5. Hipp A., Desai S., Lopez C., Sinert R.. The incidence of contrast-induced nephropathy in trauma patients. Eur. J. Emerg. Med. (2008) 15 134-139. DOI: 10.1097/MEJ.0b013e328270367d 6. Owen R.J., Hiremath S., Myers A., Fraser-Hill M., Barrett B.J.. Canadian Association of Radiologists consensus guidelines for the prevention of contrast-induced nephropathy: Update 2012. Can. Assoc. Radiol. J. (2014) 65 96-105. DOI: 10.1016/j.carj.2012.11.002 7. Chong E., Poh K.K., Shen L., Chai P., Tan H.C.. Diabetic patients with normal baseline renal function are at increased risk of developing contrast-induced nephropathy post-percutaneous coronary intervention. Singap. Med. J. (2009) 50 250-254 8. McCullough P.A., Wolyn R., Rocher L.L., Levin R.N., O’Neill W.W.. Acute renal failure after coronary intervention: Incidence, risk factors, and relationship to mortality. Am. J. Med. (1997) 103 368-375. DOI: 10.1016/S0002-9343(97)00150-2 9. Bang J.-Y., Jun I.-G., Lee J.-B., Ko Y.-S., Kim K.-W., Jeong J.-H., Kim S.-H., Song J.-G.. Impact of Sarcopenia on Acute Kidney Injury after Infrarenal Abdominal Aortic Aneurysm Surgery: A Propensity Matching Analysis. Nutrients (2021) 13. DOI: 10.3390/nu13072212 10. Demirjian S., Schold J.D., Navia J., Mastracci T.M., Paganini E.P., Yared J.P., Bashour C.A.. Predictive models for acute kidney injury following cardiac surgery. Am. J. Kidney Dis. (2012) 59 382-389. DOI: 10.1053/j.ajkd.2011.10.046 11. Tao S.M., Wichmann J.L., Schoepf U.J., Fuller S.R., Lu G.M., Zhang L.J.. Contrast-induced nephropathy in CT: Incidence, risk factors and strategies for prevention. Eur. Radiol. (2016) 26 3310-3318. DOI: 10.1007/s00330-015-4155-8 12. Becquemin J.P., Pillet J.C., Lescalie F., Sapoval M., Goueffic Y., Lermusiaux P., Steinmetz E., Marzelle J.. A randomized controlled trial of endovascular aneurysm repair versus open surgery for abdominal aortic aneurysms in low–to moderate-risk patients. J. Vasc. Surg. (2011) 53 1167-1173. DOI: 10.1016/j.jvs.2010.10.124 13. De Bruin J.L., Baas A.F., Buth J., Prinssen M., Verhoeven E.L., Cuypers P.W., van Sambeek M.R., Balm R., Grobbee D.E., Blankensteijn J.D.. Long-term outcome of open or endovascular repair of abdominal aortic aneurysm. N. Engl. J. Med. (2010) 362 1881-1889. DOI: 10.1056/NEJMoa0909499 14. Greenhalgh R.M., Brown L.C., Kwong G.P., Powell J.T., Thompson S.G.. Comparison of endovascular aneurysm repair with open repair in patients with abdominal aortic aneurysm (EVAR trial 1), 30-day operative mortality results: Randomised controlled trial. Lancet (2004) 364 843-848. DOI: 10.1016/S0140-6736(04)16979-1 15. Saratzis A., Melas N., Mahmood A., Sarafidis P.. Incidence of acute kidney injury after endovascular abdominal aortic aneurysm repair and impact on outcome. Eur. J. Vasc. Endovasc. Surg. (2015) 49 534e40. DOI: 10.1016/j.ejvs.2015.01.002 16. Pisimisis G.T., Bechara C.F., Barshes N.R., Lin P.H., Kougias P.. Risk factors and impact of proximal fixation on acute and chronic renal dysfunction after endovascular aortic aneurysm repair using glomerular filtration rate criteria. Ann. Vasc. Surg. (2013) 27 16-22. DOI: 10.1016/j.avsg.2012.05.006 17. Saratzis A., Nduwayo S., Pantelis S., Sayers R., Brown M.. Renal function is the main predictor of acute kidney injury after endovascular abdominal aortic aneurysm repair. Ann. Vasc. Surg. (2016) 31 52-59. DOI: 10.1016/j.avsg.2015.10.010 18. Saratzis A., Shakespeare J., Jones O., Brown M.J., Mahmood A., Imray C.H.E.. Pre-operative functional cardiovascular reserve is associated with acute kidney injury after intervention. Eur. J. Vasc. Endovasc. Surg. (2017) 53 717-724. DOI: 10.1016/j.ejvs.2017.01.014 19. Saratzis A.N., Goodyear S., Sur H., Saedon M., Imray C., Mahmood A.. Acute Kidney Injury After Endovascular Repair of Abdominal Aortic Aneurysm. J. Endovasc. Ther. (2013) 20 315-330. DOI: 10.1583/12-4104MR2.1 20. Lee J., Park K.M., Jung S., Cho W., Hong K.C., Jeon Y.S., Cho S.G., Lee J.B.. Occurrences and Results of Acute Kidney Injury after Endovascular Aortic Abdominal Repair. Vasc. Spec. Int. (2017) 33 135-139. DOI: 10.5758/vsi.2017.33.4.135 21. Brulotte V., Leblond F.A., Elkouri S., Therasse E., Pichette V., Beaulieu P.. Bicarbonates for the prevention of post-operative renal failure in endovascular aortic aneurysm repair: A randomized pilot trial. Anesth. Res. Pract. (2013) 2013 467326 22. Krasznai A.G., Sigterman T.A., Bouwman L.H.. Contrast Free Duplex-Assisted EVAR in Patients with Chronic Renal Insufficiency. Ann. Vasc. Dis. (2014) 7 426-429. DOI: 10.3400/avd.cr.14-00084 23. Andreucci M., Faga T., Pisani A., Sabbatini M., Russo D., Michael A.. The choice of the iodinated radiographic contrast media to prevent contrast-induced nephropathy. Adv. Nephrol. (2014) 2014 11. DOI: 10.1155/2014/691623 24. Yang J.S., Peng Y.R., Tsai S.C., Tyan Y.S., Lu C.C., Chiu H.Y., Chiu Y.J., Kuo S.C., Tsai Y.F., Lin P.C.. The molecular mechanism of contrast- induced nephropathy (CIN) and its link to in vitro studies on iodinated contrast media (CM). Biomedicine (2018) 8 1. DOI: 10.1051/bmdcn/2018080101 25. Mun J.H., Kwon S.K., Park J.H., Chu W., Kim D.H., Jung H.J., Lee S.S.. Renal function-adjusted contrast medium volume is a major risk factor in the occurrence of acute kidney injury after endovascular aneurysm repair. Medicine (2021) 100 e25381. DOI: 10.1097/MD.0000000000025381 26. Blankensteijn J.D., de Jong S.E.C.A., Prinssen M., van der Ham A.C., Buth J., van Sterkenburg S.M.M., Grobbee D.E.. Two-Year Outcomes after Conventional or Endovascular Repair of Abdominal Aortic Aneurysms. N. Engl. J. Med. (2005) 352 2398-2405. DOI: 10.1056/NEJMoa051255 27. Rastogi V., de Bruin J.L., Bouwens E., Hoeks S.E., Ten Raa S., van Rijn M.J., Fioole B., Schermerhorn M.L., Verhagen H.J.M.. Incidence, Prognostic Significance, and Risk Factors of Acute Kidney Injury Following Elective Infrarenal and Complex Endovascular Aneurysm Repair. Eur. J. Vasc. Endovasc. Surg. (2022) 64 621-629. DOI: 10.1016/j.ejvs.2022.08.024 28. Arbănași E.M., Mureșan A.V., Coșarcă C.M., Arbănași E.M., Niculescu R., Voidăzan S.T., Ivănescu A.D., Hălmaciu I., Filep R.C., Mărginean L.. Computed Tomography Angiography Markers and Intraluminal Thrombus Morphology as Predictors of Abdominal Aortic Aneurysm Rupture. Int. J. Environ. Res. Public Health (2022) 19. DOI: 10.3390/ijerph192315961
--- title: Genome-Wide Gene-Set Analysis Identifies Molecular Mechanisms Associated with ALS authors: - Christina Vasilopoulou - Sarah L. McDaid-McCloskey - Gavin McCluskey - Stephanie Duguez - Andrew P. Morris - William Duddy journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966913 doi: 10.3390/ijms24044021 license: CC BY 4.0 --- # Genome-Wide Gene-Set Analysis Identifies Molecular Mechanisms Associated with ALS ## Abstract Amyotrophic lateral sclerosis (ALS) is a fatal late-onset motor neuron disease characterized by the loss of the upper and lower motor neurons. Our understanding of the molecular basis of ALS pathology remains elusive, complicating the development of efficient treatment. Gene-set analyses of genome-wide data have offered insight into the biological processes and pathways of complex diseases and can suggest new hypotheses regarding causal mechanisms. Our aim in this study was to identify and explore biological pathways and other gene sets having genomic association to ALS. Two cohorts of genomic data from the dbGaP repository were combined: (a) the largest available ALS individual-level genotype dataset ($$n = 12$$,319), and (b) a similarly sized control cohort ($$n = 13$$,210). Following comprehensive quality control pipelines, imputation and meta-analysis, we assembled a large European descent ALS-control cohort of 9244 ALS cases and 12,795 healthy controls represented by genetic variants of 19,242 genes. Multi-marker analysis of genomic annotation (MAGMA) gene-set analysis was applied to an extensive collection of 31,454 gene sets from the molecular signatures database (MSigDB). Statistically significant associations were observed for gene sets related to immune response, apoptosis, lipid metabolism, neuron differentiation, muscle cell function, synaptic plasticity and development. We also report novel interactions between gene sets, suggestive of mechanistic overlaps. A manual meta-categorization and enrichment mapping approach is used to explore the overlap of gene membership between significant gene sets, revealing a number of shared mechanisms. ## 1. Introduction Amyotrophic lateral sclerosis (ALS) is the most common type of motor neuron disorder, characterized by the loss of both upper and lower motor neurons. This progressive motor neuron disease causes swallowing problems, paralysis, and eventually, death from neuromuscular respiratory failure [1,2,3]. Patients normally live for two to five years following the onset of symptoms, with 5–$10\%$ living for more than ten years [1,2,4]. ALS can affect individuals of any age; however, its peak onset is at 54–67 years of age [2,3,5,6]. Our current understanding of the etiology, genetic architecture and the underlying biological mechanisms of ALS is still elusive, slowing the development of prevention and treatment. In recent years, genome-wide association studies (GWAS) have enabled the discovery of numerous associations of single nucleotide polymorphisms (SNPs) to complex diseases, including ALS. Previous genetic inheritance and genome-wide studies have identified numerous variants mapped to 46 genes as monogenic causes of ALS [7,8,9,10,11,12]. In European ancestry populations, the most common monogenic cause is the intronic hexanucleotide GGGGCC (G4C2) repeat expansion (HRE) in the C9ORF72 gene [13,14]. Cu/Zn superoxide dismutase 1 SOD1, fused in sarcoma FUS, and transactive response DNA-binding protein of 43 kD TARDBP/TDP-43 are further genes associated with ALS with great reproducibility [7]. However, genotype–phenotype studies have not fully elucidated the genetic contribution to familial and sporadic ALS. While a monogenic cause can be identified in around two-thirds of familial ALS cases [7,8], the majority of sporadic cases have no genetic cause identified [14,15]. Furthermore, it has been proposed that standard GWAS approaches are unlikely to fully unravel the genetic architecture of ALS and present a number of challenges and limitations [16]. One of the reasons for this is that GWAS is a single-marker analysis in which the contribution of each SNP is tested independently. Under this hypothesis, association p-values must be adjusted using strict multiple-testing methods, such as Bonferroni, in order to control for family-wise type I errors (false positives). As a result, GWAS has limited power to detect potential risk variants with weak genetic effects, which fail to pass the multiple testing correction, leading to family-wise type II errors (false negatives). In complex diseases, such as ALS, multiple genes and biological pathways are expected to be implicated, making it necessary to understand the functional underpinnings of the disease. However, the gene products of the 30 or more known ALS-associated genes may interact with one another, participating in different molecular pathways. As such, traditional single-gene and single-marker studies are expected to present a number of challenges and limitations to the functional curation and interpretation [7,16,17]. Furthermore, univariate approaches, such as GWAS and single-gene analysis, do not consider the joint effects of multiple loci, events that are likely to be present in complex diseases such as ALS [16,18]. Understanding the functional mechanisms that underpin ALS has proven to be a challenging and complex task, made more difficult by the involvement of genes with diverse functional roles. Gene-set analysis (GSA) has been employed successfully by numerous genome-wide association studies as a method to understand the functional involvement of groups of genes in the phenotype of a complex disease. In gene-set analysis, individual SNPs are summarized into whole genes, taking into account multiple genetic associations, and genes are then summarized into gene sets [19]. A gene set is any group of genes that share a common attribute. A gene set can represent, inter alia, a biological pathway, a network module, or a group of interacting components. Each gene set is tested as a whole to investigate whether the gene set property is associated with the phenotype [19]. In a recent review [20], we describe and compare the results and the methodology of collected published gene-set analysis studies utilizing ALS GWAS cohorts which aim to identify functional pathways that are associated with ALS. We note several limitations of the collected ALS GSA studies, including the use of small ALS cohort sizes, and limited or under-documented genomic quality controls and GSA methodology. We further note the use of dimensionality reduction approaches applying arbitrary thresholds to SNPs or genes in order to ease functional interpretation, reducing reproducibility across different studies—as different thresholds may lead to different biological results and interpretations [21]—as well as risking the exclusion of false negative genes and gene sets or subtle associations. The present study combines (a) the largest currently available ALS individual-level genomic study with a European descent from the dbGaP repository ($$n = 12$$,319) [9] and (b) an aging control cohort ($$n = 13$$,210), also from dbGaP, followed by comprehensive and careful quality control and batch effect correction strategies, which in our knowledge have not been previously utilized in an ALS genome-wide gene-set analysis study. Although there is no gold standard in genome-wide gene-set analysis, previous publications have demonstrated the increased power of competitive and mean-based gene-set analysis models [19,22]. By application of the MAGMA multi-model, which uses an aggregate test statistic combining multiple gene representations, the present analysis does not make an assumption about the underlying genetic architecture [22]. Two methods are employed, a combination that to our knowledge has not been used in previous ALS GSA studies: competitive and interaction gene-set analysis. Competitive gene-set analysis tests whether the genes within a given gene set are no more strongly associated with the phenotype than the genes that do not belong to this gene set [19]. We further investigate a more complex hypothesis that combinations of multiple gene sets are more highly associated with ALS than the individual gene sets, applying post hoc interaction gene-set analysis to all the significant gene sets [23]. Furthermore, we tested our gene-set analysis approach using the majority of the available gene sets from the molecular signatures database (MSigDB) [24,25]. Lastly, we sort associated gene sets into major biological categories after manual curation, and visualize the relationships between them using enrichment maps, where each node represents an associated gene set to ALS, and each edge represents the proportion of shared genes between two gene sets. The experimental design of the present study is shown in Figure 1. The aim of this study is to identify statistically significant ALS-associated biological pathways using gene-set analysis on the largest available release of individual-level genomic ALS-control cohort data [9,26]. A large ALS-control cohort is created following detailed quality control and batch effect correction strategies [27,28]. Results are further visualized and interpreted using enrichment maps, and the relationships of the most significant gene sets are further investigated through interaction analysis. These analyses enable an in-depth exploration of the pathways associated with ALS. ## 2.1. Gene-Level Meta-Analysis Gene-level meta-analysis using MAGMA resulted in a total of 19,242 genes, from 22,039 samples of 9244 ALS cases and 12,795 healthy controls. The MAGMA multi-model gene-level results yielded 6 genes that reached high statistical significance based on false-discovery rate (FDR < 0.01) and 4 genes that passed the stricter multiple testing Bonferroni correction (alpha = 0.05; p-value <2.58×10−6), shown in Table 1. The top 6 genes of our analysis include MOB3B (MOB kinase activator 3B; FDR = 1.77 ×10−12), IFNK (interferon kappa; FDR = 3.98 ×10−10), C9ORF72 (chromosome 9 open reading frame 72; FDR = 2.50 ×10−8), UNC13A (unc-13 homolog A; FDR = 1.14 ×10−7), ADARB1 (adenosine deaminase RNA specific B1; FDR = 0.001) and KIF5A (kinesin family member 5A; FDR = 0.008). The top three ALS-associated genes MOB3B, IFNK and C9ORF72 have been reported in previous GWAS [29]. In addition, three of the six most strongly associated genes in this analysis have been established in previous studies as ALS-associated, including C9ORF72 [30], UNC13A [31] and KIF5A [9]. These three genes were also reported as significantly associated with ALS in the original GWAS study reporting on these same data [9]. Our analysis also revealed two previously identified ALS-associated genes TBK1 (TANK binding kinase) (FDR = 0.063) and FUS (fused in sarcoma; FDR = 0.155) [31] with a marginal statistical significance. ## 2.2. Gene-Set Analysis We then carried out a gene-set analysis testing for the association of ALS to each of the 31,454 collected gene sets from the molecular signatures database (version 7.5). Twenty-four gene sets were associated to ALS with p-value < 0.05 and false discovery rate < 0.05 (Table 2). These included gene sets belonging to various functional categories: the nervous system (the BioCarta CREB, DREAM, CK1, AGPCR and Shh pathways); the immune system (the BioCarta CSK, CTFR, TCR and VIP pathways); developmental pathways (the BioCarta Shh and mPR pathways); the cytoskeleton and cell cycle (the BioCarta Stathmin pathway); cell signaling (the GO cyclic nucleotide-dependent protein kinase activity, cyclic nucleotide binding, and the BioCarta IGF1R and AGPCR pathways); apoptosis and cell survival (the BioCarta BAD and IGF1R pathways); lipid metabolism and homeostasis (the BioCarta PPAR-alpha and CFTR pathways); and muscle tissue (the BioCarta IGF1R pathway). Gene-level summary statistics for the genes of each significant gene set are provided (Table S1). Figure 2 shows an enrichment map containing only the strongly ALS-associated gene sets. Sixteen of these gene sets present a highly dense cluster, meaning that they often share genes with each other: their 16 nodes are connected to neighbors through 113 edges (representing shared gene membership). The eight nodes of the remaining gene sets have just one or no first neighbor among the strongly associated gene sets. The nervous system is the most represented functional category within the cluster, for which five pathways are associated with the disease. To explore in greater depth the functional relationships of ALS-associated gene sets, we carried out an exhaustive enrichment map exploration based on p-value cut-off of 0.05 and FDR cut-off of 0.25. Enrichment maps are a way of representing the overlapping relationships between gene sets. This approach allows visualizing not only the direct overlap between the strongly ALS-associated gene sets described above (i.e., the gene sets which we would consider to be the primary observation of the study), but also their wider functional context, drawing in a total of 145 gene sets. The following sections first consider an enrichment map containing only the strongly ALS-associated gene sets, then describe enrichment map explorations for each of the five categories representing common functions of the associated gene sets (immune response, developmental, nervous system, muscle, and lipid metabolism). High-resolution versions of the enrichment map figures are also provided (Figures S1–S6). ## 2.3. Mechanistic Relationships of ALS-Associated Gene Sets The degree score (i.e., the number of neighboring nodes) for each of the 24 strongly associated gene sets is shown in Table 2. The CREB pathway is the most popular node of the enrichment map, having a degree of 50. The functional category having the largest number of gene sets in the enrichment map is the immune response, for which 7 gene sets pass FDR 0.05, 3 of which have low degrees and are not members of the main cluster. In the following subsections, each functional category is explored separately, beginning with the immune system. ## 2.3.1. Immune-Response Pathways We report 26 significant gene sets that are related to the immune response. To further investigate these immune-response gene sets, we created a sub-network (shown in Figure 3) containing 38 nodes and 132 edges, including the 26 significant immune-response gene sets and first neighbours having statistical significance (FDR < 0.05), i.e., significant gene sets of any biological category that share a proportion of genes with any of the immune response gene sets. In this enrichment map network, an edge represents the overlapping genes that two gene sets share; the higher the width of an edge, the greater the overlap (edge similarity cut-off > 0.1). In Table 3, we summarize the immune-response related gene sets and their associated p-values, FDR values, degree, i.e., the number of edges of each node within the immune-response ALS enrichment network, and their corresponding number of genes. It is noteworthy that the most popular immune response nodes (i.e., having the greater number of first neighbors) in this subnetwork are also the highest associated gene sets to ALS of the immune response category. The 8 most popular nodes (summarized in Table 3) include the BioCarta CSK pathway (p-value = 5.02 ×10−5, FDR = 0.009), BioCarta CFTR pathway (p-value = 0.002, FDR = 0.046), BioCarta TCR pathway (p-value = 0.002, FDR = 0.048), BioCarta VIP pathway (p-value = 0.003, FDR = 0.049), BioCarta CTLA4 pathway (p-value = 0.004, FDR = 0.054), BioCarta NFAT pathway (p-value = 0.005, FDR = 0.07), BioCarta GATA3 pathway (p-value = 0.010, FDR = 0.099), and BioCarta CDMAC pathway (p-value = 0.024, FDR = 0.179). Several of these popular nodes and some non-popular nodes of the immune-response subnetwork are implicated in the T-cell receptor (TCR) signaling pathway (BioCarta TCR pathway), a key immune response mechanism, and concern lymphoid cell pathways, e.g., BioCarta TCR pathway, BioCarta CSK pathway, BioCarta VIP pathway, BioCarta TCRA pathway (p-value = 0.010, FDR = 0.099), BioCarta Lymphocyte pathway, BioCarta CTLA4 pathway, BioCarta NFAT pathway and BioCarta IL17 pathway (p-value = 4.31 ×10−2, FDR = 0.22). Specifically, the highest ALS-associated gene set in our analysis BioCarta CSK pathway (FDR = 0.009) plays a role in the inhibition of T-cell receptor signaling and T-cell activation [24,25,32]. In the CSK pathway, the activated CSK kinase (COOH-terminal Srk kinase), which is transported to the plasma membrane through lipid rafts, phosphorylates the kinase Lck, which leads to the inhibition of the T-cell activation and T-cell signaling [24,25,32]. CTLA-4 (cytotoxic T-lymphocyte antigen-4) is a receptor expressed on the surface of T cells that leads to decreased T-lymphocyte activity and is considered a key immune response regulator [33]. The CTLA-4 pathway induces co-stimulatory signals during T-cell activation, providing additional control mechanisms that prevent inappropriate and hazardous T-cell activation that could lead to autoimmune disease pathogenesis [24,25,33]. In addition, the TCRA (T-cell receptor activation) pathway involves the activation of the T-Cell receptor through the Lck and Fyn tyrosine kinases [24,25]. Furthermore, the VIP (vasoactive intestinal peptide) pathway inhibits the apoptosis of activated T cells through two neuropeptides VIP and PACAP (pituitary adenylate cyclase-activating polypeptide) present in the lymphoid microenvironment, which have been known for their neuroprotective and immunomodulatory roles [24,25,34]. The BioCarta IL17 pathway (p-value = 4.31 ×10−2, FDR = 0.22) concerns the secretion of the cytokine IL-17 by activated T cells as part of the inflammatory response and has been associated with autoimmune disorders [24,25]. NFAT is a transcriptional regulator that is also associated with the activation of the T cells, and plays a crucial role in the development and function of the immune system [35]. Lastly, the GATA-3 pathway is another popular node of the immune subnetwork. GATA-3 is a transcription factor which influences the development and differentiation of peripheral T cells, is specifically involved in the activation of the Th2 cytokine genes expression, and has been associated with allergic and lymphoproliferative disorders [24,25,36]. A number of the immune-response popular nodes and their first neighbors are associated with the innate/non-specific immune system, i.e., the first-line immune protection against foreign substances, viruses, bacteria, etc. These pathways include the BioCarta Monocyte pathway (p-value = 9.34 ×10−4, FDR = 0.040), BioCarta CFTR pathway (p-value = 0.002, FDR = 0.046), KEGG RIG-I-like receptor signaling pathway (p-value = 0.002, FDR = 0.115), BioCarta CDMAC pathway (p-value = 0.024, FDR = 0.179), BioCarta NK cells pathway (p-value = 0.028, FDR = 0.190), BioCarta Neutrophil pathway (p-value = 0.040, FDR = 0.226), BioCarta Granulocytes pathway (p-value = 0.040, FDR = 0.226) and BioCarta RNA pathway (p-value = 0.042, FDR = 0.229). The monocyte pathway plays an important role in the innate immune response. Monocytes belong to the class of phagocytes. They can form macrophages or dendritic cells, and their role is to protect against bacterial, viral and fungal infections. The CFTR (cystic fibrosis transmembrane conductance regulator) protein is a chloride channel in the plasma membrane of epithelial cells, and certain CFTR mutations are the monogenic cause of cystic fibrosis [37]. The CFTR pathway has also been involved, among others, with the innate immune system and antimicrobial host defense, and the CFTR protein is expressed in macrophages and neutrophils [37,38]. The RIG-I-like (retinoic acid-inducible gene I) receptor signaling pathway involves the recognition of intracellular viral replication in the cells, as well as the initiation of inflammatory responses to eliminate the virus infection [39]. The aforementioned gene set includes, among others, the previously ALS-associated gene TBK1 (TANK binding kinase 1) with a statistical significance in our ALS gene-level analysis of FDR = 0.06. In addition, the CDMAC pathway, a popular node in our immune response subnetwork, stimulates the inappropriate proliferation and the DNA synthesis of the macrophages through the interaction of cadmium and G-protein coupled receptors [40]. Cadmium has been also associated with the inhibition of DNA repair and of the immune system, as well as activation of stress genes [40]. Neutrophils belong to the categories of granulocytes and phagocytes and play an essential role in the innate immune system and in inflammation. The natural killer cells pathway involves cytotoxicity mediated by natural killer cells, granular lymphocytes that are critical in the innate immune system. Lastly, the RNA pathway concerns the defense against viral infection by the PKR (protein kinase R), an interferon-induced protein kinase, which is activated by double-stranded RNA. Its anti-viral role mainly concerns blocking the viral translation mechanism and inducing apoptosis in the infected cells [41]. Several immune-response gene sets are linked to inflammatory regulation, a physiological part of the innate immune response, interleukins (ILs), a group of cytokines, and oxidative stress. Our results show a strong ALS association to oxidized phospholipids, the fourth most highly ALS-associated gene set of our analysis (and second in the immune-response category), Gargalovic response to oxidized phospholipids black up (p-value = 9.54 ×10−6, FDR = 0.032), and the less significantly associated gene set, Gargalovic response to oxidized phospholipids cyan up (p-value = 2.38 ×10−4, FDR = 0.187). Both of these gene sets derive from a study by Gargalovic et al. [ 42] of genes regulated by a specific oxidized phospholipid (the colors black and cyan refer to gene correlation clusters defined in that study). Oxidized phospholipids have been linked to pro-inflammatory and anti-inflammatory responses and atherogenesis through the stimulation of endothelial cells to produce inflammatory cytokines [24,25,42,43]. Furthermore, it has been shown that oxidized phospholipids are involved in acute and chronic microbial infections, metabolic disorders, and neurodegenerative diseases [43]. The BioCarta IL17 pathway (p-value = 4.31 ×10−2, FDR = 0.22) concerns the secretion of the cytokine IL-17 by activated T cells, as part of the inflammatory response [24,25]. While the inflammatory cytokine IL-17 has been associated with pro-inflammatory properties, it has also been proven to play a critical role in autoimmune diseases, cancer progression and immunopathology [44,45]. The GOBP negative regulation of interleukin 5 production (p-value = 6.38 ×10−4, FDR = 0.24) gene set is associated with the negative regulation of the cytokine IL-5, which is primarily known as a key mediator in the differentiation, growth, survival, and degranulation of eosinophils [46]. IL-5 is mainly produced by T helper-2 (Th2) lymphocytes, mainly involved in the response to parasites and allergies, and group 2 innate lymphoid cells (ILC2), and its expression is regulated by several transcription factors, such as GATA3 (BioCarta GATA3 pathway (p-value = 0.009, FDR = 0.099) [46,47]. Lastly, integrins—cell adhesion transmembrane receptors—(PID integrin cs pathway p-value = 0.005, FDR = 0.148; PID AVB3 integrin pathway p-value = 0.012034, FDR = 0.235) have been previously associated with cytokine activation as well as playing a critical role in infection, leukocyte recruitment, inflammation, angiogenesis and immunological signaling [48]. ## 2.3.2. Developmental Pathways Another prominent biological category is developmental pathways. We report 22 ALS-associated gene sets that are related to human development (see Table 4). To investigate the significance of the developmental pathways in more depth, we constructed a developmental subnetwork (shown in Figure 4), containing 39 nodes and 101 edges, including the 22 significant developmental gene sets and their first neighbors, which show a statistically significant association to ALS (FDR < 0.05), i.e., significant gene sets of any biological category that share a number of genes with any of the developmental gene sets in distance one. In this enrichment map network, an edge represents the overlapping genes that two gene sets share; the higher the width of each edge, the larger the overlap (edge similarity cut-off > 0.1). The BioCarta mPR (membrane progesterone receptor) pathway (p-value = 6.27 ×10−5, FDR = 0.009), which involves the oocyte maturation by progesterone, is the highest ALS-associated gene set among the developmental-related gene sets, along with the immune-response BioCarta CSK pathway (p-value = 5.02 ×10−5, FDR = 0.009), covered in the previous section. The BioCarta mPR pathway concerns the oocyte maturation by progesterone [24,25]. Progesterone is an endogenous steroid hormone that plays multiple roles, including oocyte meiotic maturation, embryogenesis, maintenance in pregnancy, and neural functions [49,50,51]. The binding of progesterone to intracellular and plasma membrane progesterone G protein-coupled receptors initiates a cascade of signaling pathways, including the indirect activation of the MAPK signaling. MPRs are G protein-coupled receptors (GPCRs) that have been previously associated with neuroprotective, neurosteroid and neuroendocrine functions in neurons [51]. The second most highly associated pathway in the developmental category is the Sonic Hedgehog (Shh) pathway (BioCarta Shh pathway p-value = 6.60 ×10−4, FDR = 0.040). The Shh protein plays a critical role in the development processes in multi-cellular organisms through complex signaling cascades [52]. The Shh pathway has a key role in the cellular differentiation of multiple organs, in embryonic development, repair processes and especially in neuronal development [52,53]. Specifically, the Shh signaling pathway has been strongly associated, among others, with the development of the neural tube, motor neurons, the regulation of CNS polarity, neuronal regeneration and proliferation, stem cell renewal, and patterning of the developing thalamus and ventral forebrain [52,53,54]. Shh is also involved with proliferation-linked signaling cascades after binding to a receptor complex, including Ptc-1 and smoothed G-protein coupled receptor [24,25]. It is noteworthy that, in the developmental subnetwork, the two most highly associated developmental gene sets, BioCarta mPR and Shh pathways, are the most popular nodes but also share an almost complete overlap of statistically significant first neighbors (as shown in Figure 4 and Table 5). The most highly ALS-associated common neighbor is the immune-response BioCarta CSK pathway (p-value = 5.02 ×10−5, FDR = 0.009) related to the inhibition of T-cell receptor signaling and T-cell activation [32]. Other immune-response common neighbors of the mPR and the Shh pathways include the BioCarta VIP pathway, which is also implicated in the T-cell signaling [34] and the BioCarta CFTR pathway, which has been associated among others with the innate immune system, B-cell activation and proliferation [37,38]. In addition, we observe cell signaling pathways as common neighbors for the two developmental gene sets, such as the GOMF cyclic nucleotide dependent protein kinase activity, GOMF cyclic nucleotide binding, BioCarta IGF1R pathway and the BioCarta AGPCR pathway. Cyclic nucleotides are secondary messengers that play a central role in intracellular signal transduction, responding to hormonal stimuli and intra- or extracellular environmental changes [55,56]. The BioCarta AGPCR pathway relates to the signaling attenuation of the G-protein coupled receptors (GPCR) [24,25], a transmembrane protein family which is well known for its major role in the signal transduction of extracellular stimuli across the plasma membranes. Failure to attenuate the rapid GPCR signaling leads to acute and chronic overstimulation of the receptors [57]. The IGF-1R pathway involves multiple anti-apoptotic pathways through the IGF-1R (type 1 receptor for insulin-like growth factor), promoting cell survival and growth, as well as blocking apoptotic pathways such as the BioCarta BAD pathway, another statistically significant common neighbor. Furthermore, we note the presence of several nervous system-specific signaling pathways, including the BioCarta DREAM pathway, BioCarta CREB pathway, BioCarta CK1 pathway and BioCarta AGPCR pathway gene sets. The DREAM (downstream regulatory element antagonistic modulator) pathway involves the repression of the pain sensation by the DREAM transcriptional regulator [58]. DREAM has been linked to pain signaling and is expressed in spinal cord neurons [58]. The CREB (cyclic AMP-responsive element binding) pathway mediates the activation of transcription in response to extracellular stimuli including neurotransmitters, hormones, membrane depolarization, and growth and neurotrophic factors, by the transcription factor CREB [59]. The CK1 (casein kinase 1) pathway includes the dopaminergic signaling in the neostriatum, elevating cAMP (a type of cyclic nucleotide) and activating PKA (protein kinase A) [24,25]. It has been argued that CK1 family members are signal transduction regulators in the Shh and Wnt developmental signaling pathways [60]. Lastly, we observe the presence of gene sets relating to lipid metabolism and homeostasis, such as the BioCarta CFTR pathway and the BioCarta PPAR-alpha pathway. The PPAR-alpha (peroxisome proliferator activated receptor alpha) pathway relates to the gene regulation by peroxisome proliferators via the PPAR-alpha phosphoprotein and has been linked to fatty acid metabolism regulation, and autophagy in human microglia and hepatic cells [61]. The third most highly associated gene set in the developmental pathways category concerns the signaling pathways induced by N-cadherin (PID N-cadherin pathway p-value = 5.77 ×10−4, FDR = 0.057) (shown in Table 4). Neural cadherin (N-cadherin) is an adhesion receptor mainly known for its key role in the organization of the synaptic complex, ensuring the adhesion between synaptic membranes and organizing the actin cytoskeleton, as well as being involved in cell type specific adhesion processes during embryonic development [62]. N-cadherin is ubiquitously expressed in the neuronal synapses and in the vascular smooth muscle, participating in the development and plasticity of the adult neural tissue [63,64,65]. N-cadherin has also been associated with the mediating of signal transduction events during bone development and vasomotor control [66,67]. The PID N-Cadherin pathway gene set is connected with the ALS-associated gene set PID RhoA pathway (p-value = 1.16 ×10−4, FDR = 0.022). Ras homolog gene family member A (RhoA) is a small GTPase that has an essential role in regulating the development, differentiation, survival, and death of neurons in the central nervous system [68]. The link between N-cadherin and RhoA has been previously reported by other studies, highlighting the cell signaling capabilities of N-cadherin modulating the voltage activated calcium influx by RhoA activity mechanisms and its downstream effects on the cytoskeleton [69]. In addition, we observe that the N-cadherin pathway is linked to the immune-response T-cell receptor signaling pathway gene set BioCarta TCR pathway (p-value = 0.002, FDR = 0.048). Furthermore, in the developmental subnetwork, we observe the N-cadherin gene set to be connected to the GOBP anterograde dendritic transport of neurotransmitter receptor complex (p-value = 5.60 ×10−5, FDR = 0.110), which relates to the directed neurotransmitter receptor complex movement toward the post-synapse through dendritic transport [24,25]. The PID AVB3 integrin pathway (p-value = 0.012, FDR = 0.236) is another immediate neighbor of the N-cadherin pathway. Integrins, such as cadherins, are well-studied cell adhesion molecules that mediate cell–cell and/or cell–extracellular matrix (ECM) adhesion, involved in several signaling pathways, and immune-response processes as well as having developmental roles. The PID AVB3 integrin pathway represents specifically the role of the integrins in angiogenesis [24,25]. We further observe two highly interconnected clusters in the developmental subnetwork. The first concerns pathways related to abnormalities in the morphology of the rib cage/thorax, including the gene sets HP thoracic hypoplasia (p-value = 5.01 ×10−5, FDR = 0.101), HP abnormal rib cage morphology (p-value = 6.03 ×10−5, FDR = 0.101), HP thoracic dysplasia (p-value = 9.63 ×10−5, FDR = 0.122) and HP abnormality of the ribs (p-value = 2.65 ×10−4, FDR = 0.223). All four gene sets share the previously associated ALS gene NEK1 (NIMA-related kinase 1) (p-value = 0.007, FDR = 0.27). In addition, we note a second cluster which is linked with gastrulation, an early embryonic developmental process of the blastocyst forming a multilayer gastrula, including the endoderm, mesoderm, and ectoderm (GOBP gastrulation; p-value = 5.13 ×10−4, FDR = 0.218; GOBP Formation of primary germ layer p-value = 4.68 ×10−4, FDR = 0.218) [70]. Gastrulation is also associated with embryonic placenta development, including cell–cell adhesion processes, such as chorioallantoic fusion (GOBP chorio allantoic fusion p-value = 7.28 ×10−4, FDR = 0.242) as well as the formation and the morphogenesis of anatomical structures that derive from the mesoderm—the second germ layer that develops, among others, into smooth, cardiac, and skeletal muscle, bone, reproductive organs, microglia, adrenal cortex, cartilage, blood cells, vascular endothelium and connective tissue (GOBP mesoderm morphogenesis p-value = 1.65 ×10−5, FDR = 0.111; GOBP mesoderm development p-value = 7.52 ×10−5, FDR = 0.111; GOBP cardiac muscle cell fate commitment p-value = 1.99 ×10−4; FDR = 0.190) [70]. We further observe multiple links of the gastrulation-related gene sets to statistically significant immune-response gene sets (i.e., BioCarta Monocyte pathway, BioCarta lymphocyte pathway, and PID AVB3 integrin pathway). We note the central role of the developmental gene set PID lymph angiogenesis pathway (p-value = 7.28 ×10−4, FDR = 0.095) (shown in Figure 4) referring to the VEGFR3 (Vascular Endothelial Growth Factor Receptor 3) signaling in the lymphatic endothelium [24,25,71]. The lymph angiogenesis pathway shares many of the common neighbors of the mPR and the Shh top developmental pathways, including the nervous system-specific cell-signaling pathways, BioCarta IGF1R pathway, BioCarta DREAM pathway, BioCarta CREB pathway, apoptotic BioCarta BAD pathway and BioCarta PPAR-alpha pathway which is linked to lipid metabolism and homeostasis. In addition, the lymph angiogenesis gene set is a bridging node between the two distinct interconnected cliques that were described in the previous subsubsection, the one relating to rib cage dysplasias, and the second one including gene sets linked to gastrulation and mesoderm morphogenesis. Lastly, the PID lymph angiogenesis pathway is linked to several immune-response significant gene sets, such as the BioCarta lymphocyte, monocyte and TCR pathways. ## 2.3.3. Nervous System Pathways In this subsection, we focus on nervous system-specific pathways that are associated with ALS based on the results of our genome-wide gene-set analysis. We report 19 ALS-associated gene sets that have been known to be involved in nervous system pathways and biological processes (shown in Table 6). The nervous system-specific subnetwork contains, in total, 32 nodes and 142 edges, including 19 ALS-associated nervous system related gene sets and their significant first neighbors (with FDR < 0.05) which is shown in Figure 5. The seven most popular ALS-associated gene sets within the nervous system subnetwork (with a ranging degree from 16 to 21) form an interconnected cluster (clique) (shown in Table 6). This popular clique includes the BioCarta CREB pathway (p-value = 4.05 ×10−4, FDR = 0.039), BioCarta DREAM pathway (p-value = 0.001103, FDR = 0.040), BioCarta Shh pathway (p-value = 6.6 ×10−4, FDR = 0.040), BioCarta CK1 pathway (p-value = 0.001, FDR = 0.046), BioCarta AGPCR pathway (p-value = 0.002, FDR = 0.048), BioCarta NOS1 pathway (p-value = 0.003, FDR = 0.055) and BioCarta NFAT pathway (p-value = 0.005, FDR = 0.068). The CREB (cyclic AMP response element-binding protein 1) is a stimulus-induced transcription factor that mediates the activation of transcription in response to extracellular stimuli. Stimulus-induced phosphorylation of CREB affects a variety of signaling cascades, including the ERK$\frac{1}{2}$, MAPK, PI3K/AKT, CaMK, PKC, and PKA [59]. The DREAM (downstream regulatory element antagonistic modulator) transcriptional regulator is expressed in spinal cord neurons and is involved in pain signaling [58]. The Sonic Hedgehog (Shh) pathway, the second most highly associated pathway in the developmental category, comes up as the third top gene set in the nervous system category as well. As mentioned in the developmental subsection, the Shh pathway plays a key role in neuronal development, among others [52,53]. Specifically, the Shh signaling pathway has been strongly associated among others with the development of the neural tube, differentiation of the floor plate cells, motor neurons, and inter-neurons, the regulation of CNS polarity, neuronal regeneration and proliferation, stem cell renewal, survival of ventral progenitors, specification of ventral neurons, midbrain dopaminergic differentiation, proliferative signaling cascades in the developing cerebellum and other tissues and patterning of the developing thalamus and ventral forebrain [52,53,54]. The CK1 (casein kinase 1) pathway includes dopaminergic signaling in the neostriatum through the activation of a G-protein coupled dopamine receptor by dopamine, elevating cAMP and activating PKA (protein kinase A) [24,25]. Protein kinase CK1 has also been associated with the glutamatergic synaptic transmission regulation mediated by N-methyl-D-aspartic acid (NMDA) receptors as well as with the signal transduction regulation in the Shh and Wnt developmental signaling pathways [60,72]. The AGPCR pathway refers to the signaling attenuation of the G-protein coupled receptors (GPCR), avoiding acute and chronic overstimulation of the receptors [24,25,57]. The NOS1 pathway concerns the glutamatergic-mediated nitric oxide (NO) production mediated by the NMDA postsynaptic density protein 95 (PSD95)-neuronal nitric oxide synthase (nNOS1) complex [24,25]. NO has an important role in inflammation through the up-regulation of NOS in microglia as well as in cardiovascular, reproductive, neuromuscular and nervous system functions [73]. Finally, the NFAT (nuclear factor of activated T cells) BioCarta pathway relates to hypertrophy of the cardiac muscle, but also plays a key role in the immune system, development of the nervous system and skeletal muscle [24,25,35]. The NFAT transcription factor has been also associated with the regulation of pro-inflammatory responses in cultured murine microglia [74]. The highly associated nervous system-related clique is also interconnected with six statistically significant gene sets that form links solely with this particular clique. These gene sets include immune response (i.e., BioCarta CSK pathway, BioCarta CFTR pathway and BioCarta VIP pathway), cell signaling (i.e., GOMF cyclic nucleotide-dependent protein kinase activity, and GOMF cyclic nucleotide binding) and cell cycle and cytoskeleton pathways (i.e., BioCarta Stathmin pathway). The majority of the former gene sets belong to the top seven ALS-associated gene sets of our analysis. Furthermore, the nervous system clique is connected to more popular nodes, including developmental pathways, such as the top gene set BioCarta mPR pathway, apoptotic gene sets, such as the BioCarta BAD pathway and the BioCarta IGF1R pathway (which is also closely associated with cell signaling and muscle-related processes), pathways related to homeostasis and lipid metabolism, such as the BioCarta PPAR-alpha pathway, and immune response-related gene sets, such as the BioCarta TCR pathway. The BioCarta Prion pathway (p-value = 0.003, FDR = 0.054) is the sixth most highly ALS-associated gene set within the nervous system response-related category (shown in Table 6). The ALS-associated Gargalovic response to oxidized phospholipids black up (p-value = 9.54 ×10−6, FDR = 0.032) gene set, which is related to immune response and oxidative stress, is linked with the prion pathway, which has been associated with neurotoxicity and neurodegeneration [75,76]. ## 2.3.4. Muscle Pathways Several of the ALS-associated processes and pathways could be of particular relevance to muscle cell functions. We report 13 ALS-associated gene sets that have been known to be involved in muscle pathways and biological processes (shown in Table 7). The muscle-specific subnetwork contains in total 30 nodes and 68 edges, including the 13 ALS-associated muscle system-related gene sets and their significant first neighbors (with FDR < 0.05), shown in Figure 6. Several of the ALS-associated processes and pathways could be of particular relevance to muscle cell functions (shown in Table 7). These include a number of processes that are quite specific to muscle, such as agrin-mediated organization of the skeletal muscle cytoskeleton (BioCarta AGR pathway p-value = 0.007, FDR = 0.088), the AKT/mTOR pathway (BioCarta IGF1/mTOR pathway p-value = 0.012, FDR = 0.109) that has a key role in regulation of skeletal muscle mass, cardiac muscle cell fate commitment (GOBP cardiac muscle cell fate commitment p-value = 1.99 ×10−4, FDR = 0.190), and muscle cell migration (GOBP muscle cell migration p-value = 4.75 ×10−4, FDR = 0.218), as well as processes that are of particular importance to muscle but also have more broad roles in other tissues. Among the latter are: the phosphorylation of myosin (BioCarta Myosin pathway p-value = 0.048, FDR = 0.246), which is known to regulate smooth muscle contraction and platelet release; serum response factor-mediated transcriptional regulation of genes involved in the actin cytoskeleton and cell adhesion [77], including in cardiac and smooth muscle [78], as well as several other molecular pathways described in BioCarta or in the pathway interaction database—notably, due to their high significance, pathways of the IGF1-receptor (BioCarta IGF1R pathway p-value = 9.71 ×10−4, FDR = 0.040), N-Cadherin (PID N-Cadherin pathway p-value = 5.77 ×10−4, FDR = 0.057), NFAT (BioCarta NFAT pathway p-value = 0.005, FDR = 0.069), and the angiotensin receptor (BioCarta AT1R pathway p-value = 0.007, FDR = 0.077). The IGF1 pathway has a broad role in many tissues, including the induction of growth or differentiation of target cells, cell survival and maintenance of cell function, and is notable in muscle for the regulation of muscle mass [79]. The BioCarta IGF1R pathway involves multiple antiapoptotic pathways from IGF-1R signaling, leading to BAD phosphorylation [24,25]. N-cadherin is a cell–cell adhesion glycoprotein required for left–right asymmetry during gastrulation, with roles in the central nervous system, and in cardiac and skeletal muscle development [80]. The BioCarta NFAT pathway specifically relates to hypertrophy of the cardiac muscle, but includes all four members of the NFAT transcription factors that are important to the immune system and to the development of the nervous system and skeletal muscle, including myoblast fusion [81]. The AT1R pathway relates to the angiotensin receptor, and angiotensin II mediated activation [24,25]. Angiotensin II regulates many aspects of the cardiac muscle [82]. Most of these muscle-related gene sets overlap to a limited extent with one or more of the others, in terms of sharing genes in common, and most of the shared genes relate to core cellular signaling cascades (shown in Figure 6). There are only minor differences in gene membership between the muscle cell migration and vascular-associated smooth muscle cell migration gene sets. The BioCarta NFAT pathway and BioCarta IGF1R pathway gene sets have $42\%$ similarity, both making use of the Ras/Raf/MEK/ERK signaling cascade. The BioCarta AKT/mTOR, AT1R, and MAL pathways each overlap both NFAT and IGF1R, again due to all of these pathways sharing the Ras/Raf/MEK/ERK signaling cascade. The phosphorylation of myosin and the angiotensin pathway both feature protein kinase C and the G protein subunit alpha. Certain G protein components as well as Rho-associated coiled-coil containing protein kinase 1 (ROCK1) are in common between myosin phosphorylation and muscle cell migration. The BioCarta AT1R pathway and BioCarta MAL pathway share not only the Ras/Raf/MEK/ERK signaling cascade, but other signaling components such as RAC1 and MAPK8 (JNK). ERK, RAC1, and JNK are each also core components of the AGR pathway, explaining its overlap with the MAL and ATR1 pathways. The BioCarta MAL pathway and the PID N-Cadherin pathway share some of these same signaling components, including JNK, ROCK1, RhoA, and RAC1. A number of other gene sets that have robustly significant (FDR < 0.05) genome-wide association to ALS overlap with these muscle-related gene sets, the greatest overlap being with the BioCarta NFAT pathway and BioCarta IGF1R pathway (shown in Figure 6). Core signaling cascades are again prominent among the shared genes. For example, the NFAT and IGF1R pathways each share the following: cAMP-dependent protein kinases and some calmodulin signaling with the Stathmin pathway; protein kinases with the VIP pathway (also calmodulin signaling, in the case of NFAT); several protein kinases and phosphatases with the CK1 pathway; protein kinases, AKT1, and ERK1, with the CREB pathway (along with much more in the case of IGF1R, the CREB and IGF1R pathways having $63\%$ overlap); and numerous protein kinases with the AGPCR pathway. The BAD pathway overlaps with the NFAT and the AKT/mTOR pathways, sharing components of each, while the AKT/mTOR pathway also overlaps with the CREB pathway. IGF1 and IGF binding proteins 3 and 5 are shared by muscle cell migration and the ghrelin pathway. Components (though not always the same components) of the PID RhoA pathway are shared by the N-Cadherin, AT1R, AGR, MAL, and myosin pathways. ## 2.3.5. Lipid Metabolism Pathways We report 9 ALS-associated gene sets that are related to lipid metabolism (shown in Table 8). The lipid metabolism subnetwork is a small network compared to the previous subnetworks, containing, in total, 24 nodes and 33 edges, including the 9 ALS-associated lipid metabolism-related gene sets and their significant first neighbors (with FDR < 0.05) as shown in Figure 7. The BioCarta PPAR-alpha pathway (p-value = 0.002, FDR = 0.046) and BioCarta CFTR pathway (p-value = 0.002, FDR = 0.046) are the top two gene sets within the lipid metabolism category but also the most popular nodes within the lipid metabolism subnetwork, with a common degree of 14 (shown in Table 8). Both gene sets are related to lipid metabolism but also to homeostasis/cell metabolism processes, while the CFTR pathway is also linked with immune response processes. Specifically, the CFTR (cystic fibrosis transmembrane conductance regulator) protein is a chloride channel in the plasma membrane of epithelial cells, and certain CFTR mutations are the monogenic cause of cystic fibrosis [37]. The CFTR pathway has also been involved among others with the innate immune system and antimicrobial host defence [37,38]. Furthermore, the BioCarta PPAR-alpha (peroxisome proliferator activated receptor alpha) pathway relates to the gene regulation by peroxisome proliferators via the PPAR-alpha phosphoprotein [24,25]. Peroxisomes are subcellular metabolic organelles found in almost all eukaryotic cells that play a key role in lipid metabolism and redox homeostasis, ensuring the proper cellular response to endogenous and exogenous stimuli [83,84]. The peroxisome proliferator-activated receptors (PPARs) belong to the superfamily of nuclear hormone receptors. PPAR-alpha, a PPAR isoform, affects the expression of target genes involved in cell proliferation, cell differentiation, immune and inflammation responses [61,85]. Specifically, PPAR-alpha has been established as a major regulator of the fatty acid metabolism in hepatic cells and has also been linked to autophagy in human microglia, oxidative phosphorylation and regulation of energy homeostasis cells [61,85]. PPAR-alpha expression is ubiquitous with high expression levels in tissues with a high level of fatty acid catabolism, such as the liver, heart, and muscle, but has been also detected in immune cells and specifically in T cells and macrophages [85,86]. Lastly, PPARs have been previously associated with the regulation of the oxidative metabolism of skeletal muscle and the promotion of fiber-type conversion by mediating the activity of peroxisome proliferator [87]. The vast majority of the first neighbors (12 nodes out of 14) of the two popular lipid-related gene sets BioCarta PPAR-alpha pathway and BioCarta CFTR pathway, are mutually shared, as shown in Figure 7. The most frequent biological category among the commonly shared first neighbors (FDR < 0.05) is related to nervous system pathways and processes (i.e., BioCarta CREB pathway, BioCarta Shh pathway, BioCarta DREAM pathway, BioCarta CK1 pathway and BioCarta AGPCR pathway). The rest of the biological categories concern developmental gene sets (i.e., BioCarta Shh pathway and BioCarta mPR pathway), cell signaling gene sets (i.e., BioCarta AGPCR pathway, GOMF Cyclic nucleotide dependent protein kinase activity and BioCarta IGF1R pathway) and immune response gene sets mostly related to T-cell activation (i.e., BioCarta CSK pathway BioCarta VIP pathway). Lastly, the BioCarta TCR pathway, a significant T-cell receptor immune-response node, is solely linked to the BioCarta PPAR-alpha pathway gene set. Apart from the 12 mutual neighbors of these two top lipid metabolism gene sets, the two gene sets also overlap (combined overlap similarity = 0.33), sharing 6 genes in total, including the PRKACG, PRKAR1A, PRKAR2B, PRKACB, PRKAR1B and PRKAR2A. The genes shared in common by BioCarta CFTR and PPAR-alpha pathways are protein-coding genes that are members of the serine/threonine protein kinase family, coding various subunits of the cyclic AMP (cAMP) dependent protein kinase, important to processes such as cell proliferation and differentiation [32]. Of these 6 shared genes, only PRKACG (protein kinase cAMP-activated catalytic subunit gamma) reaches statistical significance (FDR = 0.129). ## 2.4. Interaction Analysis We performed interaction gene-set analysis, using MAGMA (v1.10), across all 145 gene sets found in our earlier analysis with FDR < 0.25, aiming to investigate how the genes shared between pairs of gene sets (interaction term) may contribute to the significance of individual gene sets, i.e., whether the statistical significance of two sets may be derived from the genes shared between them. Upon running interaction analysis, MAGMA yielded 71 valid interactions between all possible pairs of gene sets in the analysis, of which 4 reached marginal statistical significance (p-value < 0.05), shown in Table 9. To interpret the interaction results, Enrichr [88,89,90] was used to investigate the functional enrichment in gene ontology terms of the shared genes between the interacting gene sets. The interactions of the BioCarta PPAR-alpha and GPCR pathways, as well as the BioCarta PPAR-alpha and CREB pathways, yielded the same top enriched gene sets (sorted by their adjusted p-values using the Benjamini–Hochberg multiple testing correction method) in gene ontology biological processes (BP), molecular functions (MF) and cellular components (CC), including activation of protein kinase A activity (GOBP:0034199) (PPAR-alpha ∩ GPCR p-value = 2.92 ×10−17, q-value = 7.50 ×10−15; PPAR-alpha ∩ CREB p-value = 1.28 ×10−16, q-value = 4.20 ×10−14), cAMP-dependent protein kinase inhibitor activity (GOMF:0004862) (PPAR-alpha ∩ GPCR p-value = 2.20 ×10−12, q-value = 9.47 ×10−11; PPAR-alpha ∩ CREB p-value = 5.19 ×10−12, q-value = 2.34 ×10−10) and plasma membrane raft (GOCC:0044853) (PPAR-alpha ∩ GPCR p-value = 7.80 ×10−6, q-value = 1.24 ×10−4; PPAR-alpha ∩ CREB p-value = 1.27 ×10−7, q-value = 3.04 ×10−5). In addition, the shared genes by the KEGG Hematopoietic cell lineage and ECM receptor interaction were enriched in: extracellular structure organization (GOBP:0043062) (p-value = 3.23 ×10−15, q-value = 4.75 ×10−13), external encapsulating structure organization (GOBP:0045229) (p-value = 3.37 ×10−15, q-value = 4.75 ×10−13), collagen binding involved in cell-matrix adhesion (GOMF:0098639) (p-value = 4.54 ×10−5, q-value = 4.09 ×10−4), neuregulin binding (GOMF:0038132) (p-value = 4.54 ×10−5, q-value = 4.09 ×10−4) and focal adhesion (GOCC:0005925) (p-value = 4.96 ×10−11, q-value = 1.00 ×10−9). Lastly, the top ontology enrichment results for the genes shared by the KEGG Gap junction and the BioCarta GPCR pathway, are activation of protein kinase A activity (GOBP:0034199) (p-value = 6.10 ×10−8, q-value = 1.27 ×10−5), protein serine/threonine kinase activity (GOMF:0004674) (p-value = 3.43 ×10−7, q-value = 1.24 ×10−5), late endosome (GOCC:0005770) (p-value = 0.004, q-value = 0.079) and early endosome (GOCC:0005769) (p-value = 0.007, q-value = 0.079). ## 3. Discussion This study has identified genes and functional gene sets that are associated with amyotrophic lateral sclerosis, and the functions of these gene sets have been explored using network approaches aiming to understand the underlying pathology of the disease. Twenty-four gene sets were observed to have robust (FDR < 0.05) association to ALS. The functional roles of these gene sets mainly concerned neuron and embryonic development, the immune response, lipid metabolism, and nervous and muscle system processes (Table 2 and Figure 2). These main findings are also summarized in Figure 8. ## 3.1. Gene Level Confirmation The MAGMA multi-model gene-level results of the ALS-control cohort yielded 6 genes that reached high statistical significance based on false-discovery rate (FDR < 0.01) and 4 genes that passed the stricter multiple testing correction method of Bonferroni (alpha = 0.05; p-value < 2.58 ×10−6). The reported genes have been previously associated with ALS by numerous studies, supporting the reproducibility of the present analysis. Some of the more well-established ALS-associated genes in our analysis include C9ORF72 [30], UNC13A [31] and KIF5A [9], which were also reported in the original GWAS study reporting the ALS cohort that we used [9]. The analysis also revealed two previously identified ALS-associated genes TBK1 (TANK binding kinase) (FDR = 0.063) and FUS (fused in sarcoma; FDR = 0.155) [31], with marginal statistical significance. ## 3.2. Gene Set Association To interpret the gene set results a combination of manual curation, Cytoscape, and Enrichment Mapping, was used to visualize biological category-specific subnetworks. It is noteworthy that the majority ($67\%$) of the 24 gene sets having robust (FDR < 0.05) association to ALS have also a very high degree (i.e., number of undirected edges shared with other gene sets/nodes) within the ALS enrichment network. These gene sets also form a highly interconnected cluster and are present in every subnetwork category, usually as one of the more popular nodes. An example of this concerns the BioCarta PPAR-alpha and CFTR gene sets within the lipid metabolism network, which are simultaneously the most popular nodes and the most highly associated with ALS within the lipid metabolism category. Gene sets/nodes with a high degree (also termed as hub nodes) are often of high functional relevance, indicating an important topological role in functionally connecting multiple biological pathways within a complex biological network [91]. This high connectivity and popularity among the most ALS-associated gene sets, which also expands to the rest of the subnetworks, may suggest important underlying roles of these particular pathways in ALS pathology. Among the most significant GSA results, pathways related to neuronal and immune systems were the most abundant, with relatively larger network size (in terms of nodes) and density (in terms of the number of edges) in comparison to other categories. A driving rationale in this study was to include a general view of ALS mechanisms to enable the identification of overlapping functions while being as data-driven as possible. To satisfy this, our analysis tested some 31,454 gene sets, representing a huge range of functions, both specific but also quite generic, and thereby identifying a large number of gene set associations, which then presented a challenge for interpretation. Several of the identified gene sets represent processes or structures that are known or thought to be involved in secondary/downstream cellular pathology of ALS, including apoptosis, the cytoskeleton, homeostasis, the cell cycle, and immune response. However, care should be taken in interpreting these observed associations: the direction of causal relationship for a genomic association can only be from the genome to the disease, not the reverse. The mechanistic consequence of a genetic variant may indeed be downstream of other molecular-cellular events: in this case, certain characteristics (environmental or other genetic factors) might predispose an individual to ALS, but some of the identified disease-associated genomic features could be causal on top of (i.e., having downstream consequences of) those characteristics. The potential causal involvement of the identified gene sets, especially of pathways having less directly studied relevance to ALS (for example, neurodevelopmental processes and pathways such as gastrulation, neural tube development, and Shh pathways), may be of interest for further investigation. It may be of interest to note that certain processes, despite their heavy implication in ALS, were not more prominent among gene set associations. For example, mechanisms of DNA damage and repair, although implicated in several common and less frequent ALS mutations [92], were only indirectly represented in the gene set associations: the CDMAC pathway, reported above in the immune response findings, relates to cadmium genotoxicity. Processes relating to RNA/protein transport, ER stress, and autophagy, were also not prominent, although several gene sets related to these were present within other functional categories: for example, vesicle-mediated transport to the plasma membrane, the anterograde dendritic transport of neurotransmitter receptor complex and the PPAR-alpha pathway, all have association to ALS (each being listed among one or more of the functional categories reported in Table 3, Table 4, Table 5, Table 6 and Table 7), as well as an enrichment to early and late endosomes in the shared genes of the interacting terms KEGG Gap junction and BioCarta GPCR pathways. ALS is a neurodegenerative disease characterized primarily by the loss of the upper and lower motor neurons. However, the mechanisms that affect motor neurodegeneration have not been fully elucidated. A pathological hallmark of ALS is the aberrant misfolding, aggregation, and deposition of protein inclusions formed by TAR DNA-binding protein of 43 kDa (TDP-43), Cu/Zn superoxide dismutase (SOD1), or fused in sarcoma (FUS) in motor neurons [76]. The significant gene sets show several connections to motor neuron pathology and the major ALS disease-associated protein of TAR DNA-binding protein of 43 kDa (TDP-43). The nervous system category included five gene sets that reached robust statistical significance (FDR < 0.05), including the BioCarta CREB, Shh, DREAM, CK1 and AGPCR pathways. The BioCarta CREB (cyclic AMP-responsive element binding) gene set is highly associated with ALS (FDR = 0.039) and highly popular in the ALS network presented here, and has been recently linked to altered expression of TDP-43 [93]. A recent study suggested that TDP-43 RNA targets are enriched in signaling pathways of the CREB transcription factor, and that TDP-43 dysfunction inhibits the activation of CREB, restricting dendritic complexity [93]. Altered TDP-43 expression results in reduced dendritic growth [94] and in vivo experiments inducing altered dendritic growth result in disrupted neuronal connectivity and cell communication, which may be causal of ALS [95,96]. A progressive inhibition of CREB activation has been observed in human C9ORF72-mutant motor neurons while also exhibiting an underexpression of synaptic genes and synaptic loss [97]. The DREAM (downstream regulatory element antagonistic modulator) (FDR = 0.040) transcriptional regulator is expressed in spinal cord neurons and has been linked to ALS-associated neuronal death and has been found to be up-regulated in SOD1 mice and ALS patients [98]. DREAM has been associated with increased apoptotic activity towards motor neurons and astrocytes induced by a progressive calcium-dependent excitotoxicity that ultimately leads to neuronal damage and motor neuron loss in ALS [98]. In addition, the highly associated Casein kinase 1 (CK1) gene set in our analysis, a dopaminergic signaling pathway in the neostriatum, has been reported to have links with motor neuron degenerative diseases including ALS [72]. Specifically, there is evidence of abnormal TDP-43 hyperphosphorylation in the brain of ALS patients and it has been shown that CK1 mediates this hyperphosphorylation on TDP-43, highlighting an important role of CK1 in ALS and other neurodegenerative diseases [72,99]. Prominent patterns of ALS-associated gene sets were observed within the immune response category, related mainly to inflammation, T cell activation/regulation, T cell receptor signaling processes, cytokine activities and innate immunity. Despite that ALS has not been generally considered a disease affected primarily by autoimmunity and/or immunodeficiency, there is increasing evidence that immune dysregulation and neuroinflammation affect its onset and progression [100,101]. Neuroinflammation is a process that is generally observed in the context of infection, injury or degeneration, and it is described by the reaction of glial (astrocytes, microglia) and infiltrating immune cells (monocytes, neutrophils, lymphocytes) with cells of the central nervous system [102]. Previous studies have highlighted neuroprotective and neurotoxic phenomena that both derive from inflammation and appear to be specific to the progression phase of ALS [101]. We identified key immune response regulation pathways that have neuroprotective and immunomodulatory roles that prevent inappropriate and hazardous T-cell activation that could lead to autoimmune pathogenesis. T cells play a role in ALS pathology and have been identified in autopsy tissues from ALS patients [102,103]. Such gene sets include the most strongly ALS-associated gene set in our analysis, the BioCarta CSK pathway (FDR = 0.009), which is implicated in the inhibition of T-cell receptor signaling and T-cell activation, BioCarta VIP (vasoactive intestinal peptide) (FDR = 0.049) and CTLA-4 (cytotoxic T-lymphocyte antigen-4) (FDR = 0.054). Interestingly, the VIP pathway inhibits the apoptosis of activated T cells through two neuropeptides VIP and PACAP (pituitary adenylate cyclase-activating polypeptide); these two neuropeptides are key regulators of the function of microglial cells during myelin degeneration and have been associated with a number of neurodegenerative diseases [34]. Microglia and astrocytes have been well studied for their role in neurodegeneration in ALS and their inflammatory and potentially neuroprotective attributes after activation, which consist of one of the hallmarks of ALS pathology [100,104]. The ALS-associated BioCarta NFAT pathway (FDR = 0.069) is also associated with the activation of T cells, and the NFAT transcription factor has been found to be involved with the regulation of pro-inflammatory responses in cultured murine microglia [74]. It has been previously proposed that an innate immune response (rather than an adaptive immune response) is responsible for ALS-specific neuroinflammation [102,103]. Innate immune cells including monocytes, neutrophils, dendritic cells, macrophages and mast cells have been linked to ALS pathology, while lymphocytes and monocytes have been linked to immune dysregulation [100,102]. Our GS-GWAS analysis identified several ALS-associated pathways related to the innate immune system and to circulating immune cells, including the BioCarta monocyte (FDR = 0.040), lymphocyte (FDR = 0.046), NK (natural killing) cells (FDR = 0.019) and neutrophil (FDR = 0.226) pathways. In addition, the CFTR pathway (FDR = 0.046) was also found to be significantly enriched, associated among others with the innate immune system and expressed in macrophages and neutrophils [37,38]. This may be relevant to previous suggestions that immune cell infiltration is increased in ALS patients [100]. Several significant gene sets in the analysis were also related to inflammatory cytokines. Cytokine activities have been previously associated with ALS neuroinflammation pathology and to a possibly autoinflammatory state [100,101,103]. Another interesting observation was a strong ALS association to oxidized phospholipid pathways, which have been recently proposed as novel mediators of neurodegeneration [105]. Oxidized phospholipids have been linked to pro-inflammatory and anti-inflammatory responses through the stimulation of endothelial cells producing inflammatory cytokines as well as in acute and chronic microbial infections and oxidative stress [42,43]. Inflammation and oxidative stress can lead to a reactive oxygen species (ROS)-induced lipid peroxidation, generating oxidized phospholipids [105,106]. Oxidized phosphatidylcholines (OxPCs) have been proposed as novel neurotoxins requiring neutralization by microglia, and biomarkers of oxidative stress [105]. In addition, OxPCs have been detected in lesions of ALS patients, among other neurodegenerative disorders, including multiple sclerosis, frontotemporal lobe dementia and spinal cord injury [105]. The fourth most strongly ALS-associated gene set of our analysis (and second in the immune-response category) was the Gargalovic response to oxidized phospholipids black up (FDR = 0.032). This gene set refers to a network module containing genes that are up-regulated in primary aortic endothelium cells after exposure to oxidized phospholipids [42]. The authors who generated this curated gene set ((black module) found that it was functionally enriched in genes involved in Prion disease [42]. This is apparent also from our enrichment map, as the only connection in the ALS network (as shown in Figure 5) of the Gargalovic response to oxidized phospholipids black up gene set is to the BioCarta Prion pathway (FDR = 0.054). The Prion pathway has been associated with neurotoxicity and neurodegeneration. Prion diseases are a group of fatal neurodegenerative diseases, caused by the misfolding and self-propagation of prion proteins and it has been proposed that ALS shares parallel prion-like mechanisms involved in the disease pathogenesis [75,76]. The third most highly significant gene set of the GS-GWAS analysis, PID RhoA (Ras homolog gene family member A) pathway (FDR = 0.022), which is primarily linked to the cytoskeleton and to cell-signaling processes, has also been associated with prion-related neurodegeneration [68]. We also report a noteworthy ALS association to the peroxisome proliferator-activated receptor alpha (PPAR-alpha) pathway, which displays multiple interesting patterns in the current analysis. PPARs belong to the family of ligand-regulated nuclear receptors and can be activated, among other lipid categories, by oxidized phospholipids. The lipid metabolism gene set BioCarta PPAR-alpha pathway (FDR = 0.046) was among the top ALS-associated results and has the second highest degree within the ALS enrichment network. Specifically, the PPAR-alpha phosphoprotein has been linked to immune and inflammation responses, autophagy in human microglia, oxidative phosphorylation, regulation of the oxidative metabolism of skeletal muscle and energy homeostasis [61,85,87]. PPAR-alpha has been proposed as a putative novel therapeutic target in various diseases, including ALS, slowing the progression of the disease [107,108]. Specifically, an in vivo ALS study demonstrated that the activation of PPAR-alpha results in neuroprotection, neuroinflammation reduction, and neurodegeneration blocking [108]. PRKCB was the highest ALS-associated gene within the BioCarta PPAR-alpha pathway (FDR = 0.037). PRKCB is a member of the serine- and threonine-specific protein kinases family, and it has been associated with various processes, including, among others, immune response homeostasis and initiation [109], lipid peroxidation and induced ferroptosis [110] and Alzheimer’s disease pathogenesis [111]. The BioCarta PPAR-alpha pathway is connected with the equally highly significant BioCarta cystic fibrosis transmembrane conductance regulator (CFTR) pathway (FDR = 0.046), sharing an almost complete overlap of first significant neighbors within the lipid subnetwork. Putative links between ALS and cystic fibrosis (CF) have been noted previously, with TAR DNA-binding protein 43 (TDP-43) dysfunction being linked to the underlying pathology of both diseases [112]. Lastly, our results show several neurodevelopmental pathways that may have a role in ALS pathology. In fact, the most strongly significant ALS-associated gene set was the BioCarta mPR (membrane progesterone receptor) pathway (FDR = 0.009) which involves oocyte maturation by progesterone. MPRs have been previously associated with neuroprotective functions in neurons [51]. Progesterone is a pleiotropic regulator of neurons and glial cells and has been linked with decreased neuroinflammation, neuroprotection, neuronal survival and antioxidant effects in ALS mouse models [113]. In addition, the BioCarta Sonic Hedgehog (Shh) pathway (FDR = 0.040) is involved in several important neuron developmental processes, including, among others, the development of the neural tube, motor neurons, the regulation of CNS polarity, neuronal regeneration and proliferation as well as exhibiting a cytoprotective role against oxidative and excitotoxic stress [52,53,54]. There is evidence of inhibition of the Shh signaling pathway in ALS cerebrospinal fluid samples [114], as well as cytoprotective effects against oxidative stress in in vitro models of ALS, suggesting a potential role of the Shh pathway in ALS pathology [115]. ## 3.3. Gene Set Interaction Analysis The interaction analysis showed that the BioCarta PPAR-alpha pathway interacts with CREB (cyclic AMP response element-binding protein 1) (p-value = 0.014) and GPCR (G-protein coupled receptor) pathways (p-value = 0.019). The CREB pathway is the most highly enriched gene set within the nervous system category (FDR = 0.039) and the most popular node within the ALS network. CREB affects a variety of signaling cascades including the ERK$\frac{1}{2}$, MAPK, PI3K/AKT, CaMK, PKC, and PKA pathways [59]. The shared genes within both interactions as well as the shared genes between the BioCarta PPAR-alpha and the CFTR pathways are protein-coding genes that are members of the serine/threonine protein kinase family, coding various subunits of cyclic AMP (cAMP) dependent protein kinase. This set of genes has been found to be significantly enriched among others to the gene ontology biological processes, “activation of protein kinase A activity” and “negative regulation of cAMP-dependent protein kinase activity”. The cAMP-dependent protein kinase A (PKA) pathway has been associated with the rescued mislocalization of TDP-43—one of the main pathological hallmarks for neurodegeneration in ALS [116]. The same study proposed PKA as a novel drug target for ALS [116]. In addition, an earlier study showed a statistically significant elevated expression of the protein kinase A in fractions of spinal cord tissue of ALS patients [117]. ## 3.4. Limitations Several limitations should be taken into account in interpreting the results of the current work. Despite the fact that we employed the largest ALS individual-level genomic data with European descent releases to date, the power of GWAS is highly dependent on sample size, so the genotyping of larger ALS cohorts could potentially unravel further genetic associations and consequently in-depth knowledge about the pathology of the disease. We did not consider very rare variants (MAF < $0.5\%$), as such variants are typically removed from GWAS analyses, being considered putative false positives. Rare variants have been proposed to play a key role in ALS. As with any GWAS study, it is possible to identify spurious associations, and much of our SNP-level methodology, including extensive quality control filtering, and careful consideration of population structure, was devoted to the minimization of this risk. In addition, the MAGMA tool has features taking into account gene density (representing the relative level of linkage disequilibrium between SNPs in each gene), gene size (number of SNPs), and gene–gene correlations, but these do not eliminate the possibility of identifying spurious associations. Finally, the gene-set analysis results are highly dependent on the current knowledge of biological pathways stored in gene-set annotation databases. Our gene-level analysis considers only protein-coding genes and their neighboring regulatory regions by including upstream and downstream windows of 20 kb. In addition, our analysis does not include multi-locus events in the genome. The incorporation of non-coding loci, regulatory elements, as well as epistatic events in the genome, has been proposed as a way to gain insight into disease mechanisms [16]. ## 4.1. Datasets To form the ALS-control cohort, two restricted access dbGaP projects have been used. The first project refers to the largest release of European descent of ALS individual-level genotype data to date, with accession number phs000101.v5.p1 and contains genotype data of multiple genomic platforms for a total of 15,480 samples including patients suffering from amyotrophic lateral sclerosis diagnosed using the El Escorial criteria and healthy controls [9,26]. From this project, we collected individual-level genotype data of 471,303 SNPs from 12,319 people (7015 males, 5214 females, 90 ambiguous) of which 10,047 were cases, 2181 were controls and 91 phenotypes were missing. These data were further processed and filtered using quality control strategies described in subsequent sub-chapters. To increase the power of our analysis, we collected a second control cohort from dbGaP with accession number phs000428.v2.p2 (Health and Retirement Study). In order to keep a homogeneous cohort, we excluded from the latter study all the self-reported African-American samples, leaving 13,210 samples containing 2,315,518 variants for subsequent analysis. The age of the recruited individuals from the Health and Retirement Study ranged from 55 to 79 years of age. ## 4.2. Genomic Quality Control Analysis We used snpQT (v0.1.7) as the main software for quality control (QC), population stratification, association analysis, pre-imputation and post-imputation QC. snpQT is an automatic Nextflow software that ensures reproducibility and scalability, with an easy-to-use design [118]. To avoid potential batch effects, we applied sample and variant QC to each cohort, separately, using the snpQT --qc workflow. Sample QC workflow in both cohorts included first the removal of variants with a very poor quality (call rate < $90\%$)—which would be removed in subsequent analysis anyway to avoid removing extra samples—and then the removal of samples with sex discrepancies, a missing phenotype (i.e., case/control), poor quality (call rate < $98\%$), and extreme heterozygosity (deviating 3 standard deviation units from the mean) as well as duplicated and cryptic relatedness samples (relatedness > $12.5\%$). When sample QC was completed, the next step was variant QC workflow, which included the removal of variants with poor quality (call rate < $98\%$), deviation from the Hardy–Weinberg equilibrium (p-value < 10−7), with a minor allele frequency (MAF) less than $1\%$ and with a significant difference in call rate between cases and controls (p-value < 10−7). The latter check was not carried out for the aging cohort, as all the samples are controls. When sample and variant QC were completed, the next step was to merge the two cohorts keeping only common variants between the two datasets. Before merging the datasets, we assured that they were aligned in the same human genome build, flipped variants to the forward strand where necessary, and removed any palindromic SNPs, using PLINK [119] and snpflip (https://github.com/biocore-ntnu/snpflip accessed on 15 January 2023, version 0.0.6). After the ALS and aging cohorts were merged, we used snpQT (v0.1.7) [118] to perform sample QC, population stratification to remove sample outliers, and variant QC, combining --qc and --pop_strat workflows. The checks and the applied thresholds on the merged cohort included the removal of samples with the following variants: very poor quality (call rate < $90\%$); poor quality (call rate < $98\%$); extreme heterozygosity (deviating 3 standard deviation units from the mean); duplicated samples; cryptic relatedness (PLINK king-ship coefficient > 0.125); outlier samples using EIGENSOFT (sigma threshold = 4) and variants with poor quality (call rate < $98\%$); variants that deviate from the Hardy–Weinberg equilibrium (p-value < 10−7); rare variants with a minor allele frequency (MAF) less than $1\%$ and with a significant difference in call rate between cases and controls (p-value < 10−7). While a thorough quality control procedure was carried out for each dataset separately, it was important to address errors and batch effects that usually emerge when merging multiple datasets deriving from different analytical platforms genotyped at different times and places [120]. To further handle potential batch effects between the two cohorts, we followed two main strategies. First, we labeled cases as “missing”, controls from the ALS cohort as “cases” and controls from the aging cohort as “controls”, performed logistic regression to this artificially labeled cohort and removed 3519 variants with high statistical association (p-value < 10−6). Lastly, we removed 62,309 variants with a statistically significant differential case-control call rate (p-value < 10−6) on a second artificial merged dataset, labeling all the samples from the ALS cohort and the aging cohort as “cases” and “controls”, respectively. To maximize the size of our final cohort through meta-analysis, we collected the excluded outlier samples after population stratification and followed a separate cycle of quality control analyses, and we refer to these as the minor case/control cohort. This minor case/control cohort consists of 399,837 variants and 3058 samples of which 672 and 2386 samples are cases and controls, respectively. For the minor case/control cohort, we removed any samples which overlapped with non-European ancestry cohorts. Then, we followed the same sample/variant QC, population stratification and batch effects correction strategies as we did for our major case/control cohort, described above. ## 4.3. Imputation To increase the number of SNPs in the two processed ALS-control datasets, we used the Sanger imputation service provided by the Wellcome Sanger Institute [121]. Before uploading the ALS-control cohorts to the Sanger imputation server, we employed the snpQT pre-imputation quality control workflow [118]. The snpQT pre-imputation quality control aligns with the Sanger imputation service prerequisites, including checks for the chromosome codes, removing duplicated variants, and lastly, using bcftools’s plug-in +fixref to correct for remaining discrepancies (e.g., fixing the reference allele with respect to a FASTA 1000 genome reference dataset). In the Sanger imputation service, we chose the Haplotype Reference Consortium [121] as a reference panel, EAGLE2 [122] as a phasing software and positional Burrows–Wheeler transform (PBWT) [123] as an imputation algorithm. We further applied post-imputation quality control on the imputed dataset, using the snpQT post-imputation quality control workflow. This workflow included the removal of rare variants with a minor allele frequency (MAF) less than $0.5\%$; variants with an INFO score less than 0.4; duplicated variants in terms of their ID; and variants that corresponded to the same SNP (duplicated records, merged variant ids among multiple dbGaP versions). ## 4.4. Genome-Wide Association Analysis We used the snpQT GWAS workflow for both imputed cohorts to acquire SNP associations to the ALS phenotype [118]. For this purpose, snpQT uses PLINK2’s generalized linear regression model and offers the option of covariates. We used gender and the first three principal components to adjust for fine-scale population structure as covariates for all cohorts. Following post-imputation quality control, the major case/control cohort included 9,056,962 variants genotyped from 20,981 samples of which 8802 were cases and 12,179 controls. In terms of gender, the major case/control cohort included 10,510 females and 10,471 males. The minor case/control cohort consisted of 9,492,483 variants and 1058 samples: 442 cases and 616 controls; 475 females and 583 males. In Figure 9 and Figure 10, the Q-Q (Quantile-Quantile) and Manhattan plots for the two imputed ALS-control cohorts, respectively, are shown, generated from snpQT. The Manhattan plot illustrates the association p-values (shown on the y-axis), and the chromosomal positions of the tested genetic variants (shown on the x-axis), whereas the Q-Q plot shows the relationship between the observed and the expected quantile p-values of the ALS/control cohort under a normal distribution. Moreover, the Q-Q plot shows the lambda genomic inflation coefficient (calculated as median 1df chi-square stat/0.456), used to control false positive results and thus, reduce prediction errors. ## 4.5. Annotation and Gene Analysis To discover ALS-associated genes and gene sets, we used the command-line software multi-marker analysis of genomic annotation (MAGMA version 1.10) [22]. Prior to the gene analysis, we performed an annotation step for all cohorts, mapping SNPs into genes based on their genomic location, using the NCBI human genome build 37 location file available at MAGMA website and the resulting.bim PLINK files of the ALS-control cohorts. We chose a window of 20 kb upstream and downstream of each gene in order to also include regulatory elements in our analysis. The gene-level annotation step, mapping each genetic variant to a gene, yielded 4,788,553 and 5,023,961 variants for the major and minor ALS-control cohorts, which were mapped to at least one gene, respectively. MAGMA does not support analyses including the Y chromosome; therefore, no SNPs were mapped to chromosome Y. The next step was gene analysis, where multiple SNP p-values are summarized into genes using the MAGMA SNP-wise multi-model. The multi SNP-wise model combines two different genetic architectures of SNP-wise mean and SNP-wise top,1 models, generating one aggregate multi p-value. The SNP-wise mean model is more sensitive to the mean association of the mapped SNPs of a gene, whereas SNP-wise top,1 considers only the highest associated SNP mapped to each gene. Additional to the calculation of the gene p-values, the gene correlations are estimated while accounting for linkage disequilibrium (LD) phenomena [22,124]. Gene analysis was performed on the two GWAS datasets, providing the total number of samples, as well as the number of female and male samples, for each GWAS cohort. ## 4.6. Gene Meta-Analysis To utilize the power of both of the collected and analyzed ALS-control cohorts, we performed a meta-analysis at the gene level to combine the gene level results of the two cohorts (major and minor case/control cohorts) and then gene-set analysis, using the command-line software multi-marker analysis of genomic annotation (MAGMA version 1.10) [22]. MAGMA uses the weighted Stouffer’s Z method in order to combine the Z-scores for each gene across cohorts. MAGMA does not require a perfect overlap of genes between cohorts. The software checks that each gene across the cohorts has the same genomic locations, and then the meta-analysis for each gene is performed using any cohort available for the particular gene [22]. The meta-analysis yielded a total of 19,242 genes, from 22,039 samples of 9244 ALS cases (3852 females and 5392 males) and 12,795 healthy controls (7133 females and 5662 males). ## 4.7. Gene-Set Analysis For gene-set analysis, the MAGMA competitive model was employed, testing if there is a statistically significant combined association of the genes within a gene set with the phenotype of interest, in comparison with the genes not in the gene set [19,22]. To extract groups of gene sets, we used the Molecular Signatures Database (MSigDB v7.5) [24,25]. In Table 10, the categories of the 31,454 collected gene sets are summarized. ## 4.8. Interaction Analysis Interaction gene-set analysis was performed on all possible pairs of the 145 gene sets that were statistically significant in the previous analysis, and the subset of common genes that they share. By performing interaction gene-set analysis for each pair of these gene sets using MAGMA [22], we aimed to determine whether each of these gene sets had a statistically significant association with the ALS trait in its own right, or whether the significance of one or both of each pair may be derived from the genes shared by the two. In order for MAGMA to carry out interaction analysis between a pair of gene sets, there must be an overlap between the sets without one set being completely contained within the other. By default, MAGMA will consider an interaction term between two gene sets as valid (i.e., testable) if it contains at least 25 genes, with each gene set containing at least 25 unique genes not within the interaction term, and with these unique genes representing at least $10\%$ of its gene set. Given that this analysis is exploratory, these parameters were set wider than the default in order to enable MAGMA to identify potential significant interactions, even for relatively small overlaps between gene sets. The parameters were set so that the interaction terms contained at least 10 genes, with each interacting gene set containing at least 10 unique genes which are not present within the interaction term, and that represented at least $1\%$ of the original gene set. ## 4.9. Enrichment Networks For the visualization of the gene-set analysis results, we used Cytoscape (version 3.9.0) to create enrichment maps [125]. We applied a cut-off of 0.25 false discovery rate (FDR) to retain the highest statistically significant gene sets. An enrichment map is a network where each node represents an ALS-associated gene set. Each node has a specific size depicting the number of genes that are included in each gene set and that have been used for the gene-level analysis in MAGMA. Furthermore, we colored the nodes with their corresponding FDR as well as a meta-annotation category based on major biological processes: apoptosis and cell survival, homeostasis/cell metabolism, immune response, cell cycle, regulation of transcription/translation, muscle, cytoskeleton, lipid metabolism, nervous system, and developmental pathways. The biological categorization of the gene sets was carried out after curation of related literature and biological information resources such as UniProt [126], GeneCards [127], the molecular signatures database [24,25], AmiGO 2 [128,129,130] and KEGG [131]. Two nodes are connected through an edge that represents a minimum $10\%$ overlap (similarity) between the two gene sets. The similarity is represented by the average of the Jaccard coefficient and the overlap coefficient. The formulas of the two similarity coefficients for gene sets A and B are the following:Jaccard coefficient = [size of (A intersect B)]/[size of (A union B)].Overlap coefficient = [size of (A intersect B)]/[size of (minimum(A, B))]. Each edge is characterized by a certain width, based on the overlap of the connected gene sets/ nodes, expressed by the combined similarity coefficient. Lastly, we performed a network topology analysis using the Cytoscape (v3.9.0) analyze network tool, through which we obtained information about the number of nodes and edges within each network, and the nodes’ centrality, expressed by the degree of each node. The degree of each node represents the number of undirected edges that it contains and reflects its popularity within the network. ## 5. Conclusions This study identified significant gene set associations to ALS that are related to a variety of biological categories, including, immune response, neuron and embryonic development, lipid metabolism, nervous and muscle system processes, summarized in Figure 8. Prominent patterns of gene set significance relating to autoimmunity, immune dysregulation, and neuroinflammation were observed. We report the neuroprotective and immunomodulatory activities of T-cell receptor signaling and T-cell activation, inflammatory cytokines, innate immune responses and immune cell infiltration processes as associated with ALS. These results are in accordance with previous studies that provided supporting evidence for ALS-specific neuroinflammation and immune dysregulation mechanisms [102,103]. The potential causal implications of the gene set association may have relevance to the ongoing discussion of the roles of immune dysregulation and neuroinflammation in ALS, and whether they contribute actively to the progression of the disease, or if they are a consequence of motor neuron degeneration and injury [100,101]. Oxidative stress and protein degradation are established ALS-associated processes. The GS-GWAS results provide potential links of these to motor neurodegeneration and to the ALS disease-associated protein, TAR DNA-binding protein of 43 kDa (TDP-43), mainly deriving from nervous system-related gene sets, such as the BioCarta CREB, Shh, DREAM, CK1 and AGPCR pathways. Oxidized phospholipid pathways, generated by a ROS-induced lipid peroxidation and prion-related mechanisms, are ALS-associated, and may also play a role in the neurotoxicity and neurodegeneration of the ALS pathology. Competitive and interaction gene-set analyses revealed an intriguing pattern of functional relationships of the proliferator-activated receptor alpha (PPAR-alpha) pathway. PPAR-alpha has been linked among others to immune and inflammation responses, autophagy in human microglia, regulation of the oxidative metabolism of skeletal muscle, as well as key roles in neuroprotection, neuroinflammation reduction, and blocking of neurodegeneration [61,85,87,108]. The interaction analysis revealed an interaction term of the PPAR-alpha gene set with the CREB and GPCR pathways. PPAR-alpha, CREB and GPCR pathways were among the top associated gene sets in our analysis, with CREB and PPAR-alpha having the highest number of neighbors in the ALS network. The shared genes of the three gene sets were enriched with protein kinase A (PKA) and cAMP-dependent protein kinase activities. The cAMP PKA pathway has been associated with rescued mislocalization of TDP-43—one of the main pathological hallmarks for neurodegeneration in ALS [116]. Lastly, we note several neurodevelopmental pathways that our analysis suggests could have potential causal roles in ALS pathology, and have previously been linked to the disease through molecular studies. The top associated gene set, the mPR (membrane progesterone receptor) pathway (FDR = 0.009), has been associated with neuroprotective functions in neurons [51], decreased neuroinflammation, neuroprotection, neuronal survival and antioxidant effects in ALS mouse models [113]. The Sonic Hedgehog (Shh) pathway (FDR = 0.040) was present within the generated ALS gene set networks, and has been suggested to have a cytoprotective role against oxidative and excitotoxic stress in ALS pathology [52,53,54,115]. A future direction of this work could be to combine expression quantitative trait locus (eQTL) data and genomic data, using summary data-based Mendelian randomization (SMR), to test the chance that genetic variants that increase risk of a disease do so through modifying gene expression [132]. Finally, employing machine learning models to predict multi-locus interactions that are associated with ALS could contribute toward understanding the underlying mechanisms of this devastating disease [16]. ## References 1. Niedermeyer S., Murn M., Choi P.J.. Respiratory Failure in Amyotrophic Lateral Sclerosis. Chest (2019) 155 401-408. DOI: 10.1016/j.chest.2018.06.035 2. Chiò A., Logroscino G., Traynor B., Collins J., Simeone J., Goldstein L., White L.. Global Epidemiology of Amyotrophic Lateral Sclerosis: A Systematic Review of the Published Literature. Neuroepidemiology (2013) 41 118-130. DOI: 10.1159/000351153 3. Al-Chalabi A., Van Den Berg L.H., Veldink J.. Gene discovery in amyotrophic lateral sclerosis: Implications for clinical management. Nat. Rev. Neurol. (2017) 13 96-104. DOI: 10.1038/nrneurol.2016.182 4. Chiò A., Logroscino G., Hardiman O., Swingler R., Mitchell D., Beghi E., Traynor B.G.. Prognostic factors in ALS: A critical review. Amyotroph Lateral Scler. (2009) 10 310-323. DOI: 10.3109/17482960802566824 5. Arthur K.C., Calvo A., Price T.R., Geiger J.T., Chiò A., Traynor B.J.. Projected increase in amyotrophic lateral sclerosis from 2015 to 2040. Nat. Commun. (2016) 7 12408. DOI: 10.1038/ncomms12408 6. Rowland L.P., Shneider N.A.. Amyotrophic Lateral Sclerosis. N. Engl. J. Med. (2001) 344 1688-1700. DOI: 10.1056/NEJM200105313442207 7. Vijayakumar U.G., Milla V., Stafford M.Y.C., Bjourson A.J., Duddy W., Duguez S.M.R.. A systematic review of suggested molecular strata, biomarkers and their tissue sources in ALS. Front. Neurol. (2019) 10 400. DOI: 10.3389/fneur.2019.00400 8. Turner M.R., Al-Chalabi A., Chio A., Hardiman O., Kiernan M.C., Rohrer J.D., Rowe J., Seeley W., Talbot K.. Genetic screening in sporadic ALS and FTD. J. Neurol. Neurosurg. Psychiatry (2017) 88 1042-1044. DOI: 10.1136/jnnp-2017-315995 9. Nicolas A., Kenna K., Renton A.E., Ticozzi N., Faghri F., Chia R., Dominov J.A., Kenna B.J., Nalls M.A., Keagle P.. Genome-wide Analyses Identify KIF5A as a Novel ALS Gene. Neuron (2018) 97 1268-1283. DOI: 10.1016/j.neuron.2018.02.027 10. Chia R., Chiò A., Traynor B.J.. Novel genes associated with amyotrophic lateral sclerosis: Diagnostic and clinical implications. Lancet Neurol. (2018) 17 94-102. DOI: 10.1016/S1474-4422(17)30401-5 11. Volk A.E., Weishaupt J.H., Andersen P.M., Ludolph A.C., Kubisch C.. Current knowledge and recent insights into the genetic basis of amyotrophic lateral sclerosis. Med. Genet. (2018) 30 252-258. DOI: 10.1007/s11825-018-0185-3 12. Smukowski S.N., Maioli H., Latimer C.S., Bird T.D., Jayadev S., Valdmanis P.N.. Progress in Amyotrophic Lateral Sclerosis Gene Discovery. Neurol. Genet. (2022) 8 e669. DOI: 10.1212/NXG.0000000000000669 13. Zou Z.Y., Zhou Z.R., Che C.H., Liu C.Y., He R.L., Huang H.P.. Genetic epidemiology of amyotrophic lateral sclerosis: A systematic review and meta-analysis. J. Neurol. Neurosurg. Psychiatry (2017) 88 540-549. DOI: 10.1136/jnnp-2016-315018 14. Connolly O., Le Gall L., McCluskey G., Donaghy C.G., Duddy W.J., Duguez S.. A Systematic Review of Genotype–Phenotype Correlation across Cohorts Having Causal Mutations of Different Genes in ALS. J. Pers. Med. (2020) 10. DOI: 10.3390/jpm10030058 15. McLaughlin L.R., Vajda A., Hardiman O.. Heritability of amyotrophic lateral sclerosis insights from disparate numbers. JAMA Neurol. (2015) 72 857-858. DOI: 10.1001/jamaneurol.2014.4049 16. Vasilopoulou C., Morris A.P., Giannakopoulos G., Duguez S., Duddy W.. What Can Machine Learning Approaches in Genomics Tell Us about the Molecular Basis of Amyotrophic Lateral Sclerosis?. J. Pers. Med. (2020) 10. DOI: 10.3390/jpm10040247 17. Gall L.L., Anakor E., Connolly O., Vijayakumar U.G., Duguez S.. Molecular and cellular mechanisms affected in ALS. J. Pers. Med. (2020) 10. DOI: 10.3390/jpm10030101 18. Du Y., Wen Y., Guo X., Hao J., Wang W., He A., Fan Q., Li P., Liu L., Liang X.. A Genome-wide Expression Association Analysis Identifies Genes and Pathways Associated with Amyotrophic Lateral Sclerosis. Cell. Mol. Neurobiol. (2018) 38 635-639. DOI: 10.1007/s10571-017-0512-2 19. De Leeuw C.A., Neale B.M., Heskes T., Posthuma D.. The statistical properties of gene-set analysis. Nat. Rev. Genet. (2016) 17 353-364. DOI: 10.1038/nrg.2016.29 20. Vasilopoulou C., Duguez S., Duddy W.. Genome-Wide Gene-Set Analysis Approaches in Amyotrophic Lateral Sclerosis. J. Pers. Med. (2022) 12. DOI: 10.3390/jpm12111932 21. Maleki F., Ovens K., Hogan D.J., Kusalik A.J.. Gene Set Analysis: Challenges, Opportunities, and Future Research. Front. Genet. (2020) 11 654. DOI: 10.3389/fgene.2020.00654 22. de Leeuw C.A., Mooij J.M., Heskes T., Posthuma D.. MAGMA: Generalized Gene-Set Analysis of GWAS Data. PLoS Comput. Biol. (2015) 11 1-19. DOI: 10.1371/journal.pcbi.1004219 23. de Leeuw C.A., Stringer S., Dekkers I.A., Heskes T., Posthuma D.. Conditional and interaction gene-set analysis reveals novel functional pathways for blood pressure. Nat. Commun. (2018) 9 3768. DOI: 10.1038/s41467-018-06022-6 24. Liberzon A., Birger C., Thorvaldsdóttir H., Ghandi M., Mesirov J.P., Tamayo P.. The Molecular Signatures Database (MSigDB) hallmark gene set collection. Cell Syst. (2015) 1 417. DOI: 10.1016/j.cels.2015.12.004 25. Subramanian A., Tamayo P., Mootha V.K., Mukherjee S., Ebert B.L., Gillette M.A., Paulovich A., Pomeroy S.L., Golub T.R., Lander E.S.. Gene set enrichment analysis: A knowledge-based approach for interpreting genome-wide expression profiles. Proc. Natl. Acad. Sci. USA (2005) 102 15545-15550. DOI: 10.1073/pnas.0506580102 26. Mailman M.D., Feolo M., Jin Y., Kimura M., Tryka K., Bagoutdinov R., Hao L., Kiang A., Paschall J., Phan L.. The NCBI dbGaP database of genotypes and phenotypes. Nucleic Acids Res. (2007) 42 D975-D979. DOI: 10.1038/ng1007-1181 27. Anderson C.A., Pettersson F.H., Clarke G.M., Cardon L.R., Morris P., Zondervan K.T.. Data quality control in genetic case-control association studies. Nat. Protoc. (2011) 5 1564-1573. DOI: 10.1038/nprot.2010.116 28. Laurie C.C., Doheny K.F., Mirel D.B., Pugh E.W., Bierut J.L., Bhangale T., Boehm F., Caporaso N.E., Cornelis M.C., Edenberg H.J.. Quality control and quality assurance in genotypic data for genome-wide association studies. Genet. Epidemiol. (2011) 34 591-602. DOI: 10.1002/gepi.20516 29. Li C.Y., Yang T.M., Ou R.W., Wei Q.Q., Shang H.F.. Genome-wide genetic links between amyotrophic lateral sclerosis and autoimmune diseases. BMC Med. (2021) 19 27. DOI: 10.1186/s12916-021-01903-y 30. Renton A.E., Majounie E., Waite A., Simón-Sánchez J., Rollinson S., Gibbs J.R., Schymick J.C., Laaksovirta H., van Swieten J.C., Myllykangas L.. A hexanucleotide repeat expansion in C9ORF72 is the cause of chromosome 9p21-linked ALS-FTD. Neuron (2011) 72 257. DOI: 10.1016/j.neuron.2011.09.010 31. Van Es M.A., Veldink J.H., Saris C.G., Blauw H.M., Van Vught P.W., Birve A., Lemmens R., Schelhaas H.J., Groen E.J., Huisman M.H.. Genome-wide association study identifies 19p13.3 (UNC13A) and 9p21.2 as susceptibility loci for sporadic amyotrophic lateral sclerosis. Nat. Genet. (2009) 41 1083-1087. DOI: 10.1038/ng.442 32. Vang T., Torgersen K.M., Sundvold V., Saxena M., Levy F.O., Skålhegg B.S., Hansson V., Mustelin T., Taskén K.. Activation of the Cooh-Terminal Src Kinase (Csk) by Camp-Dependent Protein Kinase Inhibits Signaling through the T Cell Receptor. J. Exp. Med. (2001) 193 497-508. DOI: 10.1084/jem.193.4.497 33. Fife B.T., Bluestone J.A.. Control of peripheral T-cell tolerance and autoimmunity via the CTLA-4 and PD-1 pathways. Immunol. Rev. (2008) 224 166-182. DOI: 10.1111/j.1600-065X.2008.00662.x 34. Jansen M.I., Broome S.T., Castorina A.. Exploring the Pro-Phagocytic and Anti-Inflammatory Functions of PACAP and VIP in Microglia: Implications for Multiple Sclerosis. Int. J. Mol. Sci. (2022) 23. DOI: 10.3390/ijms23094788 35. Macian F.. NFAT proteins: Key regulators of T-cell development and function. Nat. Rev. Immunol. (2005) 5 472-484. DOI: 10.1038/nri1632 36. Murga-Zamalloa C., Wilcox R.A.. GATA-3 in T-cell lymphoproliferative disorders. IUBMB Life (2020) 72 170-177. DOI: 10.1002/iub.2130 37. Hanssens L.S., Duchateau J., Casimir G.J.. CFTR Protein: Not Just a Chloride Channel?. Cells (2021) 10. DOI: 10.3390/cells10112844 38. Yoshimura K., Nakamura H., Trapnell B.C., Chu C.S., Dakemans W., Pavirani A., Lecocq J.P., Crystal R.G.. Expression of the cystic fibrosis transmembrane conductance regulator gene in cells of non-epithelial origin. Nucleic Acids Res. (1991) 19 5417-5423. DOI: 10.1093/nar/19.19.5417 39. Ren Z., Yu Y., Chen C., Yang D., Ding T., Zhu L., Deng J., Xu Z.. The Triangle Relationship Between Long Noncoding RNA, RIG-I-like Receptor Signaling Pathway, and Glycolysis. Front. Microbiol. (2021) 12 807737. DOI: 10.3389/fmicb.2021.807737 40. Misra U.K., Gawdi G., Akabani G., Pizzo S.V.. Cadmium-induced DNA synthesis and cell proliferation in macrophages: The role of intracellular calcium and signal transduction mechanisms. Cell. Signal. (2002) 14 327-340. DOI: 10.1016/S0898-6568(01)00268-6 41. Cesaro T., Michiels T.. Inhibition of PKR by Viruses. Front. Microbiol. (2021) 12 757238. DOI: 10.3389/fmicb.2021.757238 42. Gargalovic P.S., Imura M., Zhang B., Gharavi N.M., Clark M.J., Pagnon J., Yang W.P., He A., Truong A., Patel S.. Identification of inflammatory gene modules based on variations of human endothelial cell responses to oxidized lipids. Proc. Natl. Acad. Sci. USA (2006) 103 12741. DOI: 10.1073/pnas.0605457103 43. Freigang S.. The regulation of inflammation by oxidized phospholipids. Eur. J. Immunol. (2016) 46 1818-1825. DOI: 10.1002/eji.201545676 44. McGeachy M.J., Cua D.J., Gaffen S.L.. The IL-17 family of cytokines in health and disease. Immunity (2019) 50 892. DOI: 10.1016/j.immuni.2019.03.021 45. Amatya N., Garg A.V., Gaffen S.L.. IL-17 Signaling: The Yin and the Yang. Trends Immunol. (2017) 38 310-322. DOI: 10.1016/j.it.2017.01.006 46. Pelaia C., Paoletti G., Puggioni F., Racca F., Pelaia G., Canonica G.W., Heffler E.. Interleukin-5 in the Pathophysiology of Severe Asthma. Front. Physiol. (2019) 10 1514. DOI: 10.3389/fphys.2019.01514 47. Lambrecht B.N., Hammad H.. The immunology of asthma. Nat. Immunol. (2015) 16 45-56. DOI: 10.1038/ni.3049 48. Mezu-Ndubuisi O.J., Maheshwari A.. The role of integrins in inflammation and angiogenesis. Pediatr. Res. (2021) 89 1619. DOI: 10.1038/s41390-020-01177-9 49. Holesh J.E., Bass A.N., Lord M.. Physiology, Ovulation (2022) 50. Fair T., Lonergan P.. The role of progesterone in oocyte acquisition of developmental competence. Reprod. Domest. Anim. Zuchthyg. (2012) 47 142-147. DOI: 10.1111/j.1439-0531.2012.02068.x 51. Thomas P., Pang Y.. Membrane Progesterone Receptors (mPRs): Evidence for Neuroprotective, Neurosteroid Signaling and Neuroendocrine Functions in Neuronal Cells. Neuroendocrinology (2012) 96 162. DOI: 10.1159/000339822 52. Choudhry Z., Rikani A.A., Choudhry A.M., Tariq S., Zakaria F., Asghar M.W., Sarfraz M.K., Haider K., Shafiq A.A., Mobassarah N.J.. Sonic hedgehog signalling pathway: A complex network. Ann. Neurosci. (2014) 21 28. DOI: 10.5214/ans.0972.7531.210109 53. Jha N.K., Chen W.C., Kumar S., Dubey R., Tsai L.W., Kar R., Jha S.K., Gupta P.K., Sharma A., Gundamaraju R.. Molecular mechanisms of developmental pathways in neurological disorders: A pharmacological and therapeutic review. Open Biol. (2022) 12. DOI: 10.1098/rsob.210289 54. Echelard Y., Epstein D.J., St-Jacques B., Shen L., Mohler J., McMahon J.A., McMahon A.P.. Sonic hedgehog, a member of a family of putative signaling molecules, is implicated in the regulation of CNS polarity. Cell (1993) 75 1417-1430. DOI: 10.1016/0092-8674(93)90627-3 55. Maronde E.. Cyclic Nucleotide (cNMP) Analogues: Past, Present and Future. Int. J. Mol. Sci. (2021) 22. DOI: 10.3390/ijms222312879 56. Linder J.U., Schultz J.E.. Use of Chimeric Adenylyl Cyclases to Study Cyclic Nucleotide Signaling. Handbook of Cell Signaling (2010) 1537-1542. DOI: 10.1016/B978-0-12-374145-5.00188-1 57. Atwood B.K., Lopez J., Wager-Miller J., Mackie K., Straiker A.. Expression of G protein-coupled receptors and related proteins in HEK293, AtT20, BV2, and N18 cell lines as revealed by microarray analysis. BMC Genom. (2011) 12. DOI: 10.1186/1471-2164-12-14 58. Cheng H.Y.M., Pitcher G.M., Laviolette S.R., Whishaw I.Q., Tong K.I., Kockeritz L.K., Wada T., Joza N.A., Crackower M., Goncalves J.. DREAM is a critical transcriptional repressor for pain modulation. Cell (2002) 108 31-43. DOI: 10.1016/S0092-8674(01)00629-8 59. Steven A., Friedrich M., Jank P., Heimer N., Budczies J., Denkert C., Seliger B.. What turns CREB on? And off? And why does it matter?. Cell. Mol. Life Sci. (2020) 77 4049. DOI: 10.1007/s00018-020-03525-8 60. Jiang J.. CK1 in Developmental signaling: Hedgehog and Wnt. Curr. Top. Dev. Biol. (2017) 123 303. DOI: 10.1016/BS.CTDB.2016.09.002 61. Luo R., Su L.Y., Li G., Yang J., Liu Q., Yang L.X., Zhang D.F., Zhou H., Xu M., Fan Y.. Activation of PPARA-mediated autophagy reduces Alzheimer disease-like pathology and cognitive decline in a murine model. Autophagy (2020) 16 52. DOI: 10.1080/15548627.2019.1596488 62. Brusés J.L.. N-cadherin signaling in synapse formation and neuronal physiology. Mol. Neurobiol. (2006) 33 237-252. DOI: 10.1385/MN:33:3:237 63. Lelièvre E.C., Plestant C., Boscher C., Wolff E., Mège R.M., Birbes H.. N-cadherin mediates neuronal cell survival through Bim down-regulation. PLoS ONE (2012) 7. DOI: 10.1371/journal.pone.0033206 64. Redies C., Treubert-Zimmermann U., Luo J.. Cadherins as regulators for the emergence of neural nets from embryonic divisions. J. Physiol. Paris (2003) 97 5-15. DOI: 10.1016/j.jphysparis.2003.10.002 65. Takeichi M., Matsunami H., Inoue T., Kimura Y., Suzuki S., Tanaka T.. Roles of cadherins in patterning of the developing brain. Dev. Neurosci. (1997) 19 86-87. DOI: 10.1159/000111189 66. Sun Z., Parrish A.R., Hill M.A., Meininger G.A.. N-cadherin, a vascular smooth muscle cell-cell adhesion molecule: Function and signaling for vasomotor control. Microcirculation (2014) 21 208-218. DOI: 10.1111/micc.12123 67. Guntur A.R., Rosen C.J., Naski M.C.. N-cadherin adherens junctions mediate osteogenesis through PI3K signaling. Bone (2012) 50 54-62. DOI: 10.1016/j.bone.2011.09.036 68. Kim H.J., Choi H.S., Park J.H., Kim M.J., Lee H.G., Petersen R.B., Kim Y.S., Park J.B., Choi E.K.. Regulation of RhoA activity by the cellular prion protein. Cell Death Dis. (2017) 8 e2668. DOI: 10.1038/cddis.2017.37 69. Marrs G.S., Theisen C.S., Brusés J.L.. N-cadherin modulates voltage activated calcium influx via RhoA, p120-catenin, and myosin-actin interaction. Mol. Cell. Neurosci. (2009) 40 390-400. DOI: 10.1016/j.mcn.2008.12.007 70. Muhr J., Ackerman K.M.. Embryology, Gastrulation (2022) 71. Zhang Y., Ulvmar M.H., Stanczuk L., Martinez-Corral I., Frye M., Alitalo K., Mäkinen T.. Heterogeneity in VEGFR3 levels drives lymphatic vessel hyperplasia through cell-autonomous and non-cell-autonomous mechanisms. Nat. Commun. (2018) 9 1296. DOI: 10.1038/s41467-018-03692-0 72. Perez D.I., Gil C., Martinez A.. Protein kinases CK1 and CK2 as new targets for neurodegenerative diseases. Med. Res. Rev. (2011) 31 924-954. DOI: 10.1002/med.20207 73. Kopp-Scheinpflug C., Forsythe I.D.. Nitric Oxide Signaling in the Auditory Pathway. Front. Neural Circuits (2021) 15 12. DOI: 10.3389/fncir.2021.759342 74. Nagamoto-Combs K., Combs C.K.. Microglial phenotype is regulated by activity of the transcription factor, NFAT (nuclear factor of activated T cells). J. Neurosci. (2010) 30 9641-9646. DOI: 10.1523/JNEUROSCI.0828-10.2010 75. Yoshida S., Hasegawa T.. Deciphering the prion-like behavior of pathogenic protein aggregates in neurodegenerative diseases. Neurochem. Int. (2022) 155 105307. DOI: 10.1016/j.neuint.2022.105307 76. Polymenidou M., Cleveland D.W.. The Seeds of Neurodegeneration: Prion-like Spreading in ALS. Cell (2011) 147 498. DOI: 10.1016/j.cell.2011.10.011 77. Gau D., Roy P.. SRF’ing and SAP’ing—The role of MRTF proteins in cell migration. J. Cell Sci. (2018) 131 jcs218222. DOI: 10.1242/jcs.218222 78. Cen B., Selvaraj A., Prywes R.. Myocardin/MKL family of SRF coactivators: Key regulators of immediate early and muscle specific gene expression. J. Cell. Biochem. (2004) 93 74-82. DOI: 10.1002/jcb.20199 79. Signaling I., Hakuno F., Takahashi S.I., Hakuno F., Takahashi S.I.. 40 YEARS OF IGF1: IGF1 receptor signaling pathways. J. Mol. Endocrinol. (2018) 61 T69-T86. DOI: 10.1530/JME-17-0311 80. Krauss R.S., Joseph G.A., Goel A.J.. Keep Your Friends Close: Cell–Cell Contact and Skeletal Myogenesis. Cold Spring Harb. Perspect. Biol. (2017) 9 a029298. DOI: 10.1101/cshperspect.a029298 81. Lehka L., Rȩdowicz M.J.. Mechanisms regulating myoblast fusion: A multilevel interplay. Semin. Cell Dev. Biol. (2020) 104 81-92. DOI: 10.1016/j.semcdb.2020.02.004 82. De Mello W.C., Danser A.H.. Angiotensin II and the Heart. Hypertension (2000) 35 1183-1188. DOI: 10.1161/01.HYP.35.6.1183 83. Islinger M., Voelkl A., Fahimi H.D., Schrader M.. The peroxisome: An update on mysteries 2.0. Histochem. Cell Biol. (2018) 150 443-471. DOI: 10.1007/s00418-018-1722-5 84. Trachootham D., Lu W., Ogasawara M.A., Valle N.R.D., Huang P.. Redox Regulation of Cell Survival. Antioxidants Redox Signal. (2008) 10 1343. DOI: 10.1089/ars.2007.1957 85. Rakhshandehroo M., Knoch B., Müller M., Kersten S.. Peroxisome proliferator-activated receptor alpha target genes. PPAR Res. (2010) 2010 393-416. DOI: 10.1155/2010/612089 86. Yoon M.. The role of PPARalpha in lipid metabolism and obesity: Focusing on the effects of estrogen on PPARalpha actions. Pharmacol. Res. (2009) 60 151-159. DOI: 10.1016/j.phrs.2009.02.004 87. Zhang J.Q., Long X.Y., Xie Y., Zhao Z.H., Fang L.Z., Liu L., Fu W.P., Shu J.K., Wu J.H., Dai L.M.. Relationship between PPAR. Bioengineered (2017) 8 723. DOI: 10.1080/21655979.2017.1346757 88. Chen E.Y., Tan C.M., Kou Y., Duan Q., Wang Z., Meirelles G.V., Clark N.R., Ma’ayan A.. Enrichr: Interactive and collaborative HTML5 gene list enrichment analysis tool. BMC Bioinform. (2013) 14. DOI: 10.1186/1471-2105-14-128 89. Kuleshov M.V., Jones M.R., Rouillard A.D., Fernandez N.F., Duan Q., Wang Z., Koplev S., Jenkins S.L., Jagodnik K.M., Lachmann A.. Enrichr: A comprehensive gene set enrichment analysis web server 2016 update. Nucleic Acids Res. (2016) 44 W90-W97. DOI: 10.1093/nar/gkw377 90. Xie Z., Bailey A., Kuleshov M.V., Clarke D.J., Evangelista J.E., Jenkins S.L., Lachmann A., Wojciechowicz M.L., Kropiwnicki E., Jagodnik K.M.. Gene Set Knowledge Discovery with Enrichr. Curr. Protoc. (2021) 1 e90. DOI: 10.1002/cpz1.90 91. Charitou T., Bryan K., Lynn D.J.. Using biological networks to integrate, visualize and analyze genomics data. Genet. Sel. Evol. (2016) 48 27. DOI: 10.1186/s12711-016-0205-1 92. Sun Y., Curle A.J., Haider A.M., Balmus G.. The role of DNA damage response in amyotrophic lateral sclerosis. Essays Biochem. (2020) 64 847-861. DOI: 10.1042/EBC20200002 93. Herzog J.J., Xu W., Deshpande M., Rahman R., Suib H., Rodal A.A., Rosbash M., Paradis S.. TDP-43 dysfunction restricts dendritic complexity by inhibiting CREB activation and altering gene expression. Proc. Natl. Acad. Sci. USA (2020) 117 11760-11769. DOI: 10.1073/pnas.1917038117 94. Herzog J.J., Deshpande M., Shapiro L., Rodal A.A., Paradis S.. TDP-43 misexpression causes defects in dendritic growth. Sci. Rep. (2017) 7 15656. DOI: 10.1038/s41598-017-15914-4 95. Kweon J.H., Kim S., Lee S.B.. The cellular basis of dendrite pathology in neurodegenerative diseases. BMB Rep. (2017) 50 5-11. DOI: 10.5483/BMBRep.2017.50.1.131 96. Stuart G.J., Spruston N.. Dendritic integration: 60 years of progress. Nat. Neurosci. (2015) 18 1713-1721. DOI: 10.1038/nn.4157 97. Catanese A., Rajkumar S., Sommer D., Freisem D., Wirth A., Aly A., Massa-L Opez D., Olivieri A., Torelli F., Ioannidis V.. Synaptic disruption and CREB-regulated transcription are restored by K+ channel blockers in ALS. EMBO Mol. Med. (2021) 13 e13131. DOI: 10.15252/emmm.202013131 98. Larrodé P., Calvo A.C., Moreno-Martínez L., De La Torre M., Moreno-García L., Molina N., Castiella T., Iñiguez C., Pascual L.F., Mena F.J.M.. DREAM-dependent activation of astrocytes in amyotrophic lateral sclerosis. Mol. Neurobiol. (2018) 55 1-12. DOI: 10.1007/s12035-017-0713-1 99. Kametani F., Nonaka T., Suzuki T., Arai T., Dohmae N., Akiyama H., Hasegawa M.. Identification of casein kinase-1 phosphorylation sites on TDP-43. Biochem. Biophys. Res. Commun. (2009) 382 405-409. DOI: 10.1016/j.bbrc.2009.03.038 100. Xie Y., Luo X., He H., Tang M.. Novel Insight Into the Role of Immune Dysregulation in Amyotrophic Lateral Sclerosis Based on Bioinformatic Analysis. Front. Neurosci. (2021) 15 657465. DOI: 10.3389/fnins.2021.657465 101. Morello G., Spampinato A.G., Cavallaro S.. Neuroinflammation and ALS: Transcriptomic Insights into Molecular Disease Mechanisms and Therapeutic Targets. Mediat. Inflamm. (2017) 2017 7070469. DOI: 10.1155/2017/7070469 102. Prinz M., Priller J.. The role of peripheral immune cells in the CNS in steady state and disease. Nat. Neurosci. (2017) 20 136-144. DOI: 10.1038/nn.4475 103. McCauley M.E., Baloh R.H.. Inflammation in ALS/FTD pathogenesis. Acta Neuropathol. (2019) 137 715. DOI: 10.1007/s00401-018-1933-9 104. Geloso M.C., Corvino V., Marchese E., Serrano A., Michetti F., D’Ambrosi N.. The dual role of microglia in ALS: Mechanisms and therapeutic approaches. Front. Aging Neurosci. (2017) 9 242. DOI: 10.3389/fnagi.2017.00242 105. Dong Y., Yong V.W.. Oxidized phospholipids as novel mediators of neurodegeneration. Trends Neurosci. (2022) 45 419-429. DOI: 10.1016/j.tins.2022.03.002 106. Catalá A.. Lipid peroxidation of membrane phospholipids generates hydroxy-alkenals and oxidized phospholipids active in physiological and/or pathological conditions. Chem. Phys. Lipids (2009) 157 1-11. DOI: 10.1016/j.chemphyslip.2008.09.004 107. Lee D., Tomita Y., Allen W., Tsubota K., Negishi K., Kurihara T.. PPAR. Life (2021) 11. DOI: 10.3390/life11111168 108. Esmaeili M.A., Yadav S., Gupta R.K., Waggoner G.R., Deloach A., Calingasan N.Y., Flint Beal M., Kiaei M.. Preferential PPAR-. Hum. Mol. Genet. (2016) 25 317-327. DOI: 10.1093/hmg/ddv477 109. Spitaler M., Cantrell D.A.. Protein kinase C and beyond. Nat. Immunol. (2004) 5 785-790. DOI: 10.1038/ni1097 110. Zhang H.L., Hu B.X., Li Z.L., Du T., Shan J.L., Ye Z.P., Peng X.D., Li X., Huang Y., Zhu X.Y.. PKC. Nat. Cell Biol. (2022) 24 88-98. DOI: 10.1038/s41556-021-00818-3 111. Zhou Z., Chen F., Zhong S., Zhou Y., Zhang R., Kang K., Zhang X., Xu Y., Zhao M., Zhao C.. Molecular identification of protein kinase C beta in Alzheimer’s disease. Aging (2020) 12 21798-21808. DOI: 10.18632/aging.103994 112. Lo C., Cooper-Knock J., Garrard K., Martindale J., Williams T., Shaw P.. Concurrent amyotrophic lateral sclerosis and cystic fibrosis supports common pathways of pathogenesis. Amyotroph. Lateral Scler. Front. Degener. (2013) 14 473-475. DOI: 10.3109/21678421.2012.746989 113. De Nicola A.F., Meyer M., Garay L., Kruse M.S., Schumacher M., Guennoun R., Gonzalez Deniselle M.C.. Progesterone and Allopregnanolone Neuroprotective Effects in the Wobbler Mouse Model of Amyotrophic Lateral Sclerosis. Cell. Mol. Neurobiol. (2021) 42 23-40. DOI: 10.1007/s10571-021-01118-y 114. Drannik A., Martin J., Peterson R., Ma X., Jiang F., Turnbull J.. Cerebrospinal fluid from patients with amyotrophic lateral sclerosis inhibits sonic hedgehog function. PLoS ONE (2017) 12. DOI: 10.1371/journal.pone.0171668 115. Peterson R., Turnbull J.. Sonic hedgehog is cytoprotective against oxidative challenge in a cellular model of amyotrophic lateral sclerosis. J. Mol. Neurosci. (2012) 47 31-41. DOI: 10.1007/s12031-011-9660-x 116. Liu Y.J., Ju T.C., Chen H.M., Jang Y.S., Lee L.M., Lai H.L., Tai H.C., Fang J.M., Lin Y.L., Tu P.H.. Activation of AMP-activated protein kinase. Hum. Mol. Genet. (2015) 24 787-801. DOI: 10.1093/hmg/ddu497 117. Hu J.H., Zhang H., Wagey R., Krieger C., Pelech S.L.. Protein kinase and protein phosphatase expression in amyotrophic lateral sclerosis spinal cord. J. Neurochem. (2003) 85 432-442. DOI: 10.1046/j.1471-4159.2003.01670.x 118. Vasilopoulou C., Duddy W., Wingfield B., Morris A.P.. snpQT: Flexible, reproducible, and comprehensive quality control and imputation of genomic data. F1000Research (2021) 10 567. DOI: 10.12688/f1000research.53821.2 119. Purcell S., Neale B., Todd-Brown K., Thomas L., Ferreira M.A., Bender D., Maller J., Sklar P., De Bakker P.I., Daly M.J.. PLINK: A tool set for whole-genome association and population-based linkage analyses. Am. J. Hum. Genet. (2007) 81 559-575. DOI: 10.1086/519795 120. Zuvich R.L., Armstrong L.L., Bielinski S.J., Bradford Y., Carlson C.S., Crawford D.C., Crenshaw A.T., de Andrade M., Doheny K.F., Haines J.L.. Pitfalls of Merging GWAS Data: Lessons Learned in the eMERGE Network and Quality Control Procedures to Maintain High Data Quality. Genet. Epidemiol. (2011) 35 887. DOI: 10.1002/gepi.20639 121. McCarthy S., Das S., Kretzschmar W., Delaneau O., Wood A.R., Teumer A., Kang H.M., Fuchsberger C., Danecek P., Sharp K.. A reference panel of 64,976 haplotypes for genotype imputation. Nat. Genet. (2016) 48 1279-1283. DOI: 10.1038/ng.3643 122. Loh P.R., Danecek P., Palamara P.F., Fuchsberger C., Reshef Y.A., Finucane H.K., Schoenherr S., Forer L., McCarthy S., Abecasis G.R.. Reference-based phasing using the Haplotype Reference Consortium panel. Nat. Genet. (2016) 48 1443-1448. DOI: 10.1038/ng.3679 123. Durbin R.. Efficient haplotype matching and storage using the positional Burrows-Wheeler transform (PBWT). Bioinformatics (2014) 30 1266-1272. DOI: 10.1093/bioinformatics/btu014 124. Tsunoda T., Lathrop G.M., Sekine A., Yamada R., Takahashi A., Ohnishi Y., Tanaka T., Nakamura Y.. Variation of gene-based SNPs and linkage disequilibrium patterns in the human genome. Hum. Mol. Genet. (2004) 13 1623-1632. DOI: 10.1093/hmg/ddh177 125. Shannon P., Markiel A., Ozier O., Baliga N.S., Wang J.T., Ramage D., Amin N., Schwikowski B., Ideker T.. Cytoscape: A Software Environment for Integrated Models of Biomolecular Interaction Networks. Genome Res. (2003) 13 2498-2504. DOI: 10.1101/gr.1239303 126. Bateman A., Martin M.J., Orchard S., Magrane M., Agivetova R., Ahmad S., Alpi E., Bowler-Barnett E.H., Britto R., Bursteinas B.. UniProt: The universal protein knowledgebase in 2021. Nucleic Acids Res. (2021) 49 D480-D489. DOI: 10.1093/NAR/GKAA1100 127. Safran M., Rosen N., Twik M., BarShir R., Stein T.I., Dahary D., Fishilevich S., Lancet D.. The GeneCards Suite. Practical Guide to Life Science Databases (2021) 27-56. DOI: 10.1007/978-981-16-5812-9_2 128. Carbon S., Ireland A., Mungall C.J., Shu S., Marshall B., Lewis S., Lomax J., Mungall C., Hitz B., Balakrishnan R.. AmiGO: Online access to ontology and annotation data. Bioinformatics (2009) 25 288-289. DOI: 10.1093/bioinformatics/btn615 129. Carbon S., Douglass E., Good B.M., Unni D.R., Harris N.L., Mungall C.J., Basu S., Chisholm R.L., Dodson R.J., Hartline E.. The Gene Ontology resource: Enriching a GOld mine. Nucleic Acids Res. (2021) 49 D325-D334. DOI: 10.1093/NAR/GKAA1113 130. Ashburner M., Ball C.A., Blake J.A., Botstein D., Butler H., Cherry J.M., Davis A.P., Dolinski K., Dwight S.S., Eppig J.T.. Gene ontology: Tool for the unification of biology. Nat. Genet. (2000) 25 25-29. DOI: 10.1038/75556 131. Kanehisa M., Furumichi M., Tanabe M., Sato Y., Morishima K.. KEGG: New perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. (2017) 45 D353-D361. DOI: 10.1093/nar/gkw1092 132. Jacobs B.M., Taylor T., Awad A., Baker D., Giovanonni G., Noyce A.J., Dobson R.. Summary-data-based Mendelian randomization prioritizes potential druggable targets for multiple sclerosis. Brain Commun. (2020) 2 fcaa119. DOI: 10.1093/braincomms/fcaa119
--- title: Evidence and Metabolic Implications for a New Non-Canonical Role of Cu-Zn Superoxide Dismutase authors: - Ziqiao Sun - Xin-Gen Lei journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966940 doi: 10.3390/ijms24043230 license: CC BY 4.0 --- # Evidence and Metabolic Implications for a New Non-Canonical Role of Cu-Zn Superoxide Dismutase ## Abstract Copper–zinc superoxide dismutase 1 (SOD1) has long been recognized as a major redox enzyme in scavenging superoxide radicals. However, there is little information on its non-canonical role and metabolic implications. Using a protein complementation assay (PCA) and pull-down assay, we revealed novel protein–protein interactions (PPIs) between SOD1 and tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation protein zeta (YWHAZ) or epsilon (YWHAE) in this research. Through site-directed mutagenesis of SOD1, we studied the binding conditions of the two PPIs. Forming the SOD1 and YWHAE or YWHAZ protein complex enhanced enzyme activity of purified SOD1 in vitro by $40\%$ ($p \leq 0.05$) and protein stability of over-expressed intracellular YWHAE ($18\%$, $p \leq 0.01$) and YWHAZ ($14\%$, $p \leq 0.05$). Functionally, these PPIs were associated with lipolysis, cell growth, and cell survival in HEK293T or HepG2 cells. In conclusion, our findings reveal two new PPIs between SOD1 and YWHAE or YWHAZ and their structural dependences, responses to redox status, mutual impacts on the enzyme function and protein degradation, and metabolic implications. Overall, our finding revealed a new unorthodox role of SOD1 and will provide novel perspectives and insights for diagnosing and treating diseases related to the protein. ## 1. Introduction Copper–zinc superoxide dismutase 1 (SOD1 or Cu-Zn-SOD) is present mainly in the cytoplasm, with a small portion in the nucleus, of nearly all eukaryotes [1]. As a crucial component of cellular defense against oxidative stress, SOD1 breaks down O2•− through a redox cycling of the copper ion in the Cu-Zn active site to dismutate O2•− to H2O2 and O2 [2]. In addition to this enzymatic role in free radical scavenging, SOD1 has recently been unraveled with unconventional roles. Lu et al. discovered SOD1 as an RNA binding protein or regulator of RNA stability, in which mutant SOD1 (G93A) specifically altered the stability of ribonucleoprotein complex associated with 3′UTR mRNA of vascular endothelial growth factor but not wild-type SOD1 [3,4]. Tsang’s group revealed another novel, unorthodox role of SOD1 in nuclear translocation. In the nucleus, phosphorylated SOD1 bound genomic DNA promoter and mediated the expression of genes in oxidative resistance and DNA repair, which was regulated by the ATM/Mec1 oxidative sensor in response to increased H2O2 [5]. Liu’s group recently discovered that SOD1 could bind double-stranded DNA (dsDNA) by small-angle X-ray scattering, and that the SOD1-dsDNA complex responded to different oxidants by changing its shape, providing a new potential mechanism for SOD1 to protect cells from oxidative stress [6]. SOD1 could also function as a protein interactor with copper chaperone for SOD1 (CCS) to regulate itself maturation, dimerization, and activity [7,8]. Loss of SOD1 protein not only led to increased reactive oxygen species (ROS) levels and oxidative damage [9,10,11,12] but also induced resistance to acetaminophen toxicity, defects in femoral mechanical performance, injury of islets beta-cell function, and exacerbated lipogenesis with elevated non-esterified fatty acids (NEFA), total cholesterol (TC), and triglycerides (TG) [13,14,15,16,17]. Because some of these phenotypes were not readily explained by the SOD1 enzyme activity loss per se, there might be unrevealed, unorthodox roles of SOD1 involved in these abnormalities. Tyrosine 3-monooxygenase/tryptophan 5-monooxygenase activation proteins zeta (YWHAZ) and epsilon (YWHAE) belong to the 14-3-3 protein family that has other five isoforms (YWHAB, YWHAG, YWHAH, YWHAQ, and YWHAS). These proteins are ubiquitously produced in human tissues, with the highest abundance in the brain [18]. They are widely regarded as molecular adapters modulating over 200 diverse signaling proteins, especially in brain neuronal development and neurological disorders [19,20]. The multifunctions of 14-3-3 family proteins make them hard to study, due to a likely complicated network complex instead of a simple role in a single pathway. YWHAZ protein is a crucial regulator of cell growth and apoptosis pathways [21,22] and is closely related to several types of cancer [23,24]. Additionally, knockout of YWHAZ (14-3-3ζ) led to conspicuously lean in mouse pups and diminished visceral adipose accumulation, while overexpression of YWHAZ induced obesity-like phenotypes [25]. Loss of YWHAZ also increased fasting insulin levels with unaltered beta-cell glucose sensitivity [26]. However, specific functions of YWHAE (14-3-3ε) in metabolism remain unclear, although it was implicated with a latent role in insulin signaling as it interacted with insulin-like growth factor I receptor and insulin receptor substrate I [27]. Putative protein–protein interactions (PPIs) between SOD1 and YWHAE or YWHAZ, along with other two-family members (YWHAQ and YWHAG), were proposed based on a pool of protein pairs identified by affinity purification mass spectrometry (AP/MS) [28,29]. However, the AP/MS analysis results were confounded with a high background of non-specific protein binders [30]. To overcome this constraint, we applied the specific protein complementation assay (PCA) followed by GST pull-down assay to provide the first direct evidence for these PPIs. Moreover, we prepared several SOD1 single-site mutants (A4V, H46R, G85R, and G93A) associated with amyotrophic lateral sclerosis (ALS) disease [31,32,33,34] and another mutation D124N in the SOD1 zinc active site [35] to compare disruptions of the PPIs between SOD1 and YWHAE or YWHAZ at different levels of remaining SOD1 enzymatic activities. Subsequently, we overexpressed and purified the three proteins and determined the binding conditions and effects of their PPIs on stability of the three proteins and the activity of SOD1. Thereafter, we explored if these PPIs modulated lipid metabolism, cell growth, and cell survival of HEK293T cells. Altogether, we discovered two new PPIs between SOD1 and YWHAZ or YWHAE and revealed a novel, unorthodox role of SOD1 beyond its well-known redox function. ## 2.1. Evidence for Protein–Protein Interactions between SOD1 and YWHAE or YWHAZ We performed PCA to determine the postulated PPIs between human SOD1 with YWHAE or YWHAZ in HEK293T cells. The principle of PCA (Figure 1A) is that bait and prey proteins are fused to two complementary fragments (F1 and F2) of a yellow fluorescent protein (YFP) [36]. If the bait and prey proteins interact, the two fragments will covalently link to generate a yellow fluorescent signal. We prepared the plasmid constructs in four orientations to measure each protein–protein pair as F1 or F2 tagged in the N (F1N or F2N) or C (F1C or F2C) terminal of the two proteins (Figure 1B). The fluorescence signal intensity was normalized to the cell group transfecting the vector control in the same orientation. The YWHAZ and YWHAE had pronounced fluorescence signals compared with a negative control protein ATPG (fold change close to 1 and almost no fluorescence signal under the microscope) that was not supposed to interact with SOD1 from our preliminary screening (Figure 1C). Relatively, the strongest interaction of SOD1 with YWHAE occurred in the N-N orientation (17.4-fold, $p \leq 0.01$) and with YWHAZ in the C-N orientation (3.72-fold, $p \leq 0.001$) compared with the same orientation of ATPG negative control. Subcellular location and relative intensity of the fluorescent cell imaging suggested cytosol as the primary site for these PPIs, along with moderate occurrence in the nucleus. We expressed mouse proteins in human HEK293T for PCA to test whether the PPIs were independent of species. Whereas YWHAE and YWHAZ sequences are highly homologous between humans and mice, their SOD1 sequences have only $84\%$ identity (Figure S1B–D). The PCA results indicated that the PPIs were shown in the C-N orientation of mSOD1-mYWHAE (2.06-fold, $p \leq 0.01$) and in the N-C orientation of mSOD1-mYWHAZ (39.1-fold, $p \leq 0.001$) compared with the same orientation of ATPG negative control (Figure S1E,F). These results proved the PPIs existed between SOD1 and YWHAE or YWHAZ and indicated that the PPIs were relatively independent and could be cross-species. ## 2.2. Protein Complex between the Purified SOD1 and YWHAE or YWHAZ To determine if SOD1-YWHAZ and SOD1-YWHAE could form protein complexes, we used Pichia Pastrois X33 to express the His-tag conjugated SOD1-His and YWHAZ-His and E. coli BL21 (DE3) to express the GST-conjugated GST-YWHAE and GST-YWHAZ. As shown in Figure 2A, GST-YWHAE and GST-YWHAZ could bind and pulled-down SOD1-His, despite a relatively low amount of the complex. Because GST-YWHAZ and GST-YWHAE proteins were expressed intracellularly, certain non-specific proteins remained after the purification might interfere with the interaction in the molecular weight region of SOD1. To confirm the target protein bands, we performed immunoblotting against the SOD1 and YWHAZ antibodies (Figure 2B). While setting the SOD1 protein as constant, different ratios of input proteins affected the complex formation, where increasing YWHAZ or YWHAE protein facilitated the complex formation (Figure 2C). As the large molecular size of the GST tag might impede the PPI, we tested YWHAZ tagged to only six histidines (YWHAZ-His, expressed and purified from Pichia Pastrois X33). Due to both proteins having his tag, the purified SOD1 was conjugated with a SOD1 antibody and immobilized on agarose beads G for pull down assay. The YWHAZ-His protein was detected from the immunoprecipitation eluates on lane 3 (Figure 2D), indicating the PPI between YWHAZ and SOD1 and in vitro protein complex formation. Summarizing all evidence, we could confirm the formation of protein complexes between SOD1 and YWHAE or YWHAZ. ## 2.3. Impacts of SOD1 Mutations and Oxidative Stress on Protein–Protein Interactions between SOD1 and YWHAE or YWHAZ To determine impacts of SOD1 mutations on the focused PPIs, we performed PCA using two-orientated (F1N-F2N, F1C-F2N) constructs of YWHAE or YWHAZ with the wild-type (WT) or mutants of human SOD1. The upper part of Table 1 shows the absolute fold changes in fluorescence intensity for the interactions of SOD1 mutants with YWHAE or YWHAZ. The lower sectiont of Table 1 converts the fluorescent intensity fold changes to the percentage of remaining PPI relative to that of the WT-SOD1 by method Formula [1]. The transfection efficiency of various plasmids was assumed at similar levels (Figure S1). Compared with the WT, four SOD1 mutants (H46R, G85R, G93A, D124N) had decreased affinity to interact with YWHAE and YWHAZ proteins by over $50\%$ in either orientation. However, overexpressing most of these mutants of SOD1 did not increase total intracellular SOD1 activity, but some mutants resulted in a moderate decrease (up to $30\%$, $p \leq 0.05$) in the activity (Figure S3). Other studies showed similar decreases in total SOD1 activities (0–$45\%$ to wild-type SOD1) among these purified SOD1 mutant proteins [31,32,37,38]. However, the impaired or disrupted PPIs were largely independent of total intracellular SOD1 activity. Notably, the strongest breakdown in the PPIs mediated by the mutated SOD1 was different between the interactions with YWHAE and YWHAZ protein. The SOD1-D124N mutant decreased the interaction with YWHAE by $96.0\%$ in F1C-F2N orientation ($p \leq 0.01$), whereas the SOD1-G85R mutant reduced the interaction with YWHAZ by $99.7\%$ in F1N-F2N orientation ($p \leq 0.05$). Plausibly, there were different binding conditions of SOD1 with YWHAE and YWHAZ. To determine impacts of redox status on the PPI of SOD1 with YWHAE (as a representative), we treated the transfected cells with various oxidants and antioxidants. The treatments of diquat (a superoxide generator, 50 μM) and TBHP (a hydroperoxide generator, 50 μM) attenuated the interaction at 1 h and 2 h (1h: $16\%$ and $28\%$ and 2h: $11\%$ and $20\%$, respectively, $p \leq 0.05$, Figure S2A). However, those effects became insignificant after a longer incubation, which made us speculate that the effects of diquat and TBHP could be just acute. Among three ROS inhibitors (50 μM), NAC (a GSH precursor in glutathione elevation biosynthesis) slightly rescued the SOD1-H46R (as a representative) and YWHAE interaction (Figure S2B), while CuDIPs (a SOD mimic) almost blocked the interaction formation, and ebselen (a GPX mimic) made no significant effect. The same effect trend was also observed in the SOD1-D124N and YWHAE interaction (Figure S2C). In the same way that oxidants impaired PPIs, the disrupted PPIs between SOD1 mutants (H46R and D124N) and YWHAE could be restored by antioxidant compensation. ## 2.4. Impacts of Protein–Protein Interactions between SOD1 and YWHAE or YWHAZ on SOD1 Activity Incubating purified SOD1 protein with purified YWHAE protein or YWHAZ protein in the SOD1 activity assay buffer for 4 h, at 4 °C, increased the SOD1 activity by $40\%$ compared with SOD1 plus buffer group ($p \leq 0.05$, Figure 3A). The SOD1 activity kept increasing over the incubation time. A 15 h incubation with the YWHAZ protein at a ratio of 1:3 (1 SOD1 to 3 YWHAZ protein amount) caused the highest elevation of SOD1 activity (~2-fold increase, $p \leq 0.01$) (Figure 3B). To verify if such enhancement could be replicated in cells, we transfected the YWHAZ vector into HEK293T and found a $20\%$ increase ($p \leq 0.05$) in SOD1 activity (Figure 3C). We also generated YWHAZ-knockout (YWHAZ-KO) and SOD1-knockdown-YWHAZ-knockout (SOD1-KD&YWHAZ-KO) cells by CRISPR-Cas9 genome-editing technology (Figure S4F). The YWHAZ-KO cells had a $15.2\%$ decrease ($p \leq 0.05$) in SOD1 activity (Figure 3D) but unchanged SOD2 activity or SOD1 protein level (Figure S4G). Inactivating the YWHAZ gene on top of SOD1 knockdown (SOD1-KD&YWHAZ-KO cells) showed an additional decrease in SOD1 activity over the SOD1-KD cells without affecting SOD2 activity (Figure 3E). Both in vitro and in vivo experiments illustrated that PPIs between SOD1 and YWHAE or YWHAZ could specifically promote the activity of SOD1. ## 2.5. Impacts of Protein–Protein Interactions between SOD1 and YWHAE or YWHAZ on Their Protein Stability To determine effects of PPIs between SOD1 and YWHAE or YWHAZ on their relative turnover or resistance to protein degradation, we performed a cycloheximide-mediated protein degradation assay and a series of immunoblots in HEK293T cells transfected with plasmids expressing SOD1, YWHAE, and YWHAZ codon sequence. The SOD1 stability in HEK293T cells was increased by co-expressing the WT-SOD1 with YWHAE ($17.7\%$, $p \leq 0.01$) or YWHAZ ($13.5\%$, $p \leq 0.05$) (Figure 4A,B). Meanwhile, the protein stability of YWHAE (Figure 4A,C) and YWHAZ (Figure 4A,D) was increased by $40.4\%$ ($p \leq 0.01$) and $14\%$ ($p \leq 0.01$), respectively, via co-expressing WT-SOD1 with respective gene. The SOD1 mutants (G85R and D124N) had lower stability than WT-SOD1 (Figure 4F). However, the protein stability of SOD1 mutants (G85R and D124N) was unchanged by the co-expression with YWHAE (Figure 4E,F). Likewise, co-expressing SOD1 mutants (G85R and D124N) with YWHAE, which showed disrupted PPIs, did not improve their protein stability (Figure 4E,G). Strikingly, overproducing all three proteins (SOD1, YWHAE, and YWHAZ) simultaneously in the cells removed the protein stability benefits resulted from the co-expression of SOD1 with YWHAE or YWHAZ (Figure 4B–D). Overall, we identified that both proteins increased stability or were more resistant to their intracellular protein degradation when having a PPI with each other. ## 2.6. Impacts of Protein–Protein Interactions between SOD1 and YWHAE or YWHAZ on Lipid Metabolism and Relative Gene Expression Both SOD1 and YWHAZ have shown important roles in lipid metabolism [16,17,25]. To investigate impacts of the focused PPIs on lipid metabolism, we used genome editing to generate SOD1-KD (SOD1 knockdown) and SOD1-KD&YWHAZ-KO HEK293T cells (SOD1 knockdown and YWHAZ knockout) (Figure S4). Knocking down SOD1 increased cellular lipid droplet accumulation, while knockout of YWHAZ protein restored the status (Figure 5A,B). Treating the SOD1KO cells with the SOD1 mimic (CuDIPs) did not prevent the abnormal lipid accumulation, but increased (by 1μM CuDIPs) the lipid accumulation ($30\%$) when additional knocked out YWHAZ gene in SOD1-KD cells (SOD1-KD&YWHAZ-KO, Figure 5C). In searching for a plausible mechanism, we found that SREBP1 (sterol regulatory element-binding protein 1) and ACACA (acetyl-CoA carboxylase alpha) were upregulated, while LPL (lipoprotein lipase), FASN (fatty acid synthase), and HMGCR (HMG-CoA reductase) were downregulated in SOD1-KD cells. These changes could enhance lipogenesis and attenuate lipolysis (Figure 5D). Other related genes (HMGCS1, HMGCS2, LIPE, and PPARG) showed similar downregulations, but the changes were not statistically significant (Figure S5A). Interestingly, expression of these genes was restored to the control levels (WT) by knocking out YWHAZ in SOD1KO cells (Figure 5D). The responses of HMGCS1, PNPLA2, ACACA, FASN, LPL, and LIPE mRNA levels to CuDIPs varied with different genotypes of cells (Figure S5D). We also measured lipid and fatty acids profiles in HepG2 cells overexpressing WT-SOD1 or SOD1 mutants (H46R and G93A as representatives). We first demonstrated that overexpressing WT-SOD1, SOD1-H46R, or SOD1-G93A could dominate the PPIs between endogenous YWHAE or YWHAZ with WT-SOD1 since the co-immunoprecipitation bands under overexpressing (Lanes 6–8, Figure 5E) were stronger in intensity than vector control F1N (Lane 5, Figure 5E). Compared with the vehicle control, transfecting SOD1 mutants (H46R and G93A) showed no effect on TC content in HepG2 cells. The SOD1-G93R significantly elevated the TG concentration, while the WT-SOD1 and the SOD1-H46R transfection moderately decreased TG in HepG2 cells. The SOD1-H46R and SOD1-G93A transfection in HepG2 cells decreased palmitic acid (C16:0), stearic acid (C18:0), and oleic acid (C18:1n-9) levels. In particular, the C18:1n-9 was eliminated by the SOD1-H46R transfection (Figure 5G). HepG2 cells tended to accumulate lipid droplets after transfecting mutant SOD1-H46R and SOD1-G93A (Figure 5F). The mRNA levels of SREBP1, SREBP2, HMGCS2 (3-hydroxy-3-methylglutaryl-CoA synthase 2), and LPL were decreased by transfecting SOD1-G93A, whereas ACACA was upregulated by overexpressing SOD1-G93A (Figure 5H). Changes in mRNA levels of HMRCS1, HMRCR, and FASN were not statistically significant (Figure S5B) by overexpressing WT-SOD1 or mutant SOD1. Interestingly, transfecting YWHAE increased lipid droplets accumulation, and the increase was blocked by cotransfecting SOD1 and YWHAE (Figure 5I and Figure S5C). In summary, knocking down SOD1 in HEK293T increased lipid accumulation, along with upregulating several lipogenesis-related genes, which is similar to hepatic-steatosis-like phenotypes we found in Sod1−/− mice [16]. When overexpressing the SOD1-G93A mutant, it turned to bind more YWHAE or YWHAZ, and the PPIs altered the fatty acid profiles and affected the gene expression to elevate lipolysis. We also discovered the potentially correlated roles between SOD1 and YWHAZ or YWHAE in mediating lipid metabolism since the increasing trend was rescued when additional knockout YWHAZ or co-overexpressing YWHAE. ## 2.7. Impacts of Protein–Protein Interactions between SOD1 and YWHAE or YWHAZ on Cell Growth and Survival Because the 14-3-3 protein family is well known for affecting cell growth and death, including YWHAE and YWHAZ [21,22,39], potential roles of PPIs between SOD1 and YWHAE or YWHAZ could play in these pathways. We used CRISPR-editing to alter expressions of endogenous SOD1 and YWHAZ in HEK293T cells to investigate impacts of their PPIs on cell growth (proliferation) and survival. Knockdown of SOD1 (SOD1-KD) or producing the SOD1 mutant (SOD1-A4V, alanine to valine at the fourth position) decreased cell growth (Figure 6A), and cell survival was also reduced after removal of FBS from the culture media at indicated time points (Figure 6B). Overexpressing YWHAE and SOD1 alone or together in the WT or SOD1-KD cells rescued cell growth (Figure 6C), but not cell survival (Figure 6D). Overexpressing YWHAZ and SOD1 also gave a moderate restoration of cell growth (Figure 6E) in both WT and SOD1-KD cells, and only overexpressing YWHAZ improved cell survival in SOD1-KD cells (Figure 6F). Cell growth was enhanced by overexpressing SOD1 in the SOD1-KD&YWHAZ-KO cells (Figure 6G), but cell survival was decreased (Figure 6H). Similar to SOD1-KD cells, SOD1-A4V cells also had declined SOD1 activity and protein levels (Figure S4C,D). Overexpressing YWHAE or YWHAZ and SOD1 alone or together in SOD1-A4V cells rescued cell growth (Figure S6A, C) but not cell survival (Figure S6B,D). Compared with WT controls, mRNA levels of CDKN1B (cyclin-dependent kinase inhibitor 1B) and CCND1 (cyclin D1) were downregulated by knockdown of SOD1 (SOD1-KD vs. WT, Figure 6I) but rescued by an additional knockout of YWHAZ (SOD1-KD vs. SOD1-KD&YWHAZ-KO, Figure 6I). Overexpressing WT-SOD1 instead of SOD1 mutants enhanced CDKN1B and CCND1 mRNA levels (Figure S6E) that were also increased by the graded CuDIPs treatment (Figure S6F,G). However, the latter effect disappeared by knocking down SOD1 and knocking out YWHAZ. This evidence suggests that the impaired SOD1 (KD or A4V mutant) slackened cell proliferation and accelerated cell death. Moreover, YWHAE and YWHAZ might interact with SOD1 to affect cell growth and survival, and the relative ratio of protein amounts of YWHAE and YWHAZ to SOD1 could be a critical determinant. ## 2.8. Expression Correlation between SOD1 and YWHAE or YWHAZ Protein If two proteins interact and have a co-function or regulatory relationship, they may be correlated in protein expression profiles [40]. To find co-function or regulatory relationships for the putative PPIs between SOD1 and YWHAE or YWHAZ, we used a breast cancer quantitative proteome database [41] to analyze latent correlations of their protein expression profiles [41]. There was a positive correlation between SOD1 and YWHAE ($r = 0.71$, $p \leq 0.01$, Figure 7A), but a negative correlation between SOD1 and YWHAZ (r = −0.40, $p \leq 0.02$, Figure 7B). Consistent with the negative correlation between SOD1 and YWHAZ, knockout of Sod1 in mice elevated ($p \leq 0.05$) YWHAZ protein levels in the liver, brown adipose tissue (BAT) and muscle (Figure S7A,B), but did not significantly decrease (positive correlation) in the protein expression level of YWHAE (Figure S7A,C). The enhanced protein levels of both YWHAE and YWHAZ in the brown adipose tissue, muscle, and/or liver, but not other tissues, of Sod1−/− mice reinforce that the opposite correlations between SOD1 and YWHAE (+) vs. YWHAZ (−) in the progression of the human breast cancer represented a specific interdependence or interaction of these proteins. ## 3. Discussion In this study, we have provided direct evidence for the PPIs of SOD1 with YWHAE and YWHAZ under in vivo and in vitro conditions. Although these PPIs were proposed based on prior analysis of protein pairs by affinity purification mass spectrometry (AP/MS) [26,27], our research represents a significant progress. Comparatively, our methods of PCA were more specific and reliable than the AP/MS analysis that was confounded with a high background of non-specific protein binders [28]. Convincingly, we demonstrated the structural and redox environment dependences of the PPIs. Disruptions of the PPIs by mutagenesis of the binding SOD1 (H46R, G85R, G93A, D124N) imply the importance of an intact SOD1 structure or involvements of these mutated sites in the PPIs. Additionally, the PPIs were readily impaired by external oxidants (diquat and TBHP). However, three antioxidants (NAC, CuDIPs, and ebselen) exhibited different potencies in rescuing disrupted PPIs between SOD1-H46R or SOD1-D124N with YWHAE, despite their common roles in suppressing oxidative stress [42,43]. Apparently, the PPIs between SOD1 and YWHAE or YWHAZ could be modulated by the intrinsic activity of SOD1 itself and exogenous antioxidants and oxidants in their own specific ways. It is fascinating for us to unearth an enhanced SOD1 activity by the formation of protein complexes in the PPIs. The enhancement was proportional to the incubation time and YWHAE or YWHAZ protein inputs. Overexpressing or knocking out YWHAZ in HEK293T cells led to increased or decreased intracellular activity of SOD1, replicating the same in vitro impact of the PPIs on the SOD1 enzymatic function in vivo (cell). Based on 3D structure analysis of the PPI by AlphaFold2 [44], we have found that the constitutions of SOD1-YWHAE and SOD1-YWHAZ dimer complexes are highly similar (Figure S8). The high accuracy and confidence of this new advanced computational tool have been tested for predicting protein–protein interactions [45]. In dimer complexes, the YWHAE or YWHAZ dimer holds the SOD1 dimer with four alpha-helixes, especially enclosing the active site and copper–zinc metal binding loop (residues 49–84) of SOD1 within the protein–protein interface [37]. The SOD1 mutated sites (H46R, G85R, and G93A) were close to the metal binding loop of the enzyme. The protein complex model suggested binding domains between two proteins and provided additional structural-based evidence to explain why these mutations disrupted the PPIs between SOD1 and YWHAE or YWHAZ. The predicted configuration might also explain the PPIs potentially promoted the SOD1 enzymatic activity by protecting the copper–zinc active enzymatic center. Meanwhile, both YWHAE and YWHAZ proteins are widely recognized as adaptor proteins, and no existing methods could directly determine their activity. Our data confer a new, specific function for YWHAE and YWHAZ proteins. However, more actual experiments, including site-directed mutagenesis, co-crystallization, and enzyme kinetics, will be required to explain the mechanism for the enhanced activity of SOD1 by the putative PPIs with YWHAE or YWHAZ. It is equally exciting for us to demonstrate enhanced stabilities of all three proteins (YWHAE, YWHAZ, and SOD1) by the formation of protein complexes in the PPIs. These enhancements could be attributed to a modified protein-exposed surface when YWHAE or YWHAZ bind with their interactors [46]. Nonetheless, the SOD1 mutants (G85R and D124N) failed to improve the protein stability of YWHAE or YWHAZ, probably due to impaired PPIs between them. Likewise, both SOD1 mutants (G85R and D124N) had lower protein stability than wild-type SOD1 under cycloheximide treatment, consistent with others reported on decreased half-lives (protein stability) on SOD1 mutants (A4V and G85R) [32]. The 14-3-3 family proteins and SOD1 are degraded through the ubiquitin–proteasome system [46,47]. The PPIs between SOD1 and YWHAE or YWHAZ could shield the lysine residues from ubiquitination and ubiquitin-mediated degradation. While the SOD1 mutants disrupted these PPIs, an accelerated breakdown of associated proteins might occur. In the future, the impacts of these PPIs on stability of the involved proteins to resist urea or guanidine treatments should be determined for a comprehensive understanding of protein stability. The most relevant metabolic implication of our findings is the impact of the PPIs between SOD1 and YWHAE or YWHAZ on lipid metabolism, proliferation, and survival of HEK293T and/or HepG2 cells. SOD1 knockdown (SOD1-KD) induced dysregulated lipid accumulation by altering mRNA levels of SREBP1, LPL, FASN, ACACA, and HMGCR. Because the dysregulation could not be rescued by CuDIPs (SOD1 mimic) treatment, it was likely more related to PPIs of SOD1 rather than its enzymatic function. Remarkably, when an additional knockout of YWHAZ in SOD1-KD cells, the altered mRNA levels of SREBP1, FASN, and HMGCR were restored. It has been reported that SOD1 was involved in DNA repair, bound to RNA in the nucleus, and functioned as a transcription factor [5]. Meanwhile, 14-3-3 family proteins interacted with transcription factors and could mediate their binding to gene promoters [48,49]. Notably, the PPIs between SOD1 and YWHAE or YWHAZ were localized in the nucleus under PCA fluorescent images. Likely, YWHAE and YWHAZ could chaperone SOD1, promote its nuclear localization, and subsequently mediate its regulation of gene expression related to lipid metabolism. The SOD1 mutants (H46R, G85R, and G93A) tested in our study are associated with ALS [31,32,33,34]. It is widely believed that the etiology of these mutations is due to gaining toxic properties by misfolding and aggregation of the mutant SOD1 proteins rather than loss of SOD1 enzymatic activity [34,50]. In fact, overexpressing SOD1-G93A in HepG2 cells in the present study led to decreased stearic acid and oleic acid concentrations and SREBP1, HMGCS2, and LPL mRNA levels, suggesting elevated lipolysis. The loss of Sod1 was associated with abnormal lipid metabolism [16,17], and SOD1 human polymorphisms were also linked to incidences of diabetes and obesity [51,52,53,54]. Besides, Sod1-G93A mutant mice exhibited hypolipidemia with lowered LDL/HDL ratio and promoted lipolysis with pathological acidosis [55,56], while the Sod1-G93A ALS mutant mice exhibited reduced insulin-stimulated glucose uptake in skeletal muscle [57]. Similarly, knockout of YWHAZ also led to increased lipolysis with decreased visceral adipose accumulation [25]. Our discoveries of PPIs’ (SOD1 and YWHAE or YWHAZ) impact on lipid metabolism could provide new clues to those unsettled issues or unexplained phenotypes [25,26,27,55,56,57]. Because 14-3-3 proteins have recently been linked to ALS [58,59,60], our finding on altered cellular lipid metabolism by disrupted PPIs between SOD1 and YWHAE or YWHAZ may reveal a latent factor for the onset and development of ALS. Similarly, overexpression of YWHAZ is a prognostic biomarker for breast tumors and ovarian cancer, and a reduction in YWHAE levels has been seen in gastric carcinogenesis [61,62,63]. In addition, 14-3-3 family proteins were reported to co-localize with SOD1-A4V aggregates [64], but the impact of the co-localization was unillustrated. Contributions and mechanisms of their PPIs with SOD1 in cancer diseases also remain to be uncovered. Likewise, disrupted or modified PPIs in the SOD1-KD or SOD1-A4V cells might lead to retarded cell proliferation and exacerbated cell death or impaired cell survival. Furthermore, maintaining an appropriate ratio of intracellular SOD1 vs. YWHAE or YWHAZ seemed to be vital for controlling cell growth and survival. Lowering SOD1 activity and protein level (as SOD1-KD and SOD1-A4V) decreased cell growth and survival. Elevating YWHAZ: SOD1 ratio (as SOD1-KD or SOD1-A4V cells transfecting YWHAZ protein) improved cell survival and cell growth, while lowering YWHAZ: SOD1 ratio (as SOD1-KD&YWHAZ-KO overexpressing the SOD1 protein) only enhanced cell proliferation but exacerbated cell death. Furthermore, overproducing and decreasing SOD1 in HEK293T cells fostered and impeded cell proliferation by upregulating or downregulating the CCND1 gene, respectively. A recent study also discovered elevated SOD1 expression levels in lung cancer cells, facilitating cell proliferation, invasion, and migration [65]. Moreover, YWHAZ is closely associated with cell growth, apoptosis and survival in cancer cells [21,22,66]. A decreased YWHAZ expression inhibited cell proliferation and migration in gastric cancer cells, which frequently detected YWHAZ protein was upregulated [23]. In contrast, a decreased YWHAE expression aggravated tumor growth by potentially inducing cell proliferation, invasion, and migration [67]. The cell growth and survival differences between the SOD1-KD cells overexpressing YWHAE vs. YWHAZ might help explain differential expressions of these two 14-3-3 proteins in tumors, suggesting that PPIs with SOD1 could be a potential modulator of tumorigenesis. Meanwhile, the opposite correlations between YWHAE and YWHAZ with SOD1 from the breast cancer proteome data imply that these PPIs might drive different metabolic impacts in breast cancer development. The novel perspective of our newly elucidated, unorthodox roles of SOD1 may offer alternative mechanisms to explain paradoxical phenotypes of metabolic disorders. The PPIs could help explain partially a puzzling observation that knockouts of Sod1 and Gpx1, two genes encoding redox enzymes with similar functions in scavenging free radicals, led to different severities in many phenotypes, including hepatic steatosis [17]. Likewise, the PPIs may also help explain that CuDIPs, a SOD mimic, rescued the dysregulated glucose-stimulated insulin secretion in only Sod1−/− but not Gpx1−/− islets [42]. Nevertheless, there were several technical limitations in our study. Although we detected the protein complex by GST pull-down assay, immunoprecipitation, and Western blot, the amount of protein complex was low because of the non-specific interacting proteins within the purified YWHAE and YWHAZ proteins. We were not able to perform size-exclusion chromatography and co-crystallization. Further research will be conducted to optimize the condition for forming the protein complex, measuring binding constants between SOD1 and YWHAE or YWHAZ for analyzing the co-crystalline structure to map protein binding domains. Although part of our data was derived from analyses of tissues and cells with stably knocked out or knocked down of genes, many experiments were run with cells overproducing the WT and mutated proteins. Future research is needed to validate the PPIs at physiological conditions of the involved proteins. Although we performed multiple in vitro and in vivo tests and obtained consistent elevations of SOD1 activity due to its interactions with YWHAE or YWHAZ, we hope that our initial findings will lead to more systematic and basic research to characterize the molecular mechanisms and metabolic relevance of the demonstrated PPIs across various cell types, tissues, and species. To distinguish impacts of designated mutations on the enzymatic activity from those on the protein-interaction-based functions, we will prepare mutants depriving the SOD1 activity but maintaining the native structure to evaluate the phenotypes induced by PPIs only. As intracellular PPIs are not necessarily limited to only pairs of two proteins, multi-protein complexes and sophisticated functions should be considered in the future. Overall, we have elucidated two-way PPIs between SOD1 and YWHAE or YWHAZ and characterized their structural dependences and responses to redox modulation. After revealing the mutual impacts of the PPIs on the enzymatic function (activity) of SOD1 and the protein stability of YWHAE and YWHAZ, we have also demonstrated the effects of the PPIs on lipid metabolism, proliferation, and survival of cultured cells. Overall, our study unveils a new non-canonical role of SOD1 and provides novel perspectives for diagnosing and treating paradoxical diseases related to the protein. ## 4.1. Cells Culture, Animals, Plasmids, and Protein Expression and Purification HEK293T cells (gifted by Dr. Haiyuan Yu, Cornell University, Ithaca, NY, USA) were maintained in DMEM medium (Cat No. 11995065, Thermo Fisher Scientific, Waltham, MA, USA), and HepG2 cells (gifted by Dr. Ruihai Liu, Cornell University, Ithaca, NY, USA) were maintained in William’s E medium (Cat No. 12551032, Thermo Fisher Scientific, Waltham, MA, USA), supplemented with $10\%$ fetal bovine serum (FBS, Hyclone, Logan, UT, USA) and $1\%$ Antibiotic-Antimycotic (Sigma-Aldrich, Burlington, MA, USA) and incubated, at 37 °C, under air with $5\%$ CO2. The experimental cells were cultured with 5–20 passages. The Sod1−/− and WT mice were generated from the 129SVJ × C57BL/6 strain [68]. Deletion of the sod1 gene was verified by genotyping using PCR. All mice used in this study were 8-week-old males ($$n = 3$$–5 replicates). Tissue samples were frozen in liquid nitrogen and stored, at −80 °C, before proceeding to analysis. Our animal experiments were approved by the Institutional Animal Care and Use Committee at Cornell University and conducted following National Institutes of Health guidelines for animal care. The DNA fragments of human YWHAZ, YWHAE, and SOD1, and murine Ywhaz, Ywhae, and Sod1 (gifted by Dr. Haiyuan Yu, Cornell University, Ithaca, NY, USA) were subcloned into pcDNA3.1-F1 or pcDNA3.1-F2 vectors in N or C terminal (F1N or F1C and F2N or F2C). The cloning and construction of plasmids were performed as described previously [69] or following the Gateway cloning method provided by Thermo Fisher Scientific (Waltham, MA, USA). For constitutive protein expression, SOD1, YWHAE, and YWHAZ were also PCR-amplified from the cDNA library and subcloned into the pPICZαA vector. Sequenced recombinant plasmids were transformed into Pichia Pastoris X33 yeast, followed by colony PCR confirmation. Positive transformants were grown, and the produced proteins were His-tag purified, following the expression system’s instructions. The YWHAZ and YWHAE genes were PCR-amplified and subcloned into the pGEX-6P-1 vector (gifted by Dr. Yuxin Mao, Cornell University). The sequenced recombinant plasmids were transformed into E. coli strain BL21 (DE3) cells, and the transformation was confirmed using colony PCR. The correct transformants were grown, and the produced proteins were purified by GST affinity beads, following the manufacturer’s instructions. Other plasmids used in this study are listed in Table S1. ## 4.2. Protein Complementation Assay (PCA) To perform PCA, we seeded HEK293T cells (100 μL, 2 × 104 cells/well) in DMEM media without phenol red (Cat. No 21063029, Thermo Fisher Scientific, Waltham, MA, USA) onto a black 96-well cell culture plate 1 d before transfection (Cat. No 3603, Costar, Washington, DC). On the transfection day, cells were grown to 60–$70\%$ confluency and then cotransfected with 100 ng of bait vector (candidate proteins tagged by F2) plus 100 ng of prey vector (target protein SOD1 tagged by F1) or negative control group F1, F2 vectors (Details in Table S1), using lipofectamine 3000 plus p3000 reagents (Invitrogen™, Carlsbad, CA, USA) mixed with Opti-MEM medium (Cat. No11058021, Thermo Fisher Scientific, Waltham, MA, USA). At 72 h after transfection, fluorescence in the plate was measured using Tecan M1000 spectrophotometer (Zürich, Switzerland) at excitation = 512 nm and emission = 529 nm, or imaged by LSM880 Confocal multiphoton inverted i880 (Core Facilities, Cornell University, Ithaca, NY, USA) with setting laser wavelength 514 nm and detection wavelength 517–642 nm, gain = 650. The measured fluorescent intensity ($$n = 3$$–5 replicates) was normalized and analyzed against negative vector controls in the same protein terminal orientation. The fold change indicated the relative fluorescence intensity to the background. If the value was close to 1 or lower, there was no PPI. When the value was above 1, the higher the value, the stronger and more abundant the PPIs. In the oxidative stress assay, diquat and tert-butyl hydroperoxide (TBHP) were used as oxidant generators, and N-acetyl cysteine (NAC), copper diisopropylsalicylate (CuDIPs), and ebselen were used as antioxidants to treat cells at post-transfection. See Supplemental Method S1 for detailed methods. ## 4.3. Protein Complex Formation and Pull-Down Assay The GST-fusion protein pull-down was performed as previously described [70,71]. Purified GST-YWHAE or GST-YWHAZ proteins were first run through a 50kDa cutoff centrifuge concentrator (Amicon, Charlotte, NC, USA) to remove non-specific binding proteins. The leftover (>50kDa) GST-YWHAE or GST-YWHAZ protein was immobilized on glutathione sepharose 4B beads (GE Healthcare, Piscataway, NJ, USA), at 4 °C, for 1 h, then rinsed with PBS to remove the non-binding proteins. Thereafter, beads were incubated with purified SOD1-His protein (200 μg) at protein amount ratios of GST-YWHAE or GST-YWHAZ to SOD1-His: 2:1, 1:1, and 1:2, at 4 °C, for 1 h in a pull-down buffer containing 20 mM Tris, 300 mM NaCl, 2 mM β-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride (PMSF), pH 8.0. After the beads were rinsed with the pull-down buffer 3 times and with PBS 3 times, the eluates were analyzed by SDS-PAGE ($12\%$ gel), followed by Coomassie blue staining or Western blot. The SOD1-His, GST-YWHAE or GST-YWHAZ protein was replaced with the same volume of the aforementioned pull-down buffer solution as negative controls. Since the large molecular weight of GST might block the binding sites, we expressed and purified YWHAZ-His protein via Pichia Pastoris system. For immunoprecipitation assay, purified SOD1-His protein (100 μg) was mixed with the anti-SOD1 antibody (2 μg, Table S3) and incubated, at 4 °C, for 12 h. Subsequently, 50 μL of pre-washed agarose beads G slurry (Sigma, Burlington, MA, USA) was added to the protein–antibody complex solution and made up to 500 μL by binding buffer (20 mM Tris, 150 mM NaCl, 1x protease inhibitor (100x Halt™ Protease Inhibitor Cocktail EDTA-free, Thermo Fisher Scientific, Waltham, MA, USA)) and incubated for 2 h, at 4 °C. The beads were collected by centrifuge and washed 5 times with 500 μL of washing buffer (TBST (20 mM Tris, 150 mM NaCl, $0.1\%$ Tween), 1x protease inhibitor), then centrifuged at 500× g for 1 min to remove the supernatant. In the negative controls, the purified SOD1-His protein was replaced by the same volume of binding buffer, or the same amount of rabbit IgG isotype was used to replace the anti-SOD1 antibody. Subsequently, the same amount of purified YWHAZ-His protein (100 μg) was added to the beads to prepare a mixture of 500 μL with fresh-made pull-down buffer (20 mM Tris, pH 8.0, 300 mM NaCl, 2 mM β-mercaptoethanol, 1 mM PMSF) as well as the negative control groups. After incubation at 4 °C for 1 h and 5 times washings with the washing buffer, the beads were eluted with 1x SDS loading buffer, incubated at 70 °C for 10 min, and run for Western blot analysis against the anti-His antibody. ## 4.4. Mutagenesis of the SOD1 Gene and Analyses of the Disrupted PPIs The mutant human SOD1 (ALS-related mutants [31,32,33,34]: H46R, G85R, G93A, and zinc active center mutant D124N [35]) gene was generated by Q5® Site-Directed Mutagenesis Kit (New England Biolabs, Ipswich, MA, USA) using the primers listed in Table S2 using the template plasmids WT-SOD1-F1N and WT-SOD1-F1C. These mutants were chosen because: [1] they have different impacts on SOD1 activity, allowing us to test their protein–protein interactions (PPI) with YWHAZ/E at different levels of SOD1 enzymatic function or individually; [2] The AlphaFold structure model suggested these mutations were located in the domain in SOD1 that interacted with YWHAZ or YWHAE; or [3] they (H46R, G85R, and G93A) are involved in ALS, and their PPIs with YWHAZ/E might help explain why SOD1 activity was not the primary etiological factor of ALS. The sequenced mutagenesis-produced plasmids were used to perform the standard PCA mentioned above. An amount of 100 ng mutant SOD1-F1N or mutant SOD1-F1C was used for co-transfection with F2N vectors (YWHAE-F2N or YWAHZ-F2N) into HEK293T cells. At 72 h after transfection, we measured the whole-plate PCA fluorescence. Co-transfection of WT-SOD1-F1N and WT-SOD1-F1C plasmids with YWHAE-F2N or YWAHZ-F2N in every plate was used for normalization purposes. The measured disruptive mutant interaction is normalized to the percentage change compared with the wild-type interaction, following the Formula [1]:[1]Relative percentage% = Fold changemutant − 1Fold changewild type − 1 × $100\%$ ## 4.5. Measurements of SOD1 and SOD2 Activities Supernatants were collected from cell pellets lysed in SOD activity buffer (20 mM HEPES, pH 7.2, containing 1 mM EGTA, 210 mM mannitol, and 70 mM sucrose) using an ice sonication cooling cycle. The collected supernatants were diluted (with a buffer of 50 mM Tris-HCL, pH 8.0) or concentrated using a 10 kDa cutoff centrifuge concentrator (Amicon, Charlotte, NC, USA) to adjust enzymatic activity into a linear-curve range. Total SOD activity was measured using a superoxide dismutase assay kit (706002, Cayman, Ann Arbor, MI, USA). The SOD2 activity was measured by adding 10 μL of potassium cyanide solution (freshly prepared) into the sample to a final concentration of 3 mM in the assay to explicitly inhibit the SOD1 activity [72]. The actual SOD1 activity was calculated by subtracting the SOD2 activity from the total SOD activity. In the assay of effect of protein complex formation on SOD1 activity, 20 μg of His-tag-purified SOD1 and YWHAE or YWHAZ protein each were added into 100 μL of the mixture (with the SOD activity assay buffer). The mixture was incubated, at 4 °C, for 0, 1, 2, 4, and 15 h with SOD1 to YWHAE or YWHAZ at 1:1 ratio or for 15 h with SOD1 to YWHAE or YWHAZ at 1:3 ratio. At the end of the incubations, the mixture was used to assay for SOD1 activity. ## 4.6. CRISPR/Cas9 Editing Genome SOD1 and YWHAZ in HEK293T Cells HEK293T cells (106 seeded on T-25 culture flask) were transiently transfected using lipo3000 with the pX459 vector (gifted by Dr. John Schimenti, Cornell University, Ithaca, NY, USA) integrated with the guide RNA (gRNA) targeting 5′- CTAGCGAGTTATGGCGACGA-3′ sequence in exon 1 of the SOD1 gene (as pX459-SOD1), or pDG459 (gifted by Dr. Tudorita Tumbar, Cornell University, Ithaca, NY, USA) integrated the dual gRNAs targeting 5′-CATGACTGGATGTTCTGCAG-3′ and 5′-AGATATCTGCAATGATGTAC-3′ sequences in exon 4 of the YWHAZ gene (as pDG459-YWAHZ) following standard cloning method described in the protocol [73,74] to generate stable SOD1 knockdown (SOD1-KD) or YWHAZ knockout (YWHAZ-KO) cells. The SOD1-A4V (alanine to valine at fourth position) genome mutation was created by homologous recombination repairing of CRISPR-Cas knock-in through adding single-stranded oligodeoxynucleotides (ssODNs) as a donor template during the pX459-SOD1 transfection period. We only screened out SOD1-KD cells because SOD1 knockout seemed lethal to our HEK293T cells. We constructed the SOD1-A4V mutant in the cell line because the A4V mutant is also related to ALS, which depleted SOD1 activity by $97\%$ (Figure S4D) and is the most common mutant with rapid disease progression in US ALS patients [75]. In this way, we could generate both SOD1-KD and SOD1-A4V as protein-loss and ALS-related cell lines to test intracellular conditions under impaired SOD1 protein and activity. Primer sets used for the annealing formation of 20 bp gRNA oligonucleotides were listed in Table S2. At 48 h after transfection, cells were selected using puromycin (0.8 μg/mL) for another 72 h. The cells were harvested, washed 3 times with PBS, and resuspended in FACS buffer (1L PBS, 1g BSA, 2mM EDTA) to a cell density of 106. The cells were run through SONY MA900 flow cytometry (Core Facility, Cornell University, Ithaca, NY, USA) and sorted out as single cells into the 96-well plate. After the single colony grew, part was used to extract DNA by an alkaline lysis method [76], followed by colony PCR (primers in Table S2) and sequencing of exon 1 of the SOD1 gene or exon 4 of the YWHAZ gene. The rest of the colony was maintained under standard culture and passage. The heterozygotes could continue to perform the TA cloning (TOPO™ TA Cloning™ kit, Thermo Fisher Scientific, Waltham, MA, USA) on colony PCR products to confirm the sequence information for each chromosome. Successfully edited colonies were maintained for subsequent experiments. The SOD1 protein knockdown cells were further transfected with the pDG459-YWHAZ vector and followed the same process to select SOD1 protein knockdown and YWHAZ protein knockout cells. ## 4.7. Co-Immunoprecipitation HepG2 cells (seeded on T-25 culture flasks) were transfected with human WT-SOD1-F1N, SOD1-H46R-F1N, SOD1-G93A-F1N, or F1N as a control vector. After transfection for 72 h, the cells were harvested and lysed in 400 μL of lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, $1\%$ NP-40, $5\%$ glycerol). Cell lysates (1 mg protein) were incubated with 2 μg of SOD1 antibody (Santa Cruz Biotechnology, Dallas, TX, USA) for 16 h, at 4 °C. The Pierce Protein A Magnetic beads (100 μL per sample, Thermo Fisher Scientific, Waltham, MA, USA) were pre-washed by a washing buffer (25 mM Tris-HCl, pH 7.4, 0.5 M NaCl, $0.05\%$ Tween) following the user guide. Afterward, the antigen sample/antibody mixture was incubated with pre-washed magnetic beads, at room temperature, for 1 h in an end-to-end rotator. The beads were washed for 3 times with the washing buffer and 1 time with ultrapure water, and eluted into 80 μL of SDS-PAGE sample loading buffer boiled at 90 °C for 10 min. The supernatants were used for SDS-PAGE analysis and immunoblotting against SOD1, YWHAE, and YWHAZ primary antibodies and a specific IP detection secondary antibody to minimize the background (VeriBlot, ab131366, Abcam, Waltham, MA, USA). ## 4.8. Protein Degradation Assay The cycloheximide (Sigma, Burlington, MA, USA) chase assay was applied for the protein degradation assay, where cycloheximide could entirely block protein synthesis [77]. It is also called a protein stability assay described in the previous paper [46], indicating the protein’s resistance to itself degradation. HEK293T cells (seeded on 12-well plates overnight) were cotransfected with SOD1-F1N or mutant SOD1-F1N (G85R and D124N) and YWHAE-F2N or YWHAZ-F2N, or F1N or F2N control vectors as negative controls, respectively. After 24 h of transfection, cells were treated with freshly prepared cycloheximide (20 μg/mL) or PBS as the control. The cells were harvested at 0 or 24 h after the treatment to determine relative protein amounts of YWHAE, YWHAZ, and SOD1 by Western blot. The protein stability was reflected by quantifying protein band intensity of 24 h normalized to its 0 h, respectively. ## 4.9. Staining of Lipid Droplets and Detection of Lipid Profiles For lipid droplets staining, HepG2 cells were seeded onto a black 96-well cell culture plate 1 d before transfection. The cells were transfected with WT-SOD1-F1N, mutant SOD1-F1N (H46R and G93A), and F1N control vector, or cotransfected with YWHAE-F2N and SOD1-F1N or F1N control vector. The lipid droplets in the cells were stained using Nile Red Staining Kit (Abcam, Waltham, MA, USA). After 48 h of transfection, culture media was aspirated before the addition of 100 μL of Nile Red Staining solution per well. Cells were incubated, at 37 °C, under air with $5\%$ CO2 for 30 min. The whole plate was measured with Nile red fluorescence using Tecan M1000 spectrophotometer (Zürich, Switzerland) at excitation = 550 nm and emission = 640 nm, or the cells were imaged using LSM880 Confocal multiphoton inverted i880 (Core Facilities, Cornell University, Ithaca, NY, USA) with setting laser wavelength excitation = 514 nm, emission = 646 nm, detection wavelength 539–753 nm, gain = 532. The WT and CRISPR-genome-edited SOD1-KD (SOD1 knockdown) and SOD1-KD&YWHAZ-KO (SOD1 knockdown and YWHAZ knockout) cells (104 per well) were plated in black 96-well plates for 24 h and were treated with CuDIPs (1 or 10 μM of final concentrations) or an identical amount of DMSO vehicle. After the treatment for 48 h, cell culture media was aspirated before the addition of 100 μL of Nile Red Staining solution per well. Cells were incubated, at 37 °C, under air with $5\%$ CO2 for 30 min, followed by standard Nile red fluorescence measurement or fluorescence microscope imaging mentioned above. Total cholesterol (TC), total triglycerides (TG), and non-esterified fatty acid (NEFA) in HepG2 cells (106 cells/flask) were determined at 72 h after transfection. Cells were sonicated in 500 μL of PBS, and the supernatant protein concentration was determined using the Pierce Bicinchoninic Acid (BCA) protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA). The remaining supernatant (450 μL) was mixed with 1 mL of chloroform: isopropanol: NP-40 = 7:11:0.1 solution and centrifuged. The lower layer was transferred into a glass tube, dried under nitrogen gas, and dissolved into 200 mL of ethanol containing $1\%$ Triton. Concentrations of TC, TG, and NEFA in the ethanol solution were measured using kits (Wako Chemicals, Richmond, VA, USA). Fatty acid profiling in the cell homogenates, along with lipid extraction, was performed as described previously [78]. ## 4.10. Cell Growth and Survival HEK293T (WT), SOD1-A4V, SOD1-KD (SOD1 knockdown), and SOD1-KD&YWHAZ-KO (SOD1 knockdown and YWHAZ knockout) cells (104 cells/well) were seeded in 96-well plates overnight and transfected with 100 ng of SOD1-F1N, YWHAZ-F2N, YWHAE-F2N or their respective vector controls (F1N, F2N) using lipofectamine 3000 plus p3000 reagents. After 48 h of transfection, cells were further cultured in 100 μL of fresh DMEM supplemented with $10\%$ FBS for 0, 24, or 48 h for proliferation assay. At each time-point, 10 μL of MTT (12 mM) solution was added into each well to measure cell growth following a standard CyQUANT MTT cell assay protocol by Thermo Fisher Scientific (Waltham, MA, USA). Following incubation with MTT for 4 h, at 37 °C, 100 μL SDS-HCl (1 g SDS in 10 mL 0.01 M HCl) was added to each well. Incubated the microplate, at 37 °C, for another 4 h to completely dissolve the blue crystal formed. Mixed each sample thoroughly by pipetting up and down. The final absorbance was then measured at 570 nm using the SpectraMax M2e spectrophotometer (Silicon Valley, CA, USA). Cell survival was also determined using the MTT with the assay previously described [79]. After 48 h of transfection, cells were cultured and stressed in 100 μL of DMEM without FBS for 24, 48, or 72 h before adding 10 μL of MTT (12 mM) to each well. Cell survival was set as $100\%$ at the initial time point and relative survival was calculated for each time point thereafter. ## 4.11. Quantitative Real-Time PCR (qPCR) Total RNA was extracted from cells (~106 cells) using an RNeasy Mini kit (QIAGEN, Germantown, MD, USA). The high-capacity cDNA reverse transcription kit (Cat No. 43-688-14, Applied Biosystems, Waltham, MA, USA) was used for reverse transcription. The Real-time qPCR used an iTaq Universal SYBR Green Supermix (Cat No. 1725124, Bio-Rad Laboratories, Hercules, CA, USA) according to the manufacturer’s instructions on QuantStudio 7 Pro (Thermo Fisher Scientific, Waltham, MA, USA), and 2–delta delta Ct (∆∆Ct) equation [80] was used to quantify relative mRNA levels normalized to levels of the housekeeping gene GAPDH. Primers used for qPCR are listed in Table S2. ## 4.12. Western Blotting Cells (106 cells) or mouse tissues (50mg) were lysed in RIPA buffer with 1x Halt™ Protease Inhibitor Cocktail, EDTA-free (Thermo Fisher Scientific, Waltham, MA, USA). Protein concentration was assayed by the BCA protein assay kit described above. Supernatants (40–60 µg of protein per lane) of the lysates were subjected to Western blot analyses as described previously [81]. Relative densities of protein bands in individual blots were quantified using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific, Waltham, MA, USA), and the membrane was imaged by ChemiDocTM MP Imaging System (Bio-Rad). In addition, all immunoblot band intensities in this study were analyzed by ImageJ (National Institutes of Health, Bethesda, MD, USA). Antibodies used in this study are listed in Table S3. ## 4.13. Correlation Analysis of the Protein Expression Levels The protein expression data were acquired from a breast cancer quantitative proteome database (Henrik J. Johansson, 2019) [41]. The abundance of the proteins was further processed by the tools suggested by Dr. Nathaniel Vacanti (Cornell University, Ithaca, NY, USA). The final protein data was relative to an internal standard, log2 transformed, and then standardized to z-score. The missing data and outliers were removed during data processing. The x-axis was set as SOD1, and the y-axis was set as YWHAE or YWHAZ. Pearson correlation coefficient (r) with p-value was calculated for SOD1-YWHAE and SOD1-YWHAZ pairs. ## 4.14. Statistical Analysis Data are presented as mean ± SE ($$n = 3$$–6). Data generated from experiments with more than 2 treatment groups were analyzed with R software (version 4.0.3, R Core Team, Vienna, Austria) using one-way ANOVA followed by Duncan’s test. Data generated from experiments with only 2 treatment groups were analyzed using Student’s t-test. The Pearson correlation was analyzed with GraphPad Prism 8.0.1 (GraphPad Software, Inc., San Diego, CA, USA). Statistical significance of differences was set at and indicated as : $p \leq 0.05$; : $p \leq 0.01$; and : $p \leq 0.001.$ ## References 1. Chang L.Y., Slot J.W., Geuze H.J., Crapo J.D.. Molecular immunocytochemistry of the CuZn superoxide dismutase in rat hepatocytes*. J. Cell Biol.* (1988) 107 2169-2179. DOI: 10.1083/jcb.107.6.2169 2. Trist B.G., Hilton J.B., Hare D.J., Crouch P.J., Double K.L.. Superoxide Dismutase 1 in Health and Disease: How a Frontline Antioxidant Becomes Neurotoxic. Angew. Chem. Int. Ed. Engl. (2021) 60 9215-9246. DOI: 10.1002/anie.202000451 3. Lu L., Zheng L., Viera L., Suswam E., Li Y., Li X., Estevez A.G., King P.H.. Mutant Cu/Zn-superoxide dismutase associated with amyotrophic lateral sclerosis destabilizes vascular endothelial growth factor mRNA and downregulates its expression. J. Neurosci. (2007) 27 7929-7938. DOI: 10.1523/JNEUROSCI.1877-07.2007 4. Li X., Lu L., Bush D.J., Zhang X., Zheng L., Suswam E.A., King P.H.. Mutant copper-zinc superoxide dismutase associated with amyotrophic lateral sclerosis binds to adenine/uridine-rich stability elements in the vascular endothelial growth factor 3’-untranslated region. J. Neurochem. (2009) 108 1032-1044. DOI: 10.1111/j.1471-4159.2008.05856.x 5. Tsang C.K., Liu Y., Thomas J., Zhang Y., Zheng X.F.. Superoxide dismutase 1 acts as a nuclear transcription factor to regulate oxidative stress resistance. Nat. Commun. (2014) 5 3446. DOI: 10.1038/ncomms4446 6. Wang H., Wang M., Nie Z., Qiu S., Huang X., Li X., Cui Y., Liu C., Liu C.. SAXS Examinations of the Redox-Dependent Formation of a DNA-SOD1 Complex. Int. J. Mol. Sci. (2022) 23. DOI: 10.3390/ijms232012673 7. Casareno R.L., Waggoner D., Gitlin J.D.. The copper chaperone CCS directly interacts with copper/zinc superoxide dismutase. J. Biol. Chem. (1998) 273 23625-23628. DOI: 10.1074/jbc.273.37.23625 8. Sala F.A., Wright G.S.A., Antonyuk S.V., Garratt R.C., Hasnain S.S.. Molecular recognition and maturation of SOD1 by its evolutionarily destabilised cognate chaperone hCCS. PLoS Biol. (2019) 17. DOI: 10.1371/journal.pbio.3000141 9. Hamilton R.T., Bhattacharya A., Walsh M.E., Shi Y., Wei R., Zhang Y., Rodriguez K.A., Buffenstein R., Chaudhuri A.R., Van Remmen H.. Elevated Protein Carbonylation, and Misfolding in Sciatic Nerve from db/db and Sod1−/− Mice: Plausible Link between Oxidative Stress and Demyelination. PLoS ONE (2013) 8. DOI: 10.1371/journal.pone.0065725 10. Van Der Paal J., Neyts E.C., Verlackt C.C.W., Bogaerts A.. Effect of lipid peroxidation on membrane permeability of cancer and normal cells subjected to oxidative stress. Chem. Sci. (2016) 7 489-498. DOI: 10.1039/C5SC02311D 11. Li X., Weng H., Reece E.A., Yang P.. SOD1 overexpression in vivo blocks hyperglycemia-induced specific PKC isoforms: Substrate activation and consequent lipid peroxidation in diabetic embryopathy. Am. J. Obstet. Gynecol. (2011) 205 84.e1-84.e6. DOI: 10.1016/j.ajog.2011.02.071 12. Li J., Song M., Moh S., Kim H., Kim D.H.. Cytoplasmic Restriction of Mutated SOD1 Impairs the DNA Repair Process in Spinal Cord Neurons. Cells (2019) 8. DOI: 10.3390/cells8121502 13. Lei X.G., Zhu J.H., McClung J.P., Aregullin M., Roneker C.A.. Mice deficient in Cu,Zn-superoxide dismutase are resistant to acetaminophen toxicity. Biochem. J. (2006) 399 455-461. DOI: 10.1042/BJ20060784 14. Wang X., Gillen E.A., van der Meulen M.C., Lei X.G.. Knockouts of Se-glutathione peroxidase-1 and Cu,Zn superoxide dismutase exert different impacts on femoral mechanical performance of growing mice. Mol. Nutr. Food Res. (2008) 52 1334-1339. DOI: 10.1002/mnfr.200700351 15. Wang X., Vatamaniuk M.Z., Roneker C.A., Pepper M.P., Hu L.G., Simmons R.A., Lei X.G.. Knockouts of SOD1 and GPX1 Exert Different Impacts on Murine Islet Function and Pancreatic Integrity. Antioxid. Redox Signal. (2011) 14 391-401. DOI: 10.1089/ars.2010.3302 16. Wang L., Jiang Z., Lei X.G.. Knockout of SOD1 alters murine hepatic glycolysis, gluconeogenesis, and lipogenesis. Free. Radic. Biol. Med. (2012) 53 1689-1696. DOI: 10.1016/j.freeradbiomed.2012.08.570 17. Wang S.K.. Impact of Copper, Zinc Superoxide Dismutase Knockout on Lipid Metabolism in Mice. Ph.D. Thesis (2009) 18. Dougherty M.K., Morrison D.K.. Unlocking the code of 14-3-3. J. Cell Sci. (2004) 117 1875-1884. DOI: 10.1242/jcs.01171 19. Fu H., Subramanian R.R., Masters S.C.. 14-3-3 Proteins: Structure, Function, and Regulation. Annu. Rev. Pharmacol. Toxicol. (2000) 40 617-647. DOI: 10.1146/annurev.pharmtox.40.1.617 20. Foote M., Zhou Y.. 14-3-3 proteins in neurological disorders. Int. J. Biochem. Mol. Biol. (2012) 3 152-164. PMID: 22773956 21. Liang R., Chen X.Q., Bai Q.X., Wang Z., Zhang T., Yang L., Dong B.X., Gao G.X., Gu H.T., Zhu H.F.. Increased 14-3-3zeta expression in the multidrug-resistant leukemia cell line HL-60/VCR as compared to the parental line mediates cell growth and apoptosis in part through modification of gene expression. Acta. Haematol. (2014) 132 177-186. DOI: 10.1159/000357377 22. Weerasekara V.K., Panek D.J., Broadbent D.G., Mortenson J.B., Mathis A.D., Logan G.N., Prince J.T., Thomson D.M., Thompson J.W., Andersen J.L.. Metabolic-Stress-Induced Rearrangement of the 14-3-3ζ Interactome Promotes Autophagy via a ULK1- and AMPK-Regulated 14-3-3ζ Interaction with Phosphorylated Atg9. Mol. Cell. Biol. (2014) 34 4379-4388. DOI: 10.1128/MCB.00740-14 23. Nishimura Y., Komatsu S., Ichikawa D., Nagata H., Hirajima S., Takeshita H., Kawaguchi T., Arita T., Konishi H., Kashimoto K.. Overexpression of YWHAZ relates to tumor cell proliferation and malignant outcome of gastric carcinoma. Br. J. Cancer (2013) 108 1324-1331. DOI: 10.1038/bjc.2013.65 24. Matta A., Siu K.M., Ralhan R.. 14-3-3 zeta as novel molecular target for cancer therapy. Expert Opin. Ther. Targets (2012) 16 515-523. DOI: 10.1517/14728222.2012.668185 25. Lim G.E., Albrecht T., Piske M., Sarai K., Lee J.T.C., Ramshaw H.S., Sinha S., Guthridge M.A., Acker-Palmer A., Lopez A.F.. 14-3-3ζ coordinates adipogenesis of visceral fat. Nat. Commun. (2015) 6 7671. DOI: 10.1038/ncomms8671 26. Lim G.E., Piske M., Lulo J.E., Ramshaw H.S., Lopez A.F., Johnson J.D.. Ywhaz/14-3-3ζ Deletion Improves Glucose Tolerance Through a GLP-1-Dependent Mechanism. Endocrinology (2016) 157 2649-2659. DOI: 10.1210/en.2016-1016 27. Craparo A., Freund R., Gustafson T.A.. 14-3-3 (ϵ) Interacts with the Insulin-like Growth Factor I Receptor and Insulin Receptor Substrate I in a Phosphoserine-dependent Manner. J. Biol. Chem. (1997) 272 11663-11669. DOI: 10.1074/jbc.272.17.11663 28. Malty R.H., Aoki H., Kumar A., Phanse S., Amin S., Zhang Q., Minic Z., Goebels F., Musso G., Wu Z.. A Map of Human Mitochondrial Protein Interactions Linked to Neurodegeneration Reveals New Mechanisms of Redox Homeostasis and NF-kappaB Signaling. Cell Syst. (2017) 5 564-577.e12. DOI: 10.1016/j.cels.2017.10.010 29. Marcon E., Ni Z., Pu S., Turinsky A.L., Trimble S.S., Olsen J.B., Silverman-Gavrila R., Silverman-Gavrila L., Phanse S., Guo H.. Human-chromatin-related protein interactions identify a demethylase complex required for chromosome segregation. Cell Rep. (2014) 8 297-310. DOI: 10.1016/j.celrep.2014.05.050 30. Meyer K., Selbach M.. Quantitative affinity purification mass spectrometry: A versatile technology to study protein–protein interactions. Front. Genet. (2015) 6 237. DOI: 10.3389/fgene.2015.00237 31. Ratovitski T., Corson L.B., Strain J., Wong P., Cleveland D.W., Culotta V.C., Borchelt D.R.. Variation in the biochemical/biophysical properties of mutant superoxide dismutase 1 enzymes and the rate of disease progression in familial amyotrophic lateral sclerosis kindreds. Hum. Mol. Genet. (1999) 8 1451-1460. DOI: 10.1093/hmg/8.8.1451 32. Borchelt D.R., Lee M.K., Slunt H.S., Guarnieri M., Xu Z.S., Wong P.C., Brown R.H., Price D.L., Sisodia S.S., Cleveland D.W.. Superoxide dismutase 1 with mutations linked to familial amyotrophic lateral sclerosis possesses significant activity. Proc. Natl. Acad. Sci. USA (1994) 91 8292-8296. DOI: 10.1073/pnas.91.17.8292 33. Sau D., De Biasi S., Vitellaro-Zuccarello L., Riso P., Guarnieri S., Porrini M., Simeoni S., Crippa V., Onesto E., Palazzolo I.. Mutation of SOD1 in ALS: A gain of a loss of function. Hum. Mol. Genet. (2007) 16 1604-1618. DOI: 10.1093/hmg/ddm110 34. Saccon R.A., Bunton-Stasyshyn R.K., Fisher E.M., Fratta P.. Is SOD1 loss of function involved in amyotrophic lateral sclerosis?. Brain (2013) 136 2342-2358. DOI: 10.1093/brain/awt097 35. Nedd S., Redler R.L., Proctor E.A., Dokholyan N.V., Alexandrova A.N.. Cu,Zn-superoxide dismutase without Zn is folded but catalytically inactive. J. Mol. Biol. (2014) 426 4112-4124. DOI: 10.1016/j.jmb.2014.07.016 36. Kodama Y., Hu C.D.. An improved bimolecular fluorescence complementation assay with a high signal-to-noise ratio. Biotechniques (2010) 49 793-805. DOI: 10.2144/000113519 37. Deng H.X., Hentati A., Tainer J.A., Iqbal Z., Cayabyab A., Hung W.Y., Getzoff E.D., Hu P., Herzfeldt B., Roos R.P.. Amyotrophic lateral sclerosis and structural defects in Cu,Zn superoxide dismutase. Science (1993) 261 1047-1051. DOI: 10.1126/science.8351519 38. Banci L., Bertini I., Cabelli D.E., Hallewell R.A., Tung J.W., Viezzoli M.S.. A characterization of copper/zinc superoxide dismutase mutants at position 124. Zinc-deficient proteins. Eur. J. Biochem. (1991) 196 123-128. DOI: 10.1111/j.1432-1033.1991.tb15794.x 39. Hermeking H., Benzinger A.. 14-3-3 proteins in cell cycle regulation. Semin. Cancer Biol. (2006) 16 183-192. DOI: 10.1016/j.semcancer.2006.03.002 40. Vella D., Zoppis I., Mauri G., Mauri P., Di Silvestre D.. From protein–protein interactions to protein co-expression networks: A new perspective to evaluate large-scale proteomic data. EURASIP J. Bioinform. Syst. Biol. (2017) 2017 6. DOI: 10.1186/s13637-017-0059-z 41. Johansson H.J., Socciarelli F., Vacanti N.M., Haugen M.H., Zhu Y., Siavelis I., Fernandez-Woodbridge A., Aure M.R., Sennblad B., Vesterlund M.. Breast cancer quantitative proteome and proteogenomic landscape. Nat. Commun. (2019) 10 1600. DOI: 10.1038/s41467-019-09018-y 42. Wang X., Yun J.W., Lei X.G.. Glutathione peroxidase mimic ebselen improves glucose-stimulated insulin secretion in murine islets. Antioxid. Redox Signal. (2014) 20 191-203. DOI: 10.1089/ars.2013.5361 43. Yun J.W., Zhao Z., Yan X., Vatamaniuk M.Z., Lei X.G.. Glutathione peroxidase-1 inhibits transcription of regenerating islet-derived protein-2 in pancreatic islets. Free. Radic. Biol. Med. (2019) 134 385-393. DOI: 10.1016/j.freeradbiomed.2019.01.024 44. Evans R., O’Neill M., Pritzel A., Antropova N., Senior A., Green T., Žídek A., Bates R., Blackwell S., Yim J.. Protein Complex Prediction with AlphaFold-Multimer (2021) 45. Bryant P., Pozzati G., Elofsson A.. Improved prediction of protein–protein interactions using AlphaFold2. Nat. Commun. (2022) 13 1265. DOI: 10.1038/s41467-022-28865-w 46. Dar A., Wu D., Lee N., Shibata E., Dutta A.. 14-3-3 proteins play a role in the cell cycle by shielding cdt2 from ubiquitin-mediated degradation. Mol. Cell Biol. (2014) 34 4049-4061. DOI: 10.1128/MCB.00838-14 47. Kabuta T., Suzuki Y., Wada K.. Degradation of Amyotrophic Lateral Sclerosis-linked Mutant Cu,Zn-Superoxide Dismutase Proteins by Macroautophagy and the Proteasome. J. Biol. Chem. (2006) 281 30524-30533. DOI: 10.1074/jbc.M603337200 48. Kong Q., Ma W.. WRINKLED1 as a novel 14-3-3 client: Function of 14-3-3 proteins in plant lipid metabolism. Plant Signal. Behav. (2018) 13 e1482176. DOI: 10.1080/15592324.2018.1482176 49. Obsilova V., Vecer J., Herman P., Pabianova A., Sulc M., Teisinger J., Boura E., Obsil T.. 14-3-3 Protein Interacts with Nuclear Localization Sequence of Forkhead Transcription Factor FoxO4. Biochemistry (2005) 44 11608-11617. DOI: 10.1021/bi050618r 50. Huai J., Zhang Z.. Structural Properties and Interaction Partners of Familial ALS-Associated SOD1 Mutants. Front. Neurol. (2019) 10 527. DOI: 10.3389/fneur.2019.00527 51. Yin J., Wang X., Li S., Zhu Y., Chen S., Li P., Luo C., Huang Y., Li X., Hu X.. Interactions between plasma copper concentrations and SOD1 gene polymorphism for impaired glucose regulation and type 2 diabetes. Redox Biol. (2019) 24 101172. DOI: 10.1016/j.redox.2019.101172 52. Saremi L., Taghvaei S., Feizy F., Ghaffari M.E., Babaniamansour S., Saltanatpour Z.. Association study between superoxide Dismutases gene polymorphisms and development of diabetic retinopathy and cataract in Iranian patients with type two diabetes mellitus. J. Diabetes Metab. Disord. (2021) 20 627-634. DOI: 10.1007/s40200-021-00790-7 53. Vats P., Sagar N., Singh T.P., Banerjee M.. Association of Superoxide dismutases (SOD1 and SOD2) and Glutathione peroxidase 1 (GPx1) gene polymorphisms with type 2 diabetes mellitus. Free. Radic. Res. (2015) 49 17-24. DOI: 10.3109/10715762.2014.971782 54. Hernandez-Guerrero C., Parra-Carriedo A., Ruiz-de-Santiago D., Galicia-Castillo O., Buenrostro-Jauregui M., Diaz-Gutierrez C.. Genetic polymorphisms of antioxidant enzymes CAT and SOD affect the outcome of clinical, biochemical, and anthropometric variables in people with obesity under a dietary intervention. Genes Nutr. (2018) 13 1. DOI: 10.1186/s12263-017-0590-2 55. Kim S.M., Kim H., Kim J.E., Park K.S., Sung J.J., Kim S.H., Lee K.W.. Amyotrophic lateral sclerosis is associated with hypolipidemia at the presymptomatic stage in mice. PLoS ONE (2011) 6. DOI: 10.1371/journal.pone.0017985 56. Dodge J.C., Treleaven C.M., Fidler J.A., Tamsett T.J., Bao C., Searles M., Taksir T.V., Misra K., Sidman R.L., Cheng S.H.. Metabolic signatures of amyotrophic lateral sclerosis reveal insights into disease pathogenesis. Proc. Natl. Acad. Sci. USA (2013) 110 10812-10817. DOI: 10.1073/pnas.1308421110 57. Smittkamp S.E., Morris J.K., Bomhoff G.L., Chertoff M.E., Geiger P.C., Stanford J.A.. SOD1-G93A mice exhibit muscle-fiber-type-specific decreases in glucose uptake in the absence of whole-body changes in metabolism. Neurodegener Dis. (2014) 13 29-37. DOI: 10.1159/000351606 58. Wong N.K.Y., He B.P., Strong M.J.. Characterization of Neuronal Intermediate Filament Protein Expression in Cervical Spinal Motor Neurons in Sporadic Amyotrophic Lateral Sclerosis (ALS). J. Neuropathol. Exp. Neurol. (2000) 59 972-982. DOI: 10.1093/jnen/59.11.972 59. Ge W.-W., Volkening K., Leystra-Lantz C., Jaffe H., Strong M.J.. 14-3-3 protein binds to the low molecular weight neurofilament (NFL) mRNA 3’ UTR. Mol. Cell. Neurosci. (2007) 34 80-87. DOI: 10.1016/j.mcn.2006.10.001 60. Miao L., Teng J., Lin J., Liao X., Chen J.. 14-3-3 proteins interact with neurofilament protein-L and regulate dynamic assembly of neurofilaments. J. Cell Sci. (2013) 126 427-436. DOI: 10.1242/jcs.105817 61. Neal C.L., Yao J., Yang W., Zhou X., Nguyen N.T., Lu J., Danes C.G., Guo H., Lan K.-H., Ensor J.. 14-3-3ζ Overexpression Defines High Risk for Breast Cancer Recurrence and Promotes Cancer Cell Survival. Cancer Res. (2009) 69 3425-3432. DOI: 10.1158/0008-5472.CAN-08-2765 62. Hong L., Chen W., Xing A., Wu D., Wang S.. Inhibition of Tyrosine 3-Monooxygenase/Tryptophan 5-Monooxygenase Activation Protein Zeta (YWHAZ) Overcomes Drug Resistance and Tumorigenicity in Ovarian Cancer. Cell. Physiol. Biochem. (2018) 49 53-64. DOI: 10.1159/000492839 63. Leal M.F.. Clinical implication of 14-3-3 epsilon expression in gastric cancer. World J. Gastroenterol. (2012) 18 1531. DOI: 10.3748/wjg.v18.i13.1531 64. Park J.-H., Jang H.R., Lee I.Y., Oh H.K., Choi E.-J., Rhim H., Kang S.. Amyotrophic lateral sclerosis-related mutant superoxide dismutase 1 aggregates inhibit 14-3-3-mediated cell survival by sequestration into the JUNQ compartment. Hum. Mol. Genet. (2017) 26 3615-3629. DOI: 10.1093/hmg/ddx250 65. Liu S., Li B., Xu J., Hu S., Zhan N., Wang H., Gao C., Li J., Xu X.. SOD1 Promotes Cell Proliferation and Metastasis in Non-small Cell Lung Cancer via an miR-409-3p/SOD1/SETDB1 Epigenetic Regulatory Feedforward Loop. Front. Cell Dev. Biol. (2020) 8 213. DOI: 10.3389/fcell.2020.00213 66. Gan Y., Ye F., He X.-X.. The role of YWHAZ in cancer: A maze of opportunities and challenges. J. Cancer (2020) 11 2252-2264. DOI: 10.7150/jca.41316 67. Leal M.F., Ribeiro H.F., Rey J.A., Pinto G.R., Smith M.C., Moreira-Nunes C.A., Assumpção P.P., Lamarão L.M., Calcagno D.Q., Montenegro R.C.. YWHAE silencing induces cell proliferation, invasion and migration through the up-regulation of CDC25B and MYC in gastric cancer cells: New insights about YWHAE role in the tumor development and metastasis process. Oncotarget (2016) 7 85393-85410. DOI: 10.18632/oncotarget.13381 68. Ho Y.S., Gargano M., Cao J., Bronson R.T., Heimler I., Hutz R.J.. Reduced fertility in female mice lacking copper-zinc superoxide dismutase. J. Biol. Chem. (1998) 273 7765-7769. DOI: 10.1074/jbc.273.13.7765 69. Rodriguez E., Han Y., Lei X.G.. Cloning, sequencing, and expression of an Escherichia coli acid phosphatase/phytase gene (appA2) isolated from pig colon. Biochem. Biophys. Res. Commun. (1999) 257 117-123. DOI: 10.1006/bbrc.1999.0361 70. Sambrook J., Russell D.W.. Detection of Protein–protein Interactions Using the GST Fusion Protein Pulldown Technique. CSH Protoc. (2006) 2006 pdb-prot3757. DOI: 10.1101/pdb.prot3757 71. Chen L.L., Huang J.Q., Wu Y.Y., Chen L.B., Li S.P., Zhang X., Wu S., Ren F.Z., Lei X.G.. Loss of Selenov predisposes mice to extra fat accumulation and attenuated energy expenditure. Redox Biol. (2021) 45 102048. DOI: 10.1016/j.redox.2021.102048 72. Ghneim H.K., Al-Sheikh Y.A., Alshebly M.M., Aboul-Soud M.A.. Superoxide dismutase activity and gene expression levels in Saudi women with recurrent miscarriage. Mol. Med. Rep. (2016) 13 2606-2612. DOI: 10.3892/mmr.2016.4807 73. Ran F.A., Hsu P.D., Wright J., Agarwala V., Scott D.A., Zhang F.. Genome engineering using the CRISPR-Cas9 system. Nat. Protoc. (2013) 8 2281-2308. DOI: 10.1038/nprot.2013.143 74. Adikusuma F., Pfitzner C., Thomas P.Q.. Versatile single-step-assembly CRISPR/Cas9 vectors for dual gRNA expression. PLoS ONE (2017) 12. DOI: 10.1371/journal.pone.0187236 75. Armon C.. SOD1 A4V familial ALS in North America: Can understanding the past lead to a better future?. Neurology (2009) 72 1628-1629. DOI: 10.1212/01.wnl.0000344651.66120.ea 76. Truett G.E., Heeger P., Mynatt R.L., Truett A.A., Walker J.A., Warman M.L.. Preparation of PCR-quality mouse genomic DNA with hot sodium hydroxide and tris (HotSHOT). Biotechniques (2000) 29 52-54. DOI: 10.2144/00291bm09 77. Buchanan B.W., Lloyd M.E., Engle S.M., Rubenstein E.M.. Cycloheximide Chase Analysis of Protein Degradation in Saccharomyces cerevisiae. J. Vis. Exp. (2016) 110 e53975 78. Kim J., Barcus M., Magnuson A., Tao L., Lei X.G.. Supplemental defatted microalgae affects egg and tissue fatty acid composition differently in laying hens fed diets containing corn and flaxseed oil. J. Appl. Poult. Res. (2016) 25 528-538. DOI: 10.3382/japr/pfw034 79. Hu X., Wang J., He W., Zhao P., Ye C.. MicroRNA-433 targets AKT3 and inhibits cell proliferation and viability in breast cancer. Oncol. Lett. (2018) 15 3998-4004. DOI: 10.3892/ol.2018.7803 80. Livak K.J., Schmittgen T.D.. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods (2001) 25 402-408. DOI: 10.1006/meth.2001.1262 81. Cheng W., Fu Y.X., Porres J.M., Ross D.A., Lei X.G.. Selenium-dependent cellular glutathione peroxidase protects mice against a pro-oxidant-induced oxidation of NADPH, NADH, lipids, and protein. FASEB J. (1999) 13 1467-1475. DOI: 10.1096/fasebj.13.11.1467 82. Jumper J., Evans R., Pritzel A., Green T., Figurnov M., Ronneberger O., Tunyasuvunakool K., Bates R., Žídek A., Potapenko A.. Highly accurate protein structure prediction with AlphaFold. Nature (2021) 596 583-589. DOI: 10.1038/s41586-021-03819-2
--- title: 'Changes in Faecal and Plasma Amino Acid Profile in Dogs with Food-Responsive Enteropathy as Indicators of Gut Homeostasis Disruption: A Pilot Study' authors: - Cristina Higueras - Rosa Escudero - Almudena Rebolé - Mercedes García-Sancho - Fernando Rodríguez-Franco - Ángel Sainz - Ana I. Rey journal: Veterinary Sciences year: 2023 pmcid: PMC9966949 doi: 10.3390/vetsci10020112 license: CC BY 4.0 --- # Changes in Faecal and Plasma Amino Acid Profile in Dogs with Food-Responsive Enteropathy as Indicators of Gut Homeostasis Disruption: A Pilot Study ## Abstract ### Simple Summary Food-responsive enteropathy (FRE) has the greatest prevalence among the different groups of chronic enteropathies. However, information is lacking on the specific amino acid profile for FRE in dogs and its diagnostic utility. This study evaluated differences in the plasma and faecal amino acid profile between control and FRE in dogs as possible indicators of disease. We also searched for correlations between amino acids and parameter indicators of gut health, including body condition score (BCS), and indices, such as canine inflammatory bowel disease activity index (CIBDAI), to evaluate whether the amino acid profile could serve as an indicator of disease severity. Several alterations were observed in plasma and faecal amino acid profiles in sick dogs, and high correlations were found between amino acids and disease activity index or faecal characteristics. More information on the amino acid profile in dogs with FRE could help with diagnoses and lead to more precise and specific amino acid formulation, dietary interventions, better response to diet, and recovery of animals. ### Abstract Dogs suffering from food-responsive enteropathy (FRE) respond to an elimination diet based on hydrolysed protein or novel protein; however, studies regarding the amino acid profile in FRE dogs are lacking. The aim of this pilot study was to evaluate whether the plasma and faecal amino acid profiles differed between control and FRE dogs and whether these could serve as indicators of severity of illness. Blood, faecal samples, body condition score, and severity of clinical signs based on the canine inflammatory bowel disease activity index were collected before starting the elimination diet. FRE dogs had lower proportions of plasma Asparagine, Histidine, Glycine, Cystine, Leucine, and branched-chain/aromatic amino acids; however, Phenylalanine increased. In faecal samples, Cystine was greater whereas Phenylalanine was lesser in sick dogs compared to control. Leucine correlated negatively with faecal humidity (r = −0.66), and Leucine and Phenylalanine with faecal fat (r = −0.57 and r = −0.62, respectively). Faecal Phenylalanine ($r = 0.80$), Isoleucine ($r = 0.75$), and Leucine ($r = 0.92$) also correlated positively with total short-chain fatty acids, whereas a negative correlation was found with Glycine (r = −0.85) and Cystine (r = −0.61). This study demonstrates the importance of Leucine and Phenylalanine amino acids as indicators of the disease severity in FRE dogs. ## 1. Introduction Amino acids are important compounds in the organism as they constitute the main components of proteins and various bioactive molecules [1,2]. Recent studies have proved that amino acids play an important role in the gut and regulation of inflammation [3]. They participate in the proliferation and apoptosis of intestinal epithelial cells (IECs), expression of tight junction proteins (TJPs), inflammatory processes, and oxidative stress through the regulation of signalling pathways [4]. Consequently, amino acids can regulate this common process taking place in inflammatory diseases such as chronic enteropathies (CEs). Some studies carried out in human medicine have shown alterations of the amino acid profile in patients suffering from gastrointestinal diseases, such as inflammatory bowel disease (IBD), with decreased levels of several amino acids in serum samples and increased levels in faecal samples of these patients [5]. This finding also means that some amino acids could be used to monitor clinical disease activity [6] or even serve as part of the treatment by improving clinical symptoms [7]. In veterinary medicine, studies regarding metabolomics in CEs are focused on the search for potential biomarkers. It has been observed that cobalamin, folate, C-reactive protein or dysbiosis index (DI) could help in diagnostic evaluation, prognosis, and monitoring clinical activity [8]. However, studies regarding the nutritional profile of food-responsive enteropathy (FRE) in dogs are lacking. In previous research, we evaluated the short-chain and total fatty acid profiles in FRE dogs, finding alterations and correlations between some of these compounds and the intestinal disease activity index [9]. Currently, scientific reports concerning the amino acid profile in dogs with CEs are very limited, although a few studies have been conducted in dogs with immunosuppressant-responsive enteropathy (IRE) and protein-losing enteropathy (PLE), finding alterations in the amino acid profile when compared to healthy individuals [10,11,12,13]. Since FRE seems to have the greatest prevalence (60–$70\%$ of cases of CE) [14] and dogs suffering from this disease respond to an elimination diet based on hydrolysed protein or novel protein, the amino acid profile evaluation deserves more attention. Thus, the aim of this study was, first, to compare the plasma and faecal amino acid profiles between control and FRE dogs in order to expand our current knowledge and characterize the disease. Our second aim was to search for correlations between amino acids and metabolites such as short-chain fatty acids (SCFAs) or other parameter indicators of gut health, including body condition score (BCS) or indices such as canine inflammatory bowel disease activity index (CIBDAI), to evaluate whether the amino acid profile could serve as an indicator of disease severity. ## 2.1. Animals and Sample Collection All procedures and protocols were approved by the Animal Research Committee of the Veterinary Medicine Teaching Hospital, Complutense University of Madrid (reference number $\frac{11}{2021}$). The owners of all the patients accepted their participation in the study through informed consent. The criteria inclusion for healthy dogs ($$n = 6$$) were a normal physical examination, blood test, and the absence of any clinical signs, including digestive signs, for at least four months. Asymptomatic dogs with chronic diseases were excluded from the study. Only sick dogs ($$n = 9$$) that had been suffering from digestive clinical signs (weight loss, anorexia, hiporexia, vomiting, or diarrhoea) for at least three weeks were to be enrolled in the study. Moreover, they had to respond to an elimination diet based on novel protein or hydrolysed protein after one month of administration. Based on the successful response to the diet, they were diagnosed as dogs with FRE. No dog included in the study suffered from hypoproteinemia. Data including sex, age, breed, sexual status, body weight, and BCS were collected from every patient (Table 1). Breeds of dogs with FRE were three mongrel dogs, one of each breed of Labrador Retriever, Cocker Spaniel, Miniature Schnauzer, Maltese, short-haired Dachshund, and Chihuahua. Breeds of healthy dogs included five mongrel dogs and one Gordon Setter. Information about specific digestive clinical signs was measured based on the CIBDAI, as previously described [15]. Blood samples (2 mL) were obtained following the regular procedure by jugular or cephalic venipuncture and collected in heparine tubes in fasted dogs for a period of at least 8 h. Plasma was obtained after centrifugation and stored at −80 °C. Faecal samples were collected by the owners after defecation (the same morning under fasting conditions) and brought to the clinic in less than three hours where they were stored at −20 °C. Blood and faecal samples of dogs with FRE were collected before starting the elimination diet. The commercial diets consumed by the dogs before the dietary treatment consisted of cereals, animal proteins, and vegetable/animal fats (averaged % according to manufacturer’s composition: humidity, 9.5 ± 0.0; crude protein, 26.8 ± 3.4; crude fat, 11.7 ± 4.4; ash, 5.9 ± 1.7; crude fibre, 1.8 ± 0.5; soluble fibre, 6.0 ± 0.7; nitrogen-free extractives, 38.3 ± 12.5; Ca, 0.9 ± 0.1; P, 0.7 ± 0.1; C18:2, 2.9 ± 1.1; ∑n-6, 2.7 ± 1.2; ∑n-3, 0.7 ± 0.1; mg/kg vitamin E: 619.2 ± 245.5; metabolic energy/1000 g: 3332.3 ± 645.2). ## 2.2. Concentration of Free Amino Acids in Plasma Samples Plasma-free amino acids were extracted as described elsewhere [16]. Essentially, plasma samples (100 µL) were mixed with 500 µL of a mixture of acetronile:methanol:acetone. After centrifugation, the supernatant was removed, leaving the dry residue. The supernatant was evaporated in N2 stream, redissolved in 500 µL water (MilliQ) and stored at −20 °C until analysis. The plasma amino acids and their standards were then derivatised with OPA (o-phtalaldehyde) as described by Jones et al. [ 17]. Samples were derivatised in an HPLC (Hewlet-Packard 1100 Agilent Technologies GmbH, Walbronn, Germany) equipped with a fluorescent detector, a phase reverse column Porshell HPH-C18 (4.6 × 100 mm, 2.7 µm, Agilent Technologies, Walbronn, Germany), and a pre-column HPH-C18 (Infinitylab Porshell 120, 3.0 mm, UHPLC, Agilent Technologies, Germany). Two mobile phases were used: phase A, a dilution of 10 mM Na2HPO4, 10 mM Na2B4O7 pH 8.2, and 0.5 mM NaN3; and phase B, a mixture of acetonitrile:methanol:water. The detector was adjusted at 340 nm for excitation and 450 nm for emission. The determination of amino acids was made by comparing their retention times with those of a standard sample of nineteen amino acids: aspartic acid (Asp), glutamic acid (Glu), serine (Ser), alanine (Ala), arginine (Arg), cystine (Cys-Cys), histidine (His), glycine (Gly), leucine (Leu), isoleucine (Ile), lysine (Lys), methionine (Met), threonine (Thr), phenylalanine (Phe), tyrosine (Tyr), and valine (Val) (1 nm/µL in 0.1 M HCl, Agilent Technologies), along with a dilution of asparagine (Asn), glutamine (Gln) (0.01 M HCl), and tryptophan (Trp) (0.1 M HCl, Agilent Technologies). ## 2.3. Concentration of Amino Acids in Faecal Samples by Acid Hydrolysis Lyophilized samples (50–80 mg) (Lyoquest, Telstar, Tarrasa, Spain) were placed in screw-capped glass tubes and hydrolysed with 15 mL of 6 M HCl. These tubes were then flushed with N2 and heated to 110 °C for 22 h. After cooling at room temperature, samples were filtered through filter paper to a beaker, and the pH was adjusted to 5.6 by the addition of NaOH solution (phmeter Crison Basic 20+). The solution was placed in a 100 mL volumetric flask and levelled up to that volume. Then, 20 mL were collected with a syringe and filtered by Sep-pak silica cartridge. Subsequently, 2 mL of the sample extract was isolated in a vial and stored at −20 °C. Protein hydrolysates and an amino acid calibration mixture were derivatised by o-phtaldialdehyde. Finally, an analysis of these samples was properly carried out by HPLC under the same conditions previously described by plasma samples analysis. ## 2.4. Statistical Analysis For the analysis of variance, data were analysed following a completely randomised design using the general linear model (GLM) procedure contained in SAS (version 9; SAS Inst. Inc., Cary, NC, USA) following the model: Yij = µ + Ti + ξij (where Y is the data observed of the dog j of the status i, µ is the average, T is the dog status ($i = 1$, 2), and ξ is the residual error). Data were presented as the mean of each group and the standard deviation of the mean (SD) together with significance levels (p values). Differences were considered significant at $p \leq 0.05.$ Pearson correlation among different amino acids and condition indices or other components of plasma/faeces, such as humidity, fat, and α-tocopherol (determined in a previous paper [9]), were carried out using the Statgraphics-19 program. The linear adjustments between amino acids and faecal characteristics or SCFAs (analysed in a previous study) [9] were quantified by Statgraphics-19. ## 3. Results The proportion of plasma amino acids is shown in Figure 1. FRE dogs had lower proportions of Asn ($$p \leq 0.034$$), His ($$p \leq 0.009$$), Gly ($$p \leq 0.005$$), Cys-Cys ($$p \leq 0.028$$), Leu ($$p \leq 0.017$$), and ratio branched-chain amino acids/aromatic amino acids (BCAA/AAA) ($$p \leq 0.018$$) when compared to control dogs. However, FRE dogs had a greater proportion of Phe ($$p \leq 0.013$$). Correlations between plasma-free amino acids and plasma fat content, α-tocopherol concentrations, BCS, and CIBDAI indices are presented in Table 2. Total plasma fat content correlated negatively with Trp (r = −0.58, $$p \leq 0.014$$) and Phe (r = −0.50, $$p \leq 0.040$$). Next, α-Tocopherol (as an indicator of the oxidative status) correlated positively with Asn ($r = 0.56$, $$p \leq 0.020$$), Gly ($r = 0.51$, $$p \leq 0.035$$), Arg ($r = 0.65$, $$p \leq 0.004$$), and Cys-Cys ($r = 0.60$, $$p \leq 0.024$$). On the contrary, Ala correlated negatively with α-tocopherol (r = −0.49, $$p \leq 0.045$$). In addition, BCS correlated positively with Leu ($r = 0.48$; $p \leq 0.05$). Finally, CIBDAI correlated positively with Phe ($r = 0.53$, $$p \leq 0.027$$), whereas it correlated negatively with Leu (r = −0.69, $$p \leq 0.002$$), Lys (r = −0.59, $$p \leq 0.012$$), BCAA (r = −0.49, $$p \leq 0.043$$), and BCAA/AAA ratio (r = −0.67, $$p \leq 0.003$$). Proportions of faecal amino acids are shown in Figure 2. Faecal Cys-Cys proportion was greater ($$p \leq 0.005$$), whereas Phe was lesser ($$p \leq 0.032$$) in sick dogs compared to the control. The other faecal amino acids were not statistically affected. Correlations between faecal amino acids and faecal parameters (fat and humidity), BCS, and CIBDAI indices are shown in Table 3. Faecal fat percentage was negatively correlated with Phe (r = −0.62, $$p \leq 0.030$$), Lys (r = −0.65, $$p \leq 0.022$$), and Leu (r = −0.57, $$p \leq 0.050$$). Faecal humidity percentage correlated positively with Gly ($r = 0.59$, $$p \leq 0.045$$) and negatively with Leu (r = −0.66, $$p \leq 0.018$$). No correlation was found with BCS or CIBDAI. The correlation between faecal amino acids and SCFAs was also investigated (Table 4). The concentration of SCFAs between FRE and healthy dogs was quantified in a previous study [9]. Total SCFAs correlated positively with Phe ($r = 0.80$, $$p \leq 0.002$$), Ile ($r = 0.75$, $$p \leq 0.005$$), and Leu ($r = 0.92$, $$p \leq 0.0001$$); whereas a negative correlation was observed with Gly (r = −0.85, $$p \leq 0.0005$$) and Cys-Cys (r = −0.61, $$p \leq 0.035$$). The SCFA that presented the highest number of correlations was butyric acid (C4), which correlated positively with amino acids Val ($r = 0.67$, $$p \leq 0.017$$), Met ($r = 0.68$, $$p \leq 0.015$$), Ile ($r = 0.77$, $$p \leq 0.003$$), and Leu ($r = 0.80$, $$p \leq 0.001$$), while negatively with Gly (r = −0.78, $$p \leq 0.002$$). Valeric acid (C5) and isovaleric acid (IC5) also presented a high number of correlations. Isovaleric acid correlated negatively with Gly (r = −0.59, $$p \leq 0.042$$), Tyr (r = −0.59, $$p \leq 0.042$$), and Cys-Cys (r = −0.60, $$p \leq 0.039$$), and C5 correlated positively with Leu ($r = 0.59$, $$p \leq 0.043$$) and negatively with Gly (r = −0.61, $$p \leq 0.034$$). However, the C2, C3, and IC4 presented a lower number of correlations with faecal amino acids. Thus, C2 correlated negatively with Ser (r = −0.58, $$p \leq 0.049$$), and IC4 correlated positively with His ($r = 0.64$, $$p \leq 0.026$$). However, C3 correlated positively with Phe ($r = 0.62$, $$p \leq 0.030$$), and a tendency was observed with Leu ($r = 0.56$, $$p \leq 0.060$$). Finally, linear adjustments were observed between faecal characteristics (faecal fat or humidity), faecal Leu (R2 = 0.32), Phe (R2= 0.41), and Gly (R2= 0.42) (Figure 3a–c, respectively). These faecal amino acids also presented high linear responses with SCFAs. Thus, more than $70\%$ of the variation of these amino acids in faeces was linearly explained by the concentration of SCFAs: R2 = 0.84; R2 = 0.70; and R2 = 0.70 for Leu, Phe, and Gly, respectively (Figure 3d–f). ## 4. Discussion Academic literature concerning the amino acid profile in dogs with gastrointestinal disorders is currently quite limited. Some research has been actually carried out in dogs with IRE or PLE [10,11,12,13]. However, current information on the amino acid profile in FRE dogs is not yet available. In the present study, FRE dogs had lower proportions of Asn, His, Gly, Cys-Cys, Leu, and ratio BCAA/AAA in plasma. Studies carried out in human medicine have also shown alterations in the amino acid profile of patients suffering from IBD [5,7], and the benefits of supplementing some amino acids on the reduction of symptomatology have been described [7,18,19]. Also in human studies, some authors have suggested the utility of amino acids like His as monitoring tools for predicting the risk of relapse in patients with ulcerative colitis (UC) [20]. In addition, prior research points to Gly and Cys as important amino acids for the maintenance of oxidative status linked to the inflammatory process, as they are part of antioxidant enzymes such as glutathione [21]. In this study, a positive correlation was found between these amino acids and vitamin E, which is one of the most important antioxidants that participate to a great extent in cell oxidative control in connection with other antioxidant systems to ensure the homeostasis of the individual [16]. Finding lower proportions of Gly and Cys-Cys in FRE dogs could be due in part to the higher use of these amino acids to synthesize glutathione and control the augmented reactive oxygen species (ROS) production that takes place in the inflammatory process. Moreover, these amino acids, together with His, Asn, and Leu, regulate intestinal inflammation, downregulating the production of proinflammatory cytokines [4]. In contrast, Cys, Gly, Asn, and Leu are also responsible for maintaining the normal functioning of the intestinal epithelial barrier by enhancing tight junction proteins [4]. Therefore, the lack of adequate long-term levels would exacerbate the inflammatory process and aggravate the integrity of the mucosal barrier. This lack would thus lead to bacterial adhesion and alteration of transporters responsible for the absorption of nutrients which, in turn, would result in nutritional deficits [22]. It is worth emphasizing that, in the present study, plasma Phe proportion was the only one increased in dogs with FRE. A recent study by Walker et al. [ 13] also found greater Phe serum concentration in dogs with CEs. It has also been reported that both inflammation and infection often lead to increased levels of Phe in human patients [23] since cytokines induce a strong metabolic disruption, muscle tissue breakdown, and a catabolic state. This state is associated with a higher release and increased Phe plasma levels in demand of the high metabolic rate [24], with Phe being a good indicator of body protein breakdown [25,26]. It is interesting, therefore, to observe that Phe correlated positively with CIBDAI in the present study, indicating a severe state of the disease based on weight and muscle loss. Moreover, the BCAA/AAA ratio decreased in FRE in comparison to that of healthy dogs. Phe is considered an AAA that is converted into Tyr. Branched-chain amino acids (BCAA), including Leu, Ile, and Val, are responsible for regulating the metabolism of glucose, lipid and protein synthesis, intestinal health, and immunity. Thus, BCAAs represent the major nitrogen source for the synthesis of Ala, Gln, and Glu [27] which are essential components for rapidly dividing cells such as enterocytes and immune cells [28]. Other authors reported lower BCAA/AAA in gastrointestinal or hepatic diseases [26] in association with the malnutrition process or with increased protein catabolism [29]. This result was confirmed in the present study by the negative correlation between AAA (Phe) and plasma fat. Some studies carried out in humans found that plasma AAA were higher in obese patients and were positively correlated to adiposity [30,31,32]. In addition, in the present study, Leu, Lys, BCAA, and the BCAA/AAA ratio correlated negatively with CIBDAI. Leu and Lys are essential amino acids, having a significant role in protein anabolism. The amino acid Leu [33] has especially been considered the major regulator of muscle protein synthesis in neonates [34]. Moreover, it has been confirmed that, among the BCAA, the response of muscle protein synthesis is unique to Leu, whereas Val and Ile failed to stimulate protein synthesis activation [35]. This finding is in line with the present results since one of the clinical signs evaluated by the CIBDAI index is weight loss that is associated with muscle loss, a typical sign of CE. Therefore, dogs with greater weight loss would have lower levels of these essential amino acids in plasma and, consequently, a greater CIBDAI classification. According to the results of the present study, Leu was the amino acid that showed the highest correlation, together with BCAA/AAA and the illness state, followed by Lys and Phe. These might then represent potential novel biomarkers for FRE. Commercially available diets containing hydrolysed protein formulated for dogs with CEs do not specify, in most cases, any amino acid profile. More information on the amino acid profile of dogs with FRE could lead to more precise and specific amino acid formulation in dietary interventions, better response to diet, and the recovery of the animal. Contrary to what was observed in plasma samples, FRE dogs had a lower proportion of Phe in the stool, which might indicate a greater metabolic use of this compound. However, faecal Cys-Cys was high in sick dogs which could be a consequence of an increase of Cys metabolism at this level as reported previously in IBD patients, possibly due to perturbed gut microbiota [36]. The proportions of amino acids were lower in faeces than in plasma as expected, except for the amino acids Asp and Glu. It has been reported that, during acid hydrolysis, the amino acids Asn and Gln are completely converted to Asp and Glu, respectively, while Trp is destroyed [37]. The correlations observed between faecal amino acids and faeces characteristics confirm again the importance of Phe and Leu as possible indicators of intestinal disease severity. Faecal Phe reached the highest negative correlation with the fat proportion in the stool, whereas faecal Leu was negatively correlated with proportions of fat and humidity. Balasubramanian et al. [ 38] also found lesser levels of Leu, as part of BCAA (Ile, Leu, Val), in the colonic mucosa of IBD patients compared with healthy subjects and considered these as potential biomarkers. Other amino acids involved in the endogenous antioxidant capacity that could be associated with illness, such as Gly, showed a positive correlation with stool humidity in the present study. Bjerrum et al. [ 39] also found increased levels of Gly in the faecal samples of IBD individuals compared to healthy controls, since this amino acid plays a key role in oxidative homeostasis and the regulation of inflammation [40]. It seems that gut alterations could induce a higher proportion of amino acids involved in oxidative functions such as Cys or Gly, since Cys-Cys was also greater in the stool of FRE dogs. However, no correlations were observed between Cys-Cys and faecal characteristics, although it should be pointed out that both presented a negative correlation with SCFAs. It has been reported the importance of SCFAs for keeping intestinal health and their levels are reduced in the faeces of adults suffering from IBD [41] or other CEs [9] with numerous studies suggesting that they play an important role in the treatment of inflammation-related diseases [9,41]. Although SCFA production comes mainly from the fermentation of carbohydrates, bacterial fermentation of protein sources serves as well for their obtention [42,43]. It has been reported that protein fermentation by intestinal bacteria in humans could account for 17 % of SCFAs found in the caecum and $38\%$ of SCFAs produced in the rectum [42]. The ratio of available carbohydrates to protein determines substrate utilization by the gut microbiota. Therefore, when energy is scarce, proteins are catabolized by bacteria to produce amino acid-derived end products [43]. In the present research, the amino acid that presented the highest positive correlation with total SCFA proportion was Leu, which was mainly positively related to faecal C4 proportion, followed by Phe, and Ile. Leu and Ile, as BCAA, play an important role in gut health by promoting intestinal development, nutrient transporters, and immune-related function [27]. To the best of our knowledge, there is no previous information on the correlation between faecal SCFAs and faecal amino acids in dogs with CE. According to our results, the higher proportion of faecal Leu and Phe, the higher the faecal SCFAs, which would be associated with better gut health. These results confirm again their importance as indicators of disease severity and faecal characteristics. ## 5. Conclusions Our results show that dogs with FRE had different plasma and faecal amino acid profiles than control dogs. The high correlation observed between plasma Leu and Phe with CIBDAI suggests that they could be used as disease biomarkers. Furthermore, statistically significant correlations observed for Leu, Phe, Gly, and Cys-Cys with SCFAs might indicate gut microbiota functionality, as well as homeostasis disruption. Consequently, these amino acids might have a role to play in food-responsive enteropathy. ## References 1. Kamei Y., Hatazawa Y., Uchitomi R., Yoshimura R., Miura S.. Regulation of skeletal muscle function by amino acids. Nutrients (2020) 12. DOI: 10.3390/nu12010261 2. Tolbert M.K., Murphy M., Gaylord L., Witzel-Rollins A.. Dietary management of chronic enteropathy in dogs. J. Small Anim. Pract. (2022) 63 425-434. DOI: 10.1111/jsap.13471 3. Aggarwal R., Bains K.. Protein, lysine and vitamin D: Critical role in muscle and bone health. Crit. Rev. Food Sci. Nutr. (2022) 62 2548-2559. DOI: 10.1080/10408398.2020.1855101 4. He F., Wu C., Li P., Li N., Zhang D., Zhu Q., Ren W., Peng Y.. Functions and signaling pathways of amino acids in intestinal inflammation. BioMed Res. Int. (2018) 2018 9171905. DOI: 10.1155/2018/9171905 5. Bauset C., Gisbert-Ferrándiz L., Cosín-Roger J.. Metabolomics as a promising resource identifying potential biomarkers for inflammatory bowel disease. J. Clin. Med. (2021) 10. DOI: 10.3390/jcm10040622 6. Hisamatsu T., Okamoto S., Hashimoto M., Muramatsu T., Andou A., Uo M., Kitazume M.T., Matsuoka K., Yajima T., Inoue N.. Novel, objective, multivariate biomarkers composed of plasma amino acid profiles for the diagnosis and assessment of inflammatory bowel disease. PLoS ONE (2012) 7. DOI: 10.1371/journal.pone.0031131 7. Liu Y., Wang X., Hu C.A.. Therapeutic potential of amino acids in inflammatory bowel disease. Nutrients (2017) 9. DOI: 10.3390/nu9090920 8. Hernandez J., Dandrieux J.R.S.. From bench top to clinics. Vet. Clin. North Am. Small Anim. Pract. (2021) 51 137-153. DOI: 10.1016/j.cvsm.2020.09.008 9. Higueras C., Rey A.I., Escudero R., Díaz-Regañón D., Rodríguez-Franco F., García-Sancho M., Agulla B., Sainz A.. Short-chain and total fatty acid profile of faeces or plasma as predictors of food-responsive enteropathy in dogs: A preliminary study. Animals (2021) 12. DOI: 10.3390/ani12010089 10. Kathrani A., Allenspach K., Fascetti A.J., Larsen J.A., Hall E.J.. Alterations in serum amino acid concentrations in dogs with protein-losing enteropathy. J. Vet. Intern. Med. (2018) 32 1026-1032. DOI: 10.1111/jvim.15116 11. Tamura Y., Ohta H., Kagawa Y., Osuga T., Morishita K., Sasaki N., Takiguchi M.. Plasma amino acid profiles in dogs with inflammatory bowel disease. J. Vet. Intern. Med. (2019) 33 1602-1607. DOI: 10.1111/jvim.15525 12. Benvenuti E., Pierini A., Gori E., Bartoli F., Erba P., Ruggiero P., Marchetti V.. Serum amino acid profile in 51 dogs with immunosuppressant-responsive enteropathy (IRE): A pilot study on clinical aspects and outcomes. BMC Vet. Res. (2020) 16. DOI: 10.1186/s12917-020-02334-2 13. Walker H.K., Boag A.M., Ottka C., Lohi H., Handel I., Gow A.G., Mellanby R.. Serum metabolomic profiles in dogs with chronic enteropathy. J. Vet. Intern Med. (2022) 36 1752-1759. DOI: 10.1111/jvim.16419 14. Allenspach K., Culverwell C., Chan D.. Long-term outcome in dogs with chronic enteropathies: 203 cases. Vet. Rec. (2016) 178 368. DOI: 10.1136/vr.103557 15. Jergens A.E., Schreiner C.A., Frank D.E., Niyo Y., Ahrens F.E., Eckersall P.D., Benson T., Evans R.. A scoring index for disease activity in canine inflammatory bowel disease. J. Vet. Intern. Med. (2003) 17 291-297. DOI: 10.1111/j.1939-1676.2003.tb02450.x 16. Rey A.I., de-Cara A., Calvo L., Puig P., Hechavarría T.. Changes in plasma fatty acids, free amino acids, antioxidant defense, and physiological stress by oleuropein supplementation in pigs prior to slaughter. Antioxidants (2020) 9. DOI: 10.3390/antiox9010056 17. Jones B.N., Pääbo S., Stein S.. Amino acid analysis and enzymatic sequence determination of peptides by an improved o-phthaldialdehyde precolumn labeling procedure. J. Liq. Chromatogr. (1981) 4 565-586. DOI: 10.1080/01483918108059956 18. Kim C., Kovacs-Nolan J., Yang C., Archbold T., Fan M., Mine Y.. L-Tryptophan exhibits therapeutic function in a porcine model of dextran sodium sulfate (DSS)-induced colitis. J. Nutr. Biochem. (2010) 21 468-475. DOI: 10.1016/j.jnutbio.2009.01.019 19. Farré R., Fiorani M., Abdu Rahiman S., Matteoli G.. Intestinal permeability, inflammation and the role of nutrients. Nutrients (2020) 12. DOI: 10.3390/nu12041185 20. Hisamatsu T., Ono N., Imaizumi A., Mori M., Suzuki H., Uo M., Hashimoto M., Naganuma M., Matsuoka K., Mizuno S.. Decreased plasma histidine level predicts risk of relapse in patients with ulcerative colitis in remission. PLoS ONE (2015) 10. DOI: 10.1371/journal.pone.0140716 21. Kelly B., Pearce E.L.. Amino assets: How amino acids support immunity. Cell Metab. (2020) 32 154-175. DOI: 10.1016/j.cmet.2020.06.010 22. Giaretta P.R., Rech R.R., Guard B.C., Blake A.B., Blick A.K., Steiner J.M., Lidbury J., Cook A., Hanifeh M., Spillmann T.. Comparison of intestinal expression of the apical sodium-dependent bile acid transporter between dogs with and without chronic inflammatory enteropathy. J. Vet. Intern. Med. (2018) 32 1918-1926. DOI: 10.1111/jvim.15332 23. Huang S.S., Lin J.Y., Chen W.S., Liu M.H., Cheng C.W., Cheng M.L., Wang C.. Phenylalanine- and leucine-defined metabolic types identify high mortality risk in patients with severe infection. Int. J. Infect. Dis. (2019) 85 143-149. DOI: 10.1016/j.ijid.2019.05.030 24. Mori E., Suzuki H., Hasebe M., Kobayashi K.. Alterations in levels of plasma phenylalanine and its catabolism in the liver of stressed rats. Metabolism (1992) 41 807-813. DOI: 10.1016/0026-0495(92)90159-8 25. Matthews D.E.. An overview of phenylalanine and tyrosine kinetics in humans. J. Nutr. (2007) 137 1549S-1555S. DOI: 10.1093/jn/137.6.1549S 26. Dam G., Sørensen M., Buhl M., Sandahl T.D., Møller N., Ott P., Vilstrup H.. Muscle metabolism and whole blood amino acid profile in patients with liver disease. Scand. J. Clin. Lab. Investig. (2015) 75 674-680. PMID: 26243157 27. Nie C., He T., Zhang W., Zhang G., Ma X.. Branched Chain Amino Acids: Beyond nutrition metabolism. Int. J. Mol. Sci. (2018) 19. DOI: 10.3390/ijms19040954 28. Holeček M.. Relation between glutamine, branched-chain amino acids, and protein metabolism. Nutrition (2002) 18 130-133. DOI: 10.1016/S0899-9007(01)00767-5 29. Eghtesad S., Poustchi H., Malekzadeh R.. Malnutrition in liver cirrhosis: The influence of protein and sodium. Middle East J. Dig. Dis. (2013) 5 65-75. PMID: 24829672 30. Mikkola T.M., Salonen M.K., Kajantie E., Kautiainen H., Eriksson J.G.. Associations of fat and lean body mass with circulating amino acids in older men and women. J. Gerontol. Ser. A (2020) 75 885-891. DOI: 10.1093/gerona/glz126 31. Lischka J., Schanzer A., Baumgartner M., de Gier C., Greber-Platzer S., Zeyda M.. Tryptophan metabolism is associated with BMI and adipose tissue mass and linked to metabolic disease in pediatric obesity. Nutrients (2022) 14. DOI: 10.3390/nu14020286 32. Short K.R., Chadwick J.Q., Teague A.M., Tullier M.A., Wolbert L., Coleman C., Copeland K.. Effect of obesity and exercise training on plasma amino acids and amino metabolites in american indian adolescents. J. Clin. Endocrinol. Metab. (2019) 104 3249-3261. DOI: 10.1210/jc.2018-02698 33. Makanae Y., Fujita S.. Role of exercise and nutrition in the prevention of sarcopenia. J. Nutr. Sci. Vitaminol. (2015) 61 S125-S127. DOI: 10.3177/jnsv.61.S125 34. Columbus D.A., Fiorotto M.L., Davis T.A.. Leucine is a major regulator of muscle protein synthesis in neonates. Amino Acids (2015) 47 259-270. DOI: 10.1007/s00726-014-1866-0 35. Escobar J., Frank J.W., Suryawan A., Nguyen H.v., Kimball S.R., Jefferson L.S., Davis T.. Physiological rise in plasma leucine stimulates muscle protein synthesis in neonatal pigs by enhancing translation initiation factor activation. Am. J. Physiol. Endocrinol. Metab. (2005) 288 E914-E921. DOI: 10.1152/ajpendo.00510.2004 36. Walker A., Schmitt-Kopplin P.. The role of fecal sulfur metabolome in inflammatory bowel diseases. Int. J. Med. Microbiol. (2021) 311 151513. DOI: 10.1016/j.ijmm.2021.151513 37. Ovalles J.F., Rosenzweig Levy P., Celis M.T.. Estado del arte de la hidrólisis de proteínas y sinopsis sobre la determinación del perfil de aminoácidos. Una educación universitaria de calidad. Sello Editorial Publicaciones del Vicerrectorado Académico—ULA Mérida (2015) 879-914 38. Balasubramanian K., Kumar S., Singh R.R., Sharma U., Ahuja V., Makharia G.K., Jagannathan N.. Metabolism of the colonic mucosa in patients with inflammatory bowel diseases: An in vitro proton magnetic resonance spectroscopy study. Magn. Reson. Imaging (2009) 27 79-86. DOI: 10.1016/j.mri.2008.05.014 39. Bjerrum J.T., Wang Y., Hao F., Coskun M., Ludwig C., Günther U., Nielsen O.. Metabonomics of human fecal extracts characterize ulcerative colitis, Crohn’s disease and healthy individuals. Metabolomics (2015) 11 122-133. DOI: 10.1007/s11306-014-0677-3 40. Dawiskiba T.. Serum and urine metabolomic fingerprinting in diagnostics of inflammatory bowel diseases. World J. Gastroenterol. (2014) 20 163. DOI: 10.3748/wjg.v20.i1.163 41. Zhang Z., Zhang H., Chen T., Shi L., Wang D., Tang D.. Regulatory role of short-chain fatty acids in inflammatory bowel disease. Cell Commun. Signal. (2022) 20 64. DOI: 10.1186/s12964-022-00869-5 42. Macfarlane G.T., Gibson G.R., Beatty E., Cummings J.H.. Estimation of short-chain fatty acid production from protein by human intestinal bacteria based on branched-chain fatty acid measurements. FEMS Microbiol. Ecol. (1992) 10 81-88. DOI: 10.1111/j.1574-6941.1992.tb00002.x 43. Portune K.J., Beaumont M., Davila A.M., Tomé D., Blachier F., Sanz Y.. Gut microbiota role in dietary protein metabolism and health-related outcomes: The two sides of the coin. Trends Food Sci. Technol. (2016) 57 213-232. DOI: 10.1016/j.tifs.2016.08.011
--- title: 'Novel Hybrid Indole-Based Caffeic Acid Amide Derivatives as Potent Free Radical Scavenging Agents: Rational Design, Synthesis, Spectroscopic Characterization, In Silico and In Vitro Investigations' authors: - Ahmed Elkamhawy - Na Kyoung Oh - Noha A. Gouda - Magda H. Abdellattif - Saud O. Alshammari - Mohammed A. S. Abourehab - Qamar A. Alshammari - Amany Belal - Minkyoung Kim - Ahmed A. Al-Karmalawy - Kyeong Lee journal: Metabolites year: 2023 pmcid: PMC9966950 doi: 10.3390/metabo13020141 license: CC BY 4.0 --- # Novel Hybrid Indole-Based Caffeic Acid Amide Derivatives as Potent Free Radical Scavenging Agents: Rational Design, Synthesis, Spectroscopic Characterization, In Silico and In Vitro Investigations ## Abstract Antioxidant small molecules can prevent or delay the oxidative damage caused by free radicals. Herein, a structure-based hybridization of two natural antioxidants (caffeic acid and melatonin) afforded a novel hybrid series of indole-based amide analogues which was synthesized with potential antioxidant properties. A multiple-step scheme of in vitro radical scavenging assays was carried out to evaluate the antioxidant activity of the synthesized compounds. The results of the DPPH assay demonstrated that the indole-based caffeic acid amides are more active free radical scavenging agents than their benzamide analogues. Compared to Trolox, a water-soluble analogue of vitamin E, compounds 3a, 3f, 3h, 3j, and 3m were found to have excellent DPPH radical scavenging activities with IC50 values of 95.81 ± 1.01, 136.8 ± 1.04, 86.77 ± 1.03, 50.98 ± 1.05, and 67.64 ± 1.02 µM. Three compounds out of five (3f, 3j, and 3m) showed a higher capacity to neutralize the radical cation ABTS•+ more than Trolox with IC50 values of 14.48 ± 0.68, 19.49 ± 0.54, and 14.92 ± 0.30 µM, respectively. Compound 3j presented the highest antioxidant activity with a FRAP value of 4774.37 ± 137.20 μM Trolox eq/mM sample. In a similar way to the FRAP assay, the best antioxidant activity against the peroxyl radicals was demonstrated by compound 3j (10,714.21 ± 817.76 μM Trolox eq/mM sample). Taken together, compound 3j was validated as a lead hybrid molecule that could be optimized to maximize its antioxidant potency for the treatment of oxidative stress-related diseases. ## 1. Introduction Since the majority of our biological activities only take place in the presence of oxygen, it is necessary for life. However, the oxidation reaction may lead to cell damage, leading to the degradation of various oxygen substrates, proteins, lipids, and DNA. This may cause numerous diseases including inflammation, obesity, diabetes, arthritis, etc. [ 1,2]. The oxygen paradox contributes to the generation of free radicals which possess a single electron on an oxygen or nitrogen atom, known as reactive oxygen species (ROS) or reactive nitrogen species (RNS) [3,4]. Free radicals are highly unstable. They could be valuable when they are involved in physiological functions or harmful if there is no balance between the defense systems and ROS/RNS, or when the organism is incapable of restricting the destruction triggered by the free radicals, which is known by the oxidative stress [5]. Antioxidants are compounds able to inhibit or postpone the oxidation process via neutralizing free radicals. Some synthetic antioxidants including butylated hydroxyanisole (BHA; I, Figure 1A) and butylated hydroxytoluene (BHT; II, Figure 1A) have recently been reported to be toxic to the environment and the human health [6,7]. Recently, numerous natural antioxidants’ effectiveness has been reported [8]. Antioxidants, whether they are produced naturally or artificially, reduce the severity of oxidative damage via scavenging ROS or stopping ROS-mediated chain reactions. As a result, finding naturally occurring compounds with antioxidant activity followed by hybridizing these natural antioxidant chemical scaffolds becomes a significant scientific issue with several socioeconomic interests. Despite being found in nature with many health advantages [9], the research being done on hybridizing essential antioxidant pharmacophores is still very important today [10,11,12,13,14]. Caffeic acid ((E)-3-(3,4-dihydroxyphenyl)prop-2-enoic acid, III, Figure 1B) is one of the main hydroxycinnamic acids possessing active antioxidant activity, as sketched in Figure 1B. It was reported that caffeic acid has potent free radical scavenging activities [15]. Melatonin (IV, Figure 1C), which is also a remarkable antioxidant natural compound [16,17], can scavenge oxygen free radicals, such as super-oxide radicals, hydroxyl radicals, and others. Along with its antioxidant and neuroprotective activities [18,19,20,21,22,23], melatonin was reported to have a therapeutic potential for inflammation [24], cancer [25], pain [26], cardiovascular disorders, etc. [ 27,28,29,30,31,32]. Thus, numerous derivatives of melatonin were reported with several biological activities [33,34,35,36,37,38,39,40]. The potent activity of melatonin as an ROS-scavenging agent including 1O2, O2•−, H2O2, hydroxyl radical (HO•), and peroxyl radical (ROO•) is due to the electron-rich aromatic indole chemical scaffold (Figure 1C), which enables indoleamine to serve as an electron donor, forming an indolyl cation [23,41,42,43]. In this context, various indole derivatives with promising activities against oxidative stress and monoamine oxidase B enzyme (MAO-B) were recently reported by our research group [44,45,46]. The potent antioxidant activity of both chemical scaffolds (indole and caffeic acid) as shown in Figure 1 encourages our team to design a hybrid scaffold that could have potential antioxidant power. Consequently, development of an efficient indole–caffeic pharmacophore could be enormously important. As illustrated in Figure 2, the methoxy group of melatonin was substituted with an amino group, followed by amide formation by reacting with various caffeic acid analogues to generate the desired amide derivatives (3a–m). The selection process of the functional groups in compounds 3a–m was inspired by their existence in neuroprotectant compounds which showed powerful activities against oxidative stress [47,48,49]. For this objective, a straightforward synthetic approach was used to synthesize the new hybrid indole–caffeic amide analogues 3a–m, and their therapeutic potential against ROS was preliminary assessed (Figure 2). ## 2.1. Chemical Reagents, Purification, and Instrumentation The general protocols utilized for the chemical synthesis, structure elucidation, and purity of the newly synthesized indole–caffeic acid hybrids are provided in the Supplementary File. ## 2.2. Synthesis of Indole–caffeic Amide Analogues 3a–m 5-Aminoindole (1, 0.1 g, 0.75 mmol), 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDCI, 0.21 g, 1.1 mmol), and hydroxybenzotriazole (HOBt, 0.14 g, 1.1 mmol) were mixed in the presence of N,N-diisopropylethylamine (DIPEA, 0.19 mL, 1.1 mmol) and acetonitrile solvent (5 mL). The appropriate carboxylic acid reagent (2, 0.75 mmol, 1 eq.) was then added. The reaction was carried out at room temperature (25 °C) for 2 h. The excess acetonitrile was evaporated. Work-up was performed using ethyl acetate (EA) and water. The organic solution was evaporated, dried, and purified via flash column chromatography (SiO2, n-hexane:EA = 10:1) to obtain the indole–caffeic amide derivatives in suitable yields (Table 1). ## 2.2.1. 4-Amino-3-hydroxy-N-(1H-indol-5-yl)benzamide (3a) Yellowish-white solid. M.p.: 182–183 °C. HPLC purity: 5.614 min, $96.71\%$. 1H NMR (400 MHz, DMSO-d6) δ 10.96 (s, 1H), 9.58 (s, 1H), 9.24 (s, 1H), 7.93 (s, 1H), 7.36–7.29 (m, 5H), 6.62 (d, $J = 8.00$ Hz, 1H), 6.37 (s, 1H), 5.09 (br, 2H). 13C NMR (100 MHz, DMSO-d6) δ 165.59, 143.47, 140.75, 133.08, 131.98, 127.80, 126.06, 123.17, 120.17, 116.54, 114.44, 112.88, 112.27, 111.23, 102.00. HRMS (ESI) m/z calculated for C15H14N3O2 [M+H]+: 268.1086, found: 268.1076. ## 2.2.2. 2-Bromo-N-(1H-indol-5-yl)-4,5-dimethoxybenzamide (3b) White solid. M.p.: 256–257 °C. HPLC purity: 11.235 min, $99.73\%$. 1H NMR (400 MHz, DMSO-d6) δ 11.02 (s, 1H), 10.09 (s, 1H), 7.99 (s, 1H), 7.33–7.31 (m, 3H), 7.21 (s, 1H), 7.14 (s, 1H), 6.60 (m, 1H), 3.82 (s, 6H). 13C NMR (100 MHz, DMSO-d6) δ 165.55, 150.40, 148.38, 133.33, 132.04, 131.51, 127.82, 126.35, 116.90, 115.60, 112.65, 111.59, 111.49, 110.08, 101.57, 55.57. HRMS (ESI) m/z calculated for C17H16BrN2O3 [M+H]+: 375.0344, found: 375.0331. ## 2.2.3. 4-Bromo-3,5-dihydroxy-N-(1H-indol-5-yl)benzamide (3c) Yellow solid. M.p.: 211–212 °C. HPLC purity: 7.414 min, $99.60\%$. 1H NMR (400 MHz, DMSO-d6) δ 11.01 (s, 1H), 10.29 (br, 2H), 10.03 (s, 1H), 7.95 (s, 1H), 7.35–7.32 (m, 3H), 6.93 (s, 2H), 6.40 (s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 165.57, 155.60, 136.38, 133.35, 131.41, 127.80, 126.32, 116.28, 112.36, 111.44, 106.41, 101.57, 101.37. HRMS (ESI) m/z calculated for C15H12BrN2O3 [M+H]+: 347.0031, found: 347.0018. ## 2.2.4. 3-Amino-4-hydroxy-N-(1H-indol-5-yl)benzamide (3d) Orange solid. M.p.: 182–183 °C. HPLC purity: 12.966 min, $99.87\%$. 1H NMR (400 MHz, DMSO-d6) δ 11.63 (br, 1H), 11.04 (s, 1H), 10.15 (s, 1H), 8.57 (s, 1H), 8.17 (d, $J = 8.00$ Hz, 1H), 7.95 (s, 1H), 7.36–7.32 (m, 3H), 7.24 (d, $J = 8.00$ Hz, 1H), 6.41 (s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 163.15, 154.68, 136.93, 134.60, 133.48, 131.08, 127.82, 126.45, 126.37, 125.25, 119.31, 116.54, 112.78, 111.47, 101.59. HRMS (ESI) m/z calculated for C15H14N3O2 [M+H]+: 268.1086, found: 268.1094. ## 2.2.5. (E)-3-(4-Hydroxyphenyl)-N-(1H-indol-5-yl)acrylamide (3e) White solid. M.p.: 153–154 °C. HPLC purity: 8.983 min, $96.06\%$. 1H NMR (400 MHz, DMSO-d6) δ 10.98 (s, 1H), 9.85 (br, 2H), 7.99 (s, 1H), 7.47–7.43 (m, 3H), 7.33–7.27 (m, 3H), 6.82 (d, $J = 8.00$ Hz, 2H), 6.64 (d, $J = 16.00$ Hz, 1H), 6.38 (s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 163.98, 159.32, 139.72, 133.10, 131.92, 129.77, 127.98, 126.38, 126.30, 119.73, 116.22, 115.15, 111.65, 110.93, 101.57. HRMS (ESI) m/z calculated for C17H15N2O2 [M+H]+: 279.1134, found: 279.1126. ## 2.2.6. (E)-3-(4-Hydroxy-3-methoxyphenyl)-N-(1H-indol-5-yl)acrylamide (3f) Yellowish-white solid. M.p.: 177–178 °C. HPLC purity: 9.259 min, $97.06\%$. 1H NMR (400 MHz, DMSO-d6) δ 10.98 (s, 1H), 9.86 (s, 1H), 9.44 (s, 1H), 8.00 (s, 1H), 7.45 (d, $J = 12.00$ Hz, 1H), 7.33–7.30 (m, 3H), 7.18 (s, 1H), 7.05 (d, $J = 8.00$ Hz, 1H), 6.82 (d, $J = 8.00$ Hz, 1H), 6.67 (d, $J = 12.00$ Hz, 1H), 6.38 (s, 1H), 3.83 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 163.92, 148.82, 148.29, 140.02, 133.09, 131.95, 127.94, 126.88, 126.30, 122.11, 120.01, 116.17, 115.09, 111.65, 111.22, 110.93, 101.36, 55.93. HRMS (ESI) m/z calculated for C18H17N2O3 [M+H]+: 309.1239, found: 309.1234. ## 2.2.7. 4-Hydroxy-N-(1H-indol-5-yl)-3,5-dimethoxybenzamide (3g) White solid. M.p.: 198–199 °C. HPLC purity: 7.354 min, $97.64\%$. 1H NMR (400 MHz, DMSO-d6) δ 11.01 (s, 1H), 9.82 (s, 1H), 8.93 (s, 1H), 7.89 (s, 1H), 7.35–7.31 (m, 5H), 6.40 (s, 1H), 3.85 (s, 6H). 13C NMR (100 MHz, DMSO-d6) δ 165.01, 147.86, 139.08, 133.39, 131.36, 127.84, 126.29, 125.32, 116.94, 113.03, 111.38, 105.83, 101.52, 56.55. HRMS (ESI) m/z calculated for C17H17N2O4 [M+H]+: 313.1188, found: 313.1178. ## 2.2.8. 3,4-Dihydroxy-N-(1H-indol-5-yl)benzamide (3h) White solid. M.p.: 145–146 °C. HPLC purity: 5.485 min, $99.53\%$. 1H NMR (400 MHz, DMSO-d6) δ 10.98 (s, 1H), 9.75 (s, 1H), 9.35 (br, 2H), 7.94 (s, 1H), 7.40–7.30 (m, 5H), 6.81 (d, $J = 8.00$ Hz, 1H), 5.38 (s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 165.31, 148.86, 145.29, 133.22, 131.71, 127.80, 126.94, 126.18, 119.78, 116.56, 115.81, 115.28, 112.46, 111.31, 101.50. HRMS (ESI) m/z calculated for C15H13N2O3 [M+H]+: 269.0926, found: 269.0913. ## 2.2.9. 4-Hydroxy-N-(1H-indol-5-yl)-2-methylbenzamide (3i) White solid. M.p.: 180–181 °C. HPLC purity: 7.529 min, $98.76\%$. 1H NMR (400 MHz, DMSO-d6) δ 10.97 (s, 1H), 9.82 (s, 1H), 9.65 (s, 1H), 7.97 (s, 1H), 7.34–7.29 (m, 4H), 6.66–6.64 (m, 2H), 6.37 (s, 1H), 2.34 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 167.78, 158.70, 138.10, 133.15, 131.96, 129.63, 128.97, 127.82, 126.20, 117.59, 115.79, 112.52, 111.56, 111.38, 101.50, 20.29. HRMS (ESI) m/z calculated for C16H15N2O2 [M+H]+: 267.1134, found: 267.1121. ## 2.2.10. (E)-3-(3,4-Dihydroxyphenyl)-N-(1H-indol-5-yl)acrylamide (3j) Yellow solid. M.p.: 171–172 °C. HPLC purity: 7.262 min, $99.59\%$. 1H NMR (400 MHz, DMSO-d6) δ 10.99 (s, 1H), 9.87 (s, 1H), 9.28 (br, 2H), 8.00 (s, 1H), 7.39–7.30 (m, 4H), 7.01 (s, 1H), 6.90 (d, $J = 8.00$ Hz, 1H), 6.78 (d, $J = 8.00$ Hz, 1H), 6.58 (d, $J = 12.00$ Hz, 1H), 6.38 (d, $J = 4.00$ Hz, 1H). 13C NMR (100 MHz, DMSO-d6) δ 163.95, 147.88, 146.01, 140.11, 133.08, 131.92, 127.93, 126.87, 126.28, 121.08, 119.51, 116.27, 115.13, 114.26, 111.64, 110.91, 101.56. HRMS (ESI) m/z calculated for C17H15N2O3 [M+H]+: 295.1083, found: 295.1070. ## 2.2.11. 4-Hydroxy-N-(1H-indol-5-yl)benzamide (3k) White solid. M.p.: 168–169 °C. HPLC purity: 6.908 min, $98.94\%$. 1H NMR (400 MHz, DMSO-d6) δ 10.99 (s, 1H), 10.01 (s, 1H), 9.81 (s, 1H), 7.95 (s, 1H), 7.87 (d, $J = 8.00$ Hz, 2H), 7.36–7.31 (m, 3H), 6.86 (d, $J = 8.00$ Hz, 2H), 6.39 (s, 1H). 13C NMR (100 MHz, DMSO-d6) δ 165.17, 160.61, 133.28, 131.62, 129.92, 127.83, 126.39, 126.19, 116.64, 115.23, 112.57, 111.35, 101.54. HRMS (ESI) m/z calculated for C15H13N2O2 [M+H]+: 253.0977, found: 253.0966. ## 2.2.12. 4-Hydroxy-N-(1H-indol-5-yl)-3-methoxybenzamide (3l) White solid. M.p.: 190–191 °C. HPLC purity: 7.301 min, $99.75\%$. 1H NMR (400 MHz, DMSO-d6) δ 11.01 (s, 1H), 9.83 (s, 1H), 9.60 (s, 1H), 7.93 (s, 1H), 7.56 (d, $J = 4.00$ Hz, 1H), 7.51 (dd, $J = 8.0$, 4.0 Hz, 1H), 7.35–7.31 (m, 3H), 6.87 (d, $J = 8.00$ Hz, 1H), 6.40 (s, 1H), 3.86 (s, 3H). 13C NMR (100 MHz, DMSO-d6) δ 165.10, 150.00, 147.58, 133.33, 131.51, 127.84, 126.64, 126.22, 121.56, 116.80, 115.22, 122.78, 112.03, 111.37, 101.53, 56.12. HRMS (ESI) m/z calculated for C16H15N2O3 [M+H]+: 283.1083, found: 283.1073. ## 2.2.13. (E)-3-(4-Hydroxy-3,5-dimethoxyphenyl)-N-(1H-indol-5-yl)acrylamide (3m) Yellow solid. M.p.: 199–200 °C. HPLC purity: 8.923 min, $99.88\%$. 1H NMR (400 MHz, DMSO-d6) δ 11.00 (s, 1H), 9.90 (s, 1H), 8.84 (s, 1H), 8.03 (s, 1H), 7.48 (d, $J = 16.00$ Hz, 1H), 7.35–7.30 (m, 3H), 6.92 (s, 2H), 6.71 (d, $J = 16.00$ Hz, 1H), 6.39 (s, 1H), 3.83 (s, 6H). 13C NMR (100 MHz, DMSO-d6) δ 163.84, 148.52, 140.30, 137.82, 133.09, 131.98, 127.94, 126.31, 125.74, 120.36, 115.04, 111.66, 110.82, 105.67, 101.56, 56.37. HRMS (ESI) m/z calculated for C19H19N2O4 [M+H]+: 339.1345, found: 339.1340. ## 2.3.1. 2,2-Diphenyl-1-picrylhydrazyl Radical-Scavenging Activity (DPPH Assay) Final concentrations of 20 mM of the tested compounds and 0.1 mM DMSO were prepared to determine the range of the inhibitory concentration 50 (EC50). A solution of Trolox (a water-soluble analogue of vitamin E, 1000 μM) was prepared in DMSO, from which 5 final concentrations were prepared including 5, 10, 20, 40, and 80 μM. DPPH (2,2-diphenyl-1-picryl-hydrazyl-hydrate) free radical assay was performed as reported [50]. Further details were provided in the Supplementary Materials. ## 2.3.2. 2,2’-Azino-bis(3-ethylbenzothiazoline-6-sulfonate (ABTS) Assay The assay was performed as reported [51], with minor modifications. For details, please refer to the Supplementary Materials. ## 2.3.3. Ferric Reducing Power (FRAP) Assay A stock solution of Trolox (3 mM in methanol) was made, and the following dilutions were prepared at the concentrations of 1500, 1000, 800, 400, 200, 100, and 50 μM. Samples were initially dissolved in DMSO to obtain a 40 mM concentration (depending on the molecular weight of each compound). Then, they were diluted to reach the concentration of 0.2 mM with methanol. The assay was performed as reported [52], with minor modifications. For details, please refer to the Supplementary Materials. ## 2.3.4. Oxygen Radical Absorbance Capacity (ORAC Assay) A stock solution of Trolox (2 mM in MeOH) was prepared, and the following dilutions were prepared: 1200, 900, 600, 500, 400, 300, 200, 100, and 50 μM. Samples were initially dissolved in DMSO at concentrations of 40 mM according to the provided molecular weights. Then, samples were diluted with methanol until reaching the concentration of 0.1 mM. The assay was carried out as reported [53], with modifications. Further details are provided in the Supplementary Materials. ## 3.1. Chemical Synthesis As sketched in Scheme 1, a series of indole-based benzamide and caffeic acid amide analogues 3a–m was synthesized. The amide formation reaction was accomplished via reacting 5-aminoindole [1] in acetonitrile solvent with a variety of commercially available benzoic or caffeic acid derivatives [2]. The coupling reagents EDCI and HOBt were used, in addition to DIPEA as an organic base. As shown in Table 1, a variety of indole-based benzamide and caffeic acid amide analogues possessing different chemical substituents were acquired in acceptable yields. ## 3.2. Structure Elucidation of the Newly Synthesized Amide Derivatives 3a–m The chemical structures of the newly synthesized indole-based benzamide and caffeic acid amide analogues (3a–m) were elucidated using different spectroscopic techniques. The purity of compounds 3a–m was found to be more than $96\%$. The 1H NMR spectra of all the synthesized analogues were characterized by two major singlet peaks with high chemical shifts (>9.00 ppm); the proton of the NH group of the indole scaffold and the proton of the amide group (CONH). As provided in the Supplementary File, the 1H NMR spectrum of the indole-based benzamide analogue 3a displayed the free para NH2 group at 5.09 ppm as a broad peak (br) representing the two protons of the amino group. In addition, the amide carbon (CO) appeared clearly in the 13C NMR spectrum of compound 3a at 165.59 ppm, which confirmed the formation of the amide group. Indeed, 13C NMR spectra of all the newly synthesized indole-based benzamide and caffeic acid amide analogues (3a–m) showed signals resonating around 163.00–168.00 ppm (CO group of the amide moiety). For compound 3b, the two methoxy groups were found at 3.82 and 55.57 ppm in the 1H and 13 C NMR, respectively. The 1H NMR chart of compound 3c was characterized by a singlet long aromatic peak at 6.93 ppm representing the two para phenyl protons of the 4-bromo-3,5-dihydroxy benzoyl moiety. In addition, the protons of the two hydroxyl moieties were detected as a singlet broad peak at 10.29 ppm. In its 13C NMR chart, compound 3c showed a long peak at 155.60 attributable to the two carbons carrying the two hydroxyl moieties. The 1H NMR spectrum of the benzamide derivative 3d was characterized presence of three singlet peaks with chemical shifts higher than 9.00 ppm (the indole NH, the amide NH, and the free phenolic OH groups). The first synthesized caffeic acid amide ((E)-3-(4-hydroxyphenyl)-N-(1H-indol-5-yl)acrylamide, 3e) showed these three protons in the range of 9.85–10.98 ppm. In the meantime, one vinylic proton was successfully detected with its characteristic trans J coupling constant of 16.00 Hz at 6.64 ppm. The carbon of the amide moiety appeared at 163.98 in the 13C NMR spectrum of the caffeic acid amide 3e, while the carbon holding the free phenolic OH group was detected at 159.32 ppm. Similarly, the 1H NMR spectrum confirmed the synthesis and the final chemical structure of the second caffeic acid amide derivative in this series (3f) by the presence of three singlet peaks in the range of 9.44–10.98 ppm representing the indole NH, amide NH, and phenolic OH groups, in addition to the three protons of the meta methoxy group (m-OCH3) in the aliphatic region (3.83 ppm). Its 13C NMR chart showed the amide carbon peak at 163.92 ppm, the two carbons bearing the free OH and the methoxy groups at 148.82 and 148.29 ppm, in addition to the aliphatic carbon of the methoxy group at 55.93 ppm. The 1H NMR spectrum of compound 3g was characterized by a long peak in the aliphatic region at 3.85 representing the six protons of the two methoxy groups. Meanwhile, its 13C NMR chart showed the amide carbon chemical shift at 165.01 ppm, a long peak at 147.86 ppm attributable to the two carbons that hold the two methoxy groups, and the characteristic peak of the carbon bearing the free OH at 139.08 ppm. Similarly, 3,4-dihydroxy-N-(1H-indol-5-yl)benzamide (3h) was characterized by the two common singlet peaks at 10.98 and 9.75 ppm representing the NH protons of the amide linkage and the indole ring. In addition, a broad singlet peak of 2H was found at 9.35 ppm, attributable to the two free OH phenolic moieties. The two carbons bearing these phenolic hydroxyl groups were detected in its 13C NMR spectrum at 148.82 and 145.20 ppm. The methyl group (CH3) of analogue 3i was represented by a singlet peak (3H) in the aliphatic region at 2.34 ppm and a peak at 20.29 ppm (13C NMR). In addition to the two common singlet peaks of the NH groups of the amide linker and the indole ring at 10.99 and 9.87 ppm, the third caffeic amide analogue 3j also showed a broad peak of 2H representing the protons of the two hydroxyl groups at 9.28 ppm and a doublet peak at 6.58 with a coupling constant of 12.00 Hz attributable to a vinylic proton. The amide CO group was represented at 163.95 ppm (13C NMR), while the two carbons bearing the two hydroxyl groups were represented by two peaks at 147.88 and 146.01 ppm. A vinylic carbon of compound 3j was also successfully detected at 140.11 ppm. In the 13C NMR spectrum of compound 3k, the amide carbon and the carbon atom holding the free OH group were displayed above 160.00 ppm. In compound 3l, the amide carbon appeared at 165.10 ppm, while the two carbons bearing the free OH and the methoxy groups were displayed at 150.00 and 147.58 ppm. The chemical shifts of the methoxy group in 3l were represented by a singlet peak in the aliphatic region at 56.12 ppm (13C NMR) and 3.86 ppm (1H NMR). Finally, the chemical structure of the final indole-based caffeic acid amide analogue 3m was confirmed by detecting three singlet peaks with chemical shifts of more than 8.00 ppm representing the three protons of the free OH, the amide NH, and the indole NH. Moreover, the two vinylic protons were clearly identified as two doublet peaks with J values of 16.00 Hz at 7.48 and 6.71 ppm. In addition, the six protons of the two methoxy groups were found as a long singlet peak at 3.83 ppm. The amide carbon appeared at 163.84 ppm, the two carbons bearing the two methoxy groups appeared as a long peak at 148.52 ppm, and the two aliphatic carbons of the methoxy groups showed a singlet peak at 56.37 ppm. These data proved and confirmed the formation and purity of the desired amide derivatives 3a–m. ## 3.3. In Silico Druggability Studies of the Newly Synthesized Amide Derivatives 3a–m There is no guarantee that a small molecule that possesses a potent interaction with its target protein could be a successful therapeutic candidate. Poor absorption, distribution, metabolism, and excretion (ADME) characteristics may be the reason for this failure. Thus, many promising small molecules fail during the drug discovery process. Moreover, the drug development process is expensive. Accordingly, the pharmacokinetic (PK) characteristics of the indole-based benzamide and caffeic acid amide analogues (3a–m) were evaluated via the SwissADME platform by using distance/pharmacophore models coded as graph-based marks [54]. Using this platform, numerous crucial characteristics can be anticipated including the solubility of the final compounds, their gastrointestinal absorption, and brain entry abilities. During the different steps of the new drug development, these PK factors would constitute the foundation stone of the outcome’s anticipation [55]. Another major PK property is the topological polar surface area (TPSA) of a compound which refers to the surface sum over the entire polar atoms, mainly nitrogen and oxygen, together with their associated hydrogen atoms. The TPSA is obtained by subtracting from the molecular surface the area of carbon atoms, halogens, and hydrogen atoms bonded to carbon atoms (i.e., nonpolar hydrogen atoms). TPSA is considered a great metric to improve the ability of a drug to penetrate the cells, which could enhance the efficacy of the synthesized drug candidate. Molecules possessing TPSA > 140 Å2 are predicted to not be able to cross cell membranes. On the other hand, a TPSA value < 90 Å2 was found to be essential for a drug candidate to cross the blood–brain barrier (BBB) [56]. Furthermore, compliance with the Lipinski rule of five [57] is another important guide on whether a compound can be taken orally. The outcomes of the in silico PK study are presented in Table 2 and Figure 3. The predicted PK properties of all the newly synthesized indole-based benzamide and caffeic acid amide analogues (3a–m) revealed that all compounds would have high gastrointestinal absorption. In addition, all compounds displayed compliance to the Lipinski rule of five, indicating their great potential to be promising drug candidates with acceptable PK characteristics. Most compounds showed TPSA < 90 Å2, suggesting a potential antioxidant effect also in the brain to battle different neurodegenerative diseases, as hypothesized. These results suggest the PK stability of the indole-based benzamide and caffeic acid amide series. ## 3.4.1. DPPH Radical Scavenging Activity First, all the synthesized analogues (3a–m) were initially screened for their scavenging effects on the DPPH radical. Trolox was used as a reference (IC50 = 33.84 ± 1.01 µM). As illustrated in Table 3, our goal was to discover the antioxidant properties of incorporating the phenolic OH group(s) in addition to some other moieties such as methoxy, bromo, and amino groups in different positions on the phenyl ring, which is attached to position 5 of the antioxidant indole core via two different linkers (depending on the amide type, benzamide or caffeic acid amide). While compounds 3b–e, 3g, 3i, 3k, and 3l showed moderate to weak activity as compared to Trolox, compounds 3a, 3f, 3h, 3j, and 3m were found to have excellent radical scavenging activities with IC50 values of 95.81 ± 1.01, 136.8 ± 1.04, 86.77 ± 1.03, 50.98 ± 1.05, and 67.64 ± 1.02 µM. It was noted that compounds possessing the 3,4-dihydroxyphenyl moiety exhibited promising activities (the benzamide derivative 3h and the caffeic acid amide derivative 3j). On the other hand, the 4-amino-3-hydroxyphenyl moiety was only able to demonstrate its free radical scavenging effect in compound 3a, which has the 4-amino-3-hydroxy phenyl moiety. The 4-hydroxy-3-methoxyphenyl and 4-hydroxy-3,5-dimethoxyphenyl moieties were only able to show their activities in the caffeic acid amide analogues 3f and 3m, respectively. It was also noticed that the majority of the synthesized caffeic acid amide derivatives (three out of four) were able to show higher free radical scavenging activity compared to their benzamide analogues. It could be the double-bond moiety in these caffeic acid amide derivatives that may increase the capacity of the molecule to interact with the free radicals via enhancing the electron conjugation effect in the whole chemical structure so that they do not engage in a destructive biochemical reaction. Based on this primary screening, the position and nature of substitutions on the phenyl moiety and the presence of the double bond in the middle of the structure were found to be essential factors directly affecting the free radical scavenging activity of these new indole-based amides. Consequently, the most potent derivatives (3a, 3f, 3h, 3j, and 3m) were further evaluated. ## 3.4.2. ABTS•+ Radical Cation Scavenging Assay ABTS activity was measured in terms of percentage inhibition (%) of the ABTS•+ radical cation by each of the five most active compounds (3a, 3f, 3h, 3j, and 3m). The ABTS values of the five samples are presented in Table 3. While compound 3a (with the 4-amino-3-hydroxy phenyl moiety) was able to scavenge the radical cation ABTS•+ with an IC50 value of 33.33 ± 1.96 µM, which is almost the similar potency of the standard Trolox (29.62 ± 1.86 µM), compound 3h possessing 3,4-dihydroxy phenyl moiety showed a higher IC50 value of 39.98 ± 0.92 µM. Interestingly, three compounds out of the five (3f, 3j, and 3m) showed higher capacities to neutralize the radical cation ABTS•+ than Trolox with IC50 values of 14.48 ± 0.68, 19.49 ± 0.54, and 14.92 ± 0.30 µM, respectively. ## 3.4.3. FRAP Assay The three highly potent analogues (3f, 3j, and 3m) were considered for FRAP and ORAC assays. The FRAP assay assesses the antioxidant properties of the tested compound based on its reducing ability. The values obtained, shown in Table 3, were consistent with the DPPH and ABTS assays. In this study, compound 3j (the caffeic acid derivative possessing 3,4-dihydroxyphenyl moiety) presented the highest antioxidant capacity with a FRAP value of 4774.37 ± 137.20 μM Trolox eq/mM sample, followed by compounds 3m (4-hydroxy-3,5-dimethoxyphenyl moiety containing caffeic acid derivative) and 3f (4-hydroxy-3-methoxyphenyl moiety containing caffeic acid derivative) with values of 2308.7 ± 73.73 and 1951.45 ± 75.97 μM Trolox eq/mM sample, respectively. Based on these findings, it could be concluded that caffeic amide analogue 3j possessing the two phenolic OH groups not only offered the top free radical scavenging capability, but also the strongest reducing power among the tested compounds. Indeed, the antioxidant activity of a small molecule largely depends on both the chemical structure of the compound and the test system. Accordingly, it cannot be fully assessed by one single technique due to the various mechanisms of antioxidant action. As a result, the ORAC test was chosen to be the next further test for these three promising analogues (3f, 3j, and 3m). ## 3.4.4. ORAC Assay Through the ORAC test, the antioxidant capacity was investigated of the three highly active compounds (3f, 3j, and 3m) that had demonstrated high antioxidant activity with the previous DPPH, ABTS, and FRAP tests. The ORAC test was intended to validate the results obtained with the previous approaches and extend the activity profile for each tested derivative. All tested compounds exhibited a dynamic ability to reduce the oxidative degradation of the fluorescent molecule, caused by peroxyl radicals. Compounds 3m and 3f showed very high ORAC antioxidant power (9253.47 ± 806.00 and 7293.46 ± 208.48 μM Trolox eq/mM sample, respectively). In a similar way to the previous assay (FRAP), the best antioxidant capacity against the peroxyl radicals was observed for compound 3j (10,714.21 ± 817.76 μM Trolox eq/mM sample). ## 4. Conclusions As a step toward the development of novel free-radical scavenging hybrid agents for oxidative stress-related therapy, a new series of indole-based benzamide and caffeic acid amide analogues (3a–m) was successfully designed and synthesized. Among them, compounds 3a (4-amino-3-hydroxy benzamide derivative), 3f (4-hydroxy-3-methoxyphenyl containing caffeic acid derivative), 3h (3,4-dihydroxy benzamide), 3j (3,4-dihydroxyphenyl containing caffeic acid derivative), and 3m (4-hydroxy-3,5-dimethoxyphenyl containing caffeic acid derivative) were able to show promising DPPH radical scavenging activities with IC50 values of 95.81 ± 1.01, 136.8 ± 1.04, 86.77 ± 1.03, 50.98 ± 1.05, and 67.64 ± 1.02 µM, respectively The three caffeic acid derivatives 3f, 3j, and 3m neutralized the free radical cation ABTS•+ more than Trolox with IC50 values of 14.48 ± 0.68, 19.49 ± 0.54, and 14.92 ± 0.30 µM, respectively. Using FRAP and ORAC assays, compound 3j was the most active antioxidant agent with values of 4774.37 ± 137.20 and 10,714.21 ± 817.76 μM Trolox eq/mM sample, respectively. Most small molecules were anticipated to be soluble and to penetrate the brain. No violations of the Lipinski rule of five were noticed, indicating a pharmacokinetically stable profile. Consequently, the hybrid compound 3j is reported as a new antioxidant candidate with highly potent and promising free radical scavenging activities. ## Figures, Scheme and Tables Figure 1: (A) Chemical structures of some well-known synthetic antioxidants (BHA; I and BHT; II); (B) Chemical structure of the natural antioxidant caffeic acid (III) and a proposed mechanism for caffeic acid antioxidant activity; (C) Chemical structure of the natural antioxidant melatonin (IV) and the proposed mechanism of how the electron-rich aromatic indole ring system of melatonin scavenges the hydroxyl radical (HO•). Figure 2: Design of structural hybridization of the caffeic acid derivatives with melatonin scaffold, in addition to the synthetic planning, to obtain the novel indole–caffeic amide analogues 3a–m. Scheme 1: (a) Appropriate carboxylic acid derivative (0.75 mmol), EDCI (1.1 mmol), HOBt (1.1 mmol), DIPEA (1.1 mmol), acetonitrile (5 mL), 25 °C, 12 h. (Full description is indicated in the Materials and Methods section). Figure 3: PK properties of the newly synthesized amides (3a–m) predicted by the SwissADME platform. TABLE_PLACEHOLDER:Table 1 TABLE_PLACEHOLDER:Table 2 TABLE_PLACEHOLDER:Table 3 ## References 1. Kumar S., Gupta E., Kaushik S., Kumar Srivastava V., Mehta S.K., Jyoti A.. Evaluation of oxidative stress and antioxidant status: Correlation with the severity of sepsis. Scand. J. Immunol. (2018) 87 e12653. DOI: 10.1111/sji.12653 2. Darenskaya M.A., Kolesnikova L.I., Kolesnikov S.I.. Oxidative Stress: Pathogenetic Role in Diabetes Mellitus and Its Complications and Therapeutic Approaches to Correction. Bull. Exp. Biol. Med. (2021) 171 179-189. DOI: 10.1007/s10517-021-05191-7 3. Pohanka M.. Role of oxidative stress in infectious diseases. A review. Folia Microbiol. (2013) 58 503-513. DOI: 10.1007/s12223-013-0239-5 4. Sies H.. Oxidative stress: A concept in redox biology and medicine. Redox Biol. (2015) 4 180-183. DOI: 10.1016/j.redox.2015.01.002 5. Poprac P., Jomova K., Simunkova M., Kollar V., Rhodes C.J., Valko M.. Targeting Free Radicals in Oxidative Stress-Related Human Diseases. Trends Pharmacol. Sci. (2017) 38 592-607. DOI: 10.1016/j.tips.2017.04.005 6. Liu R., Mabury S.A.. Synthetic Phenolic Antioxidants: A Review of Environmental Occurrence, Fate, Human Exposure, and Toxicity. Environ. Sci. Technol. (2020) 54 11706-11719. DOI: 10.1021/acs.est.0c05077 7. Wang W., Xiong P., Zhang H., Zhu Q., Liao C., Jiang G.. Analysis, occurrence, toxicity and environmental health risks of synthetic phenolic antioxidants: A review. Environ. Res. (2021) 201 111531. DOI: 10.1016/j.envres.2021.111531 8. Neha K., Haider M.R., Pathak A., Yar M.S.. Medicinal prospects of antioxidants: A review. Eur. J. Med. Chem. (2019) 178 687-704. DOI: 10.1016/j.ejmech.2019.06.010 9. Kundu T., Pramanik A.. Expeditious and eco-friendly synthesis of new multifunctionalized pyrrole derivatives and evaluation of their antioxidant property. Bioorganic Chem. (2020) 98 103734. DOI: 10.1016/j.bioorg.2020.103734 10. Kundu T., Bhattacharjee B., Hazra S., Ghosh A.K., Bandyopadhyay D., Pramanik A.. Synthesis and Biological Assessment of Pyrrolobenzoxazine Scaffold as a Potent Antioxidant. J. Med. Chem. (2019) 62 6315-6329. DOI: 10.1021/acs.jmedchem.9b00717 11. Bandeira P.T., Dalmolin M.C., de Oliveira M.M., Nunes K.C., Garcia F.P., Nakamura C.V., de Oliveira A.R.M., Piovan L.. Synthesis, Antioxidant Activity and Cytotoxicity of N-Functionalized Organotellurides. Bioorganic Med. Chem. (2019) 27 410-415. DOI: 10.1016/j.bmc.2018.12.017 12. Tailor N., Sharma M.. Antioxidant hybrid compounds: A promising therapeutic intervention in oxidative stress induced diseases. Mini Rev. Med. Chem. (2013) 13 280-297. PMID: 22512588 13. Vavříková E., Křen V., Jezova-Kalachova L., Biler M., Chantemargue B., Pyszková M., Riva S., Kuzma M., Valentová K., Ulrichová J.. Novel flavonolignan hybrid antioxidants: From enzymatic preparation to molecular rationalization. Eur. J. Med. Chem. (2017) 127 263-274. DOI: 10.1016/j.ejmech.2016.12.051 14. Carullo G., Mazzotta S., Koch A., Hartmann K.M., Friedrich O., Gilbert D.F., Vega-Holm M., Schneider-Stock R., Aiello F.. New Oleoyl Hybrids of Natural Antioxidants: Synthesis and In Vitro Evaluation as Inducers of Apoptosis in Colorectal Cancer Cells. Antioxidants (2020) 9. DOI: 10.3390/antiox9111077 15. Gülçin I.. Antioxidant activity of caffeic acid (3,4-dihydroxycinnamic acid). Toxicology (2006) 217 213-220. DOI: 10.1016/j.tox.2005.09.011 16. Sofic E., Rimpapa Z., Kundurovic Z., Sapcanin A., Tahirovic I., Rustembegovic A., Cao G.J.. Antioxidant capacity of the neurohormone melatonin. J. Neural Transm. (2005) 112 349-358. DOI: 10.1007/s00702-004-0270-4 17. Cipolla-Neto J., Amaral F.G.D.. Melatonin as a Hormone: New Physiological and Clinical Insights. Endocr. Rev. (2018) 39 990-1028. DOI: 10.1210/er.2018-00084 18. Reiter R.J., Mayo J.C., Tan D.X., Sainz R.M., Alatorre-Jimenez M., Qin L.. Melatonin as an antioxidant: Under promises but over delivers. J. Pineal Res. (2016) 61 253-278. DOI: 10.1111/jpi.12360 19. Reiter R.J., Tan D.X., Cabrera J., D’Arpa D.. Melatonin and tryptophan derivatives as free radical scavengers and antioxidants. Adv. Exp. Med. Biol. (1999) 467 379-387. DOI: 10.1007/978-1-4615-4709-9_48 20. Reiter R.J., Tan D.X., Rosales-Corral S., Galano A., Zhou X.J., Xu B.. Mitochondria: Central Organelles for Melatonin’s Antioxidant and Anti-Aging Actions. Molecules (2018) 23. DOI: 10.3390/molecules23020509 21. Reiter R.J., Rosales-Corral S., Tan D.X., Jou M.J., Galano A., Xu B.. Melatonin as a mitochondria-targeted antioxidant: One of evolution’s best ideas. Cell. Mol. Life Sci. CMLS (2017) 74 3863-3881. DOI: 10.1007/s00018-017-2609-7 22. Velkov Z.A., Velkov Y.Z., Galunska B.T., Paskalev D.N., Tadjer A.V.. Melatonin: Quantum-chemical and biochemical investigation of antioxidant activity. Eur. J. Med. Chem. (2009) 44 2834-2839. DOI: 10.1016/j.ejmech.2008.12.017 23. Estevão M.S., Carvalho L.C., Ribeiro D., Couto D., Freitas M., Gomes A., Ferreira L.M., Fernandes E., Marques M.M.B.. Antioxidant activity of unexplored indole derivatives: Synthesis and screening. Eur. J. Med. Chem. (2010) 45 4869-4878. DOI: 10.1016/j.ejmech.2010.07.059 24. Esposito E., Cuzzocrea S.. Antiinflammatory activity of melatonin in central nervous system. Curr. Neuropharmacol. (2010) 8 228-242. DOI: 10.2174/157015910792246155 25. Di Bella G., Mascia F., Gualano L., Di Bella L.. Melatonin anticancer effects. Int. J. Mol. Sci. (2013) 14 2410-2430. DOI: 10.3390/ijms14022410 26. de Zanette S.A., Vercelino R., Laste G., Rozisky J.R., Schwertner A., Machado C.B., Xavier F., de Souza I.C.C., Deitos A., Torres I.L.. Melatonin analgesia is associated with improvement of the descending endogenous pain-modulating system in fibromyalgia: A phase II, randomized, double-dummy, controlled trial. BMC Pharmacol. Toxicol. (2014) 15. DOI: 10.1186/2050-6511-15-40 27. Lochner A., Marais E., Huisamen B.. Melatonin and cardioprotection against ischaemia/reperfusion injury: What’s new? A review. J. Pineal Res. (2018) 65 e12490. DOI: 10.1111/jpi.12490 28. Chen J., Xia H., Zhang L., Zhang H., Wang D., Tao X.. Protective effects of melatonin on sepsis-induced liver injury and dysregulation of gluconeogenesis in rats through activating SIRT1/STAT3 pathway. Biomed. Pharmacother. (2019) 117 109150. DOI: 10.1016/j.biopha.2019.109150 29. Rahman A., Hasan A.U., Kobori H.. Melatonin in chronic kidney disease: A promising chronotherapy targeting the intrarenal renin–angiotensin system. Hypertens. Res. (2019) 42 920-923. DOI: 10.1038/s41440-019-0223-9 30. Tordjman S., Chokron S., Delorme R., Charrier A., Bellissant E., Jaafari N., Fougerou C.. Melatonin: Pharmacology, Functions and Therapeutic Benefits. Curr. Neuropharmacol. (2017) 15 434-443. DOI: 10.2174/1570159X14666161228122115 31. Chitimus D.M., Popescu M.R., Voiculescu S.E., Panaitescu A.M., Pavel B., Zagrean L., Zagrean A.M.. Melatonin’s Impact on Antioxidative and Anti-Inflammatory Reprogramming in Homeostasis and Disease. Biomolecules (2020) 10. DOI: 10.3390/biom10091211 32. Salehi B., Sharopov F., Fokou P.V.T., Kobylinska A., Jonge L., Tadio K., Sharifi-Rad J., Posmyk M.M., Martorell M., Martins N.. Melatonin in Medicinal and Food Plants: Occurrence, Bioavailability, and Health Potential for Humans. Cells (2019) 8. DOI: 10.3390/cells8070681 33. Rodríguez-Franco M.I., Fernández-Bachiller M.I., Pérez C., Hernández-Ledesma B., Bartolomé B.. Novel tacrine—Melatonin hybrids as dual-acting drugs for Alzheimer disease, with improved acetylcholinesterase inhibitory and antioxidant properties. J. Med. Chem. (2006) 49 459-462. DOI: 10.1021/jm050746d 34. Luo X.-T., Wang C.-M., Liu Y., Huang Z.-G.. New multifunctional melatonin-derived benzylpyridinium bromides with potent cholinergic, antioxidant, and neuroprotective properties as innovative drugs for Alzheimer’s disease. Eur. J. Med. Chem. (2015) 103 302-311. DOI: 10.1016/j.ejmech.2015.08.052 35. Bautista-Aguilera O.M., Esteban G., Bolea I., Nikolic K., Agbaba D., Moraleda I., Iriepa I., Samadi A., Soriano E., Unzeta M.. Design, synthesis, pharmacological evaluation, QSAR analysis, molecular modeling and ADMET of novel donepezil-indolyl hybrids as multipotent cholinesterase/monoamine oxidase inhibitors for the potential treatment of Alzheimer’s disease. Eur. J. Med. Chem. (2014) 75 82-95. DOI: 10.1016/j.ejmech.2013.12.028 36. Rivara S., Pala D., Bedini A., Spadoni G.. Therapeutic uses of melatonin and melatonin derivatives: A patent review (2012. Expert Opin. Ther. Pat. (2015) 25 425-441. DOI: 10.1517/13543776.2014.1001739 37. Wang S.Y., Shi X.C., Laborda P.. Indole-based melatonin analogues: Synthetic approaches and biological activity. Eur. J. Med. Chem. (2020) 185 111847. DOI: 10.1016/j.ejmech.2019.111847 38. Mor M., Rivara S., Silva C., Bordi F., Plazzi P.V., Spadoni G., Diamantini G., Balsamini C., Tarzia G., Fraschini F.. Melatonin receptor ligands: Synthesis of new melatonin derivatives and comprehensive comparative molecular field analysis (CoMFA) study. J. Med. Chem. (1998) 41 3831-3844. DOI: 10.1021/jm9810093 39. Landagaray E., Ettaoussi M., Leclerc V., Traoré B., Perez V., Nosjean O., Boutin J.A., Caignard D.H., Delagrange P., Berthelot P.. New melatonin (MT1/MT2) ligands: Design and synthesis of (8,9-dihydro-7H-furo[3,2-f]chromen-1-yl) derivatives. Bioorganic Med. Chem. (2014) 22 986-996. DOI: 10.1016/j.bmc.2013.12.054 40. Angelova V.T., Rangelov M., Todorova N., Dangalov M., Andreeva-Gateva P., Kondeva-Burdina M., Karabeliov V., Shivachev B., Tchekalarova J.. Discovery of novel indole-based aroylhydrazones as anticonvulsants: Pharmacophore-based design. Bioorganic Chem. (2019) 90 103028. DOI: 10.1016/j.bioorg.2019.103028 41. Tan D.X., Reiter R.J., Manchester L.C., Yan M.T., El-Sawi M., Sainz R.M., Mayo J.C., Kohen R., Allegra M., Hardeland R.. Chemical and physical properties and potential mechanisms: Melatonin as a broad spectrum antioxidant and free radical scavenger. Curr. Top. Med. Chem. (2002) 2 181-197. DOI: 10.2174/1568026023394443 42. Stasica P., Ulanski P., Rosiak J.M.. Melatonin as a hydroxyl radical scavenger. J. Pineal Res. (1998) 25 65-66. DOI: 10.1111/j.1600-079X.1998.tb00387.x 43. Reiter R.J.. Cytoprotective properties of melatonin: Presumed association with oxidative damage and aging. Nutrition (1998) 14 691-696. DOI: 10.1016/S0899-9007(98)00064-1 44. Elkamhawy A., Kim H.J., Elsherbeny M.H., Paik S., Park J.H., Gotina L., Abdellattif M.H., Gouda N.A., Cho J., Lee K.. Discovery of 3,4-dichloro-N-(1H-indol-5-yl)benzamide: A highly potent, selective, and competitive hMAO-B inhibitor with high BBB permeability profile and neuroprotective action. Bioorganic Chem. (2021) 116 105352. DOI: 10.1016/j.bioorg.2021.105352 45. Elsherbeny M.H., Kim J., Gouda N.A., Gotina L., Cho J., Pae A.N., Lee K., Park K.D., Elkamhawy A., Roh E.J.. Highly Potent, Selective, and Competitive Indole-Based MAO-B Inhibitors Protect PC12 Cells against 6-Hydroxydopamine- and Rotenone-Induced Oxidative Stress. Antioxidants (2021) 10. DOI: 10.3390/antiox10101641 46. Elkamhawy A., Paik S., Kim H.J., Park J.H., Londhe A.M., Lee K., Pae A.N., Park K.D., Roh E.J.. Discovery of N-(1-(3-fluorobenzoyl)-1H-indol-5-yl)pyrazine-2-carboxamide: A novel, selective, and competitive indole-based lead inhibitor for human monoamine oxidase B. J. Enzym. Inhib. Med. Chem. (2020) 35 1568-1580. DOI: 10.1080/14756366.2020.1800666 47. Yeon S.K., Choi J.W., Park J.-H., Lee Y.R., Kim H.J., Shin S.J., Jang B.K., Kim S., Bahn Y.-S., Han G.. Synthesis and evaluation of biaryl derivatives for structural characterization of selective monoamine oxidase B inhibitors toward Parkinson’s disease therapy. Bioorganic Med. Chem. (2018) 26 232-244. DOI: 10.1016/j.bmc.2017.11.036 48. Cores Á., Abril S., Michalska P., Duarte P., Olives A.I., Martín M.A., Villacampa M., León R., Menéndez J.C.. Bisavenathramide Analogues as Nrf2 Inductors and Neuroprotectors in In Vitro Models of Oxidative Stress and Hyperphosphorylation. Antioxidants (2021) 10. DOI: 10.3390/antiox10060941 49. Elkamhawy A., Park J.E., Hassan A.H.E., Pae A.N., Lee J., Park B.G., Roh E.J.. Synthesis and evaluation of 2-(3-arylureido)pyridines and 2-(3-arylureido)pyrazines as potential modulators of Aβ-induced mitochondrial dysfunction in Alzheimer’s disease. Eur. J. Med. Chem. (2018) 144 529-543. DOI: 10.1016/j.ejmech.2017.12.045 50. Boly R., Lamkami T., Lompo M., Dubois J., Guissou I.. DPPH free radical scavenging activity of two extracts from Agelanthus dodoneifolius (Loranthaceae) leaves. Int. J. Toxicol. Pharmacol. Res. (2016) 8 29-34 51. Arnao M.B., Cano A., Acosta M.. The hydrophilic and lipophilic contribution to total antioxidant activity. Food Chem. (2001) 73 239-244. DOI: 10.1016/S0308-8146(00)00324-1 52. Benzie I.F.F., Strain J.J.. The Ferric Reducing Ability of Plasma (FRAP) as a Measure of “Antioxidant Power”: The FRAP Assay. Anal. Biochem. (1996) 239 70-76. DOI: 10.1006/abio.1996.0292 53. Liang Z., Cheng L., Zhong G.Y., Liu R.H.. Antioxidant and antiproliferative activities of twenty-four Vitis vinifera grapes. PLoS ONE (2014) 9. DOI: 10.1371/journal.pone.0105146 54. Daina A., Michielin O., Zoete V.. SwissADME: A free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules. Sci. Rep. (2017) 7 42717. DOI: 10.1038/srep42717 55. Alavijeh M.S., Chishty M., Qaiser M.Z., Palmer A.M.. Drug metabolism and pharmacokinetics, the blood-brain barrier, and central nervous system drug discovery. NeuroRx J. Am. Soc. Exp. NeuroTherapeutics (2005) 2 554-571. DOI: 10.1602/neurorx.2.4.554 56. Pajouhesh H., Lenz G.R.. Medicinal chemical properties of successful central nervous system drugs. NeuroRx J. Am. Soc. Exp. NeuroTherapeutics (2005) 2 541-553. DOI: 10.1602/neurorx.2.4.541 57. Lipinski C.A.. Lead- and drug-like compounds: The rule-of-five revolution. Drug Discov. Today: Technol. (2004) 1 337-341. DOI: 10.1016/j.ddtec.2004.11.007
--- title: 'Amazon Amandaba—Sociodemographic Factors, Health Literacy, Biochemical Parameters and Self-Care as Predictors in Patients with Type 2 Diabetes Mellitus: A Cross-Sectional Study' authors: - Victória Brioso Tavares - Aline Lobato de Farias - Amanda Suzane Alves da Silva - Josiel de Souza e Souza - Hilton Pereira da Silva - Maria do Socorro Castelo Branco de Oliveira Bastos - João Simão de Melo-Neto journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9966953 doi: 10.3390/ijerph20043082 license: CC BY 4.0 --- # Amazon Amandaba—Sociodemographic Factors, Health Literacy, Biochemical Parameters and Self-Care as Predictors in Patients with Type 2 Diabetes Mellitus: A Cross-Sectional Study ## Abstract Background: Health literacy (HL) and its domains (functional, critical, and communicative) appear to be related to self-care adherence in people with type 2 diabetes mellitus (DM2). This study aimed to verify if sociodemographic variables are predictors of HL, if HL and the sociodemographic factors affect biochemical parameters together, and if HL domains are predictors of self-care in DM2. Methods: We used the baseline assessment data from 199 participants ≥ 30 years in the project, “Amandaba na Amazônia: Culture Circles as a Strategy to Encourage Self-care for DM in Primary Health Care,” which took place in November and December 2021. Results: In the HL predictor analysis, women ($$p \leq 0.024$$) and higher education ($$p \leq 0.005$$) were predictors of better functional HL. The predictors of biochemical parameters were: glycated hemoglobin control with low critical HL ($$p \leq 0.008$$); total cholesterol control with female sex ($$p \leq 0.004$$), and low critical HL ($$p \leq 0.024$$); low-density lipoprotein control with female sex ($$p \leq 0.027$$), and low critical HL ($$p \leq 0.007$$); high-density lipoprotein control with female sex ($$p \leq 0.001$$); triglyceride control with low Functional HL ($$p \leq 0.039$$); high levels of microalbuminuria with female sex ($$p \leq 0.014$$). A low critical HL was a predictor of a lower specific diet ($$p \leq 0.002$$) and a low total HL of low medication care ($$p \leq 0.027$$) in analyses of HL domains as predictors of self-care. Conclusion: Sociodemographic factors can be used to predict HL, and HL can predict biochemical parameters and self-care. ## 1. Introduction Health Literacy (HL) was coined in 1970 [1] and was initially defined as the ability to deal with words and numbers in a medical setting. However, the concept of HL has evolved, and it is now considered to encompass a variety of social, personal, and cognitive skills required to obtain, process (critical thinking), and understand the information in the health context [2,3,4,5,6,7]. According to Nutbeam [8] HL has three dimensions: functional HL (FHL)—the ability to read and write; communicative HL (CHL)—the ability to absorb and apply the information obtained; and critical HL (CrHL)—the analysis and deeper understanding of information for decision-making. These different dimensions can affect the patient’s autonomy and ability to use health information, which affects self-care and treatment adherence decision-making [9]. Low levels of HL can lead to harmful health practices [10], including less knowledge, management, and self-care, resulting in lower medication adherence and increased hospitalization and mortality, especially in those with non-communicable diseases (NCD) [11,12,13]. The Test of Functional Health Literacy in Adults [14], the Rapid Estimate of Adult Literacy in Medicine [15], and the Newest Vital Sign [16] are among the HL assessment instruments that seek to measure HL more objectively and only assess specific dimensions. However, more recent instruments, including the European Health Literacy Survey Questionnaire [17], the Health Literacy Questionnaire [18], and the 14-item health literacy scale (HLS-14) [19] consider the multidimensionality of the HL and measure self-reported HL. Diabetes mellitus (DM) is an NCD that requires understanding a wide range of clinical recommendations and information to manage the disease, including self-care activities, such as healthy eating, physical activity, adherence to prescribed medications, and foot care. Diabetes affects approximately 463 million people globally (aged between 20 and 70 years). Brazil ranks sixth among the top ten countries or territories in terms of the number of adults (20–79 years) with diabetes (15.7 million in 2021) [20,21]. The most common methods for assessing self-care behaviors are questionnaires on the adoption of behaviors during a specific period, typically 24 h [22], 1 week [23], or 1 month [24], or the frequency of specific self-care behaviors during the previous week [25]. In Brazil, no studies have reported comprehensive data on HL among people with type 2 diabetes mellitus (DM2). Most studies use questionnaires that assess only one component of HL, making it difficult to understand the magnitude of HL. Nevertheless, previous studies have reported that an insufficient level of some HL domains seems to be associated with low adherence to self-care behavior in people with DM2, especially regarding glycemic monitoring and control [26,27,28]. Although few studies have used specific instruments to evaluate the direct influence of HL on different aspects of self-care in DM2, approximately $65\%$ of people with DM2 have been reported to have low HL [29]. People living in the northern region of Brazil and older adults with DM2, whom the Unified Health System assists, seem to have low HL [30]. Because the World Health Organization defines HL as a social determinant of health mediated by cultural and situational demands [31], it is necessary to understand how HL relates to sociodemographic factors, and due to the context of the DM, how it relates to biochemical parameter control and self-care. Thus, this study aimed to determine if sociodemographic factors and HL are predictors of biochemical parameters (glycated hemoglobin, triglycerides, total cholesterol, high-density lipoprotein (HDL), low-density lipoprotein (LDL), non-high-density lipoprotein cholesterol (NHDL), and microalbuminuria) and whether HL and its domains are predictors of specific self-care behaviors in patients with DM2 in Brazil. We hypothesized that other sociodemographic variables are essential for HL diagnosis and that HL influences biochemical parameter control and self-care behavior maintenance. ## 2.1. Study Design This was a cross-sectional observational study with descriptive and inferential statistics. ## 2.2. Setting and Period of Study The study used data from the participants’ baseline assessments in the project “Amandaba na Amazônia: Culture Circles as a Strategy to Encourage Self-care for DM in Primary Health Care”, which took place in November and December 2021. The project was conducted in the Belém-Pará setting of the Unified Brazilian Health System, using a survey of patients with DM2 registered in the Family Health Strategies (FHS) in two administrative health districts of the municipality, district 1: Guamá, and district 2: Bengui. ## 2.3. Population The study population consisted of individuals with DM2 aged 30 years or older. ## 2.4. Eligibility Criteria The inclusion criteria were patients aged ≥30 years with DM2 that were registered and had been followed up for at least 1 year in one of the eight selected FHS units. In addition, the participants had to be able to read and have adequate or corrected self-reported hearing and visual acuity to understand the study. ## 2.5. Sampling Probabilistic random sampling, through a drawing based on the survey of the defined population, was used to select patients. ## 2.6. Sample The sample size was calculated using the Gpower 3.1 software (HUU, Dusseldorf, Germany), based on the variable “Hb1Ac (glycated hemoglobin)” in the study by Rodrigues et al. [ 32] who compare HL in adults and older adults with diabetes between health units in two municipalities in São Paulo, Brazil, using the 14-item health literacy scale [18,33,34]. This presented an odds ratio of 4.455 and a p value < 0.05, indicating a minimum sample of 79 participants (low literacy = 34; high literacy = 45). This was based on the N2/N1 allocation ratio of 1.33, a proportion p2 of 0.125, β error probability of 0.8, α error probability of 0.05. Our initial sample consisted of 230 individuals, of whom 199 were selected based on the eligibility criteria (Figure 1). ## 2.7. Data Collection and Variables An individual standardized questionnaire was used to collect data, which included the following sociodemographic variables: sex, age, race, education (number of years of formal education), HL, and per capita income; and health-related and clinical characteristics: duration of DM2 diagnosis (years), perception of general health, smoking and alcohol consumption, number of consultations in the previous year, private health insurance, systemic arterial hypertension (SAH), physical activity, and peripheral neuropathy. Physical activity was assessed using the Brazilian national self-assessment of health status survey, which categorizes it as sedentary for those who identify with the answer options “sitting most of the day” and “does not walk much during the day”, regular: “carries light weights or climbs stairs frequently or exercises regularly” and “carries heavy weights or exercises regularly”, and irregular: “walks or stands a lot during the day, but does not carry or lift weights regularly” was replaced by “irregular physical activity” [35]. The Michigan Neuropathy Screening Instrument (MSNI) was used to assess peripheral neuropathy; MSNI values ≤5 = no neuropathy, ≥5.5 = neuropathy [36]; biochemical parameters: triglycerides (normal: 36–149; high ≥150 mg/dL), total cholesterol (normal: 87–189; high ≥190 mg/dL), HDL (normal ≥40; low: 20–39 mg/dL), LDL (normal: 15–99; high ≥100 mg/dL), NHDL (normal: 59–129; high ≥130 mg/dL), HbA1c (normal: 4.8–6.9; high ≥$7\%$), and microalbuminuria (low <30; normal: 30–300; high >300 mg/g), were obtained through laboratory tests. HL was assessed using the 14-item health literacy scale developed by Suka et al. [ 18] and validated in Brazilian Portuguese by Batista et al. [ 34]. It is one of the few instruments with a fair quality assessment that considers the expanded concept of a relatively short assessment, and has been translated and validated into Brazilian Portuguese. It is scored on a 5-point Likert scale, with responses ranging from “totally agree” to “strongly disagree”, organized into three dimensions: functional (five items), communicative (five items), and critical (four items), according to the theoretical model of HL proposed by Nutbeam [8]. The total score ranges from 14 to 70, and higher average scores for each item are associated with higher literacy, except in the functional dimension, where the result is inverted. Based on the average of the total score of the participants, the classification of HL was divided into low (<46) and high (≥46), representing a high or low ability to access, absorb and apply, and understand health information in accordance with the HLS-14 specific domains and overall score [8,34]. The Brazilian version of the Diabetes Self-Care Activities Questionnaire (QAD) [37] was used to assess self-care, which covered five aspects of the diabetes treatment regimen: general diet, specific diet, physical activity, glycemic monitoring, foot care, and medication. These dimensions represent various diabetes treatment activities performed independently by patients, with questions about the frequency of activities performed in the previous 7 days and their agreement with the doctor’s or other health professional’s prescription. Thus, a score of 7 means ideal self-care adherence, while a score of <3 indicates minor care. Each aspect’s median number of days is examined. The invitation signature of the consent form and the administration of the questionnaires and laboratory tests were all done at the participant’s home during a previously scheduled visit. All questionnaires were administered personally by 12 researchers who had been trained on a protocol for approaching and using the instruments. ## 2.8. Primary Outcomes The primary outcomes were defined according to the following objectives: [1] HL, [2] biochemical parameters, and [3] self-care. ## 2.9. Bias The study design made it susceptible to non-response bias, which was attempted to be minimized through prior scheduling of visits and the benefit of returning the results of laboratory tests to participants to encourage participation in the study, so no questionnaire was incomplete. However, in our study direct contact was made with the participant drawn from a list, with each participant being replaced by the next in sequence, so only those who accepted the invitation after the drawing were included. The selection bias was minimized by obtaining a sample from a defined population and reporting the population selection steps and recruitment/inclusion criteria to minimize the unwarranted generalization of the findings. ## 2.10. Statistical Analysis Descriptive statistical analysis was performed to calculate the frequency (absolute and relative), mean and standard deviation (parametric), or medians with interquartile range (IQR, non-parametric) as measures of central and dispersion tendency, respectively. The data underwent the Kolmogorov–Smirnov normality test, Pearson’s chi-square test, Fisher’s exact test, and the Mann–Whitney U test. A Zcrit value of ≥1.96 was considered for the post hoc residual adjustment test Multiple linear regression (for microalbuminuria and QAD) and multivariate logistic regression models with an enter approach were developed to delineate the relationships between the continuous and categorical variables, and the various domains of HL and self-care and between HL and laboratory tests. The entry criteria for the final logistic regression model were a p-value of <0.20 [38], the absence of multicollinearity (Tolerance >0.10, VIF <10) and the normality of residuals (Durbin–Watson: 1.5–2.5) were analyzed for both linear and logistic regressions. A p-value of < 0.05 was considered statistically significant. The IBM SPSS Statistics 26.0 software was used for the analysis. ## 3.1. Analysis of The Sociodemographic, Health-Related, and Clinical Characteristics between Groups Table 1 shows the general characteristics of the participants and their attributes according to the HLS-14. In terms of sociodemographic variables, men were predominant ($53.8\%$), and those above 60 years old ($54.3\%$), black and brown ($88.4\%$), high FHL ($56.3\%$), CHL ($50.3\%$), and CrHL ($56.8\%$), with a mean education of 7.6 years and a mean per capita income of BRL 482.4, with the majority receiving between BRL 300 and 600 ($38.7\%$). Regarding the health-related and clinical characteristics, there was a predominance of a common perception of general health status ($50.3\%$), a diagnosis time of 5–10 years ($38.7\%$), one to five consultations per year ($54.3\%$), no private health insurance ($88.4\%$), non-smokers ($50.3\%$), alcohol consumers ($67.3\%$), those with irregular physical activity ($46.2\%$), SAH ($68.3\%$), and peripheral neuropathy ($66.3\%$). Among the biochemical parameters, there was a predominance of high levels of total cholesterol ($50.8\%$), LDL ($61.8\%$), NHDL ($63.3\%$), triglycerides ($63.8\%$), and glycated hemoglobin ($83.4\%$), with levels within the parameters indicated for HDL ($52.8\%$) and microalbuminuria ($71.4\%$). In the QAD, the averages were low (less than 3.5 days) for the following domains: general diet (mean: 3.0, SD: 2.5), specific diet (mean: 1.5, SD: 0.9), physical activity (mean: 1.6, SD: 2.2), and glycemic monitoring (mean: 1.0, SD: 1.9), and high for the domain’s foot care (mean: 3.7, SD: 2.2), and medication (mean: 4.4, SD: 1.8). Differences were observed between the high and low HL groups for the following variables: FHL, CHL, and CrHL ($p \leq 0.0001$), with a predominance of high HL in all domains in G2, diagnosis time ($$p \leq 0.005$$), with a predominance of high HL diagnosed between 11 and 20 years, private health insurance ($$p \leq 0.025$$), and high HL among those with private health insurance. In the QAD, G2 scored higher in the foot care domain ($$p \leq 0.009$$) and medication domain ($$p \leq 0.013$$). ## 3.2. Sociodemographic Characteristics as Predictors of Health Literacy In the univariate analysis (Table 2), education ($$p \leq 0.058$$) and per capita income ($$p \leq 0.092$$) had the lowest p-values for inclusion in the predictor model of total literacy, and for functional literacy, sex ($$p \leq 0.039$$) and education ($$p \leq 0.005$$). No variable provided the necessary value for analyzing the effect on communicative literacy; only per capita income provided the cut-off value ($$p \leq 0.069$$) in critical literacy. There was no multicollinearity between the selected variables. Multivariate analysis showed that being a woman ($$p \leq 0.024$$) and having a higher level of education ($$p \leq 0.005$$) were predictors of better functional literacy. ## 3.3. Sociodemographic Characteristics and Health Literacy as Predictors of Biochemical Parameters The final analysis included 199 participants (Table 3). In the univariate analysis, the variables that presented the minimum p-value for inclusion in the predictor model of glycated hemoglobin were age ($$p \leq 0.017$$), critical literacy ($$p \leq 0.011$$), and per capita income ($$p \leq 0.066$$); for total cholesterol, sex ($$p \leq 0.003$$), age ($$p \leq 0.092$$), and critical literacy ($$p \leq 0.024$$); for LDL, sex ($$p \leq 0.022$$), age ($$p \leq 0.112$$), and critical literacy ($$p \leq 0.007$$); for HDL, sex ($$p \leq 0.000$$), age ($$p \leq 0.000$$), education ($$p \leq 0.197$$), and functional literacy ($$p \leq 0.082$$); for NHDL, sex ($$p \leq 0.066$$), age ($$p \leq 0.120$$), and critical literacy ($$p \leq 0.106$$); for triglycerides, functional literacy only ($$p \leq 0.005$$), and for microalbuminuria sex ($$p \leq 0.032$$ and $$p \leq 0.011$$), age ($$p \leq 0.028$$), education ($$p \leq 0.179$$), functional literacy ($$p \leq 0.130$$ and $$p \leq 0.192$$), communicative literacy ($$p \leq 0.026$$), and per capita income ($$p \leq 0.163$$ and $$p \leq 0.089$$). There was no multicollinearity between the selected variables. Multivariate analysis showed that low critical literacy was a predictor of glycated hemoglobin control ($$p \leq 0.008$$); female sex ($$p \leq 0.004$$) and low critical literacy ($$p \leq 0.024$$) were predictors of total cholesterol control; the same was found for LDL, female sex ($$p \leq 0.027$$), and low critical literacy ($$p \leq 0.007$$). Only women predicted HDL control ($$p \leq 0.001$$), and low functional literacy predicted triglyceride control ($$p \leq 0.039$$). Women were predictors of high microalbuminuria levels ($$p \leq 0.014$$). ## 3.4. Health Literacy as Predictors of Self-Care The final analysis included 199 participants (Table 4). All assumptions of the multiple linear regression were observed; among the variables, low critical literacy was a predictor of a lower specific diet ($$p \leq 0.002$$) and low total literacy of minor care with medication ($$p \leq 0.027$$). There were no effects of literacy on general diet, physical activity, foot care, and blood glucose monitoring. ## 4. Discussion In this study we aimed to investigate the relationship between HL, sociodemographic factors, and because of the context of DM, the control of biochemical parameters and self-care in 199 patients with DM2 who were randomly selected from two health districts of a municipality in the northern region of Brazil. According to the HLS-14, $50.7\%$ of the people in our sample had a high total level of HL. The assessment of HL in patients with DM2 varied greatly across countries [39]. Previous studies in Brazil using the HLS-14 reported percentages of adequate HL ranging from 51.4 to $56.2\%$ [40,41,42], and low HL ranging from $33.8\%$ to $51.6\%$, specifically among public health service users [32,38]. A higher level of education and being a woman were identified as independent factors of a high FHL level based on sociodemographic information. This finding is consistent with Nutbeam’s functional HL definition of “basic skills in reading and writing to enable individuals to function effectively in everyday situations” [8]. Findings on the association between sex and HL are inconsistent worldwide. The disparities in HL between men and women may be related to the fact that women outperformed men in basic educational indicators, including the “adjusted rate of net school attendance to the initial and final grades of elementary school”, and the “adjusted rate of net school attendance in the high school for 15- to 17-year-olds” between 2016 and 2019 in Brazil, particularly in the Northern region [43]. Furthermore, it may be related to women having a greater familiarity with navigating the health system to deal with health issues, which may provide more opportunities to build their knowledge base [44]. In the analysis of the biochemical control predictors, women were independent predictors of the control of total cholesterol, LDL, and HDL levels, but a predictor of levels above the control of microalbuminuria. Women tend to eat healthier than men, consuming more fruit and vegetables and less meat. However, women may be more sedentary and have a lower success rate of glucose-lowering therapy [45], thereby increasing the risk of dual therapy failure. In contrast to the findings of previous studies that reported that women with DM2 had significantly higher glycated hemoglobin levels [46,47] and poorer glycemic control than men, our findings indicate the opposite. Although data from high-income countries indicate that women are less likely than men to receive the care recommended by guidelines, to adhere to glycemia-lowering therapy, and to meet treatment targets for glycemia and lipids and that women are frequently reported as hurting diabetes self-management [48], studies of the indicators of the line of care for people with diabetes in Brazil reported that in 2019 women used the Popular Pharmacy Program more often to obtain medication ($53.4\%$), had a higher proportion of medical assistance in the past year ($81.0\%$), had their last appointment for DM follow-up at a PHC center ($51.1\%$), and were hospitalized less often due to DM or complications ($13.1\%$) [21]. Additionally, the regulation of glucose homeostasis, treatment response, and psychological factors [46] may contribute to the difference between men’s and women’s biochemical control. Lack of control in biochemical parameters was frequent in both groups of our sample, which may have affected the result that low CrHL was an independent predictor of glycated hemoglobin and total and LDL cholesterol control, and that low FHL was an independent predictor of triglyceride control. Considering the HL concept, it is arguable that HL mediates the control of biochemical parameters indirectly through health decision-making, as the CrHL requires “skills to critically analyze and use the information to exert control over life events and situations” [49]. However, the findings are still inconsistent because HL is significantly more associated with behaviors such as diet, physical activity, and medication use. Therefore, it may not be a direct determinant of biochemical parameters as the control involves a collection of physiological factors that are unique to each individual and can interfere with the effect of adopting general health practices. In addition, patients with higher HL may have similar odds of achieving a clinical goal as those with low HL, regardless of their baseline HL levels. They may engage in self-care activities, including self-monitoring of blood glucose [50]. Despite this, low HL was a predictor of minimal medication care, and low CrHL was a predictor of less adherence to a specific diet. A meta-analysis [51] revealed a weak relationship between HL and medication adherence, possibly due to other factors including adherence determinants. However, in our study this was the only concept of the QAD for which the total score of the questionnaire was decisive, assuming that the set of HL domains influences self-care. The QAD-specific diet concept reflects self-care activities that require a more complex elaboration of nutritional knowledge [37], corroborating the CrHL concept, which necessitates empowering patients to assimilate and implement health-related information and knowledge [52,53,54]. Brazil has the most publications on HL ($20\%$) among South American countries, with most studies focusing on the functional aspect of HL. This study contributes to the literature by assessing the other dimensions of the HL [35]. It is noteworthy that because of the nature of the study, the results represent a specific population at a specific time. Therefore, the generalization of these findings to other sociodemographic and cultural contexts may be limited. Investigation of the potential interactions of other specific clinical aspects was beyond the objective of this study, but their relationship with biochemical parameter control can be considered a limitation. Finally, our study highlights the need for national and regional studies on HL and DM, particularly in terms of self-care activities and their impact on the control of clinical parameters. ## 5. Conclusions In this study, patients who took longer to get diagnosed had private health insurance took better care of their feet, received regular medications, and had higher total HL. We found that women and a higher level of education were the sociodemographic variables that predicted better functional HL. A low critical HL level was a predictor of high value of glycated hemoglobin, total cholesterol, and LDL levels. Women were predictors of high value of total cholesterol, HDL, LDL, and a high level of microalbuminuria control. Only low functional HL was found to be a predictor of triglyceride level. In this DM2 population, only critically low HL was a predictor of a lower specific diet, and low total HL was a predictor of less medication care. ## References 1. Simonds S.K.. Health Education as Social Policy. Health Educ. Monogr. (1974) 2 1-10. DOI: 10.1177/10901981740020S102 2. Peerson A., Saunders M.. Health literacy revisited: What do we mean and why does it matter?. Health Promot. Int. (2009) 24 285-296. DOI: 10.1093/heapro/dap014 3. Adams R.J., Stocks N.P., Wilson D.H., Hill C.L., Gravier S., Kickbusch I., Beilby J.J.. Health literacy—A new concept for general practice?. Aust. Fam. Physician (2009) 38 144-147. PMID: 19283256 4. Passamai M.P.B., Sampaio H.A.C., Dias A.M.I., Cabral L.A.. Functional health literacy: Reflections and concepts on its impact on the interaction among users, professionals and the health system. Interface-Comun. Saúde Educ. (2012) 16 301-314. DOI: 10.1590/S1414-32832012005000027 5. Sørensen K., Van den Broucke S., Fullam J., Doyle G., Pelikan J., Slonska Z., Brand H.. (HLS-EU) Consortium Health Literacy Project European. Health literacy and public health: A systematic review and integration of definitions and models. BMC Public Health (2012) 12. DOI: 10.1186/1471-2458-12-80 6. Andrus M.R., Roth M.T.. Health Literacy: A Review. Pharmacotherapy (2002) 22 282-302. DOI: 10.1592/phco.22.5.282.33191 7. Burnett N.. Education for All: Literacy for Life (2005) 8. Nutbeam D.. Health literacy as a public health goal: A challenge for contemporary health education and communication strategies into the 21st century. Health Promot. Int. (2000) 15 259-267. DOI: 10.1093/heapro/15.3.259 9. DeWalt D.A., Berkman N.D., Sheridan S., Lohr K.N., Pignone M.P.. Literacy and health outcomes. J. Gen. Intern. Med. (2004) 19 1228-1239. DOI: 10.1111/j.1525-1497.2004.40153.x 10. Berkman N.D., Dewalt D.A., Pignone M.P., Sheridan S.L., Lohr K.N., Lux L., Sutto S.F., Swinson T., Bonito A.J.. Literacy and Health Outcomes: Summary (2004) 11. Rudd R.E.. Health literacy skills of U.S. adults. Am. J. Health Behav. (2007) 31 S8-S18. DOI: 10.5993/AJHB.31.s1.3 12. Inoue M., Takahashi M., Kai I.. Impact of communicative and critical health literacy on understanding of diabetes care and self-efficacy in diabetes management: A cross-sectional study of primary care in Japan. BMC Fam. Pract. (2013) 14. DOI: 10.1186/1471-2296-14-40 13. Tomita S., Aoki T., Ohde S., Takahashi O., Kimura T., Matsushima M.. Association between health literacy and multimorbidity: A nationwide, cross-sectional study of a Japanese population. BMJ Open (2022) 12 e052731. DOI: 10.1136/bmjopen-2021-052731 14. Carthery-Goulart M.T., Anghinah R., Areza-Fegyveres R., Bahia V.S., Brucki S.M.D., Damin A., Formigoni A.P., Frota N., Guariglia C., Jacinto A.F.. Performance of a Brazilian population on the test of functional health literacy in adults. Rev. Saúde Pública (2009) 43 631-638. DOI: 10.1590/S0034-89102009005000031 15. Murphy P.W., Davis T.C., Long S.W., Jackson R.H., Decker B.C.. Rapid Estimate of Adult Literacy in Medicine (REALM): A Quick Reading Test for Patients. J. Read. (1993) 37 124-130 16. Rowlands G., Khazaezadeh N., Oteng-Ntim E., Seed P., Barr S., Weiss B.D.. Development and validation of a measure of health literacy in the UK: The newest vital sign. BMC Public Health (2013) 13. DOI: 10.1186/1471-2458-13-116 17. Sørensen K., van den Broucke S., Pelikan J.M., Fullam J., Doyle G., Slonska Z., Kondilis B., Stoffels V., Osborne R.H., Brand H.. Measuring health literacy in populations: Illuminating the design and development process of the European Health Literacy Survey Questionnaire (HLS-EU-Q). BMC Public Health (2013) 13. DOI: 10.1186/1471-2458-13-948 18. Suka M., Odajima T., Kasai M., Igarashi A., Ishikawa H., Kusama M., Nakayama T., Sumitani M., Sugimori H.. The 14-item health literacy scale for Japanese adults (HLS-14). Environ. Health Prevent. Med. (2013) 18 407-415. DOI: 10.1007/s12199-013-0340-z 19. Osborne R.H., Batterham R.W., Elsworth G.R., Hawkins M., Buchbinder R.. The grounded psychometric development and initial validation of the Health Literacy Questionnaire (HLQ). BMC Public Health (2013) 13. DOI: 10.1186/1471-2458-13-658 20. 20. International Diabetes Federation IDF Diabetes Atlas10th ed.International Diabetes FederationBrussels, Belgium2021Available online: https://www.diabetesatlas.org(accessed on 22 March 2022). IDF Diabetes Atlas (2021) 21. Malta D.C., Ribeiro E.G., Gomes C.S., Alves F.T.A., Stopa S.R., Sardinha L.M.V., Pereira C.A., Duncan B.B., Schimidt M.I.. Indicadores da linha de cuidado de pessoas com diabetes no Brasil: Pesquisa Nacional de Saúde 2013 e 2019. Epidemiol. Serviços Saúde (2022) 31 e2021382. DOI: 10.1590/ss2237-9622202200011.especial 22. Freund A., Johnson S.B., Silverstein J., Thomas J.. Assessing daily management of childhood diabetes using 24-hour recall interviews: Reliability and stability. Health Psychol. (1991) 10 200-208. DOI: 10.1037/0278-6133.10.3.200 23. Sallis J.F., Haskell W.L., Wood P.D., Fortmann S.P., Rogers T., Blair S.N., Paffenbarger R.S.. Physical activity assessment methodology in the Five-City Project. Am. J. Epidemiol. (1985) 121 91-106. DOI: 10.1093/oxfordjournals.aje.a113987 24. Weinger K., Butler H.A., Welch G.W., La Greca A.M.. Measuring diabetes self-care: A psychometric analysis of the Self-Care Inventory-Revised with adults. Diabetes Care (2005) 28 1346-1352. DOI: 10.2337/diacare.28.6.1346 25. Toobert D.J., Glasgow R.E., Bradley C.. Assessing diabetes self-management: The summary of diabetes self-care activities questionnaire. Handbook of Psychology and Diabetes (1994) 351-375 26. Lovrić B., Placento H., Farčić N., Lipič Baligač M., Mikšić Š., Mamić M., Jovanović T., Vidić H., Karabatić S., Cviljević S.. Association between Health Literacy and Prevalence of Obesity, Arterial Hypertension, and Diabetes Mellitus. Int. J. Environ. Res. Public Health (2022) 19. DOI: 10.3390/ijerph19159002 27. Marciano L., Camerini A.-L., Schulz P.J.. The Role of Health Literacy in Diabetes Knowledge, Self-Care, and Glycemic Control: A Meta-analysis. J. Gen. Intern. Med. (2019) 34 1007-1017. DOI: 10.1007/s11606-019-04832-y 28. Tang Y.H., Pang S.M.C., Chan M.F., Yeung G.S.P., Yeung V.T.F.. Health literacy, complication awareness, and diabetic control in patients with type 2 diabetes mellitus. J. Adv. Nurs. (2008) 62 74-83. DOI: 10.1111/j.1365-2648.2007.04526.x 29. De Sampaio H.A.C., Carioca A.A.F., Sabry M.O.D., dos Santos P.M., Coelho M.A.M., da Passamai M.P.B.. Letramento em saúde de diabéticos tipo 2: Fatores associados e controle glicêmico. Ciência Saúde Coletiva (2015) 20 865-874. DOI: 10.1590/1413-81232015203.12392014 30. De Santos M.I.P.O., Portella M.R.. Condições do letramento funcional em saúde de um grupo de idosos diabéticos. Rev. Bras. Enferm. (2016) 69 156-164. DOI: 10.1590/0034-7167.2016690121i 31. 31. World Health Organization Health Promotion Glossary of Terms 2021World Health OrganizationGeneva, Switzerland2021. Health Promotion Glossary of Terms 2021 (2021) 32. Rodrigues V.P., Matos L.R., Tenani C.F., Batista M.J.. Literacia em saúde em adultos e idosos diabéticos usuários de sistema público de saúde em municípios de São Paulo. Rev. Ciênc. Med. (2022) 31 e215094 33. Ishikawa H., Nomura K., Sato M., Yano E.. Developing a measure of communicative and critical health literacy: A pilot study of Japanese office workers. Health Promot. Int. (2008) 23 269-274. DOI: 10.1093/heapro/dan017 34. Batista M.J., de Marques A.C.P., Silva Junior M.F., Alencar G.P., Sousa M.D.L.R.D.. Translation, cross-cultural adaptation and psychometric evaluation of Brazilian Portuguese version of the 14-item Health Literacy Scale. Ciênc. Saúde Coletiva (2020) 25 2847-2857. DOI: 10.1590/1413-81232020257.22282018 35. Pavão A.L.B., Werneck G.L., Campos M.R.. Autoavaliação do estado de saúde e a associação com fatores sociodemográficos, hábitos de vida e morbidade na população: Um inquérito nacional. Cad. Saúde Pública (2013) 29 723-734. DOI: 10.1590/S0102-311X2013000800010 36. De Oliveira F.B., Botelho K.K.P., Bezerra A.R., de Azevedo D.I.O., dos Santos-Couto-Paz C.C., Fachin-Martins E.. Cross-cultural adaptation to Brazilian Portuguese of the Michigan Neuropathy Screening Instrument: MNSI-Brazil. Arq. Neuro-Psiquiatr. (2016) 74 653-661. DOI: 10.1590/0004-282X20160094 37. Michels M.J., Coral M.H.C., Sakae T.M., Damas T.B., Furlanetto L.M.. Questionário de Atividades de Autocuidado com o Diabetes: Tradução, adaptação e avaliação das propriedades psicométricas. Arq. Bras. Endocrinol. Metabol. (2010) 54 644-651. DOI: 10.1590/S0004-27302010000700009 38. Tenani C.F., Junior M.F.S., Lino C.M., Sousa M.d.L.R.d., Batista M.J.. The role of health literacy as a factor associated with tooth loss. Rev. Saúde Pública (2021) 55 116. DOI: 10.11606/s1518-8787.2021055003506 39. Abdullah A., Liew S.M., Salim H., Ng C.J., Chinna K.. Correction: Prevalence of limited health literacy among patients with type 2 diabetes mellitus: A systematic review. PLoS ONE (2022) 17. DOI: 10.1371/journal.pone.0261430 40. Fede B.P., Artuso L.R., Tenani C.F., Lino C.M., Batista M.J.. Health literacy and prevention practices for COVID-19: A cross-sectional study in Brazilian adults. RSD (2021) 10 e425101623839. DOI: 10.33448/rsd-v10i16.23839 41. Polonski T.C., Zanin L., Oliveira A.M.G., Dutra E.R., da Filho D.A.S., Flório F.M.. Influência da inclusão digital na alfabetização em saúde de idosos. ETD-Educ. Temática Digit. (2022) 24 584-597. DOI: 10.20396/etd.v24i3.8661696 42. Silva-Junior M.F., Rosário de Sousa M.d.L., Batista M.J.. Health literacy on oral health practice and condition in an adult and elderly population. Health Promot. Int. (2020) 36 933-942. DOI: 10.1093/heapro/daaa135 43. 43. IBGE-Instituto Brasileiro de Geografia e Estatística Pesquisa Nacional por Amostra de Domicílios Contínua: Educação 2019IBGERio de Janeiro, Brazil2020978-65-87201-09-2. Pesquisa Nacional por Amostra de Domicílios Contínua: Educação 2019 (2020) 44. Lee H.Y., Lee J., Kim N.K.. Gender Differences in Health Literacy Among Korean Adults: Do Women Have a Higher Level of Health Literacy Than Men?. Am. J. Men’s Health (2015) 9 370-379. DOI: 10.1177/1557988314545485 45. Kautzky-Willer A., Harreiter J.. Sex and gender differences in therapy of type 2 diabetes. Diabetes Res. Clin. Pract. (2017) 131 230-241. DOI: 10.1016/j.diabres.2017.07.012 46. Chiu C.-J., Wray L.A.. Gender Differences in Functional Limitations in Adults Living with Type 2 Diabetes: Biobehavioral and Psychosocial Mediators. Ann. Behav. Med. (2010) 41 71-82. DOI: 10.1007/s12160-010-9226-0 47. Duarte F.G., da Silva Moreira S., da Almeida M.C.C., de Souza Teles C.A., Andrade C.S., Reingold A.L., Moreira E.D.. Sex differences and correlates of poor glycaemic control in type 2 diabetes: A cross-sectional study in Brazil and Venezuela. BMJ Open (2019) 9 e023401. DOI: 10.1136/bmjopen-2018-023401 48. Alexandre K., Campbell J., Bugnon M., Henry C., Schaub C., Serex M., Elmers J., Desrichard O., Peytremann-Bridevaux I.. Factors influencing diabetes self-management in adults: An umbrella review of systematic reviews. JBI Evid. Synth. (2021) 19 1003-1118. DOI: 10.11124/JBIES-20-00020 49. Liu C., Wang D., Liu C., Jiang J., Wang X., Chen H., Ju X., Zhang X.. What is the meaning of health literacy? A systematic review and qualitative synthesis. Fam. Med. Community Health (2020) 8 e000351. DOI: 10.1136/fmch-2020-000351 50. Rothman R.L.. Influence of Patient Literacy on the Effectiveness of a Primary Care–Based Diabetes Disease Management Program. JAMA (2004) 292 1711. DOI: 10.1001/jama.292.14.1711 51. Zhang N.J., Terry A., McHorney C.A.. Impact of Health Literacy on Medication Adherence. Ann. Pharmacother. (2014) 48 741-751. DOI: 10.1177/1060028014526562 52. Guo X., Zhai X., Hou B.. Adequacy of health literacy and its effect on diabetes self-management: A meta-analysis. Aust. J. Prim. Health (2020) 26 458. DOI: 10.1071/PY20079 53. Lai A.Y., Ishikawa H., Kiuchi T., Mooppil N., Griva K.. Communicative and critical health literacy, and self-management behaviors in end-stage renal disease patients with diabetes on hemodialysis. Patient Educ. Couns. (2013) 91 221-227. DOI: 10.1016/j.pec.2012.12.018 54. Lee Y.-J., Shin S.-J., Wang R.-H., Lin K.-D., Lee Y.-L., Wang Y.-H.. Pathways of empowerment perceptions, health literacy, self-efficacy, and self-care behaviors to glycemic control in patients with type 2 diabetes mellitus. Patient Educ. Couns. (2016) 99 287-294. DOI: 10.1016/j.pec.2015.08.021
--- title: 'Potential Early Markers for Breast Cancer: A Proteomic Approach Comparing Saliva and Serum Samples in a Pilot Study' authors: - Indu Sinha - Rachel L. Fogle - Gizem Gulfidan - Anne E. Stanley - Vonn Walter - Christopher S. Hollenbeak - Kazim Y. Arga - Raghu Sinha journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9966955 doi: 10.3390/ijms24044164 license: CC BY 4.0 --- # Potential Early Markers for Breast Cancer: A Proteomic Approach Comparing Saliva and Serum Samples in a Pilot Study ## Abstract Breast cancer is the second leading cause of death for women in the United States, and early detection could offer patients the opportunity to receive early intervention. The current methods of diagnosis rely on mammograms and have relatively high rates of false positivity, causing anxiety in patients. We sought to identify protein markers in saliva and serum for early detection of breast cancer. A rigorous analysis was performed for individual saliva and serum samples from women without breast disease, and women diagnosed with benign or malignant breast disease, using isobaric tags for relative and absolute quantitation (iTRAQ) technique, and employing a random effects model. A total of 591 and 371 proteins were identified in saliva and serum samples from the same individuals, respectively. The differentially expressed proteins were mainly involved in exocytosis, secretion, immune response, neutrophil-mediated immunity and cytokine-mediated signaling pathway. Using a network biology approach, significantly expressed proteins in both biological fluids were evaluated for protein–protein interaction networks and further analyzed for these being potential biomarkers in breast cancer diagnosis and prognosis. Our systems approach illustrates a feasible platform for investigating the responsive proteomic profile in benign and malignant breast disease using saliva and serum from the same women. ## 1. Introduction Breast cancer is the second leading cause of mortality for women in the United States and is estimated to result in 43,250 deaths in 2022 [1]. Early detection for breast cancer can reduce breast cancer-related mortality. Among women aged 50 years and older, reports have demonstrated a 20–$40\%$ reduction in breast cancer mortality in women who underwent mammography and clinical breast examination [2]. Among women screened at younger ages (40–49 years), mortality rates decrease by 13–$23\%$. A detailed analysis of these data suggests that a survival rate of $96\%$ can be achieved if women underwent mammography every three months [3]. However, the cost and risks of mammography (such as radiation exposure) with increased frequency of use are not ideal. Furthermore, despite accurate mammography diagnoses, the screening procedure may result in relatively high rates of false-positive ($56\%$) and false-negative ($22\%$) diagnoses in women younger than 50 years, especially in women with dense parenchymal breast tissue [4,5]. Because of these shortcomings, there is a need to develop additional diagnostic methods to further enhance the sensitivity and specificity of breast cancer detection, particularly in women with dense breast tissue, and thereby reducing the need for unnecessary biopsies. As a complementary approach to mammography, determination of biomarkers in saliva and/or serum could be a critical measurement for the early detection of breast cancer. Saliva is considered an easily obtained clear fluid, which is indicative of an individual’s protein profile at the time of collection. Testing saliva as a diagnostic fluid meets the criteria for an inexpensive, non-invasive, reliable, and relatively simple procedure that can be repeated with a minimum discomfort to patients. In addition, providing a saliva sample may cause less anxiety in study participants than providing a blood sample [6]. The clinical utility of saliva as a diagnostic fluid is being recognized in several diseases, including cancer [7,8,9]. A meta-analysis revealed that salivary proteins represent good biomarkers for diagnosis of several cancer types including that of the breast [10,11]. Earlier studies focused on transcriptomic and proteomic signatures in saliva revealing sensitive and specific biomarkers for the detection of breast cancer using two-dimensional difference gel electrophoresis (2D-DIGE) [9]. A recent review systematically captures proteomics-based technologies for comparing dysregulated proteins in breast cancer in several body fluids including saliva and serum [12]. A variety of methods including surface enhanced laser desorption/ionization [13] and nano-liquid chromatography-quadrupole-time-of-flight technology [14] have been utilized for discovering biomarkers for breast cancer in saliva and plasma. Moreover, the isobaric tags for relative and absolute quantitation (iTRAQ) technique has been utilized for identifying salivary proteins as potential biomarkers for breast disease [15]. In a comparison study, the global-tagging iTRAQ technique was found to be more sensitive than the cysteine-specific Isotope-coded affinity tag (cICAT) method, which in turn was equally sensitive as the 2D-DIGE technique [16]. iTRAQ has an advantage over ICAT and other methods since several samples can be analyzed simultaneously, and helps reduce the time spent for mass spectrometry analysis [17]. Another advantage of iTRAQ is the possibility of identifying proteins with varying pI and molecular weights. In addition, using iTRAQ the relative and absolute quantification is possible across different sample states for a synchronous comparison of biological fluids such as saliva and serum from normal, benign and malignant breast disease cases. We hypothesize that protein changes occurring in breast cells and their environment will be reflected in the saliva and serum of breast cancer patients. We further hypothesize that protein changes in the benign stages will differ from those in the malignant stages of breast disease. In the present study, we compared the proteomic profile in saliva and serum samples from women without breast disease (referred to as normal in our study), with benign breast disease, and with malignant breast disease using the iTRAQ technique. Several proteins were identified in both the benign and malignant groups that could be potential biomarkers for early detection and prognosis of breast cancer in women. ## 2.1. Proteins Identified in Saliva Samples A total of 591 proteins were identified following iTRAQ analysis in the saliva samples (Table S1). Of these, the expression of 110 proteins were statistically different ($p \leq 0.05$) in samples from either benign/normal (B/N), malignant/normal (M/N) or malignant/benign (M/B) comparisons (Table 1). Proteins were considered down-regulated when the pooled summary ratio was less than 1, and up-regulated when the pooled ratio was greater than 1. Additionally, 44 proteins in B/N samples (16 up-regulated, 28 down-regulated), 67 proteins in M/N samples (26 up-regulated, 41 down-regulated) and 35 proteins in M/B samples (17 up-regulated, 18 down-regulated) were observed as differentially expressed. It was clear that there were more down-regulated proteins in saliva samples in each comparison. Eleven proteins (ANXA1, PRELP, PRDX1, H2B2F, GSTP1, PRPC, CDC42, K2C1, PRTN3, CRNN, 6PGD) were significantly down-regulated in both B/N and M/N comparisons ($p \leq 0.05$). Six proteins were significantly up-regulated in both B/N and M/N (CYTS, CAH6, CATD, LG3BP, QSOX1, AMY1B) ($p \leq 0.05$). S10A8 was greater than 1 for benign and less than 1 for malignant diagnosis ($p \leq 0.05$), while 4 proteins (CYTN, AMY2B, PIGR, PERL) were high in both M/N and M/B ($p \leq 0.05$), 8 proteins (ANXA1, PSB3, CATG, H4, TALDO, TKT, TGM3, S10A8) were low in both M/N and M/B ($p \leq 0.05$) andANXA1 was down-regulated in all three comparisons. Interestingly, 10 proteins had >2 fold change in M/B (A2MG, RN150, MYO7A, PSA1, CERU, AFAM, BPIA2, SH3L1, HIS1, TTHY), 11 proteins had >2 fold change in B/N (LEG1H, DAB2P, FAM3D, CAH6, QSOX1, RAP1B, ITLN1, C251, ACSL3, S10A8, AMY1B) and 14 proteins had >2 fold change in M/N (CYTS, AMY2B, PIGR, CAH6, PLGT3, STOM, QSOX1, RETN, CYTC, AMY1B, ACTN2, AMYP, ZA2G, STAT). On the other hand, among the down-regulated proteins, 18 were <0.5 fold or less in B/N (RUSC1, SPRL1, RN150, PSA5, PRELP, ITA1, H2B2F, SPB13, CYTA, K2C6B, PSA1, GSTP1, PRPC, K2C1, PSA, PRTN3, HIS1, 6PGD), 24 proteins in M/N (ANXA1, PRELP, PSB3, NUDT5, K2C5, MMP9, CATG, PNPH, H2B2F, H4, ANXA5, ISK7, PRPC, K2C1, TGM3, ARSA, SH3L1, DEF3, S10A8, CPPED, K2C4, PRTN3, CRNN, 6PGD) and 10 proteins in M/B (PSB3, CATG, FAM3D, H2B1L, TERA, HSP76, H4, K1C10, S10AC, S10A8). Considering the AUC values calculated from receiver operating characteristic (ROC) curve analysis for saliva proteins to distinguish between breast tumor and normal breast tissue (Table 1), 14 proteins were designated as outstanding (>$90\%$) and 22 proteins each with excellent (80–$90\%$) and acceptable (70–$80\%$) ratings for their diagnostic ability [18]. ## 2.2. Proteins Identified in Serum Samples A total of 371 proteins were identified in the serum samples by iTRAQ analysis (Table S2). Of these, the expressions of 56 proteins were significantly ($p \leq 0.05$) altered in the samples from either B/N, M/N or M/B comparisons (Table 2). In addition, 29 proteins in B/N samples (13 up-regulated, 16 down-regulated), 30 proteins in M/N samples (11 up-regulated, 19 down-regulated) and 15 proteins in M/B samples (4 up-regulated, 11 down-regulated) were observed as differentially expressed. Similar to saliva, a greater number of proteins were down-regulated in serum samples in each comparison. Seven proteins (APOB, TRFE, A2MG, HEP2, KAIN, TSP1, THBG) were significantly down-regulated in both B/N and M/N comparisons ($p \leq 0.05$) and 4 proteins (PRDX2, A1BG, FIBA, APOH) were significantly up-regulated in both B/N and M/N ($p \leq 0.05$). HBB was up-regulated while 4 proteins (TRFE, APOA1, TSP1, APOA2) were down-regulated in both M/N and M/B comparisons. TSP1 was down-regulated in all the three comparisons. In addition, 4 proteins (HBB, VINC, CD5L, PCD20) showed more than 1.5-fold change in M/B, 8 proteins (DYST, VWF, CO6, PRDX2, A1BG, LUM, CE290, APOH) were changed by >1.5 fold in B/N, and 7 proteins (HBB, PRDX2, A1AG1, A1BG, BLVRB, FIBA, APOH) were up-regulated by >1.5 fold in M/N. On the other hand, 4 proteins (TRFE, CATD, COL11, DYHC1) were down-regulated by 0.5 fold or lower in B/N, 6 proteins (TRFE, SOX, TSP1, MED30, A1AT, SMC3) were down-regulated in M/N and 3 proteins (APOA1, GPKOW, ERBIN) were down-regulated in M/B. AUC values for serum samples indicated that 9 proteins demonstrated outstanding (>$90\%$), 8 proteins showed excellent (80–$90\%$) and 7 proteins showed acceptable (70–$80\%$) diagnostic performance (Table 2). ## 2.3. Enrichment Analysis of Proteins in Saliva Samples GO enrichment analysis showed that in all three comparisons (B/N, M/N and M/B), most salivary proteins were involved in exocytosis, secretion, immune response, neutrophil mediated immunity and cytokine-mediated signaling pathway, but the number of proteins associated with these processes varied between groups (Figure 1A). Most proteins were localized in the extracellular exosome, extracellular space, secretory granule lumen, secretory vesicle or cytoplasmic vesicles, and again the number of proteins varied among the groups. In terms of molecular functions, the proteins were annotated as enzyme inhibitor activity, calcium ion binding, endopeptidase regulator activity and peptidase activity (Figure 1A, Table S3). KEGG pathway analysis identified a total of 25, 9 and 16 pathways ($p \leq 0.05$) and Reactome pathway analysis identified 44, 36 and 25 pathways ($p \leq 0.05$) for the B/N, M/N and M/B groups of saliva samples, respectively. The overall comparison among the groups can be found in Table S3. The top 10 enriched Reactome pathways related to each of the group samples are shown for B/N (Figure 1B), M/N (Figure 1C) and M/B (Figure 1D) related to the significant proteins in each group. The saliva proteins identified from iTRAQ analysis of B/N, M/N and M/B groups were mainly involved in the neutrophil degranulation and innate immune response based on Reactome pathway analysis ($p \leq 0.05$). ## 2.4. Enrichment Analysis of Proteins in Serum Samples The serum proteins were mostly involved in the regulation of biological processes, were located primarily in organelles or extracellular region, and mostly displayed binding, catalytic or structural molecular activities (Figure 2A, Table S4). KEGG pathway analysis identified a total of 10, 11 and 13 pathways ($p \leq 0.05$) and Reactome pathway analysis identified 59, 45 and 43 pathways ($p \leq 0.05$) for B/N, M/N and M/B group of samples, respectively (Figure 2B–D, Table S4). The serum proteins identified from iTRAQ analysis of B/N group were mainly involved in platelet degranulation ($$p \leq 1.64$$ × 10−12), response to elevated platelet cytosolic Ca2+ ($$p \leq 2.31$$ × 10−12), platelet activation, signaling and aggregation ($$p \leq 8.61$$ × 10−10). While M/N group serum proteins were engaged in chylomicron assembly ($$p \leq 2.90$$ × 10−9), chylomicron remodeling ($$p \leq 2.90$$ × 10−9), and retinoid metabolism and transport ($$p \leq 5.34$$ × 10−8). Further, in the serum samples from M/B group the proteins were also involved in similar proteins as B/N group with lesser p values; platelet degranulation ($$p \leq 1.24$$ × 10−9), response to elevated platelet cytosolic Ca2+ ($$p \leq 1.56$$ × 10−9), platelet activation, signaling and aggregation ($$p \leq 8.13$$ × 10−8). ## 2.5. Protein-Protein Interaction (PPI) Networks for Proteins in Saliva The network of B/N consisted of 798 interactions among 28 significant saliva proteins and 602 of their first interacting neighbors. Among the major hub proteins in the B/N group, CDC42, HSP7C, PSA1 and PSA5 were down-regulated and had multiple interacting partners, whereas S10A8, CATD, FINC and LG3BP were up-regulated with a moderate number of interactions (Figure 3). The network of M/N consisted of 620 interactions among 44 significant proteins and 521 of their first interacting neighbors. The major hubs in the M/N group consisted of CDC42, H2AX, PSB3 and PDIA1 which were down-regulated and had several to moderate interacting partners while LGS3BP, STOM, ACTN2 and VPS41 were up-regulated with fewer interacting partners (Figure 3). In addition, the network of M/B consisted of 522 interactions among 19 significant proteins and 407 of their first interacting neighbors. Among the major hub proteins in the M/B group, TERA, FINC, HSP7C, PSB3, S10AB and ANXA1 were down-regulated and had several to moderate interacting partners, whereas HS71A, PSA1, A2MG and TTHY were up-regulated with a moderate number of interactions (Figure 3). All the PPIs in each of the groups in saliva are listed in Table S5. ## 2.6. Protein–Protein Interaction Networks for Proteins in Serum Overall, there were fewer interactions in serum among the smaller number of significant proteins and far fewer interacting partners compared to the respective PPI networks among saliva proteins. In particular, the network of B/N consisted of 161 interactions among 20 significant proteins and 151 of their first interacting neighbors. Among the major hub proteins in the B/N group, DYHC1, APOB, CATD, TSP1 and A2MG were down-regulated and had moderate interacting partners and were down-regulated, whereas CE290, PRDX2, CADH5 and APOC1 were up-regulated with a moderate number of interactions (Figure 3). The network of M/N consisted of 175 interactions among 20 significant proteins and 165 of their first interacting neighbors. The major hubs in the M/N group consisted of MED30, SMC3, APOB and APOA1, which were down-regulated and had several to moderate number of interacting partners, whereas PRDX2, HBB, FIBA and FETUA were up-regulated with fewer interacting partners (Figure 3). Additionally, the network of M/B consisted of 66 interactions among 11 significant proteins and 61 of their first interacting neighbors. Among the major hub proteins in the M/B group, APOA1, APOA2, GPKOW and TSP1 were down-regulated and had moderate interacting partners whereas VINC and HBB were up-regulated with a moderate number of interactions (Figure 3). All the PPIs in each of the groups for the serum samples are listed in Table S5. ## 2.7. Protein Ratios across Serum and Saliva in B/N and M/N Groups Following the iTRAQ analysis, proteins commonly identified in saliva and serum samples of the B/N and M/N groups were fitted without interaction by two-way ANOVA models. As a result, we identified 17 proteins that were significantly ($p \leq 0.05$) different among serum and saliva (Table S6). A subset of these proteins that were detected in 6–8 saliva and serum samples are shown in Figure 4. These included alpha-1B-glycoprotein precursor (A1BG), fibrinogen alpha chain isoform alpha-E preproprotein (FIBA), alpha-1-antichymotrypsin precursor (AACT), extracellular matrix protein 1 isoform 3 precursor (ECM1), peroxiredoxin-2 (PRDX2), 78 kDa glucose regulated protein precursor (ERP78) and galactin-3-binding protein precursor (LG3BP). PRDX2, A1BG, ECM1, ERP78 and FIBA showed lower ratios in saliva samples when compared to serum samples while LG3BP and AACT ratios were higher in saliva in contrast to serum samples of the same subjects. Upon further comparison between B/N and M/N in the saliva and serum samples, TSP1 was found to be significantly different in serum ($p \leq 0.05$). All the above proteins were presently measured as a ratio following iTRAQ analysis and need to be validated using actual quantitation by either Western blot analysis or ELISA in the future. ## 2.8. Prognostic Performance Analysis When the association of the expression levels of genes encoding significant proteins with prognostic outcome was investigated through survival analyses, all protein sets in B/N and M/N saliva and serum samples (Figure 5), except in the M/B group serum data, indicated high impact on overall patient survival ($p \leq 0.05$) in breast cancer. According to the parameters of HR and p-values, the prognostic performance of the protein sets in the saliva data was observed to be more significant than the protein sets in the serum data for all the groups. In addition, the comparisons of the B/N and M/N group samples had better prognostic performance than the M/B group samples in both the saliva and serum data. The prognostic performance of each gene encoding significant protein based on high-risk vs. low-risk of the dataset for invasive breast carcinoma (BRCA) obtained from The Cancer Genome Atlas (TCGA) were used to draw the Kaplan–Meier (KM) plots and are presented as box plots for the significant proteins in B/N, M/B and M/N groups in saliva as well as in serum (Figures S1–S6). ## 3. Discussion When comparing the proteomic profile of saliva and serum samples from the same women with a diagnosis of benign or malignant state of the breast disease relative to those of women with no breast disease, we have identified proteins that differed in expression levels. Further, analyzing the significant protein changes suggested involvement of several biological pathways and functionalities. We constructed potential protein–protein interaction networks among hub proteins detected in serum and saliva samples and their interacting partners to identify potential biomolecular markers to be explored for diagnosis or prognosis. Since mammography can lead to false positives and anxiety in subjects, utilizing more than one biomarker from our analysis would greatly improve early diagnosis of breast cancer using non-invasive testing in saliva and/or serum. Interestingly, several proteins in our saliva and serum proteomic analysis qualified for outstanding and excellent diagnostic power based on the AUC values (Table 1 and Table 2). However, ROC curve analysis was based on RNA-Seq data from breast tumor tissues compared to normal tissues from the TCGA database; therefore, it is worthwhile to investigate which of these secretory proteins succeed as biomarkers for early breast cancer diagnosis using a larger cohort. Several circulating proteins have been identified in the plasma and serum of patients with breast cancer [19] but we still lack highly sensitive and specific biomarkers. Below, we discuss some of the pertinent proteins that were significantly altered among the different groups (B/N, M/M and M/B) in our analysis of saliva and/or serum and their relevance for a potential biomarkers for breast cancer. Saliva: Fibronectins bind cell surfaces and various compounds including collagen, fibrin, heparin, DNA and actin. In our analysis, fibronectin isoform 11 preproprotein (FINC), was up-regulated 1.97 fold in B/N ($p \leq 0.05$) and did not change in M/N, whereas it was down-regulated at 0.56 fold in the M/B group ($p \leq 0.05$). It has been reported that a liquid biopsy detecting FINC on circulating extracellular vesicles could be a promising method to detect early breast cancer [20]. Indeed, FINC was one of the hub proteins that had 15 interacting partners and has an AUC of $93.05\%$ with an outstanding diagnostic power. The SPARK-like isoform 1 protein 1 precursor (SPRL1) is an extracellular matrix glycoprotein that has been implicated in the pathogenesis of several disorders, including cancer. In our analysis, SPRL1 was down-regulated at 0.15 fold ($p \leq 0.05$) in B/N and 0.44 fold in M/N ($p \leq 0.05$). Previously, a significantly reduced expression SPRL1 was observed in human breast cancer tissues compared to that in normal breast epithelial tissues, at both mRNA and protein levels. In addition, the down-regulation of SPRL1 was significantly correlated with lymphatic metastasis [21]. SPRL1 was found to have an outstanding diagnostic power with an AUC of $96.5\%$. Histone H2AX (H2AX) is a type of histone protein from the H2A family encoded by the H2AFX gene. An important phosphorylated form is γH2AX (S139), which forms when double-strand breaks appear. In our analysis, H2AX was marginally up-regulated at 1.16 fold in B/N ($p \leq 0.05$) but down-regulated significantly in M/N at 0.5 fold ($p \leq 0.05$) and at 0.39 fold in M/B group ($p \leq 0.05$). Evaluating the formation of γH2AX in breast tumor tissue could potentially be a sensitive means of early breast cancer detection as these levels may reflect endogenous genomic instability in breast cancerous tissues [22]. Additionally, the detection of γH2AX could benefit early cancer screening, with breast cancer included [23]. Even though in our analysis we found H2AX to be down-regulated in M/N group, it is important to note that we detected H2AX in saliva and this could be conveniently used for monitoring breast disease. H2AX was one of the hub proteins that had 102 interacting partners and had an AUC of $93.7\%$ with an outstanding diagnostic power. Cystatin-SN precursor (CYTN) belongs to the type 2 cystatin superfamily, which restricts the proteolytic activities of cysteine proteases. In our analysis, CYTN was marginally up-regulated at 1.94- and 1.96-fold in M/N and M/B groups, respectively, while only a moderate change of 1.09 was noted in benign samples ($p \leq 0.05$). CYTN promotes cell proliferation, clone formation and metastasis in breast cancer cells and has been proposed to be a potential prognostic biomarker and therapeutic target for breast cancer [24]. CYTN was found to have an outstanding diagnostic power with AUC of $93.1\%$. Serum: Hemoglobin subunit beta (HBB) is a member of the globin family, a structurally conserved group of proteins often containing the heme group, which have the ability to reversibly bind O2 and other gaseous ligands in erythrocytes [25]. In our analysis, HBB was up-regulated 1.97- and 2.04-fold in M/N and M/B groups, respectively ($p \leq 0.05$), but moderately down-regulated in B/N group. This protein has been implicated as a potential biomarker of breast cancer progression [26]. It was one of the hub proteins that had 5 interacting partners and had an AUC of $93.7\%$ with outstanding diagnostic power. Retinol-binding protein 4 (RET4) is a recently identified adipokine that is elevated in patients with obesity or type 2 diabetes [27]. The iTRAQ analysis revealed that RET4 was up-regulated 1.48 fold in M/N group ($p \leq 0.05$) and may be detectable earlier as suggested from our study (1.40 fold increase in B/N, $p \leq 0.05$). In a case control study, higher serum RET4 levels were associated with the risk of breast cancer [28]. It was one of the hub proteins with just 1 interacting partner (TTHY) and had an AUC of $93.5\%$ with outstanding diagnostic power. Cadherin-5 isoform X1 (CADH5) is a member of the cadherin family which are calcium-dependent cell adhesion proteins. Previously, using a glycoproteomic approach CADH5 emerged as a novel biomarker for metastatic breast cancer [29]. The iTRAQ analysis revealed that CADH5 was up-regulated 1.18 fold in M/N group ($p \leq 0.05$) and was most likely detected in the benign stage of breast cancer as suggested from our study (1.50 fold increase in B/N, $p \leq 0.05$). It was one of the hub proteins with 8 interacting partners and has an AUC of $91.6\%$ with outstanding diagnostic power. Von Willebrand factor preproprotein (VWF) is a large multimeric plasma glycoprotein that plays important roles in normal hemostasis. VWF can also impact cancer cell metastasis [30] and more recently it has been shown by the same group that breast cancer cells mediate endothelial cell activation and promote VWF release [31]. However, in our analysis VWF was elevated 1.57 fold in serum samples of benign breast cancer diagnosis ($p \leq 0.05$), so this may be a potential marker that may provide damage to endothelial cells early in the disease. It was one of the hub proteins with 4 interacting partners and had an AUC of $91.6\%$. Alpha-2-macroglubulin isoform X1 (A2MG) is a protease inhibitor and cytokine transporter covering a wide range of proteases, including trypsin, thrombin and collagenase. Even though it has a high AUC value for diagnosis ($92.4\%$), it was modestly down-regulated in both benign and malignant samples ($p \leq 0.05$). Others have reported it to be lower [32] or higher [14] in breast cancer. Peroxiredoxin-2 (PRDX2), and peroxiredoxins in general, catalyze the reduction reaction of peroxide and maintain the balance of intracellular H2O2 levels. In our analysis, PRDX2 was up-regulated 1.89- to 2.16-fold in B/N and M/N groups, respectively ($p \leq 0.05$), but exhibited no change in M/B group. High mRNA expression of PRDX$\frac{1}{2}$/$\frac{4}{5}$/6 was significantly associated with shorter relapse-free survival in breast cancer patients [33]. It was one of the hub proteins that had 16 interacting partners and had an AUC of $80.2\%$ with excellent diagnostics power. Among the proteins commonly identified across serum and saliva, PRDX2, LG3BP and TSP1 are promising for further investigation to distinguish the benign from the malignant stage of breast cancer in a larger cohort. Moreover, some of the proteins identified in the present study have been associated with Hallmarks of Cancer specific proteins in breast cancer [34], including FINC, proteasome subunit alpha type-1 isoform 2 (PSA1), proteasome subunit alpha type-5 isoform 1 (PSA5), proteasome subunit beta type-3 (PSB3), phosphoglycerate kinase 1 (PGK1), heat shock cognate 71 kDa protein isoform 1 (HSP7C) and glutathione S-transferase p (GSTP1) which may provide insights into the early detection of breast disease. We have further identified several proteins in saliva, including AMY1B, AMY2B, BPIB2, CPPED, DEF3, H2A2A, H2BC18, ISK7, LEG1H, PNP, PRELP, SPB13, STAT, QSOX1, RNF150 and VPS41, and in serum, namely, CEP290, CO8B, CO6, CPN2, GPKOW, HEP2 and PIPOX which have not been reported in the literature to previously be associated with breast cancer. These are suggestive potential candidate biomarkers for the early detection of breast cancer. ## 4.1. Study Subjects Subjects were recruited at the Hershey Medical Center Breast Cancer Center upon their routine visit for a mammogram. Sixty healthy adult women with no breast disease, 13 adult women with a diagnosis of benign breast disease and 15 adult women with a diagnosis of malignant breast disease were enrolled in the study. All participants provided written informed consent, following the protocol approved by the Pennsylvania State University Institutional Review Board (STUDY00005159). Subjects were recruited based on the following inclusion criteria: English-speaking female volunteers, 25–85 years of age, who had undergone mammogram examination and were currently non-smokers. Exclusion criteria included any evidence of cancer other than the breast and undergoing treatment for breast cancer prior to saliva and blood sample collection. When there was any abnormality detected on the mammogram, subjects were advised to undergo a biopsy. The diagnosis on the breast biopsy tissues following the surgical pathology reporting on Hematoxylin and Eosin-stained sections were provided by Board Certified Pathologists in the Department of Pathology, at the Penn State College of Medicine. Table 3 provides the characteristics of the subjects used for iTRAQ analysis. ## 4.2. Collection and Storage of Biological Samples Saliva and blood samples were collected in the fasting state. Saliva samples were centrifuged at 10,000 rpm for 15 min at 4 °C and the clear supernatants were aliquoted in 1 mL screw capped vials. For serum, clotted blood was separated and centrifuged at 1300 rpm for 15 min at 4 °C. The clear serum was aliquoted in 1 mL screw cap vials. All biological samples were stored at −80 °C until analyzed. ## 4.3. Sample Processing and Labeling Procedure for iTRAQ Analysis Eight saliva and serum samples from the participants in each group were processed for iTRAQ analysis as described earlier [35,36]. The serum samples but not the saliva samples were depleted of the 6 most abundant proteins including albumin, IgG, IgA, transferrin, haptoglobin and antitrypsin using a Multiple Affinity Removal System LC Column (Agilent Technologies, Santa Clara, CA). Briefly, equal amounts of protein (100 μg) from each sample were digested with trypsin and subsequently labeled with one of 8 unique isobaric tags using the iTRAQ® Reagent-8Plex Multiplex kit (AB SCIEX, Framingham, MA). Quantitative fragments, ranging from 113 to 121 Daltons, following MS/MS fragmentation shows proportionally how much of each peptide peak came from each of the individually labeled samples. The Penn State College of Medicine’s Proteomic Core Facility received the tagged samples which were subsequently resolved by two-dimensional liquid chromatography prior to triple time-of-flight (TOF) mass spectrometry. Peptide identification, protein grouping and subsequent protein quantitation were done using the Paragon algorithm as implemented in Protein Pilot 5.0 software (ProteinPilot 5.0, which contains the Paragon Algorithm 5.0.0.0, build 4632 from ABI/MDS- SCIEX), searching the NCBI human database plus a list of 389 common contaminants (see Appendix A for details). The datasets presented in this software are ratios of samples with defined diagnoses (e.g., B/N, M/N or M/B). Ratios significantly greater than 1 in a B/N ratio indicates a differential increase in protein in benign compared to normal (similarly for M/N and M/B), and ratios significantly less than 1 in a B/N ratio indicates a differential decrease in benign compared to normal (similarly for M/N and M/B). ## 4.4. Gene Set Over-Representation Analysis Functional annotations associated with the significant proteins determine as a result B/N, M/N and M/B comparisons in the saliva and serum data were identified in terms of biological processes, signaling and metabolic pathways by over-representation analyses using the Consensus PathDB [37]. As the sources for pathway databases, Kyoto Encyclopedia of Genes and Genomes (KEGG) [38] and Reactome [39] were used. While the annotation of the biological process, cellular components and molecular function were determined using Gene Ontology (GO) [40] annotations. The significance of over-representations was evaluated by adjusted-p-values via Fisher’s Exact Test, followed by Benjamini-Hochberg correction. Functional enrichment results with an adjusted p-value < 0.05 were considered statistically significant. ## 4.5. Construction of Protein–Protein Interaction Network Using physical protein–protein interaction (PPI) data consisting of 68,948 interactions among 10,835 proteins which were experimentally detected in human and stored in the BioGRID database (version 4.4.210) [41], PPI networks were constructed around the significant proteins found in three comparisons (B/N, M/N, M/B) in saliva and serum data by enriching them with their first-neighbor interactions. The visualization of the PPI networks was performed via Cytoscape (v.3.7.0) software [42]. ## 4.6. Prognostic Performance Analysis The prognostic performance of the significant proteins in all three comparisons (B/N, M/N, M/B) for saliva and serum data were assessed with survival analyses according to the established pipeline [43,44] using RNA-Sequencing (RNA-Seq) data from 1102 patients suffering from invasive breast carcinoma obtained from TCGA database. Each individual was classified into low- and high-risk groups according to their risk score, the prognostic index (PI), according to the linear component of the Cox model with the equation PI = β1×1 + β2×2 +… + βp×p [1] where xi is the expression value of each gene and βi is the coefficient obtained from the Cox fitting. Survival analyses were performed using the survival package (v.3.3.1) [45] in R (v.4.0.4). KM survival plots provided visualizations for the survival time statistics calculated by log-rank test, and the log-rank p-value < 0.05 was considered as the cut-off to describe the statistical significance of survival in each group. In addition, the HR was calculated to quantify the relative hazard of each KM plot. ## 4.7. ROC Curve Analysis ROC curve analysis was performed for each significant protein in each of the three comparisons (B/N, M/N, M/B) for saliva and serum data using RNA-Seq data of the BRCA dataset including 1102 tumor and 113 normal samples in order to assess the diagnostic capability of protein markers to discriminate between individuals. The AUC values for each ROC curve were measured to determine how well it can discriminate between two diagnostic groups (tumor and normal). In general, a ROC = 0.5 suggests no discrimination, 0.7 ≤ ROC < 0.8 suggests acceptable discrimination, 0.8 ≤ ROC < 0.9 suggests excellent discrimination and ROC ≥ 0.9 is considered outstanding discrimination [18]. ROC analyses were performed via pROC package [46] in R. ## 4.8. Statistical Analyses To combine protein ratios across separate iTRAQ runs and to determine whether protein ratios differed significantly between the three comparisons (B/N, M/N and M/B), the ratios were modeled using a random effects model described earlier [35,47]. Briefly, this procedure used a weighted average of individual ratios across multiple iTRAQ runs to estimate an overall ratio. The weights were proportional to the inverse of the variance for each individual run. The overall protein ratio was deemed statistically significant at the $5\%$ level if the $95\%$ confidence interval did not contain 1. Proteins identified in multiple iTRAQ experiments with ratios that were statistically significant after combining across runs were considered proteins of interest. Algorithms for combining proteins were programmed using the rmeta package in R. Additionally, for comparing protein ratios across serum and saliva in B/N and M/N samples, two-way ANOVA models without interaction were fitted using individual iTRAQ protein log2(ratios) obtained from saliva and serum samples for either benign (B) or malignant (M) compared to normal (N) samples. The car R package [48] was utilized to perform the marginal test comparing saliva and serum mean protein log2(ratios) while controlling for diagnosis (benign or malignant). Two sample t-tests were performed to compare iTRAQ protein log2(ratios) for select proteins in M/N vs. B/N groups after restricting to samples from either saliva or serum. Statistical significance was assessed at the α = 0.05 level. Because of the exploratory nature of this study, no adjustment for multiple testing was applied. R 4.0.5 [49] was used to create box plots and perform all statistical analyses. ## Appendix A.1.1. 2D-LC Separations First dimension separation of the peptides of the 8-plex iTRAQ reagent labeled samples was accomplished by weak anion exchange chromatography (WAX) using a PolyWAX column (4.6 × 200 mm, PolyLC, Columbia, MD, USA) on a Waters 600E HPLC System. Buffer A contained 10 mM ammonium acetate, $85\%$ acetonitrile/$15\%$ water and $0.1\%$ formic acid (FA). Buffer B contained $30\%$ acetonitrile/$70\%$ water and $0.1\%$ FA. The flow rate was 0.5 mL/min. The gradient was Buffer A at $100\%$ A (0–3 min) following sample injection, $0\%$ → $8\%$ Buffer B (17 min), $45\%$ → $100\%$ Buffer B (25 min), then isocratic $100\%$ Buffer B (5 min), then at 55 min switched back to $100\%$ Buffer A to re-equilibrate for the next injection. The first 8 mL of eluant was discarded (containing detergent), then 64 additional 0.5-mL fractions were collected. To balance peptide load injected onto the mass spec, every 9th fraction from these 58 0.5-mL fractions was combined together to make 8 “balanced” fractions for analysis, (i.e., fractions 18, 27, 35, 43 and 51 were combined; fractions 19, 28, 36, 44 and 52 were combined, …, fractions 26, 34, 42, 50 and 58 were combined). All 8 of these combined WAX fractions were dried down completely to reduce volume and to remove the volatile ammonium acetate salts, then resuspended in 15 μL of $2\%$ (v/v) acetonitrile, $0.1\%$ FA and filtered prior to reverse phase C18 nanoflow-LC separation. ## Appendix A.1.2. Mass Spectrometric (MS) Analysis For 2nd dimension separation by reverse phase nanoflow LC, each WAX fraction was auto-injected using an EKsigent NanoLC-Ultra-2D Plus and Eksigent 200 μm × 0.5 mm C18-CL 3 μm 120 A Trap Column and eluted through a Waters, 75um 25cm Nano Column. The elution gradient was $95\%$ C/$5\%$ D (300 nL per minute flowrate) to $65\%$ C/$35\%$ D in 120 min, $15\%$ C/$85\%$ D from 120 to 130 min, then (initial conditions) $95\%$ C/$5\%$ D from 130–150 min. MS spectra taken from 8 WAX fractions, using a 120 min gradient from an Eksigent NanoLC-Ultra-2D Plus system, using a 200 µm × 0.5 mm Eksigent C18-CL 3 µm 120 Å Trap Column and elution through a Waters, 75 μm 25 cm Nano Column. ABSciex 5600 Triple-TOF settings used: parent scan acquired for 250 msec, then up to 50 MS/MS spectra acquired over 2.5 sec for a total cycle time of 2.8 sec. Gas 1 (Nitrogen) = 2, Gas 3 (Nitrogen) = 25 (MS/MS spectrum taken (total: 15,492 MS/MS spectra)). ## Appendix A.2. Database Search for Protein Identification and Quantitation The method was essentially used as described earlier [35]. ## References 1. Siegel R.L., Miller K.D., Fuchs H.E., Jemal A.. Cancer statistics, 2022. CA A Cancer J. Clin. (2022) 72 7-33. DOI: 10.3322/caac.21708 2. Kerlikowske K., Grady D., Rubin S.M., Sandrock C., Ernster V.L.. Efficacy of screening mammography. A me-ta-analysis. JAMA (1995) 273 149-154. DOI: 10.1001/jama.1995.03520260071035 3. Michaelson J.S., Halpern E., Kopans D.B.. Breast cancer: Computer simulation method for estimating optimal intervals for screening. Radiology (1999) 212 551-560. DOI: 10.1148/radiology.212.2.r99au49551 4. Kerlikowske K., Barclay J.. Outcomes of modern screening mammography. J. Natl. Cancer Inst. Monogr. (1997) 22 105-111. DOI: 10.1093/jncimono/1997.22.105 5. Elmore J.G., Barton M.B., Moceri V.M., Polk S., Arena P.J., Fletcher S.W.. Ten-year risk of false positive screening mammo-grams and clinical breast examinations. N. Eng. J. Med. (1998) 338 1089-1096. DOI: 10.1056/NEJM199804163381601 6. Wormwood K.L., Aslebagh R., Channaveerappa D., Dupree E.J., Borland M.M., Ryan J.P., Darie C.C., Woods A.G.. Salivary proteomics and biomarkers in neurology and psychiatry. Proteomics Clin. Appl. (2015) 9 899-906. DOI: 10.1002/prca.201400153 7. Streckfus C.F., Bigler L.. Saliva as a diagnostic fluid. Oral Dis. (2002) 8 69-76. DOI: 10.1034/j.1601-0825.2002.1o834.x 8. Nagler R.M.. Saliva as a tool for oral cancer diagnosis and prognosis. Oral Oncol. (2009) 45 1006-1010. DOI: 10.1016/j.oraloncology.2009.07.005 9. Zhang L., Xiao H., Karlan S., Zhou H., Gross J., Elashoff D., Akin D., Yan X., Chia D., Karlan B.. Discovery and preclinical validation of salivary transcriptomic and proteomic biomarkers for the non-invasive detection of breast cancer. PLoS ONE (2010) 5. DOI: 10.1371/journal.pone.0015573 10. Rapado-González Ó., Martínez-Reglero C., Salgado-Barreira Á., Takkouche B., López-López R., Suárez-Cunqueiro M.M., Muinelo-Romay L.. Salivary biomarkers for cancer diagnosis: A meta-analysis. Ann. Med. (2020) 52 131-144. DOI: 10.1080/07853890.2020.1730431 11. Koopaie M., Kolahdooz S., Fatahzadeh M., Manifar S.. Salivary biomarkers in breast cancer diagnosis: A systematic re-view and diagnostic meta-analysis. Cancer Med. (2022) 11 2644-2661. DOI: 10.1002/cam4.4640 12. Neagu A.N., Jayathirtha M., Whitham D., Mutsengi P., Sullivan I., Petre B.A., Darie C.C.. Proteomics-Based Identification of Dysregulated Proteins in Breast Cancer. Proteomes (2022) 10. DOI: 10.3390/proteomes10040035 13. Streckfus C.F., Bigler L.R., Zwick M.. The use of surface-enhanced laser desorption/ionization time-of-flight mass spectrom-etry to detect putative breast cancer markers in saliva: A feasibility study. J. Oral Pathol. Med. (2006) 35 292-300. DOI: 10.1111/j.1600-0714.2006.00427.x 14. Delmonico L., Bravo M., Silvestre R.T., Ornellas M.H., De Azevedo C.M., Alves G.. Proteomic profile of saliva and plasma from women with impalpable breast lesions. Oncol. Lett. (2016) 12 2145-2152. DOI: 10.3892/ol.2016.4828 15. Streckfus C.F., Storthz K.A., Bigler L., Dubinsky W.P.. A comparison of the proteomic expression in pooled saliva specimens from individuals diagnosed with ductal carcinoma of the breast with and without lymph node involvement. J. Oncol. (2009) 2009 737619. DOI: 10.1155/2009/737619 16. Wu W.W., Wang G., Baek S.J., Shen R.F.. Comparative study of three proteomic quantitative methods, DIGE, cICAT, and iTRAQ, using 2D gel- or LC-MALDI TOF/TOF. J. Prot. Res. (2006) 5 651-658. DOI: 10.1021/pr050405o 17. Sur T.. iTRAQ, a New prospect of toxicogenomics. Int. J. Hum. Genet. (2013) 13 33-40. DOI: 10.1080/09723757.2013.11886194 18. Hosmer D.W., Lemeshow S.. Applied Logistic Regression (2000) 160-164 19. Veyssière H., Bidet Y., Penault-Llorca F., Radosevic-Robin N., Durando X.. Circulating proteins as predictive and prognostic biomarkers in breast cancer. Clin. Proteom. (2022) 19 25. DOI: 10.1186/s12014-022-09362-0 20. Moon P.G., Lee J.E., Cho Y.E., Lee S.J., Chae Y.S., Jung J.H., Kim I.S., Park H.Y., Baek M.C.. Fibronectin on circulating extracellular vesicles as a liquid biopsy to detect breast cancer. Oncotarget (2016) 7 40189-40199. DOI: 10.18632/oncotarget.9561 21. Cao F., Wang K., Zhu R., Hu Y.W., Fang W.Z., Ding H.Z.. Clinicopathological significance of reduced SPARCL1 expression in human breast cancer. Asian Pac. J. Cancer Prev. (2013) 14 195-200. DOI: 10.7314/APJCP.2013.14.1.195 22. Varvara P.V., Karaolanis G., Valavanis C., Stanc G., Tzaida O., Trihia H., Patapis P., Dimitroulis D., Perrea D.. gamma-H2AX: A potential biomarker in breast cancer. Tumour Biol. (2019) 41 1010428319878536. DOI: 10.1177/1010428319878536 23. Sedelnikova O.A., Bonner W.M.. Gamma-H2AX in cancer cells: A potential biomarker for cancer diagnostics, prediction and recurrence. Cell Cycle (2006) 5 2909-2913. PMID: 17172873 24. Dai D.N., Li Y., Chen B., Du Y., Li S.B., Lu S.X., Zhao Z.P., Zhou A.J., Xue N., Xia T.L.. Elevated expression of CST1 promotes breast cancer progression and predicts a poor prognosis. J. Mol. Med. (2017) 95 873-886. DOI: 10.1007/s00109-017-1537-1 25. Ponzetti M., Capulli M., Angelucci A., Ventura L., Monache S.D., Mercurio C., Calgani A., Sanità P., Teti A., Rucci N.. Non-conventional role of haemoglobin beta in breast malignancy. Br. J. Cancer (2017) 117 994-1006. DOI: 10.1038/bjc.2017.247 26. Bonaventura C., Henkens R., Alayash A.I., Banerjee S., Crumbliss A.L.. Molecular controls of the oxygenation and redox reactions of hemoglobin. Antioxid. Redox Signal. (2013) 18 2298-2313. DOI: 10.1089/ars.2012.4947 27. Janke J., Engeli S., Boschmann M., Adams F., Böhnke J., Luft F.C., Sharma A.M., Jordan J.. Retinol-binding protein 4 in human obesity. Diabetes (2006) 55 2805-2810. DOI: 10.2337/db06-0616 28. Jiao C., Cui L., Ma A., Li N., Si H.. Elevated Serum Levels of Retinol-Binding Protein 4 Are Associated with Breast Cancer Risk: A Case-Control Study. PLoS ONE (2016) 11. DOI: 10.1371/journal.pone.0167498 29. Fry S.A., Sinclair J., Timms J.F., Leathem A.J., Dwek M.V.. A targeted glycoproteomic approach identifies cadherin-5 as a novel biomarker of metastatic breast cancer. Cancer Lett. (2013) 328 335-344. DOI: 10.1016/j.canlet.2012.10.011 30. Patmore S., Dhami S.P.S., O’Sullivan J.M.. Von Willebrand factor and cancer; metastasis and coagulopathies. J. Thromb. Haemost. (2020) 18 2444-2456. DOI: 10.1111/jth.14976 31. Dhami S.P.S., Patmore S., Comerford C., Byrne C.M., Cavanagh B., Castle J., Kirwan C.C., Kenny M., Schoen I., O’Donnell J.S.. Breast cancer cells mediate endothelial cell activation, promoting von Willebrand factor release, tumor adhesion, and transendothelial migration. J. Thromb. Haemost. (2022) 20 2350-2365. DOI: 10.1111/jth.15794 32. Opstal-van Winden A.W., Krop E.J., Kåredal M.H., Gast M.C., Lindh C.H., Jeppsson M.C., Jönsson B.A., Grobbee D.E., Peeters P.H., Beijnen J.H.. Searching for early breast cancer biomarkers by serum protein profiling of pre-diagnostic serum; a nested case-control study. BMC Cancer (2011) 11. DOI: 10.1186/1471-2407-11-381 33. Mei J., Hao L., Liu X., Sun G., Xu R., Wang H., Liu C.. Comprehensive analysis of peroxiredoxins expression profiles and prognostic values in breast cancer. Biomark. Res. (2019) 7 16. DOI: 10.1186/s40364-019-0168-9 34. Kori M., Ozdemir G.E., Arga K.Y., Sinha R.. A Pan-Cancer Atlas of Differentially Interacting Hallmarks of Cancer Proteins. J. Pers Med. (2022) 12. DOI: 10.3390/jpm12111919 35. Sinha I., Karagoz K., Fogle R.L., Hollenbeak C.S., Zea A.H., Arga K.Y., Stanley A.E., Hawkes W.C., Sinha R.. “Omics” of Selenium Biology: A Prospective Study of Plasma Proteome Network Before and After Selenized-Yeast Supplementation in Healthy Men. OMICS (2016) 20 202-213. DOI: 10.1089/omi.2015.0187 36. Sinha I., Goel R., Bitzer Z.T., Trushin N., Liao J., Sinha R.. Evaluating electronic cigarette cytotoxicity and inflammatory responses in vitro. Tob. Induc. Dis. (2022) 20 45. DOI: 10.18332/tid/147200 37. Kamburov A., Stelzl U., Lehrach H., Herwig R.. The ConsensusPathDB interaction database: 2013 update. Nucleic Acids Res. (2013) 41 D793-D800. DOI: 10.1093/nar/gks1055 38. Kanehisa M., Furumichi M., Tanabe M., Sato Y., Morishima K.. KEGG: New perspectives on genomes, pathways, diseases and drugs. Nucleic Acids Res. (2017) 45 D353-D361. DOI: 10.1093/nar/gkw1092 39. Croft D., O’Kelly G., Wu G., Haw R., Gillespie M., Matthews L., Caudy M., Garapati P., Gopinath G., Jassal B.. Reactome: A database of reactions, pathways and biological processes. Nucleic Acids Res. (2011) 39 D691-D697. DOI: 10.1093/nar/gkq1018 40. Ashburner M., Ball C.A., Blake J.A., Botstein D., Butler H., Cherry J.M., Davis A.P., Dolinski K., Dwight S.S., Eppig J.T.. Gene ontology: Tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. (2000) 25 25-29. DOI: 10.1038/75556 41. Oughtred R., Stark C., Breitkreutz B.J., Rust J., Boucher L., Chang C., Kolas N., O’Donnell L., Leung G., McAdam R.. The BioGRID interaction database: 2019 update. Nucleic Acids Res. (2019) 47 D529-D541. DOI: 10.1093/nar/gky1079 42. Shannon P., Markiel A., Ozier O., Baliga N.S., Wang J.T., Ramage D., Amin N., Schwikowski B., Ideker T.. Cytoscape: A software environment for integrated models of biomolecular interaction networks. Genome Res. (2003) 13 2498-2504. DOI: 10.1101/gr.1239303 43. Aguirre-Gamboa R., Gomez-Rueda H., Martínez-Ledesma E., Martínez-Torteya A., Chacolla-Huaringa R., Rodriguez-Barrientos A., Tamez-Peña J.G., Treviño V.. SurvExpress: An online biomarker validation tool and database for cancer gene expression data using survival analysis. PLoS ONE (2013) 8. DOI: 10.1371/journal.pone.0074250 44. Gulfidan G., Turanli B., Beklen H., Sinha R., Arga K.Y.. Pan-cancer mapping of differential protein-protein interactions. Sci. Rep. (2020) 10 3272. DOI: 10.1038/s41598-020-60127-x 45. Therneau T.. A Package for Survival Analysis in R. R Package Version 3.3-1. (2022) 46. Robin X., Turck N., Hainard A., Tiberti N., Lisacek F., Sanchez J.C., Müller M.. pROC: An open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinform. (2011) 12. DOI: 10.1186/1471-2105-12-77 47. Fogle R.L., Hollenbeak C.S., Stanley B.A., Vary T.C., Kimball S.R., Lynch C.J.. Functional proteomic analysis reveals sex-dependent differences in structural and energy-producing myocardial proteins in rat model of alcoholic cardiomyopathy. Physiol. Genomics (2011) 43 346-356. DOI: 10.1152/physiolgenomics.00203.2010 48. Fox J., Weisberg S.. An R Companion to Applied Regression (2019) 49. 49. R Core Team R: A Language and Environment for Statistical ComputingR Foundation for Statistical ComputingVienna, Austria2021. R: A Language and Environment for Statistical Computing (2021)
--- title: Neonatal Orally Administered Zingerone Attenuates Alcohol-Induced Fatty Liver Disease in Experimental Rat Models authors: - Bernice Asiedu - Busisani Wiseman Lembede - Monica Gomes - Abe Kasonga - Pilani Nkomozepi - Trevor Tapiwa Nyakudya - Eliton Chivandi journal: Metabolites year: 2023 pmcid: PMC9966972 doi: 10.3390/metabo13020167 license: CC BY 4.0 --- # Neonatal Orally Administered Zingerone Attenuates Alcohol-Induced Fatty Liver Disease in Experimental Rat Models ## Abstract Alcohol intake at different developmental stages can lead to the development of alcohol-induced fatty liver disease (AFLD). Zingerone (ZO) possess hepato-protective properties; thus, when administered neonatally, it could render protection against AFLD. This study aimed to evaluate the potential long-term protective effect of ZO against the development of AFLD. One hundred and twenty-three 10-day-old Sprague–Dawley rat pups (60 males; 63 females) were randomly assigned to four groups and orally administered the following treatment regimens daily during the pre-weaning period from postnatal day (PND) 12–21: group 1—nutritive milk (NM), group 2—NM +1 g/kg ethanol (Eth), group 3—NM + 40 mg/kg ZO, group 4—NM + Eth +ZO. From PND 46–100, each group from the neonatal stage was divided into two; subgroup I had tap water and subgroup II had ethanol solution as drinking fluid, respectively, for eight weeks. Mean daily ethanol intake, which ranged from 10 to 14.5 g/kg body mass/day, resulted in significant CYP2E1 elevation ($p \leq 0.05$). Both late single hit and double hit with alcohol increased liver fat content, caused hepatic macrosteatosis, dysregulated mRNA expression of SREBP1c and PPAR-α in male and female rats ($p \leq 0.05$). However, neonatal orally administered ZO protected against liver lipid accretion and SREBP1c upregulation in male rats only and attenuated the alcohol-induced hepatic PPAR-α downregulation and macrosteatosis in both sexes. This data suggests that neonatal orally administered zingerone can be a potential prophylactic agent against the development of AFLD. ## 1. Introduction Alcohol liver disease (ALD) develops from prolonged excessive alcohol consumption. Studies show that genetic predisposition contributes substantially to alcohol use disorder [1,2]. Others show that exposure to alcohol in utero trigger epigenetic modifications that also may play a role in mediating alcohol use disorders [3,4]. Numerous studies indicate that the intrauterine environment can set either a healthy or diseased trajectory for offspring in the future; a phenomenon described as the developmental origin of health and disease [5,6]. We hypothesized that ALD may develop based on this concept [7]. Human and preclinical studies indicate that prenatal alcohol exposure (PAE) predisposes offspring to increased alcohol consumption in adolescence and adulthood [8,9]. Consumption of alcohol in adolescence has been shown to cause use disorders and abuse, which may result in the development of ALD [10,11]. Hence, early-life exposure to alcohol and/or alcohol consumption in adolescence may be a significant risk factor for ALD. Therefore, the combination of early-life alcohol exposure and alcohol consumption in adolescence constitute a double-hit effect for the development of ALD. Although the suckling growth phase is equally susceptible to neuroplasticity as the prenatal phase [12], its impact on alcohol consumption in adolescence and the subsequent development of ALD have not been thoroughly explored. Chronic liver disease affects about $10\%$ of the world’s human population and, its mortal end-stage generally follows cirrhosis and liver cancer [13]. Varied factors characterize liver disease as the fourth to the fifth cause of death worldwide [13]. Non-alcoholic fatty liver disease ranks first contributing up to $40\%$ of liver diseases and hepatitis B and C viruses and alcohol overconsumption contribute $30\%$, $15\%$ and $11\%$, respectively [13]. Alcohol abuse ranks third to smoking and hypertension as a cause of preventable death [14]. Alcohol-induced liver disease accounts for $20\%$ to $50\%$ of the prevalence of liver cirrhosis [15]. Alcohol consumption results in liver damage on a spectrum, from simple steatosis to hepatitis with fibrosis [16,17]. The development of a fatty liver is the earliest response to heavy (>40 g ethanol/day) and/or chronic consumption of alcohol and this occurs in about $90\%$ of heavy drinkers [18]. However, only 8–$20\%$ of those that develop fatty liver progress to severe forms of liver disease [18]. Steatosis, which is categorized as either macro- and or microsteatosis, occurs when more than $5\%$ fat constitutes liver tissue resulting in hepatocytes becoming distended with lipids [19]. Hepatic macrosteatosis is typified by the presence of a large fat droplet that pushes the nucleus to the periphery of the hepatocyte [19,20], while with microsteatosis, many small, less than 1 μm in diameter, cytoplasmic lipid droplets give the hepatocyte a foamy-like appearance but with the nucleus remaining in the middle of the cell [19,20]. There are two variants of macrosteatosis: large and small droplet macrosteatosis [21]. In the former, a large unilocular lipid droplet fills up the hepatocyte and pushes the nucleus to the periphery and in the latter are multilocular lipid droplets that occupy less than half of the cytoplasm with the nucleus remaining intact [21]. In the progression of steatosis, small lipid droplets coalesce into large fat droplets [22]. Studies show that macrosteatosis is common [23], but microsteatosis is rare in AFLD [22]. Macrosteatosis has a good prognosis with rare progression to fibrosis [24]. Although the pathogenesis of ALD is not entirely understood, it is known to stem from the toxic effects of ethanol and its metabolite acetaldehyde (AA), which mediate increased intestinal permeability, changes to the gut microbiome and an inflammatory response to cellular injury [25]. In vivo, alcohol dehydrogenase (ADH) metabolises ethanol to AA. The metabolism of ethanol involves the upregulation of sterol regulatory element-binding transcription factor 1 c (SREBP-1c), which upon interacting with lipid droplets membrane proteins, promotes the formation of lipid droplets in hepatocytes resulting in the development of steatosis [26]. An increase in the reduced nicotinamide adenine dinucleotide (NADH) to oxidised nicotinamide adenine dinucleotide (NAD+) ratio inhibits β-oxidation and activates triglyceride synthesis [25], which increases the quantity of fat in hepatocytes. Additionally, the inhibition of the transactivation activity and DNA-binding of peroxisome activator receptor α (PPAR-α) by AA results in inhibiting β-oxidation [26]. The metabolism of low to moderate amounts of consumed alcohol induces the ADH system, but when excess alcohol is consumed, its metabolism induces the cytochrome P450 E21 (CYP2E1) system [27]. Induction of the CYP2E1 generates excessive amounts of reactive oxygen species that then mediate lipid peroxidation [28]. The by-products of lipid peroxidation, for example, malondialdehyde and 4-hydroxynonenal, activate adaptive immunity [25] and the intrinsic apoptotic pathways [29]. Activation of these pathways result in hepatic neutrophil infiltration and liver inflammation via the myeloid differentiation primary response gene 88 (MyD88) independent signalling pathway [25]. Recruitment of the MyD88 mediates the activation of the pro-inflammatory transcription factor, nuclear factor kappa β (NF-Kβ) and its downstream inflammatory cytokines; tumour necrosis factor-alpha (TNF-α), Interleukin (IL)-6, IL-10 and IL-1β [30,31] that contribute to hepatocellular damage. Several pharmacological agents have proven effective as potential prophylaxes in managing adverse effects induced by exposure to toxic substances during periods of developmental plasticity [32,33,34]. Zingerone has been reported to possess antisteatotic properties that can protect against the development of fatty liver via the activation of AMP-activated protein kinase (AMPK) [33,35]. Furthermore, oral supplementation of 40 mg/kg body wt of zingerone via intragastric intubation to alcohol-fed Wistar rats protected against hepatoxicity [36]. We therefore determined the prophylactic potential of zingerone to protect against the development of alcohol-induced fatty liver disease using rat models. ## 2.1. Study Setting and Animal Use Ethical Clearance This study was conducted at the Wits Research Animal Facility (WRAF) of the University of the Witwatersrand, Johannesburg, South Africa. This study complied with accepted laboratory animal use and care stipulated in the South African National standard (SANS: 10386:2008) and the animal protection act 1962, Act No. 71. Ethical clearance for the experiment (ethical clearance certificate number: $\frac{2019}{10}$/57B) was granted by the Animal Research Ethics Committee of the University of the Witwatersrand. ## 2.2. Animals and Animal Management The study used one hundred and twenty-three 10-day-old suckling male and female Sprague–Dawley rat pups (60 males; 63 females) from dams with 8 to 12 rat pups. During the suckling period, from postnatal day (PND) 1 to 21, the rat litter were housed with their respective dams in acrylic cages at the WRAF. At weaning from PND 21 to PND 100, the rat pups were each individually housed in acrylic cages and allowed ad libitum access to feed and tap water. The room temperature was maintained at 24 ± 2 °C. A 12:12 h dark-and-light cycle (with lights on at 07:00 h) was kept throughout the experimental period. ## 2.3. Experimental Design This interventional study comprised three stages (Figure 1): intervention during the neonatal growth phase, a growing out phase with no intervention and intervention during the adult growth phase (Figure 1). At the neonatal stage, 123 10-day-old rat pups (60 males; 63 females) following a 2-day habituation period were then randomly assigned to four groups and administered different treatment regimens as follows: Group I: nutritive milk (NM); Group II: NM + 1 g/kg body mass ethanol (NM + Eth); Group III: NM+ 40 mg/kg body mass of ZO (NM + ZO); Group IV: NM + Eth + ZO. Interventions were administered from PND 12–21. Ethanol and zingerone dosage have previously been used in rat pups [35,37,38]. Nutritive milk was used as the vehicle for administration of the interventions. During the second stage (PND 22–45), the weaned rats were allowed to grow to adulthood without any intervention but had ad libitum access to normal rat chow and plain drinking water. The intervention during adulthood started from PND 46 and continued to PND 100. During this intervention, rats from each group at the neonatal stage were divided into two subgroups; rats in subgroup I had plain drinking water and their counterparts in subgroup II had ethanol as drinking fluids for eight weeks (Figure 1). The rats were adapted to incremental ethanol solution initially at $5\%$ (v/v) for one week, then $10\%$ (v/v) for another week, and $20\%$ (v/v) ethanol solution for the remaining six weeks per the protocol described by Ojeda et al., 2008 [39]. Rats have a natural aversion to alcohol and prefer ethanol solutions at lower concentrations [40]. Hence, they reduce their intake volumes as the ethanol concentration increases [40]. Thus, incremental ethanol at 5 and $10\%$ was used to prime them to the taste of alcohol and prevent decreased intake at $20\%$ ethanol concentration. The amount of ethanol consumed weekly by each was measured and computed in g/100 g body mass using the “Tables for determining grams values of ethanol solution” previously described [41]. The rats that were orally gavaged with alcohol during the neonatal growth phase and had ethanol solution as drinking fluid in adulthood had a double hit with alcohol, and those that were gavaged with alcohol during the neonatal growth phase only had an early single hit with alcohol. The rats in the subgroups that received alcohol solution as a drinking fluid only during adulthood had a late single hit. ## 2.4. Terminal Procedure On PND 101, following an overnight fast, the terminal body masses of the rats were measured using an electronic balance (Snowrex, Johannesburg, South Africa). The rats were then euthanised with 200 mg/kg of sodium pentobarbital (Eutha-naze®, Bayer, Johannesburg, South Africa) via intraperitoneal injection and cut open by a midline incision on the abdomen and thorax. Blood was drawn via cardiac puncture using an 18 G syringe into heparinised tubes and then centrifuged (Senova NovaFuge centrifuge, Shanghai, China) at 3000× g for 15 min. Plasma was then collected and stored at −80 °C for further biochemical assays. Each rat’s liver was carefully dissected from the abdominal cavity and weighed on an electronic scale (Presica 310 M, Presica Instruments, Dietikon, Switzerland). The liver was then divided into four parts. One part was rinsed in cold saline and stored at −20 °C to determine hepatic thiobarbituric acid (TBARS). A sample from the right lobe of each liver was immersion fixed in $10\%$ phosphate-buffered formalin solution (Merck, Johannesburg, South Africa) for histological analysis. Another part of the liver was stored in sealed ziplock plastic bags at −20 °C for total fat content determination, and the last portion was placed in RNAlater solution and kept at −80 °C for molecular analysis. ## 2.5. Computation of the Hepatosomatic Index The hepatosomatic index was computed by dividing the mass of each liver by the respective terminal body mass of each rat and expressed as a percentage (%). ## 2.6.1. Liver Tissue Homogenisation The liver sample (100 mg) was homogenised in 10 mL of phosphate buffer (0.1 M, pH = 7.4) with an ultra turrax homogeniser (T-25 basic, Janke & Kunkel Ultra Turrax, Germany). The resultant homogenate was centrifuged at 3000× g for 15 min at 4 °C. The resultant supernatant was used to determine hepatic lipid peroxidation by measuring the supernatant’s TBARS concentration. ## 2.6.2. Determination of Peroxidation in the Liver The liver TBARS concentration was estimated using the method described by Niehaus and Samuelsson [1968] [42]. Briefly, 2 mL of the supernatant from the liver sample homogenate was diluted with distilled water in a 1:1 ratio. 2.0 mL of the working reagent ((thiobarbituric acid (TBA)–trichloroacetic acid (TCA)–hydrochloric acid (HCL)) in a ratio of 1:1:1) was then added to the diluted supernatant. The mixture was boiled in a water bath for 15 min, allowed to cool on ice for 5 min, and then centrifuged (Senova NovaFuge centrifuge, Shanghai, China) at 3000× g for 5 min at room temperature. The absorbance of the supernatant obtained was read on a spectrophotometer (Beckman Coulter, CA, USA) at 532 nm. ## 2.7. Determination of Liver Lipid Content The total liver lipid content was determined using the soxhlet extraction method as described by AOAC (2005; method number 920.39) using petroleum ether as the solvent. ## 2.8. Determination of Liver Histomorphometry Liver tissue samples preserved in $10\%$ phosphate-buffered formalin were processed with an automated tissue processor (Microm STP 120 Thermo Scientific, Waltham, MA, USA), embedded in paraffin wax and sectioned at 3 µm using a microtome (Leica instruments GmbH, (Pty) Ltd., Wetzlar, Germany) for histological analysis. The sections were stained with haematoxylin and eosin (H&E) using a Gemini AS slide stainer coupled to a Clearvue cover slipper (Fisher Scientific, Waltham, MA, USA). The stained liver sections were viewed under an Olympus BH2-RFCA microscope (Olympus Corporation, Tokyo, Japan) coupled to an Olympus XC 10 HD camera (Olympus Corporation, Tokyo, Japan) for histological assessment and image capture. ## 2.9. Determination of Surrogate Plasma Biomarkers of Liver Function Plasma activities of aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were determined using an automated clinical chemistry analyser (IDEXX VetTest® Clinical Chemistry Analyser, IDEXX Laboratories Inc., Westbrook, ME, USA) according to the manufacturer’s instructions. The calibrated auto-analyser performed the analysis on pre-loaded disks for AST and ALT using 10 μL of plasma. ## 2.10. Determination of Plasma CYP2E1, TNF-α and IL-6 Concentration Rat-specific ELISA kits (Elabscience®, Rat CYP2E1, TNF-α, IL-6 ELISA kit, Wuhan, Hubei Province, China) were used to determine the plasma CYP2E1, tumour necrosis factor-alpha (TNF-α) and interleukin-6 (IL-6) concentration following the manufacturers’ instructions. The test employed a sandwich ELISA principle. The optical density of the resulting reactions was measured at 450 nm on a microplate reader (Thermo Fisher Scientific Inc., Waltham, MA, USA), and the sample concentrations were extrapolated from the standard curve. ## 2.11.1. RNA Extraction and cDNA Synthesis Liver tissue (50 mg) was finely ground with a mortar and pestle, and the RNA was extracted using Aurum™ Total RNA Mini Kit (BioRad, Hercules, CA, USA). The RNA quantity was assessed by measuring the absorption at 260 nm and the purity by the $\frac{260}{280}$ nm absorbance ratio using the NanoDrop lite spectrophotometer (Thermofisher Scientific, Johannesburg, South Africa). The volume of RNA needed to make a final concentration of 0.5 μg was calculated and synthesised to complementary DNA (cDNA) with LunaScript supermix (Inqaba biotec, Johannesburg, South Africa). Nuclease-free water was added to make a final volume of 20 μL. The preparation was gently mixed and incubated at 25 °C for 2 min, 55 °C for 10 min, and 95 °C for 1 min on the thermal cycler (PxE0.2, Thermo Fisher Scientific, Waltham, MA, USA). The cDNA samples were then stored at −20 °C until further use. ## 2.11.2. Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) Analysis Real-time PCR was performed using LunaScript master mix (Inqaba biotec, Johannesburg, South) with the master mix and primers mixed according to the manufacturer’s protocol. The primers used are provided in Supplementary Table S1. Complementary DNA was diluted to 1:20 ratio with nuclease-free water. The prepared mix was added to appropriate wells in 96-well plates (Roche, Johannesburg, South Africa). The cDNA template was added last and the plates were sealed with optical adhesive covers. Quantitative real-time PCR (qRT-PCR) was measured on the LightCycler 96 (Roche diagnostics, Basel, Switzerland) following thermal cycling conditions of the manufacturers’ protocol. Relative gene expression was analysed using the −2ΔΔCT method. Gene expression was normalised to the mRNA expression of beta-actin. ## 2.11.3. Statistical Analysis GraphPad Prism 8 software (Graph-Pad Software Inc., San Diego, CA, USA) was used to analyse data. Data are expressed as mean ± standard deviation. A one-way ANOVA was used to analyse multiple-group data, followed by mean comparison using the Tukey post hoc test for parametric data. The Kruskal–Wallis test was used to analyse multiple group nonparametric data, followed by mean comparisons by the Dunns post hoc test for nonparametric. Statistical significance was considered when $p \leq 0.050.$ ## 3.1. Effect of Neonatal Orally Administered Zingerone on Ethanol Consumption in Adult Rats Figure 2 shows the weekly mean ethanol intake of the rats. In male and female rats, early and late single hit as well as double hit of alcohol alone or together with neonatal zingerone did not affect ethanol consumption in adulthood ($p \leq 0.05$). Weekly ethanol intake ranged from 7.02 ± 2.09 g/100 body mass to 10.12 ± 3.01 g/100 g body mass (Supplementary Table S2). Male rats that had a late single hit with alcohol had the highest ethanol intake, while female rats that had a double hit with alcohol had the highest ethanol intake among counterpart male and female rats that consumed ethanol solution in adulthood. ## 3.2. Effect of Neonatal Orally Administered Zingerone on Hepatosomatic Index Treatment regimens had no effect on the absolute liver masses and hepatosomatic indices of the male rats ($p \leq 0.05$ vs. control; Table 1). The absolute liver mass and hepatosomatic indices of the female rats in each treatment group did not differ from the control counterparts ($p \leq 0.05$ vs. control; Table 1). Female rats that had an early single hit with alcohol had a significantly increased absolute liver mass compared to counterparts that had zingerone orally administered during the neonatal growth phase in combination with either a late single hit and or a double hit with alcohol ($$p \leq 0.023$$, NM + Eth + Wad vs. NM + ZO + Ethad; $$p \leq 0.049$$, NM + Eth + Wad vs. NM + Eth + ZO + Ethad). When adjusted to the body mass, the hepatosomatic index of the female rats was not significantly different ($p \leq 0.05$) compared to the other treatment groups. ## 3.3. Effect of Neonatal Orally Administered Zingerone on Liver Fat Content Treatment regimens significantly affected the liver fat content of male and female rats ($$p \leq 0.006$$, (males): $$p \leq 0.008$$, (females) Figure 3). A late single hit with alcohol significantly increased ($$p \leq 0.039$$; Figure 3A) liver fat content of male rats compared to control counterparts but a combination of neonatal orally administered zingerone with a late single hit decreased liver fat content compared to male counterparts that only had a late single alcohol hit ($$p \leq 0.036$$, NM + ZO + Ethad vs. NM + Ethad, Figure 3A). An early single and double hit with alcohol alone or together with neonatal administered zingerone had no effect on liver fat content of male rats ($p \leq 0.05$; Figure 3A). A late single and double hit with alcohol significantly increased liver fat content of female rats ($$p \leq 0.045$$, NM + Ethad vs. NM + Wad; $$p \leq 0.023$$, NM + Eth + Ethad vs. NM + Wad, Figure 3D). Neonatal orally administered zingerone in combination with either a single and or double alcohol hit resulted in similar liver fat content with that of control counterparts ($$p \leq 0.858$$, NM + ZO + Ethad vs. NM + Wad; $$p \leq 0.067$$, NM + Eth + ZO + Ethad vs. NM + Wad; Figure 3D). The liver fat content of female rats that had a combination of neonatal zingerone and double alcohol hit was similar to that of counterparts that had a late single or double alcohol hit ($p \leq 0.05$, Figure 3D). ## 3.4. Effect of Neonatal Orally Administered Zingerone on Hepatic Histomorphometric Changes Figure 4A and Figure 5A show the histomorphometric changes of the liver sections at X10. This is to provide an overview of the treatment effect. Figure 4B and Figure 5B is produced at X40 for differentiation of the steatotic cells. Hepatic lobules of male and female rats that consumed plain tap water (NM + Wad, NM + Eth + Wad, NM + ZO + Wad and NM + Eth + ZO + Ethad) contained regularly arranged hepatocytes radiating from the central vein and had no visible hepatic steatosis in the control rats. An early single hit with alcohol (NM + Eth + Wad) resulted in microsteatosis in female rats (Figure 5B; arrow D), which was not observed in the female rats that had an early single hit alcohol in combination with neonatal zingerone (Figure 5B). The late single hit (NM+ Ethad) and double hit (NM + Eth + Ethad) with alcohol resulted in small and large droplet steatosis (Figure 4 and Figure 5; arrow A & B) in male and female rats. A combination of neonatal zingerone with the late single hit (NM + ZO + Ethad) and double hit (NM + Eth + ZO + Ethad) of alcohol resulted in relatively less severe small and large droplet steatosis (Figure 4 and Figure 5). ## 3.5. Effect of Neonatal Orally Administered Zingerone on Lipid Regulatory Genes Treatment regimens had a significant ($p \leq 0.05$) effect on the mRNA expression level of PPAR-α of both male and female rats. Both late single and double hit with alcohol significantly decreased ($$p \leq 0.014$$, NM + Ethad vs. NM + Wad; $$p \leq 0.0009$$, NM + Eth + Ethad vs. NM + Wad, Figure 3B) PPAR-α expression levels in male rats and they also significantly decreased ($$p \leq 0.040$$, NM + Ethad vs. NM + Wad; $$p \leq 0.019$$, NM + Eth + Ethad vs. NM + Wad Figure 3B) that of females relative to control. Neonatal orally administered ZO in combination with either a late single or double alcohol hit had no effect on PPAR-α expression in male and female rats relative to the control ($p \leq 0.05$). Both a late single and double hit with alcohol significantly increased the SREBP1c expression in male and female rats ($p \leq 0.05$, Figure 3C,F). Neonatal orally administered zingerone in combination with a late single alcohol hit had no effect on SREBP1c expression level in male ($$p \leq 0.635$$, Figure 3C) and female ($$p \leq 0.960$$, Figure 3F) rats compared to their respective control counterparts. Neonatal orally administered zingerone in combination with a double alcohol hit had no effect on SREBP1c expression level in male rats ($p \leq 0.05$ vs. control) but it significantly increased SREBP1c expression in female rats ($$p \leq 0.005$$). ## 3.6. Effect of Neonatal Orally Administered Zingerone on Plasma Liver Enzyme Activities In male and female rats, treatment regimen had no significant effect ($p \leq 0.05$ vs. control) on plasma AST and ALT activities (Table 2). ## 3.7. Effect of Neonatal Orally Administered Zingerone on Plasma CYP2E1 and Hepatic TBARS Ethanol consumption increased ($p \leq 0.05$, Figure 6A,C) CYP2E1 concentration in male and female rats that had late single hit and double hit with ethanol. Neonatal orally administered zingerone in combination with either late single hit or double hit with alcohol significantly increased CYP2E1 concentration of male and female rats relative to control. Treatment regimens had no effect on hepatic TBARS concentration in male ($$p \leq 0.096$$, Figure 6B) and female rats ($$p \leq 0.050$$, Figure 6D). ## 3.8. Effect of Neonatal Orally Administered Zingerone on Biomarkers of Inflammation Treatment regimens had no effect ($p \leq 0.05$ vs. control) on plasma TNF-α and IL-6 concentrations and hepatic mRNA expression of NFK-β and TNF-α male rats and female rats (Figure 7A–H). ## 4. Discussion In this study, we investigated the effect of neonatal orally administered zingerone on an early and late single and a double hit with alcohol on the development of alcohol-induced fatty liver disease in adulthood. A late single hit and double hit with alcohol in male and female rats resulted in an increased liver fat content accompanied by macrosteatosis, downregulation of PPAR-α and upregulation of SREBP1c. Neonatal orally administered zingerone attenuated fat accretion by preventing an upregulation in SREBP1c in male rats. It also mitigated hepatic steatosis and the downregulation of PPAR-α induced by late single or double alcohol hit in male and female rats. Our findings showed that neonatal administered ethanol and zingerone had no effect on ethanol consumption of the rats in adulthood. Previous studies reported that exposure to alcohol during the prenatal developmental phase did not affect adolescent ethanol consumption because of aversion to alcohol developed as a result of prenatal exposure [43,44]. However, other research revealed that prenatal exposure to ethanol via food and substrate exchange between maternal and foetal blood in the placenta increases the likelihood of increasing alcohol intake in adolescence and adulthood because it enhances the brain’s reward system [3,45]. Findings from the current study suggest that the aversive as well the appetitive chemosensory stimuli to alcohol were not modified by neonatal orally administered ethanol and/or zingerone. However, we did not assess the neurochemical changes induced by the neonatal exposure to the interventions, hence there is need for further investigations into the neurochemical changes induced by exposure to alcohol and zingerone during the neonatal growth phase. Early, late single hit and double hit with alcohol had no effect on the hepatosomatic index in the present study. In line with our finding, another research group found that alcohol consumption does not affect the hepatosomatic index of rats [46]. However, in contrast to our findings, other studies showed that alcohol consumption increased liver weight [47,48]. The studies of AL-Humadi et al., 2019 [47] and Rasineni et al., 2019 [48] made use of the Lieber–DeCarli alcohol liquid diet, which contains $35.5\%$ compared to Labchef standard rat chow (Epol®, Johannesburg, South Africa) with $5\%$ fats with alcohol solution as a drinking fluid used in this study. A late single hit (both sexes) and double hit (females only) with alcohol increased the liver fat content. Alcohol disrupts several aspects of hepatic lipid flux that leads to lipid accumulation, including activation of SREBP1c to stimulate lipogenesis [49]. Studies report that alcohol causes hepatic lipid accumulation, which may result in the development of steatosis [50,51]. Hence, it was not surprising that alcohol-induced increase in liver fat content was accompanied with the formation of prominent small and large droplet macrosteatosis in both male and female rats. This observation was associated with peroxisome proliferator activator receptor-alpha (PPAR-α) downregulation and sterol regulatory element binding protein 1c (SREBP1c) upregulation. Previous studies reported that alcohol consumption caused alcohol-induced fatty liver by downregulating PPAR-α [52]. Furthermore, ethanol decreases the AMP-activated protein kinase (AMPK) Sirtuin (SIRT) signalling pathway and its downstream signalling proteins resulting in the upregulation of SREBP1c [53]. These proteins, PPAR-α and SREBP1c, are involved in regulating the hepatic fatty oxidation pathway and lipid droplet formation [26,54]. Therefore, it is also plausible that in the present study, hepatic macrosteatosis was caused by downregulation and upregulation in PPAR-α and SREBP1c, respectively. Neonatal orally administered zingerone mitigated the alcohol-induced downregulation of hepatic PPAR-α expression in male and female rats but protected against SREBP1c upregulation in male rats only, and attenuated the accretion of liver fat and the formation of large droplet macrosteatosis, especially in the male rats. Currently, we do not have an explanation for the sexual dimorphic effect of zingerone on SREBP1c expression; thus, further investigation is required. Nonetheless, preceding studies have reported that sexual dimorphism in rats can be attributed to variance in rate and pattern of early development and responses to insult in male and female rats [55,56]. In vivo zingerone activates PPAR-α [57]. It is therefore possible that it exerted its antisteatotic effects by preventing the downregulation of PPAR-α Additionally, previous research had found that zingerone can suppress the expression of SREBP1c [33] to protect against fatty liver formation and lipid accumulation therefore our findings are in tandem with other research outcomes. Steatosis is generally reversible after ethanol withdrawal [58]. However, we observed microsteatosis in the livers of female rats that had an early single hit with alcohol, which was neither accompanied with a significant increase in liver fat content nor an effect on the lipid regulatory genes. In line with our findings, Shen et al., 2014 [59] also reported that adult female rat offspring prenatally exposed to ethanol also exhibited microsteatosis. Presently, it is difficult to provide an explanation to these phenomena. In our recent study we observed that oral administration of ethanol during the suckling reduced the plasma triglyceride concentration of female rat pups [38]. We therefore speculate that the observed microsteatosis is likely due to the neonatal alcohol-induced export of triglycerides from the plasma to the liver. In the current study, hepatic lobular inflammation and necrosis were not evident in the liver sections of male and female rats, suggesting that none of the interventions caused liver inflammation. Our finding is consistent with other studies that showed that in rodents the consumption of alcohol only did not induce significant hepatic inflammation and necrosis [60], except under prolonged (29 weeks) consumption of $40\%$ ethanol [61]. Additionally, the rats only developed simple steatosis without a progression to alcoholic steatohepatitis. Simple steatosis is not associated with inflammation [62,63,64], though given more time, this could have progressed to steatohepatitis. Plasma AST and ALT activities are considered surrogate biomarkers for liver injury [65]. Ethanol consumption that caused hepatocyte inflammation and necrosis was shown to be associated with increased plasma AST and ALT activity [66] (Huang et al., 2010) and Yamasaki et al., 2019 [67] reported no change in AST and ALT activity without inflammation and necrosis. However, Radic et al., 2019 [68] observed that despite causing hepatic focal inflammation and mild necrosis, the consumption of ethanol did not affect plasma AST and ALT activity of rats suggesting poor correlation between liver enzyme activity and damage [69]. Our findings showed similarities in histological findings on hepatocytes with no evidence of necrosis and inflammation which explains similarities in plasma ALT and AST concentration of the rats since there was no liver cell damage. The present data indicate that eight weeks of ethanol consumption by the rats was not sufficient to induce liver cell injury. Importantly, we show that neonatal orally administered zingerone did not cause liver inflammation and injury and this finding is in agreement with its previously reported effects in adult rats [36,70,71]. Our findings showed that ethanol consumption in the late single and double hit groups elevated CYP2E1 concentration of the rats. CYP2E1 catalyses the metabolism of alcohol [27] in the liver. Chronic ethanol consumption induces CYP2E1 activity in order to facilitate ethanol metabolism [27]. Additionally, ethanol has been reported to induce and stabilize CYP2E1 proteins post-translationally [72]. Contrary to our finding, Kolata et al., 2020 [73] did not observe an increased CYP2E1 protein in male Wistar rats that consumed $10\%$ ethanol solution for 6 weeks. This variance might be because we used $20\%$ ethanol, a concentration that was $100\%$ higher. Furthermore, the current study used Sprague–Dawley rats, which could have also brought differences in rat strain responses. In the current study, we observed that neonatal orally administered zingerone did not affect the alcohol-induced increase in CYP2E1 proteins. Acute pharmacokinetic study of zingerone in rats via mass spectrophotometry demonstrated that it does not significantly affect of the microsomal content of P450 enzymes [74]. Therefore, zingerone cannot be used to promote alcohol abstinence as the induction of CYP2E1 is associated with greater tolerance to high ethanol intake [27]. The induction of CYP2E1 is associated with alcohol-induced oxidative stress via the generation of acetaldehyde and increased production of free radical species, which can form adducts with DNA and cause oxidative tissue injury [27]. However, our findings recorded similar hepatic TBARS concentration across treatment regimens, which suggests no initiation of significant lipid peroxidation in both male and female rats. Our findings are in agreement with the works of Kołota et al., 2020 [73] and Radic et al., 2019 [68] but at variance with Teare et al., 1994 [75] and Keegan et al., 1995 [61] due possibly to the difference in experimental duration and the development of more advanced stages of ALD. Therefore, our findings corroborate the hypothesis that oxidants from CYP2E1 play a minor role in the mechanisms involved in the early stages of ALD [76], although extended alcohol consumption or a more sensitive biomarker such as F2-isoprostanes might have yielded a significant result, as TBARs are not considered sensitive and reliable biomarkers of lipid peroxidation due to their reactivity and metabolism [77]. Late single and double hit with alcohol increased hepatic TBARs by $44.1\%$ and $52.0\%$, respectively, in the female rats. However neonatal orally administered zingerone reduced TBARS by $27.1\%$ and $23.7\%$ in the late single and double alcohol hit groups, respectively. Our study corroborates previous reports that indicate that zingerone markedly reduces lipid peroxidation via its free radical scavenging ability [36,78]. While it has been previously reported that zingerone can reduce high-fat diet induced steatosis and its associated inflammation, what is not known is its ability to programme for long-term protection against alcohol-induced fatty liver disease. ## 5. Conclusions The present study showed that a late and double hit with alcohol in rats resulted in the development of AFLD, characterised by small and large droplet macrosteatosis, with a downregulation in PPAR-α and an upregulation in SREBP1c without significant inflammatory changes or elevation in liver enzymes. Alcohol fatty liver disease renders the liver susceptible to toxic effects of alcohol and other insults. However, neonatal orally administered zingerone attenuated AFLD in a sex-dependent manner. Zingerone can therefore be strategically administered in the neonatal phase as a potential prophylactic agent for its beneficial effects against AFLD and ability to blunt the development ALD, and lessen the burden of ALD on the healthcare system. This information should be taken into consideration when developing guidelines regarding alcohol consumption during breastfeeding. Additionally, this study recommends fortifying diets of breastfeeding mothers with ginger, a rich source of zingerone, for its prophylactic benefits against diseases. ## References 1. Tawa E.A., Hall S.D., Lohoff F.W.. Overview of the Genetics of Alcohol Use Disorder. Alcohol. Alcohol. (2016.0) 51 507-514. DOI: 10.1093/alcalc/agw046 2. Siomek-Gorecka A., Dlugosz A., Czarnecki D.. The Molecular Basis of Alcohol Use Disorder (AUD). Genetics, Epigenetics, and Nutrition in AUD: An Amazing Triangle. Int. J. Mol. Sci. (2021.0) 22. DOI: 10.3390/ijms22084262 3. Nizhnikov M.E., Popoola D.O., Cameron N.. Transgenerational Transmission of the Effect of Gestational Ethanol Exposure on Ethanol Use-Related Behavior. Alcohol. Clin. Exp. Res. (2016.0) 40 497-506. DOI: 10.1111/acer.12978 4. Meroni M., Longo M., Rametta R., Dongiovanni P.. Genetic and Epigenetic Modifiers of Alcoholic Liver Disease. Int. J. Mol. Sci. (2018.0) 19. DOI: 10.3390/ijms19123857 5. Gluckman P.D., Hanson M.A., Mitchell M.D.. Developmental origins of health and disease: Reducing the burden of chronic disease in the next generation. Genome Med. (2010.0) 2 14. DOI: 10.1186/gm135 6. Ellis P.J., Morris T.J., Skinner B.M., Sargent C.A., Vickers M.H., Gluckman P.D., Gilmour S., Affara N.A.. Thrifty metabolic programming in rats is induced by both maternal undernutrition and postnatal leptin treatment, but masked in the presence of both: Implications for models of developmental programming. BMC Genom. (2014.0) 15. DOI: 10.1186/1471-2164-15-49 7. Asiedu B., Nyakudya T.T., Lembede B.W., Chivandi E.. Early-life exposure to alcohol and the risk of alcohol-induced liver disease in adulthood. Birth Defects Res. (2021.0) 113 451-468. DOI: 10.1002/bdr2.1881 8. Hannigan J.H., Chiodo L.M., Sokol R.J., Janisse J., Delaney-Black V.. Prenatal alcohol exposure selectively enhances young adult perceived pleasantness of alcohol odors. Physiol. Behav. (2015.0) 148 71-77. DOI: 10.1016/j.physbeh.2015.01.019 9. Gaztañaga M., Angulo-Alcalde A., Chotro M.G.. Prenatal Alcohol Exposure as a Case of Involuntary Early Onset of Alcohol Use: Consequences and Proposed Mechanisms from Animal Studies. Front. Behav. Neurosci. (2020.0) 14 26. DOI: 10.3389/fnbeh.2020.00026 10. Rangmar J., Dahlgren-Sandberg A., Aronson M., Fahlke C.. Cognitive and executive functions, social cognition and sense of coherence in adults with fetal alcohol syndrome. Nord. J. Psychiatry (2015.0) 69 1754-1760. DOI: 10.3109/08039488.2015.1009487 11. Hagström H., Hemmingsson T., Discacciati A., Andreasson A.. Alcohol consumption in late adolescence is associated with an increased risk of severe liver disease later in life. J. Hepatol. (2018.0) 68 505-510. DOI: 10.1016/j.jhep.2017.11.019 12. Ellsworth L., Harman E., Padmanabhan V., Gregg B.. Lactational programming of glucose homeostasis: A window of opportunity. Reproduction (2018.0) 156 R23-R42. DOI: 10.1530/REP-17-0780 13. Muriel P.. The Liver. Liver Pathophysiology (2017.0) 3-22 14. Singal A.K., Anand B.S.. Recent trends in the epidemiology of alcoholic liver disease. Clin. Liver Dis. (2013.0) 2 53-56. DOI: 10.1002/cld.168 15. WHO\|Global Status Report on Alcohol and Health 2018 16. Gao B., Bataller R.. Alcoholic Liver Disease: Pathogenesis and New Therapeutic Targets. Gastroenterology (2011.0) 141 1572-1585. DOI: 10.1053/j.gastro.2011.09.002 17. Teschke R.. Alcoholic Liver Disease: Alcohol Metabolism, Cascade of Molecular Mechanisms, Cellular Targets, and Clinical Aspects. Biomedicines (2018.0) 6. DOI: 10.3390/biomedicines6040106 18. Ohashi K., Pimienta M., Seki E.. Alcoholic liver disease: A current molecular and clinical perspective. Liver Res. (2018.0) 2 161-172. DOI: 10.1016/j.livres.2018.11.002 19. Nassir F., Rector R.S., Hammoud G.M., Ibdah J.A.. Pathogenesis and Prevention of Hepatic Steatosis. Gastroenterol. Hepatol. (2015.0) 11 167-175 20. Yeh M.M., Brunt E.M.. Pathological Features of Fatty Liver Disease. Gastroenterology (2014.0) 147 754-764. DOI: 10.1053/j.gastro.2014.07.056 21. Mostafa M., Abdelkader A., Evans J.J., Hagen C.E., Hartley C.P.. Fatty Liver Disease: A Practical Approach. Arch. Pathol. Lab. Med. (2020.0) 144 62-70. DOI: 10.5858/arpa.2019-0341-RA 22. Theise N.D.. Histopathology of alcoholic liver disease. Clin. Liver Dis. (2013.0) 2 64-67. DOI: 10.1002/cld.172 23. Sakhuja P.. Pathology of alcoholic liver disease, can it be differentiated from nonalcoholic steatohepatitis?. World J. Gastroenterol. (2014.0) 20 16474-16479. DOI: 10.3748/wjg.v20.i44.16474 24. Tandra S., Yeh M.M., Brunt E.M., Vuppalanchi R., Cummings O.W., Ünalp-Arida A., Wilson L.A., Chalasani N.. Presence and significance of microvesicular steatosis in nonalcoholic fatty liver disease. J. Hepatol. (2011.0) 55 654-659. DOI: 10.1016/j.jhep.2010.11.021 25. Dunn W., Shah V.H.. Pathogenesis of Alcoholic Liver Disease. Clin. Liver Dis. (2016.0) 20 445-456. DOI: 10.1016/j.cld.2016.02.004 26. Gu J., Zhang Y., Xu D., Zhao Z., Zhang Y., Pan Y., Cao P., Wang Z., Chen Y.. Ethanol-induced hepatic steatosis is modulated by glycogen level in the liver. J. Lipid Res. (2015.0) 56 1329-1339. DOI: 10.1194/jlr.M056978 27. Osna N.A., Donohue T.M., Kharbanda K.K.. Alcoholic Liver Disease: Pathogenesis and Current Management. Alcohol Res. Curr. Rev. (2017.0) 38 147-161 28. Stickel F., Datz C., Hampe J., Bataller R.. Pathophysiology and Management of Alcoholic Liver Disease: Update 2016. Gut Liver (2017.0) 11 173-188. DOI: 10.5009/gnl16477 29. Deaciuc I.V., D’Souza N.B., Spitzer J.J.. Tumor Necrosis Factor-alpha Cell-Surface Receptors of Liver Parenchymal and Nonparenchymal Cells during Acute and Chronic Alcohol Administration to Rats. Alcohol. Clin. Exp. Res. (1995.0) 19 332-338. DOI: 10.1111/j.1530-0277.1995.tb01511.x 30. Petrasek J., Csak T., Szabo G.. Toll-Like Receptors in Liver Disease. Adv. Clin. Chem. (2012.0) 59 155-201. DOI: 10.1016/b978-0-12-405211-6.00006-1 31. Bessone F., Razori M.V., Roma M.G.. Molecular pathways of nonalcoholic fatty liver disease development and progression. Cell. Mol. Life Sci. (2019.0) 76 99-128. DOI: 10.1007/s00018-018-2947-0 32. Nyakudya T.T., Mukwevho E., Nkomozepi P., Erlwanger K.H.. Neonatal intake of oleanolic acid attenuates the subsequent development of high fructose diet-induced non-alcoholic fatty liver disease in rats. J. Dev. Orig. Health Dis. (2018.0) 9 500-510. DOI: 10.1017/S2040174418000259 33. Mohammed H.M.. Zingerone ameliorates non-alcoholic fatty liver disease in rats by activating AMPK. J. Food Biochem. (2022.0) 46 e14149. DOI: 10.1111/jfbc.14149 34. Ajah A.A., Lembede B.W., Nkomozepi P., Erlwanger K.H., Nyakudya T.T.. Neonatal Oral Administration of Chrysin Prevents Long-Term Development of Non-Alcoholic Fatty Liver Disease in a Sexually Dimorphic Manner in Fructose Nurtured Sprague Dawley Rats. Life (2022.0) 12. DOI: 10.3390/life12060790 35. Muhammad N., Lembede B.W., Erlwanger K.H.. Zingerone Administered Neonatally Prevents the Subsequent Development of High Dietary Fructose-Induced Fatty Liver in Sprague Dawley Rats. J. Med. Food (2021.0) 24 944-952. DOI: 10.1089/jmf.2020.0189 36. Mani V., Arivalagan S., Siddique A.I., Namasivayam N.. Antioxidant and anti-inflammatory role of zingerone in ethanol-induced hepatotoxicity. Mol. Cell. Biochem. (2016.0) 421 169-181. DOI: 10.1007/s11010-016-2798-7 37. Cheslock S.J., Varlinskaya E.I., Silveri M.M., Petrov E.S., Spear L.P., Spear N.E.. Acute Effects of Ethanol and the First Suckling Episode in the Newborn Rat. Alcohol. Clin. Exp. Res. (2000.0) 24 996-1002. DOI: 10.1111/j.1530-0277.2000.tb04642.x 38. Asiedu B., Lembede B.W., Nyakudya T.T., Chivandi E.. Orally administered zingerone does not mitigate alcohol-induced hepatic oxidative stress in growing Sprague Dawley rat pups. Drug Chem. Toxicol. (2022.0) 0 1-10. DOI: 10.1080/01480545.2022.2085740 39. Ojeda M.L., Delgado-Villa M.J., Llopis R., Murillo M.L., Carreras O.. Lipid Metabolism in Ethanol-Treated Rat Pups and Adults: Effects of Folic Acid. Alcohol Alcohol. (2008.0) 43 544-550. DOI: 10.1093/alcalc/agn044 40. Aguiar A., Da-Silva V., Boaventura G.. Can calories from ethanol contribute to body weight preservation by malnourished rats?. Braz. J. Med. Biol. Res. (2004.0) 37 841-846. DOI: 10.1590/S0100-879X2004000600009 41. Veale W.L., Myers R.D.. Tables for Determining Grams and Caloric Values of Ethanol Solutions. Purdue Neuropscycological Ser. (1968.0) 1 1-10 42. Niehaus W.G., Samuelsson B.. Formation of Malonaldehyde from Phospholipid Arachidonate during Microsomal Lipid Peroxidation. JBIC J. Biol. Inorg. Chem. (1968.0) 6 126-130. DOI: 10.1111/j.1432-1033.1968.tb00428.x 43. Hilakivi L.. Effects of Prenatal Alcohol Exposure on Neonatal Sleep-wake Behaviour and Adult Alcohol Consumption in Rats. Acta Pharmacol. Toxicol. (2009.0) 59 36-42. DOI: 10.1111/j.1600-0773.1986.tb00131.x 44. Biggio F., Talani G., Locci V., Pisu M., Boero G., Ciarlo B., Grayson D., Serra M.. Low doses of prenatal ethanol exposure and maternal separation alter HPA axis function and ethanol consumption in adult male rats. Neuropharmacology (2018.0) 131 271-281. DOI: 10.1016/j.neuropharm.2017.12.005 45. Bordner K., Deak T.. Endogenous opioids as substrates for ethanol intake in the neonatal rat: The impact of prenatal ethanol exposure on the opioid family in the early postnatal period. Physiol. Behav. (2015.0) 148 100-110. DOI: 10.1016/j.physbeh.2015.02.013 46. van de Wiel J.A.G., Duijf C.M.P., Pertijs J.C.L.M., Peereboom-Stegeman J.H.J.C., Bos R.P.. Growth and liver morphology after long-term ethanol consumption of rats. Lab. Anim. (1990.0) 24 265-272. DOI: 10.1258/002367790780866263 47. Al-Humadi H., Al-Saigh R., Sahib A.. The impact of low alcohol consumption on the liver and inflammatory cytokines in diabetic rats. Adv. Clin. Exp. Med. (2018.0) 28 331-337. DOI: 10.17219/acem/77085 48. Rasineni K., Thomes P.G., Kubik J.L., Harris E.N., Kharbanda K.K., Casey C.A.. Chronic alcohol exposure alters circulating insulin and ghrelin levels: Role of ghrelin in hepatic steatosis. Am. J. Physiol. Liver Physiol. (2019.0) 316 G453-G461. DOI: 10.1152/ajpgi.00334.2018 49. You M., Arteel G.E.. Effect of ethanol on lipid metabolism. J. Hepatol. (2019.0) 70 237-248. DOI: 10.1016/j.jhep.2018.10.037 50. You M., Fischer M., Deeg M.A., Crabb D.W.. Ethanol Induces Fatty Acid Synthesis Pathways by Activation of Sterol Regulatory Element-binding Protein (SREBP). J. Biol. Chem. (2002.0) 277 29342-29347. DOI: 10.1074/jbc.M202411200 51. Ji C., Shinohara M., Vance D., Than T.A., Ookhtens M., Chan C., Kaplowitz N.. Effect of Transgenic Extrahepatic Expression of Betaine-Homocysteine Methyltransferase on Alcohol or Homocysteine-Induced Fatty Liver. Alcohol. Clin. Exp. Res. (2008.0) 32 1049-1058. DOI: 10.1111/j.1530-0277.2008.00666.x 52. Park H.J., Lee S.-J., Song Y., Jang S.-H., Ko Y.-G., Kang S.N., Chung B.Y., Kim H.-D., Kim G.-S., Cho J.-H.. J. Med. Food (2014.0) 17 103-110. DOI: 10.1089/jmf.2013.2849 53. Jiang Z., Zhou J., Zhou D., Zhu Z., Sun L., Nanji A.A.. The Adiponectin-SIRT1-AMPK Pathway in Alcoholic Fatty Liver Disease in the Rat. Alcohol. Clin. Exp. Res. (2015.0) 39 424-433. DOI: 10.1111/acer.12641 54. Straub B.K.. Lipid droplet-associated proteins. Importance in steatosis, steatohepatitis and hepatocarcinogenesis. Der Pathol. (2015.0) 36 146-152. DOI: 10.1007/s00292-015-0082-3 55. Sundrani D.P., Roy S.S., Jadhav A.T., Joshi S.R.. Sex-specific differences and developmental programming for diseases in later life. Reprod. Fertil. Dev. (2017.0) 29 2085-2099. DOI: 10.1071/RD16265 56. Lembede B.W., Joubert J., Nkomozepi P., Erlwanger K.H., Chivandi E.. Insulinotropic Effect of S-Allyl Cysteine in Rat Pups. Prev. Nutr. Food Sci. (2018.0) 23 15-21. DOI: 10.3746/pnf.2018.23.1.15 57. Chung S.W., Kim M.K., Chung J.H., Kim D.H., Choi J.S., Anton S., Seo A.Y., Park K.-Y., Yokozawa T., Rhee S.H.. Peroxisome Proliferator-Activated Receptor Activation by a Short-Term Feeding of Zingerone in Aged Rats. J. Med. Food (2009.0) 12 345-350. DOI: 10.1089/jmf.2007.0660 58. Seitz H.K., Bataller R., Cortez-Pinto H., Gao B., Gual A., Lackner C., Mathurin P., Mueller S., Szabo G., Tsukamoto H.. Alcoholic liver disease. Nat. Rev. Dis. Primers (2018.0) 4 16. DOI: 10.1038/s41572-018-0014-7 59. Shen L., Liu Z., Gong J., Zhang L., Wang L., Magdalou J., Chen L., Wang H.. Prenatal ethanol exposure programs an increased susceptibility of non-alcoholic fatty liver disease in female adult offspring rats. Toxicol. Appl. Pharmacol. (2014.0) 274 263-273. DOI: 10.1016/j.taap.2013.11.009 60. Larosche I., Choumar A., Fromenty B., Lettéron P., Abbey-Toby A., Van Remmen H., Epstein C.J., Richardson A., Feldmann G., Pessayre D.. Prolonged ethanol administration depletes mitochondrial DNA in MnSOD-overexpressing transgenic mice, but not in their wild type littermates. Toxicol. Appl. Pharmacol. (2008.0) 234 326-338. DOI: 10.1016/j.taap.2008.11.004 61. Keegan A., Martini R., Batey R.. Ethanol-related liver injury in the rat: A model of steatosis, inflammation and pericentral fibrosis. J. Hepatol. (1995.0) 23 591-600. DOI: 10.1016/0168-8278(95)80067-0 62. Greuter T., Malhi H., Gores G.J., Shah V.H.. Therapeutic opportunities for alcoholic steatohepatitis and nonalcoholic steatohepatitis: Exploiting similarities and differences in pathogenesis. J. Clin. Investig. (2017.0) 2 e95354. DOI: 10.1172/jci.insight.95354 63. Mazzolini G., Sowa J.-P., Atorrasagasti C., Kücükoglu Ö., Syn W.-K., Canbay A.. Significance of Simple Steatosis: An Update on the Clinical and Molecular Evidence. Cells (2020.0) 9. DOI: 10.3390/cells9112458 64. Heeren J., Scheja L.. Metabolic-associated fatty liver disease and lipoprotein metabolism. Mol. Metab. (2021.0) 50 101238. DOI: 10.1016/j.molmet.2021.101238 65. Morales-González J.A., Sernas-Morales M.D.L., Morales-Gonzalez A., González-López L.L., Madrigal-Santillán E.O., Vargas-Mendoza N., Fregoso-Aguilar T.A., Anguiano-Robledo L., Madrigal-Bujaidar E., Álvarez-González I.. Morphological and biochemical effects of weekend alcohol consumption in rats: Role of concentration and gender. World J. Hepatol. (2018.0) 10 297-307. DOI: 10.4254/wjh.v10.i2.297 66. Huang C.-H., Chang Y.-Y., Liu C.-W., Kang W.-Y., Lin Y.-L., Chang H.-C., Chen Y.-C.. Fruiting Body of Niuchangchih (. J. Agric. Food Chem. (2010.0) 58 3859-3866. DOI: 10.1021/jf100530c 67. Yamasaki K., Sugamoto K., Arakawa T., Nishiyama K., Yamasaki M.. Chronic intake of high-dose of blueberry leaf extract does not augment the harmful effects of ethanol in rats. PeerJ (2019.0) 7 e6989. DOI: 10.7717/peerj.6989 68. Radic I., Mijovic M., Tatalovic N., Mitic M., Lukic V., Joksimovic B., Petrovic Z., Ristic S., Velickovic S., Nestorovic V.. Protective effects of whey on rat liver damage induced by chronic alcohol intake. Hum. Exp. Toxicol. (2019.0) 38 632-645. DOI: 10.1177/0960327119829518 69. Morales-Gonzalez J.A., Gutierrez-Salinas J., Yanez L., Villagomez-Rico C., Badillo-Romero J., Hernandez-Munoz R.. Morphological and biochemical effects of a low ethanol dose on rat liver regeneration: Role of route and timing of administration. Dig. Dis. Sci. (1999.0) 44 1963-1974. DOI: 10.1023/A:1026601814082 70. Cheong K.O., Shin D.-S., Bak J., Lee C., Kim K.W., Je N.K., Chung H.Y., Yoon S., Moon J.-O.. Hepatoprotective effects of zingerone on carbon tetrachloride- and dimethylnitrosamine-induced liver injuries in rats. Arch. Pharmacal Res. (2015.0) 39 279-291. DOI: 10.1007/s12272-015-0696-2 71. Narayanan J.M., Jesudoss V.A.S.. Hepatoprotective potential of zingerone against nonalcoholic fatty liver disease in rats fed with fructose-enriched diet. Gen. Physiol. Biophys. (2016.0) 35 185-194. DOI: 10.4149/gpb_2015041 72. Roberts B.J., Song B.-J., Soh Y., Park S.S., Shoaf S.E.. Ethanol Induces CYP2E1 by Protein Stabilization. Role of Ubiquitin Conjugation in the Rapid Degradation of CYP2E1. J. Biol. Chem. (1995.0) 270 29632-29635. DOI: 10.1074/jbc.270.50.29632 73. Kołota A., Głąbska D., Oczkowski M., Gromadzka-Ostrowska J.. Oxidative Stress Parameters in the Liver of Growing Male Rats Receiving Various Alcoholic Beverages. Nutrients (2020.0) 12. DOI: 10.3390/nu12010158 74. Huang X.. Pharmacokinetics Studies Of Zingerone In Rats 75. Teare J., Greenfield S.M., Watson D., Punchard N., Miller N., Rice-Evans C.A., Thompson R.P.. Lipid peroxidation in rats chronically fed ethanol. Gut (1994.0) 35 1644-1647. DOI: 10.1136/gut.35.11.1644 76. Kono H., Bradford B.U., Yin M., Sulik K.K., Koop D.R., Peters J.M., Gonzalez F.J., McDonald T., Dikalova A., Kadiiska M.B.. CYP2E1 is not involved in early alcohol-induced liver injury. Am. J. Physiol. Liver Physiol. (1999.0) 277 G1259-G1267. DOI: 10.1152/ajpgi.1999.277.6.G1259 77. Milne G.L.. Classifying oxidative stress by F2-Isoprostane levels in human disease: The re-imagining of a biomarker. Redox Biol. (2017.0) 12 897-898. DOI: 10.1016/j.redox.2017.04.028 78. Vinothkumar R., Vinothkumar R., Sudha M., Nalini N.. Chemopreventive effect of zingerone against colon carcinogenesis induced by 1,2-dimethylhydrazine in rats. Eur. J. Cancer Prev. (2014.0) 23 361-371. DOI: 10.1097/CEJ.0b013e32836473ac
--- title: The Effect of Extraction Methods on Phytochemicals and Biological Activities of Green Coffee Beans Extracts authors: - Octavia Gligor - Simona Clichici - Remus Moldovan - Dana Muntean - Ana-Maria Vlase - George Cosmin Nadăș - Ioana Adriana Matei - Gabriela Adriana Filip - Laurian Vlase - Gianina Crișan journal: Plants year: 2023 pmcid: PMC9966978 doi: 10.3390/plants12040712 license: CC BY 4.0 --- # The Effect of Extraction Methods on Phytochemicals and Biological Activities of Green Coffee Beans Extracts ## Abstract The objectives of the present study consisted of identifying the impact of extraction methods and parameters held over the phytochemistry and biological activities of green coffee beans. Extraction processes belonging to two categories were performed: classical methods—maceration, Soxhlet extraction, and such innovative methods as turboextraction, ultrasound-assisted extraction, and a combination of the latter two. Total polyphenolic and flavonoid content, as well as in vitro antioxidant activity of the resulted extracts were spectrophotometrically determined. Extracts displaying the highest yields of bioactive compounds were subjected to High Performance Liquid Chromatography-Mass Spectrometry analysis. The extracts with the best phytochemical profiles were selected for biological activity assessment. In vivo, a model of plantar inflammation in Wistar rats was used to determine antioxidant activity, by evaluating the oxidative stress reduction potential, and anti-inflammatory activity. In vitro antimicrobial activity was also determined. The Soxhlet extraction and ultrasound-assisted extraction gave the highest bioactive compound yields. The highest total polyphenolic content was 2.691 mg/mL gallic acid equivalents and total flavonoid content was 0.487 mM quercetin equivalents for the Soxhlet extract subjected to 60 min extraction time. Regarding the antioxidant activity, ultrasound-assisted extraction reached the highest levels, i.e., 9.160 mg/mL Trolox equivalents in the DPPH (2,2-diphenyl-1-picryl-hydrazyl-hydrate) assay and 26.676 mM Trolox equivalents in the FRAP (Ferric Reducing Antioxidant Power) assay, at a 30 min extraction time and 50 °C extraction temperature. The 60 min Soxhlet extract reached the highest level for the ABTS+ (2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)) assay, 16.136 mM Trolox equivalents, respectively. Chlorogenic acid was present in the highest concentration in the same Soxhlet extract, 1657.179 µg/mL extract, respectively. Sterolic compounds were found in high concentrations throughout all the analyzed extracts. A proportional increase between yields and extraction parameter values was observed. Increased inhibition of Gram-negative bacteria was observed. The finally selected Soxhlet extract, that of 60 min extraction time, presented a significant in vivo antioxidant activity, with a slight anti-inflammatory activity. Antioxidant levels were elevated after 2 h of extract administration. Pro-inflammatory cytokine secretion was not influenced by the administration of the extract. ## 1. Introduction Coffee beans refer to the seeds of the coffee plant (the genus Coffea L.), botanically considered as a shrub or tree. The coffee plant is part of the botanical family Rubiaceae. Presently, the family includes hundreds of plant species, with the genus Coffea encompassing over 70 distinct species. Two of these species, present economic importance—Coffea arabica L. (often named Arabica) and Coffea canephora Pierre ex A. Froehner (often named Robusta). This article focuses on the species that was first mentioned, Coffea arabica L. The evergreen plant grows up to 10 m in height, consisting of a main stem, originating secondary branches, with leaves being disposed in an opposite decussate arrangement and colored dark green. The seeds of the plant are located inside the fruits, named cherries or berries. The seeds are elliptical in shape. [ 1,2,3]. The process of roasting, although holding great industrial importance, due to its implication in the palatability of the final beverage, has been noted to cause the degradation of numerous bioactive compounds beneficial for human health, such as polyphenolic compounds, polysaccharides, proteins, and many others. This matter subsequently reduces the biological potential of the plant material [4,5,6]. Green coffee beans have been reported to present numerous biological activities, such as obesity reduction, type II diabetes prevention, reducing oxidative stress, improving cardiovascular parameters, reducing the risk of chronic hepatic disorders, and also neuroprotective, antitumor, antioxidant, anti-inflammatory, and antimicrobial effects [7,8,9,10]. This wide array of biological activities is mainly considered to stem from the polyphenolic compounds found in this plant material [2]. In order to ensure the availability of such bioactive compounds, along with offering high recovery rates from any plant matrix, the extraction process holds much importance. Generally, extraction methods are divided into two main categories, based on the era in which they appeared. Names such as “conventional” or “classical” methods are attributed to the methods used prior to the end of the 20th century. Whereas the terms “innovative” or “emerging” extraction methods, are used for those methods that appeared thereafter. Conventional methods entail high temperatures, long periods of time in order for the extraction process to reach completion, large solvent volumes, large amounts of plant material, as well as often hazardous solvents. Maceration, decoction, Soxhlet extraction constitute examples of such conventional methods. These characteristics are presently regarded as unfavorable, thus urging for safer, less time-consuming, alternative means of bioactive compound extraction methods to be discovered, with lower amounts of alternative solvents and plant matrix also being required. Thus, bearing the name “green extraction methods” due to their indicative safety, microwave-assisted extraction, ultrasound-assisted extraction, pressurized liquid extraction, etc. are considered clear examples of such methods [11,12]. While several reports on green coffee beans provide outlooks over the influence of extraction method and parameters over bioactive compounds yields, few have focused on the importance of such aspects of the extraction process over the biological activities of the resulted extracts [13,14,15]. The purpose of this article was to provide a better understanding of the relationship between extraction type and extraction conditions and the final extraction yields, as well as to determine the influence such extraction parameters might hold over the biological properties of the obtained green coffee bean extracts, such as antimicrobial activity, antioxidant, and anti-inflammatory effects. ## 2. Results Sample nomenclature was based on extraction method and the parameters that were studied for each sample: M for maceration, S for Soxhlet extraction (SE), U for Ultrasound-assisted extraction (UAE), T for turboextraction (TBE), and UT for the combination for the UAE and TBE (UTE). A number of 20 different samples resulted after extraction. One sample was obtained after maceration (M), 3 samples were obtained through SE (S), 6 samples were obtained by TBE (T), 9 samples were obtained by UAE (U), and the last sample was obtained by UTE (UT). The nomenclature of the employed extraction methods, as well as that of the obtained samples is clarified in Table 1. Total polyphenolic content (TPC) and total flavonoid content (TFC), as well as antioxidant capacity were evaluated for the samples listed above, following the methods detailed in chapter 4. Materials and Methods. Chromatographic evaluation followed only for the samples which presented the highest values for the analysis processes previously stated. The final step of the study consisted of the biological activity assessment of the samples containing the highest yields of bioactive compounds. ## 2.1. Influence of Extraction Parameters on TPC and TFC values Table 2 depicts the results of TPC and TFC assessment of the green coffee bean samples. As seen in the respective table, the SE method attained the highest content of polyphenols, with the 60 min extraction period being superior to all samples. SE values were followed by maceration and TBE in regard to polyphenolic yields. UAE and UTE achieved the lowest levels of yields for both assessments. ## 2.2. Influence of Extraction Parameters on In Vitro Antioxidant Capacity Results for in vitro antioxidant capacity assessment are detailed in Table 3. Following the DPPH assay, no major differences were observed throughout the samples, excepting the UAE sample with the extraction conditions of 30 min extraction time and 50 °C extraction temperature, i.e., sample U35. This value was closely followed by that of the extract obtained through the combination of UAE and TBE, i.e., sample UT. For the FRAP assay, the same UAE sample, U35, showed the highest value. However, notable differences between the rest of the extraction methods were observed, namely SE, i.e., samples S20 and S60, as well as UAE, i.e., samples U15 and U23, which reached medium to high levels. For UAE, in this case, there was a visible decrease in levels for samples that were subjected to higher time periods and temperature values. For TBE, an increase in antioxidant levels could be observed, along with the increase in the values of the extraction parameters, up until the highest of these values were reached, namely the parameters of 4 cycles of 5 min (20 min) with 6000 rpm speed, i.e., sample T46. After which, an abrupt decrease in the antioxidant levels in this case, for sample T48. This extract was obtained at the same extraction time but at 8000 rpm. Results were evidently different for the ABTS+ assay. In this case, the Soxhlet sample with the 60 min extraction period, S60, as well as all TBE samples (of both extraction times, 10 min, and 20 min, respectively) reached the highest levels. ## 2.3. HPLC-MS Analysis of the Extracts Following the assessments of the previous subchapters, Section 2.1. Influence of extraction parameters on TPC and TFC values, and Section 2.2. In vitro antioxidant capacity of the extracts, 11 extracts were selected for further analysis. The main criterion for selection was the highest yield level in the extracts. Thus, the samples were screened for polyphenolic compounds, flavonoid compounds, and sterolic compounds (Table 4). ## 2.3.1. Analysis of Polyphenolic and Flavonoid Compounds Only one polyphenolic compound, chlorogenic acid, was identified in the samples that were selected for chromatographic evaluation. The highest yield of chlorogenic acid was recorded for the SE sample with extraction time of 60 min, sample S60. Other notable yields were obtained also through SE, with extraction time of 20 min and 40 min, samples S20 and S40, respectively. TBE samples reached comparable yields, namely the conditions of 4 cycles of 5 min and rotation speed of 4000 rpm, and 6000 rpm, samples T44 and T46, respectively. Maceration, UAE, and UTE reached only low to medium yields of chlorogenic acid. Additionally, kaempferol was the only flavonoid compound identified in one of the samples. The extract produced through maceration, sample M, was the only extract in which kaempferol reached quantifiable levels. ## 2.3.2. Analysis of Sterolic Compounds 6 sterolic compounds were found in high levels in the majority of the samples: α-tocopherol, γ-tocopherol, ergosterol, stigmasterol, β-sitosterol, and campesterol. Results are presented in Table 4. UAE was the only method that managed the extraction of α-tocopherol. All 3 UAE samples along with the sample resulted from UTE, sample UT, presented nearly equal values of α-tocopherol. γ-Tocopherol was identified in all UAE samples albeit in higher levels, depending on the extraction parameters. As in the case for its previously discussed isomer, UAE and UTE enabled the highest yields. For UAE, the extraction conditions of 30 min and 50 °C proved most advantageous, i.e., sample U35. Ergosterol was detected only partially in the samples, most notably, the S20 sample, followed by TBE samples, especially for the parameters of 4 cycles of 5 min and 4000 rpm, i.e., sample T44. The latter was followed closely by UAE (specifically 30 min, 50 °C—sample U35) and the UT sample. High levels of stigmasterol were also found in most of the samples, i.e., TBE (2 cycles of 5 min, and 4000 rpm—sample T24), UAE (30 min, 50 °C—sample U35), followed closely by maceration and SE (60 min—sample S60). β-Sitosterol, the sterolic compound present in the highest levels in the analyzed samples, was also better extracted by TBE (2 cycles of 5 min, 4000 rpm—sample T24) and UAE (30 min, 40 °C —sample U34), followed by SE (60 min extraction time—S60) and UTE. Campesterol was found in high levels in the 60 min SE sample (S60), followed by a TBE sample, i.e., 2 cycles of 5 min and 4000 rpm (T24), and a UAE sample, i.e., 30 min and 50 °C (U35). ## 2.4. Determination of Antimicrobial Activity Following the assessment of the phytochemical profile by HPLC-MS, the extracts that were considered to present the most favorable profiles were further subjected to antimicrobial activity determination. Therefore, extracts S60 and U35 were selected for this evaluation. ## 2.4.1. In Vitro Qualitative Study of Antimicrobial Activity The disk diffusion test was applied in order to assess the antimicrobial potential of the samples. The samples presented an increased efficiency against Gram-negative bacteria, a moderate efficiency against Gram-positive bacteria, and reduced activity towards Candida albicans. Values are shown in Table 5. Diameters of inhibition areas are as follows: 5.86 to 7.29 mm for Gram-positive species, 10.09 to 13.38 mm for Gram-negative bacteria, and 6.94 to 7.21 mm for Candida albicans. Results demonstrated an increased antimicrobial activity against Gram-negative bacteria. ## 2.4.2. In Vitro Quantitative Study of Antimicrobial Activity The potential against Gram-negative strains of the samples was evidenced by the first screening method. However, the MIC (minimum inhibitory concentration) method was utilized in order to assess the quantitative antimicrobial potential against all microbial species in the initial qualitative study. Table 6 presents the varied antimicrobial response of the extracts. ## 2.5. Assessment of Oxidative Stress and Inflammation Markers Extract S60 was selected for in vivo biological activity evaluation. As this extract was administered to the experimental animals solely against control groups receiving anti-inflammatory treatment (Indomethacin) or carboxymethylcellulose (CMC), the authors would like to note that extract S60 is to be referred simply by its extraction method name, that is SE, for the remainder of this chapter (S60 being the only extract obtained SE that was further studied). Lipid peroxidation marker (MDA) and endogenous antioxidant levels such as reduced glutathione (GSH), oxidated glutathione (GSSG), and their respective ratio (GSH/GSSG) were assessed in order to quantify the oxidative stress. Antioxidant enzymes were also measured by catalase (CAT) and glutathione peroxidase (GPx) activities in the plantar tissues. Results are displayed in Figure 1. MDA levels remained elevated at 2 h and 24 h for the SE treated group compared to control group while GSH levels increased significantly after 24 h compared to control and Indomethacin treated groups ($p \leq 0.05$ and $p \leq 0.01$). MDA levels decreased only in animals treated with Indomethacin, both at 2 h and 24 h compared to control ($p \leq 0.01$ and $p \leq 0.001$). GSSG levels diminished after SE treatment, at 2 h, compared to control group ($p \leq 0.01$), similar to the values recorded after Indomethacin treatment ($p \leq 0.001$). The effect of SE on GSSG formation was more pronounced at 24 h ($p \leq 0.001$) than at 2 h. Another noteworthy observation was the increase in the GSH/GSSG ratio at 2 h for Indomethacin group vs. control ($p \leq 0.05$). CAT activity was amplified significantly in group treated with Indomethacin, at 24 h after induction of inflammation, compared to control group while GPx activity decreased significantly after Indomethacin administration, both at 2 h ($p \leq 0.05$) and 24 h ($p \leq 0.05$). In SE group, GPx activity was comparable to the control group, both at 2 h and 24 h ($p \leq 0.05$). The levels of pro-inflammatory cytokines, i.e., IL-6 and TNF-α, were also evaluated in the plantar tissue, at 2 h and 24 h after induction of inflammation. Figure 2 portrays the obtained results. Thus, SE reduced IL-6 secretion in the paw tissue, at 2 h compared to control, but the results were statistically insignificant ($p \leq 0.05$). The IL-6 levels diminished significantly after Indomethacin administration ($p \leq 0.05$). SE treatment did not influence the IL-6 secretion in soft plantar tissue at 24 h after carrageenan injection ($p \leq 0.05$). TNF-α levels in the paw tissue, measured at 2 h after induction of inflammation, decreased only in Indomethacin group ($p \leq 0.05$) while SE maintained high levels of this cytokine, close to control group ($p \leq 0.05$) At 24 h, the treatments did not induce significant difference between groups in TNF-α secretion, in the soft plantar tissue. ## 3. Discussion The seeds of Coffea arabica L. comprise the majority of the worldwide production of coffee. They are situated within the fruits of the plant, which are also named cherries or berries. The fruit consists of an exocarp (skin of the fruit), a mesocarp containing pectins, an endocarp containing polysaccharides, and the silver skin, which coats the seed, containing polysaccharides, proteins, polyphenols, etc. A physical characteristic of the seeds is their elliptical shape [1,3]. Once harvested, the seeds are subjected to dehulling and processing by dry or wet methods, followed by roasting, which is responsible for conferring the aroma of the finally obtained coffee drink. However, roasting has been noted to cause the degradation of polysaccharides, lipids, chlorogenic acids, other polyphenolic compounds, trigonelline (a compound also responsible for flavor), and protein denaturation. It has been observed that these changes led to reduction in the subsequent biological activities of the plant material [4,5,6]. Pimpley et al. reviewed the effect of roasting of coffee beans over the content of chlorogenic acids (among which hydroxycinnamic acids are a part of, such as caffeic acid, ferulic acid and p-coumaric acid). The study concluded that up to $95\%$ of said compounds were degraded through intense roasting, leaving a very small percentage of these compounds in the plant material. In addition, coffee beans and pulp extracts were reported to inhibit lipid accumulation in cell cultures of adipocytes. Additionally, green coffee beans extracts were observed to reduce obesity and insulin resistance in mouse models. In human test subjects, a decaffeinated green coffee bean extract led to decreased metabolic syndrome markers, such as lipid profile, blood pressure, insulin resistance, etc. [ 7]. Martínez-López et al. reported that a moderate consumption of a green/roasted (35:65) blend improved cardiovascular parameters in human subjects with moderate hypercholesterolemia, such as serum lipid profile, blood pressure, body weight, while also increasing plasma antioxidant capacity [8]. Caro-Gómez et al. also noted ameliorated cardiometabolic syndrome parameters, such as fasting glucose, insulin resistance, liver triglyceride levels, as well as increasing IL-6 levels and positively impacting gut microbiota in ApoE−/− mice following an administration of green coffee beans extract [9]. A study conducted by Wang et al. on endothelial EA.hy926 cells pretreated with green coffee beans polyphenolic extracts, demonstrated endothelial protective effects by reducing the production of reactive oxygen species, and increasing endothelial nitric oxide synthase levels [10]. Recent scientific articles have investigated the influence of extraction methods over the biochemical profile of green coffee beans extracts. Yuniarti et al., for example, have optimized an extraction process assisted by natural deep eutectic solvents for caffeine and chlorogenic acid from the green beans pertaining to the species C. canephora L. Optimal conditions were found to be 4:1 for the mole ratio of choline chloride-sorbitol, 60 min of extraction time, and 1:30 g/mL for the liquid-solid ratio. The extraction method used as reference was maceration [13]. Menzio et al. managed to enhance mass transfer and selectivity of caffeine extraction from green coffee beans by combining supercritical CO2 extraction with UAE [14]. Gawlik-Dziki et al., apart from demonstrating the in vitro capacity of lowering lipoxygenase levels of methanolic green coffee beans extracts, also compared plant materials from different locations worldwide [15]. However, it is the authors’ opinion that further research must be conducted on the correlation between the extraction methods as well as extraction parameters and biological activity, particularly for this plant material. A tendency of increase in TPC was observed along with the increase in extraction time, for each of the extraction methods employed in this study. However, in the case of UAE, an increase in TPC was observed along with the increase in extraction temperature, as well. A similar report, although for a different version of plant material, spent filter coffee, by Pavlović et al. stated that the increase in extraction time led to a decrease in TPC, a finding not applicable in the case of this study [16]. Concerning the TFC of the samples, minor differences were observed throughout the majority of the samples, regardless of the extraction technique used. The sole exception to this observation was once again the Soxhlet extraction, namely the sample obtained after 60 min extraction time (sample S60). Al-Dhabi et al. have reported the positive influence that temperature and extraction time held over waste spent coffee grounds subjected to a UAE method. However, a further increase in these parameters was observed to obtain lower TPC and TFC values, due to the degradation of the compounds as a result of excessive exposure [17]. This observation was not found in this study, as extraction time for UAE did not exceed 30 min, and the highest temperature value was 50 °C. A previously cited study, led by Pavlović et al., also observed an increase in antioxidant capacity for DPPH and FRAP assays, with the decrease in extraction time, for the case of a microwave-assisted extraction for spent filter coffee. In addition, a proportional correlation between TPC values and antioxidant capacity (in the case of DPPH and FRAP assays) was observed for the previously mentioned study [16]. As seen in Table 3, no notable correlation between results for the employed assays could be observed within this study. At the very least, sample S60, which presented the highest TPC and TFC levels, registered the highest value for the ABTS+ assay as well. Contrary to the present HPLC findings, Nzekoue et al. have reported most of the quantified compounds mentioned above (see Table 4), for hydroalcoholic and hydromethanolic sonicated extracts, albeit the plant material of that study consisted of roasted coffee silver skin. Which is known to contain multiple unconjugated polyphenolic compounds, as opposed to green coffee beans [18]. Results for sterolic compounds were in accordance to previously reported data by Dong et al. for green coffee oil extracted by a UAE method [19]. Although an increase in extraction parameter values led to a somewhat improvement in concentrations for stigmasterol, β-sitosterol, and campesterol, the results were contrary for TBE. In that case, the extraction time of 4 cycles of 5 min (20 min) and 6000 rpm could have possibly contributed to the degradation of the compounds, see sample T46. The lowest levels for stigmasterol, β-sitosterol, and campesterol were also registered in UAE samples, namely, the conditions of 20 min extraction time and 30 °C (sample U23), which might indicate these extraction parameters as insufficient. Concerning the antimicrobial activity of the analyzed samples (samples U35 and S60), Nzekoue et al. have contrarily reported a low activity of similarly achieved extracts of roasted coffee silver skin [18]. Another noteworthy observation of the present study would consist in that of the lower MICs registered for the Gram-negative species. A possible reason for this phenomenon could be the limited diffusion on the agar surface as opposed to an appropriate overall inhibitory concentration in wells containing liquid MH medium. In terms of in vivo activity, the oxidative stress reduction potential of the analyzed extract, SE, could be a result of the present sterolic compounds which were reported to have antioxidant activity [20]. The results confirmed this finding by a good in vivo antioxidant effect of the SE extract, especially at 2 h after induction of inflammation. SE extract increased the reduced glutathione level and diminished the oxidation of glutathione in the plantar tissue, the effect having lasted even 24 h. The present results were comparable with those of recent scientific literature. Bhandarkar et al. have found that green coffee beans only ameliorated heart and liver inflammation in Wistar rats fed with high-carbohydrate, high-fat diet with green coffee extract ($5\%$ in food), with or without caffeine, compared to control groups. Results were correlated with those of rats which received a corn starch diet or a high-carbohydrate, high-fat diet [21]. Pergolizzi et al. found that the application of a C. robusta L. Linden ointment induced a prolonged and sustained anti-inflammatory effect on carrageenan-induced rat oedema [22]. In human subjects, Martínez-López et al. noticed a tendency towards the reduction in pro-inflammatory cytokines, among which IL-6 and TNF-α, associated with green coffee consumption by human subjects, albeit not significantly. However, MDA levels decreased after human coffee consumption [8]. According to the results of a meta-analysis of randomized clinical trials, performed by Asbaghi et al., such effects of green coffee beans are due to their phenolic compounds, i.e., caffeoylquinic acids. These compounds have been observed to reduce pro-inflammatory cytokines in both liver and white adipose tissue in humans, especially lowering IL-6 and TNF-α levels [23]. ## 4.1. Plant Material, Reagents, and Laboratory Equipment Ground green coffee beans (Coffea arabica L.) were procured from a local company (All For Nature, Timișoara, Timiș, Romania). Folin-Ciocâlteu reagent, sodium carbonate (Na2CO3), Aluminum chloride (AlCl3), ABTS+ (diammonium 2,2′-azino-bis(3-ethylbenzothiazoline-6-sulfonate), DPPH (2,2-Diphenyl-1-(2,4,6-trinitrophenyl)hydrazyl), TPTZ (2,4,6-Tris(2-pyridyl)-s-triazine), indomethacin, carboxymethylcellulose, o-phthalaldehyde, Lambda carrageenan type IV were purchased from Sigma–Aldrich (Taufkirchen, Germany). 2-thiobarbituric acid and Bradford reagent were acquired from Merck KGaA (Darmstadt, Germany) and ELISA cytokines tests (TNF-α and IL-6, respectively) were obtained from Elabscience (Houston, TX, USA). Bradford total protein assay was purchased from Biorad (Hercules, CA, USA). All analytical grade, HPLC reagents and standards were acquired from Sigma–Aldrich (Taufkirchen, Germany) and Decorias (Rediu, Romania). The following equipment was used for the present study: SER 148 solvent extraction unit (VELP® Scientifica, Usmate Velate, Italy), T 50 ULTRA-TURRAX® disperser (IKA®-Werke GmbH & Co. KG, Staufen, Germany), Sonic-3 ultrasonic bath (Polisonic, Warsaw, Poland), refrigerated high speed centrifuge Sigma 3-30KS (Sigma Laborzentrifugen GmbH, Osterode am Harz, Germany), Specord 200 Plus spectrophotometer (Analytik Jena, Jena, Germany), Agilent 1100 Series HPLC Value System coupled with an Agilent 1100 mass spectrometer (LC/MSD Ion Trap SL) (Agilent Technologies, Santa Clara, CA, USA), Bioblock Scientific 94200 rotary evaporator (Heidolph Instruments GmbH & Co. KG, Schwabach, Germany), vacuum controller HS-0245 (Hahnshin Scientific Co., Tongjin-eup, Gimpo-si, Gyeonggi-do, South Korea), Brinkman Polytron homogenizer (Kinematica AG, Littau-Luzern, Switzerland). ## 4.2. Extraction Methods $70\%$ ethanol was used as solvent, with the selected solvent to sample ratio being 10:1 (v/w). These parameters were chosen to remain constant throughout all extraction processes in order to allow results uniformity and to provide coherence for the comparison between extraction methods. Once each extraction process was finished, samples were separated by centrifugation at 12000 rpm, for 10 min. ## 4.2.1. Maceration The procedure for this extraction method followed the conditions provided by the Romanian Pharmacopoeia. Respectively, 50 mL $70\%$ alcohol were added to 5 g plant material, in a Falcon flask. The mixture was left for a period of 10 days, at room temperature, and submitted to periodical agitation. Following extraction, the sample was centrifugated in order to ensure separation. ## 4.2.2. Soxhlet Extraction (SE) For each sample, 5 g plant material was added in an extraction cup, along with 50 mL $70\%$ alcohol. The heating plate temperature was set to 210 °C and the studied extraction time values were: 20 min, 40 min, and 60 min. The samples were separated once the extraction process was completed. ## 4.2.3. Turboextraction (TBE) The parameters studied for this extraction process were time and rotation speed. Extraction time was divided into 2 cycles of 5 min (a total of 10 min), and 4 cycles of 5 min (a total of 20 min), respectively. The studied rotation speed values were 4000, 6000, and 8000 rpm. This manner of experimentation was considered to be advantageous as it limited the risk of solvent evaporation and device overheating. The samples were centrifugated afterwards. ## 4.2.4. Ultrasound-Assisted Extraction (UAE) The studied extraction time values were 10, 20, and 30 min. The assessed temperature values were 30°, 40°, and 50 °C. The constant parameters were frequency, 50 Hz, and power, 230 V, respectively. The samples were separated following extraction. ## 4.2.5. Combination of UAE and TBE (UTE) In order to provide an efficient combination of these two extraction methods, the selected parameters remained fixed throughout the process. Thus, the ultrasonic bath was brought to 30 °C, the disperser speed was set to 4000 rpm, and extraction time was reduced to 5 min. Fixed values were selected in this case so as to prevent solvent evaporation and overheating of the two devices. ## 4.3. Assessment of Total Phenolic Content (TPC) In order to determine the total polyphenolic content, the Folin-Ciocâlteu method was applied, following recommendations provided Csepregi et al., with several modifications [24]. 270 µL Folin-Ciocâlteu reagent were added to 60 µL plant extract, followed by 270 µL $6\%$ Na2CO3 (w/v). After 30 min incubation, in an environment devoid of light, sample absorbances were determined at 765 nm. Gallic acid was selected as standard, and results were therefore expressed as mg gallic acid equivalents per mL (GAE mg/mL). ## 4.4. Assessment of Total Flavonoid Content (TFC) An adapted version of the method employed by Pinacho et al. was used to determine the flavonoid content of the samples [25]. As such, 400 µL solution of AlCl3 20 mg/mL in $5\%$ acetic acid in ethanol 3:1 (v/v) ratio were mixed with 200 µL plant extract. Measurements were carried out at 420 nm wavelength. Quercetin was selected as standard. Finally, results were given as mM quercetin equivalents (QE mM). ## 4.5.1. DPPH Radical Scavenging Activity One of the methods used to evaluate the antioxidant potential of the samples was the DPPH assay. This experiment was carried out following the indications of Martins et al., after several adaptations [26]. 200 µL extract was mixed with 800 µL DPPH radical methanolic solution. Following an incubation of 30 min, in a dark environment, at 40 °C, the absorbances of the samples were measured at 517 nm. Trolox reagent was selected as standard. Results were expressed as mg Trolox equivalents per mL extract (TE mg/mL). ## 4.5.2. ABTS+ Scavenging Activity This assay was performed according to a method used by Erel et al. [ 27]. 200 µL acetate buffer 0.4 M, pH 5.8 were added to 20 µL ABTS+ in acetate buffer 30 mM, pH 3.6, with the addition of 12.5 µL extract to the mixture obtained earlier. Absorbances were measured at 660 nm. Trolox reagent was selected as standard, and results were therefore expressed as mM Trolox equivalents (mM TE). ## 4.5.3. FRAP Assay Experimentation was performed according to Csepregi et al. [ 24], i.e., FRAP reagent was obtained by adding 25 mL acetate buffer (300 mM, pH 3.6) to 2.5 mL TPTZ solution (10 mM TPTZ in 40 mM HCl) and 2.5 mL FeCl3 (20 mM in water). The newly prepared reagent was mixed with 30 µL extract. The mixture was incubated for 30 min, after which absorbances were determined at 620 nm. Trolox reagent was chosen as standard. Results were given as mM Trolox equivalents (TE mM). ## 4.6. Chromatographic Analysis The phytochemical profile of the extracts was investigated by liquid chromatography tandem mass spectrometry (LC-MS/MS) with two distinct analytical methods, previously validated [28,29]. The following equipment was used: Agilent Technologies 1100 HPLC Series system (Agilent, Santa Clara, CA, USA) equipped with column thermostat, auto sampler type G1313A, binary gradient pump type G13311A, degasser type G1322A, and UV detector type G1316A. A mass spectrometer from Agilent was coupled with this system (MS with Ion Trap 1100 SL (LC/MSD Ion Trap VL, Agilent, Santa Clara, CA, USA). The first analytical method was slightly modified and was applied to identify and quantify 23 polyphenols in vegetal extracts [28,29,30,31]. Chromatographic separation was performed on a reverse phase analytical column (Zorbax SB-C18, 100 mm × 3.0 mm id, 3.5 μm, Agilent Technologies, Santa Clara, CA, USA) with a mobile phase consisting of a mixture of methanol: acetic acid 0,$1\%$ (v/v) and a binary gradient. Elution began with a linear gradient, initially with $5\%$ methanol and ending with $42\%$ methanol at 35 min. For the next 3 min, isocratic elution followed with $42\%$ methanol. Further, the column was rebalanced with $5\%$ methanol for the following 7 min, as previously detailed [28,29,30,31]. Afterwards, the bioactive compounds were detected in both UV and MS mode. For detection of polyphenolic acids, the UV detector operated at λ = 330 nm (up to 17 min). Afterwards, for detection of the flavonoids and their aglycones, the UV detector operated at λ = 370 nm (up to 38 min). The MS system operated using an electrospray ionization source (ESI) in negative mode [28,29,30,31]. The second LC-MS analytical method was used to identify other 6 polyphenols from plant samples, as detailed in previous studies [32]. The same equipment and analytical column as aforementioned were used for chromatographic separation. The mobile phase consisted in a mixture of methanol: acetic acid $0.1\%$ (v/v) and a binary gradient. Briefly, at start—$3\%$ methanol; at 3 min—$8\%$ methanol; from 8.5 min to 10 min—$20\%$ methanol. The column was rebalanced with $3\%$ methanol [32]. Bioactive compounds were detected in plant samples in MS mode with the MS system operating with an ESI in negative mode. For identification of each bioactive compound from the plant extracts, spectra from library were compared with the MS spectra/traces. To quantify the compounds, after MS detection, a UV trace was used. For the identified compounds, the calibration curve of their corresponding standards was considered for quantification of their peak areas [28,29,30,31,32]. The phytosterols were determined according to a previously validated LC-UV-MS/MS method [33,34,35]. The same equipment and chromatographic analytical column were used. However, the elution of the compounds was performed in an isocratic manner. The mobile phase consisted in a mixture of acetonitrile: methanol (90:10, v/v), a flow rate of 1 mL/min at 45 °C and 5 μL injection volume. The same mass spectrometer was used, equipped with an ion trap and atmospheric-pressure chemical ionization (APCI) source, operating in positive ionization mode. The working conditions were carefully adjusted to reach maximum sensitivity [33,34,35]. Five external standards were used for complete identification of the compounds, which was performed by comparing the retention times and mass spectra. To reduce interference and for detection of bioactive compounds, multiple reaction monitoring mode (MRM) was employed. The Agilent ChemStation (vB01.03) and the DataAnalysis (v5.3) software were used for the acquisition and investigation of chromatographic data. All results were expressed as micrograms of bioactive compound per mL (μg/mL) of vegetal extract. The Supplementary Material contains the UV chromatograms of the analyzed samples (Figures S1–S11) as well as the analytical parameters of the database (Table S1). ## 4.7.1. In Vitro Qualitative Study of Antimicrobial Activity The antimicrobial potential of the samples was evaluated by means of the disk diffusion method against standards consisting of strains of Gram-positive, Gram-negative bacteria, and yeasts. The following Gram-positive strains were selected as standards: Staphylococcus aureus ATCC 6538P, Listeria monocytogenes ATCC 13932, Enterococcus faecalis ATCC 29212, and Bacillus cereus ATCC 11778. Gram-negative strains standards consisted of Escherichia coli ATCC 10536, Salmonella enteritidis ATCC 13076 and Pseudomonas aeruginosa ATCC 27853. Candida albicans ATCC 10231 was selected as a yeast strain standard. Standard antibacterial and antifungal controls were amoxicillin, for bacteria and ketoconazole for the yeast. Screening was carried out according to EUCAST standards [36]. ## 4.7.2. In Vitro Quantitative Study of Antimicrobial Activity The quantitative evaluation was performed by means of the minimum inhibitory concentration (MIC) method for the same eight standard microbial strands. The method was performed in accordance with the EUCAST protocols [36], with slight modifications. 96-wells titer plates, containing the extracts diluted in liquid MH medium and inoculated with 20 µL microbial suspension, were used. Extract stock solutions were diluted using a two-fold serial dilution system in ten consecutive wells, from the initial concentration ($\frac{1}{1}$) to the highest ($\frac{1}{512}$). The total broth volume was brought to 200 µL. Microbial inoculum in MH broth as positive control and microbial inoculum in $30\%$ ethanol as negative control were prepared and placed in wells 11 and 12, respectively. For bacteria, the plates were incubated at 37 °C for 24 h, and at 28 °C for 48 h for Candida. MIC values were determined as the lowest concentration of the extracts’ dilution that inhibited the growth of the microbial cultures (having the same OD as the negative control), compared to the positive control, as established by a decreased value of absorbance at 450 nm (HiPo MPP-96, Biosan, Latvia). MIC50 was determined as well, representing the MIC value at which ≥$50\%$ of the bacterial/yeast cells were inhibited in their growth, considered as the well with the OD value similar to the average between the positive and negative control. ## 4.8. Assessment of Biological Activites After the phytochemical profile of samples was completed, biological activities were determined in vitro for the 60 min SE. The selected sample presented the highest number of identified compounds and in the highest yields. ## 4.8.1. Carrageenan Induced Inflammation Model in Rats An animal model of plantar inflammation on Wistar rats (110–130 g mean weight) was used to evaluate the in vivo anti-inflammatory activity. Acclimatization of the animals was conducted as following: 12 h light/12 h dark cycles, $35\%$ humidity, free access to water, and a normocaloric standard diet (VRF1) and randomized in 4 groups, 8 rats each. Over a course of 4 days, treatment was administered through oral gavage, using a volume of up to 0.25 mL, namely: group 1—carboxymethylcellulose $2\%$ (positive control group—CMC); group 2—Indomethacin 5 mg/body weight (b.w.) in carboxymethylcellulose $1.5\%$ (Indom); group 3—15 mg TPC/b.w./day (60 min SE). Finally, on the fifth day, inflammation was induced by injecting 100 µL of freshly prepared $1\%$ carrageenan (λ-carrageenan, type IV, Sigma) diluted in normal saline in the right hind footpad [37]. Negative control was established by injecting an identical volume of saline solution in the left hind paw. After the administration of the carrageenan, soft paw tissue was sampled, at 2 h, and 24 h, respectively. The procedure was carried out under an intraperitoneal injection of 90 mg/kg ketamine and 10 mg/kg xylazine. Levels of oxidative stress markers and cytokines were evaluated in the tissue samples after homogenization in 50 mMTRIS–10 mM EDTA buffer (pH 7.4) [37]. Protein content was evaluated in accordance with the Bradford method [38]. ## 4.8.2. Oxidative Stress Assessment The obtained raw paw tissue homogenates were assessed for malondialdehyde (MDA), reduced glutathione (GSH) and oxidized glutathione (GSSG) levels and GSH/GSSG ratio. Spectrofluorimetry was employed in order to quantify MDA formation, using the 2-thiobarbituric acid method, while the Hu method was used to determine GSH and GSSG levels [39,40]. ## 4.8.3. Proinflammatory Cytokine Assessment The plantar tissue homogenates were subjected to TNF-α, and IL-6 level evaluation by ELISA assay. The protocol provided by the manufacturer was employed. The results were expressed as pg/mg protein. ## 4.8.4. Statistical Analysis for the Assessment of Biological Activities Data were analyzed by one-way ANOVA and Tukey’s multiple comparisons post-test, using GraphPad Prism 8 software. A p value < 0.05 was considered statistically significant. The results were expressed as mean ± standard deviation. ## 5. Conclusions The impact of extraction methods and parameters over the phytochemistry and biological activities of green coffee beans was studied. The highest bioactive compound yields were reached by Soxhlet extraction and ultrasound-assisted extraction. Bioactive compound yields were observed to increase proportionately with the increase in extraction parameters, such as extraction time, temperature, or homogenization speed. Bioactive compound levels were observed to decrease once degradation temperatures were presumably reached. The extracts presented a highly inhibitory effect against Gram-negative bacteria. The 60 min Soxhlet extract, presenting the most favorable results, was eventually selected for in vivo biological activity determination. Non-endogenous antioxidant levels were positively impacted, principally at 2 h after extract administration. Out of the enzymatic antioxidants that were studied, GPx was significantly elevated while the cytokines secretion in the paw tissue was not influenced. In conclusion, aspects such as extraction method and extraction parameters influence both the compositional and biological quality of green coffee beans extracts. In addition, the potential of green coffee beans to serve as a natural, biologically safe, and effective source of bioactive compounds has been demonstrated. This plant material may find practical applications as food supplements, adjuvant therapies, or nutraceuticals as part of the treatment of illnesses of an acute or chronic nature. ## References 1. Farah A., Dos Santos T.F., Preedy V.R.. The Coffee Plant and Beans: An Introduction. Coffee in Health and Disease Prevention (2015) 5-10 2. Garg S.K., Gupta R.C.. Green Coffee Bean. Nutraceuticals: Efficacy, Safety and Toxicity (2016) 653-667 3. Wei F., Tanokura M., Preedy V.R.. Organic Compounds in Green Coffee Beans. Coffee in Health and Disease Prevention (2015) 149-162 4. Wei F., Tanokura M., Preedy V.R.. Chemical Changes in the Components of Coffee Beans during Roasting. Coffee in Health and Disease Prevention (2015) 83-91 5. Mojica B.E., Fong L.E., Biju D., Muharram A., Davis I.M., Vela K.O., Rios D., Osorio-Camacena E., Kaur B., Rojas S.M.. The Impact of the Roast Levels of Coffee Extracts on their Potential Anticancer Activities. J. Food Sci. (2018) 83 1125-1130. DOI: 10.1111/1750-3841.14102 6. Bauer D., Abreu J., Jordão N., da Rosa J.S., Freitas-Silva O., Teodoro A.. Effect of roasting levels and drying process of. Int. J. Mol. Sci. (2018) 19. DOI: 10.3390/ijms19113407 7. Pimpley V., Patil S., Srinivasan K., Desai N., Murthy P.S.. The chemistry of chlorogenic acid from green coffee and its role in attenuation of obesity and diabetes. Prep. Biochem. Biotechnol. (2020) 50 969-978. DOI: 10.1080/10826068.2020.1786699 8. Martínez-López S., Sarriá B., Mateos R., Bravo-Clemente L.. Moderate consumption of a soluble green/roasted coffee rich in caffeoylquinic acids reduces cardiovascular risk markers: Results from a randomized, cross-over, controlled trial in healthy and hypercholesterolemic subjects. Eur. J. Nutr. (2019) 58 865-878. DOI: 10.1007/s00394-018-1726-x 9. Caro-Gómez E., Sierra J.A., Escobar J.S., Álvarez-Quintero R., Naranjo M., Medina S., Velásquez-Mejía E.P., Tabares-Guevara J.H., Jaramillo J.C., León-Varela Y.M.. Green coffee extract improves cardiometabolic parameters and modulates gut microbiota in high-fat-diet-fed ApoE. Nutrients (2019) 11. DOI: 10.3390/nu11030497 10. Wang S., Sarriá B., Mateos R., Goya L., Bravo-Clemente L.. TNF-α-induced oxidative stress and endothelial dysfunction in EA.hy926 cells is prevented by mate and green coffee extracts, 5-caffeoylquinic acid and its microbial metabolite, dihydrocaffeic acid. Int. J. Food Sci. Nutr. (2019) 70 267-284. DOI: 10.1080/09637486.2018.1505834 11. Belwal T., Ezzat S.M., Rastrelli L., Bhatt I.D., Daglia M., Baldi A., Devkota H.P., Orhan I.E., Patra J.K., Das G.. A critical analysis of extraction techniques used for botanicals: Trends, priorities, industrial uses and optimization strategies. Trends Anal. Chem. (2018) 100 82-102. DOI: 10.1016/j.trac.2017.12.018 12. Renard C.M.G.C.. Extraction of bioactives from fruit and vegetables: State of the art and perspectives. LWT-Food Sci. Technol. (2018) 93 390-395. DOI: 10.1016/j.lwt.2018.03.063 13. Yuniarti E., Saputri F.C., Mun’Im A.. Natural deep eutectic solvent extraction and evaluation of caffeine and chlorogenic acid from green coffee beans of. Indian J. Pharm. Sci. (2019) 81 1062-1069. DOI: 10.36468/pharmaceutical-sciences.604 14. Menzio J., Binello A., Barge A., Cravotto G.. Highly-efficient caffeine recovery from green coffee beans under ultrasound-assisted SC–CO. Processes (2020) 8. DOI: 10.3390/pr8091062 15. Gawlik-Dziki U., Świeca M., Dziki D., Kowalska I., Pecio Ł., Durak A., Sęczyk Ł.. Lipoxygenase inhibitors and antioxidants from green coffee-mechanism of action in the light of potential bioaccessibility. Food Res. Int. (2014) 61 48-55. DOI: 10.1016/j.foodres.2014.05.002 16. Pavlović M.D., Buntić A.V., Šiler-Marinković S.S., Dimitrijević-Branković S.I.. Ethanol influenced fast microwave-assisted extraction for natural antioxidants obtaining from spent filter coffee. Sep. Purif. Technol. (2013) 118 503-510. DOI: 10.1016/j.seppur.2013.07.035 17. Al-Dhabi N.A., Ponmurugan K., Maran Jeganathan P.. Development and validation of ultrasound-assisted solid-liquid extraction of phenolic compounds from waste spent coffee grounds. Ultrason. Sonochem. (2017) 34 206-213. DOI: 10.1016/j.ultsonch.2016.05.005 18. Nzekoue F.K., Angeloni S., Navarini L., Angeloni C., Freschi M., Hrelia S., Vitali L.A., Sagratini G., Vittori S., Caprioli G.. Coffee silverskin extracts: Quantification of 30 bioactive compounds by a new HPLC-MS/MS method and evaluation of their antioxidant and antibacterial activities. Food Res. Int. (2020) 133 109128. DOI: 10.1016/j.foodres.2020.109128 19. Dong W., Chen Q., Wei C., Hu R., Long Y., Zong Y., Chu Z.. Comparison of the effect of extraction methods on the quality of green coffee oil from Arabica coffee beans: Lipid yield, fatty acid composition, bioactive components, and antioxidant activity. Ultrason. Sonochem. (2021) 74 105578. DOI: 10.1016/j.ultsonch.2021.105578 20. Chua L.Y.W., Chua B.L., Figiel A., Chong C.H., Wojdyło A., Szumny A., Choong T.S.Y.. Antioxidant activity, and volatile and phytosterol contents of. Molecules (2019) 24. DOI: 10.3390/molecules24071397 21. Bhandarkar N.S., Mouatt P., Brown L., Panchal S.K.. Green coffee ameliorates components of diet-induced metabolic syndrome in rats. J. Funct. Foods (2019) 57 141-149. DOI: 10.1016/j.jff.2019.04.003 22. Pergolizzi S., D’Angelo V., Aragona M., Dugo P., Cacciola F., Capillo G., Dugo G., Lauriano E.R.. Evaluation of antioxidant and anti-inflammatory activity of green coffee beans methanolic extract in rat skin. Nat. Prod. Res. (2020) 34 1535-1541. DOI: 10.1080/14786419.2018.1523161 23. Asbaghi O., Kashkooli S., Mardani M., Rezaei Kelishadi M., Fry H., Kazemi M., Kaviani M.. Effect of green coffee bean extract supplementation on liver function and inflammatory biomarkers: A meta-analysis of randomized clinical trials. Complement. Ther. Clin. Pract. (2021) 43 101349. DOI: 10.1016/j.ctcp.2021.101349 24. Csepregi K., Neugart S., Schreiner M., Hideg É.. Comparative evaluation of total antioxidant capacities of plant polyphenols. Molecules (2016) 21. DOI: 10.3390/molecules21020208 25. Pinacho R., Cavero R.Y., Astiasarán I., Ansorena D., Calvo M.I.. Phenolic compounds of blackthorn (. J. Funct. Foods (2015) 19 49-62. DOI: 10.1016/j.jff.2015.09.015 26. Martins N., Barros L., Dueñas M., Santos-Buelga C., Ferreira I.C.F.R.. Characterization of phenolic compounds and antioxidant properties of. RSC Adv. (2015) 5 26991-26997. DOI: 10.1039/C5RA03963K 27. Erel O.. A novel automated direct measurement method for total antioxidant capacity using a new generation, more stable ABTS radical cation. Clin. Biochem. (2004) 37 277-285. DOI: 10.1016/j.clinbiochem.2003.11.015 28. Benedec D., Hanganu D., Filip L., Oniga I., Tiperciuc B., Olah N.K., Gheldiu A.-M., Raita O., Vlase L.. Chemical, antioxidant and antibacterial studies of Romanian. Farmacia (2017) 65 252-256 29. Toiu A., Vlase L., Gheldiu A.M., Vodnar D., Oniga I.. Evaluation of the antioxidant and antibacterial potential of bioactive compounds from. Farmacia (2017) 65 351-355 30. Toiu A., Vlase L., Vodnar D.C., Gheldiu A.M., Oniga I.. Molecules (2019) 24. DOI: 10.3390/molecules24142666 31. Epure A., Pârvu A.E., Vlase L., Benedec D., Hanganu D., Gheldiu A.M., Toma V.A., Oniga I.. Phytochemical profile, antioxidant, cardioprotective and nephroprotective activity of Romanian chicory extract. Plants (2021) 10. DOI: 10.3390/plants10010064 32. Rusu M.E., Fizesan I., Pop A., Mocan A., Gheldiu A.M., Babota M., Vodnar D.C., Jurj A., Berindan-Neagoe I., Vlase L.. Walnut (. Molecules (2020) 25. DOI: 10.3390/molecules25092187 33. Toiu A., Mocan A., Vlase L., Pârvu A.E., Vodnar D.C., Gheldiu A.M., Moldovan C., Oniga I.. Comparative phytochemical profile, antioxidant, antimicrobial and in vivo anti-inflammatory activity of different extracts of traditionally used Romanian. Molecules (2019) 24. DOI: 10.3390/molecules24081597 34. Rusu M.E., Fizeșan I., Pop A., Gheldiu A.M., Mocan A., Cirșan G., Vlase L., Loghin F., Popa D.-S., Tomuta I.. Enhanced recovery of antioxidant compounds from hazelnut (. Antioxidants (2019) 8. DOI: 10.3390/antiox8100460 35. Adehuwa-Olabode A.A., Sautreau A., Vlase L., Vlase A.M., Muntean D.. The Phytochemical Analysis and Antioxidant Capacity Determination of Five. Stud. Univ. Babes-Bolyai Chem. (2022) 67 17-35. DOI: 10.24193/subbchem.2022.3.02 36. Matuschek E., Brown D.F.J., Kahlmeter G.. Development of the EUCAST disk diffusion antimicrobial susceptibility testing method and its implementation in routine microbiology laboratories. Clin. Microbiol. Infect. (2014) 20 O255-O266. DOI: 10.1111/1469-0691.12373 37. Moldovan B., Filip A., Clichici S., Suharoschi R., Bolfa P., David L.. Antioxidant activity of Cornelian cherry (. J. Funct. Foods (2016) 26 77-87. DOI: 10.1016/j.jff.2016.07.004 38. Patil C.R., Gadekar A.R., Patel P.N., Rambhade A., Surana S.J., Gaushal M.H.. Dual effect of. Homeopathy (2009) 98 88-91. DOI: 10.1016/j.homp.2009.01.003 39. Conti M., Morand P., Levillain P., Lemonnier A.. Improved fluorometric determination of malonaldehyde. Clin. Chem. (1991) 37 1273-1275. DOI: 10.1093/clinchem/37.7.1273 40. Hu M.L.. Measurement of protein thiol groups and glutathione in plasma. Methods Enzymol. (1994) 233 380-385. PMID: 8015473
--- title: Denosumab Is Superior to Raloxifene in Lowering Risks of Mortality and Ischemic Stroke in Osteoporotic Women authors: - Ting-Chun Liu - Chien-Ning Hsu - Wen-Chin Lee - Shih-Wei Wang - Chiang-Chi Huang - Yueh-Ting Lee - Chung-Ming Fu - Jin-Bor Chen - Lung-Chih Li journal: Pharmaceuticals year: 2023 pmcid: PMC9966982 doi: 10.3390/ph16020222 license: CC BY 4.0 --- # Denosumab Is Superior to Raloxifene in Lowering Risks of Mortality and Ischemic Stroke in Osteoporotic Women ## Abstract Both osteoporosis and cardiovascular disease (CVD) share similar pathways in pathophysiology and are intercorrelated with increased morbidity and mortality in elderly women. Although denosumab and raloxifene are the current guideline-based pharmacological treatments, their impacts on cardiovascular protection are yet to be examined. This study aimed to compare mortality rate and cardiovascular events between denosumab and raloxifene in osteoporotic women. Risks of CVD development and all-cause mortality were estimated using Cox proportional hazard regression. A total of 7972 (3986 in each group) women were recruited between January 2003 and December 2018. No significant difference between denosumab and raloxifene was observed in composite CVDs, myocardial infarction, or congestive heart failure. However, comparison of the propensity score matched cohorts revealed that patients with proportion of days covered (PDC) ≥$60\%$ had lower incidence of ischemic stroke in the denosumab group than that in the raloxifene group (aHR 0.68; $95\%$ CI 0.47–0.98; $$p \leq 0.0399$$). In addition, all-cause mortality was lower in the denosumab group than in the raloxifene group (aHR 0.59; $95\%$ CI 0.48–0.72; $$p \leq 0.001$$), except in patients aged <65 y/o in this cohort study. We concluded that denosumab is superior to raloxifene in lowering risks of all-cause mortality and certain ischemic strokes in osteoporotic women. ## 1. Introduction Osteoporosis and resultant fracture have become a public health problem, leading to a heavy global economic, social, and health burden. A previous study reported the estimation that 200 million women worldwide suffer from osteoporosis [1]. In particular, global deaths and disability-adjusted life-years attributable to osteoporosis and its related fracture increased from 207,367 and 8,588,936 in 1990 to 437,884 and 16,647,466 in 2019, respectively [2]. Aging has a major impact on the arterial system and heart, leading to an increase in cardiovascular disease (CVD) including atherosclerosis, hypertension, myocardial infarction, and stroke [3]. Both osteoporosis and CVD events are associated with physical disability, higher health care costs, impaired quality of life, and increased mortality [4,5]. The relationship between osteoporosis and CVD could be explained by their common risk factors such as age, smoking, alcohol consumption, physical activity, and menopause. Recent evidence has shown that osteoporosis and CVD share similar pathophysiological pathways and that osteoporotic patients are at a higher risk of developing major CVDs, such as stroke and ischemic heart disease [6]. The receptor activator of nuclear factor-κB ligand (RANKL)/the receptor activator of nuclear factor-κB (RANK)/osteoprotegerin (OPG) is one of the key signaling pathways shared by osteoporosis and CVD. RANKL–RANK interaction on the surface of pre-osteoclasts that activates downstream signals that initiate osteoclast activation, differentiation, and function [7]. OPG acts as a decoy receptor for RANKL, preventing RANK–RANKL interactions and blocking the resulting downstream osteoclastogenic cascade, and exerts an anti-calcific effect within the vasculature [8]. Denosumab is a fully human monoclonal antibody against human RANKL that mimics OPG, thus preventing it from activating RANK, and inhibits osteoclastogenesis, thereby decreasing bone resorption and increasing bone density. According to the Fracture Reduction Evaluation of Denosumab in Osteoporosis Every 6 Months (FREEDOM) trial, denosumab reduces the risk of vertebral, nonvertebral, and hip fractures. It can increase bone mineral density (BMD) at the lumbar spine and the total hip [9]. Treatment with a recombinant fusion protein, Fc-OPG, has been shown to inhibit vascular calcification in animal studies [10]. A study conducted with human RANKL knock-in (huRANKL-KI) mice showed that treatment with denosumab reduced aortic calcium deposits in prednisolone-treated huRANKL-KI mice by up to $50\%$ based on calcium measurement [11]. However, no related study was conducted in humans. Considering the above information, we hypothesized that denosumab may have an influence in reducing cardiovascular events by slowing the progression of vascular calcification and decreasing the risk of CVD or heart failure. Estrogen deprivation is associated with decreased bone mass and increased osteoclast formation [12,13,14]. Raloxifene is a benzothiophene non-steroidal derivative that is used as a second-generation selective estrogen receptor modulator (SERM). It binds to estrogen receptor and produces estrogen-like effects on bone, reducing resorption and increasing BMD in postmenopausal women. Raloxifene also acts as an estrogen agonist in pre-osteoclasts, inhibiting their proliferative capacity [15]. The Multiple Outcomes of Raloxifene Evaluation (MORE) trial, the pivotal treatment trial of raloxifene, demonstrated significant reductions in the risk of vertebral fractures after three years. [ 16]. Although raloxifene therapy for four years did not adversely affect the risk of cardiovascular events in the overall cohort, it reduced the risk of cardiovascular events in the subset of women with increased cardiovascular risk [17]. Denosumab has been shown to exert protective effects against CVD [18]. Chen et al. [ 19] reported that denosumab suppress the progression of coronary arterial calcification. Raloxifene may offer cardiovascular protection by normalizing the lipid profile, reducing oxidative stress, and improving endothelial function via increased nitric oxide production [20,21,22]. A recently published meta-analysis of randomized controlled trials demonstrated favorable effects of raloxifene on the lipid profile in women [23,24]. However, no direct comparison of the mortality and cardiovascular protective effects between denosumab and raloxifene has been made. As denosumab and raloxifene are currently recommended in selected patients by several osteoporosis management guidelines [25,26,27,28], it is urgently necessary to obtain evidence of these two medications regarding their cardiovascular benefits. The present study aimed to compare the effects of denosumab and raloxifene on mortality and CVD prevention. ## 2.1. Patient Characteristics A total of 33,576 adult osteoporotic patients receiving denosumab or raloxifene were identified, of whom 30,434 met the inclusion criteria (denosumab: 16,754 patients; raloxifene: 13,680 patients) (Figure 1). Compared with patients initiated with raloxifene, those initiated with denosumab were older (73.18 ± 9.72 vs. 69.58 ± 10.76 years old) and had more comorbidities, including history of dementia, liver diseases, diabetes, renal diseases, hypertension, hyperlipidemia, and obstructive sleep apnea (Table 1). In the propensity score matched (PSM) cohort, 7972 denosumab- and raloxifene-matched pairs were analyzed over a six-year follow-up period. The baseline characteristics were well balanced in the matched groups and are presented in Table 1. Approximately $45\%$ of the study cohort had a history of hypertension, $27\%$ of hyperlipidemia, and $23\%$ of diabetes. During the follow-up period, most patients were concomitantly administered medications for calcium ($29\%$)/VitD ($21\%$), followed by antihypertensive agents ACEi/ARB/Aliskiren ($27\%$), statins ($20\%$), anti-diabetics ($18\%$), and thrombotic prevention. The frequency of these medications was similar between the denosumab and raloxifene groups (Table 1). ## 2.2. CVD Incidence During the six-year follow-up period, the incidence of all CVDs was $7.80\%$ ($$n = 311$$) and $10.71\%$ ($$n = 427$$), and the incidence of all-cause mortality was $4.47\%$ ($$n = 178$$) and $8.10\%$ ($$n = 323$$) in the denosumab and raloxifene groups, respectively. Among the major CVDs, the highest event rate was ischemic stroke ($4.59\%$ in denosumab, $7.12\%$ in raloxifene), followed by congestive heart failure ($3.24\%$ vs. $4.04\%$ in denosumab vs. raloxifene groups) and myocardial infarction ($1.00\%$ vs. $1.56\%$ in denosumab vs. raloxifene groups) (Table 2). The cumulative incidence of mortality and CVDs between the denosumab and raloxifene groups is illustrated in Figure 2. No significant difference was observed in composite CVD incidence (Figure 2a: log-rank test, $$p \leq 0.3362$$) and heart failure (Figure 2c: log-rank test, $$p \leq 0.4819$$) between the two groups. Notably, the incidence of ischemic stroke was significantly lower in the denosumab group than in the raloxifene group (Figure 2b; log-rank test, $$p \leq 0.0304$$). Following adjustment for the baseline characteristics, comorbidities, and the use of concomitant medications, no significant difference was observed between denosumab and raloxifene in MI, ischemic stroke, heart failure, and all-cause CVDs. However, in the stratified analyses, we found that among patients with proportion of days covered (PDC) ≥$60\%$, the incidence of ischemic stroke was significantly lower in the denosumab group than in the raloxifene group (adjusted hazard ratio (aHR) 0.68; $95\%$ CI 0.47–0.98; $$p \leq 0.0399$$) (Figure 3). ## 2.3. All-Cause Mortality Of note, all-cause mortality was lower in the denosumab group than in the raloxifene group. Kaplan–Meier analysis showed that the all-cause mortality was significantly lower in the denosumab group than in the raloxifene group (Figure 2d; log-rank test, $$p \leq 0.0022$$). The stratified analyses further confirmed the superiority of denosumab to raloxifene in lowering the risk of all-cause mortality (aHR 0.59; $95\%$ CI 0.48–0.72; $$p \leq 0.001$$). In patients less than 65 years of age, denosumab users showed lower all-cause mortality, though the difference was not significant. Among osteoporotic women with varying degrees of renal impairment, the superiority of denosumab to raloxifene in lowering the risk of all-cause mortality was apparent (Figure 3). ## 3. Discussion Multimorbidity is a growing concern, especially in elderly patients and patients with chronic diseases [29,30,31,32,33]. Organ protection therefore emerges as a key concept in the management of chronic diseases including hypertension, diabetes, and chronic kidney disease. Osteoporosis and resultant fracture have been well-recognized as a silent threat to elderly patients and postmenopausal women. Both osteoporosis and CVD are common in these patient populations. As CVD and osteoporosis share some pathophysiological similarities, we quantified the cardiovascular benefits of denosumab and raloxifene based on the concept of organ protection. Denosumab and raloxifene are recommended for osteoporosis treatment in postmenopausal women. Dose adjustment of both medications is not required in renal function impairment. In the present study, we showed that denosumab is superior to raloxifene in reducing the risk of ischemic stroke, in particular, in patients with PDC ≥$60\%$ and with eGFR 30~59.9 mL/min/1.73 m2. In line with our findings, the results of the Raloxifene Use for The Heart (RUTH) trial revealed that raloxifene may carry increased risks of venous thromboembolism and fatal stroke [34]. Comparative event rates of MI or heart failure in the denosumab and raloxifene groups were found in our study. Raloxifene has been reported to improve the levels of serum low-density lipoprotein cholesterol, total cholesterol, and non-high-density lipoprotein cholesterol in the post-hoc analysis of the MORE trial [35]. However, in the RUTH trial, a large randomized controlled trial, raloxifene did not significantly affect the risk of coronary heart disease [34]. On the other hand, in recent years, the RANKL/RANK/OPG pathway has been discovered not only as a regulator of bone remodeling but also as a regulator of vascular calcification [36]. Despite the fact that Samelson et al. [ 37] reported that denosumab had no effect on the progression of aortic calcification or the incidence of CV adverse events in postmenopausal women, $24\%$ patients did not finish complete doses of denosumab, and the follow-up period was only three years. Recently, Suzuki et al. [ 38] found that denosumab ameliorated aortic arch calcification in dialysis patients. The calcified area apparently diminished through 30 months of treatment and it was deemed necessary for ≥2 years for decalcification due to denosumab. Hsu et al. [ 18] demonstrated that denosumab use is associated with a significant reduction in cardiovascular events compared to alendronate use with better adherence of >$60\%$ medication possession ratio. Despite no significant differences in MI and heart failure between the denosumab and raloxifene groups, our study did not obtain evidence of vascular calcification in X-ray or CT, owing to data limitations. Further study is necessary to elucidate whether denosumab can improve vascular calcification. We demonstrated the lower risk of all-cause mortality in the denosumab group compared with the raloxifene group. Effective osteoporosis treatment has been reported to reduce mortality, per se, because in addition to decreasing bone fracture, treating osteoporosis might improve the ability of an individual to cope with or recover from an acute illness, probably by maintaining physiological reserve and preventing frailty [39]. Head-to-head comparison of denosumab and raloxifene in various bone fractures is lacking. However, denosumab was proven to effectively reduce risks of vertebral, non-vertebral, and hip fracture in the FREEDOM trail [9], whereas raloxifene only reduced the risk of vertebral fracture in the MORE trial [16]. Wu et al. also demonstrated that high adherence users of denosumab, defined by receiving three or four doses, had lower risks of all-cause mortality than low adherence users of denosumab, defined by one or two doses (aHR 0.64, $95\%$ CI 0.48–0.86) [40]. Moreover, the RANKL/RANK/OPG pathway plays numerous roles in various organs [41]. This pathway has been shown to participate in intestinal immunity [42], central nervous system inflammation [43], skin inflammation [44], and fever that occurs during infection [45]. This preclinical evidence supports the potential benefit of denosumab treatment in patient populations suffering from various infectious and inflammatory diseases. Although these potential benefits might explain the lower risks of all-cause mortality shown in our study, additional clinical investigations are required to examine the therapeutic benefit of denosumab in these settings. Although we performed analysis using one of the largest and most representative medical research databases in Taiwan [46], there are certain limitations to the current study. First, from the database studied, we were unable to identify specific causes of death in individual patients. Second, there were certain unmeasured confounding variables, such as smoking, alcohol, and home nutrient supplement. Finally, laboratory results, such as 1,25(OH)2 vitamin D and parathyroid hormone, are not part of routine care nor health insurance imbursed tests. Despite these limitations, the present study was valuable because we demonstrated the differences in lowering risks of all-cause mortality and ischemic stroke in denosumab compared to raloxifene in osteoporotic women. ## 4.1. Data Source This retrospective cohort study was conducted using electronic health record (EHR) data from the Chang Gung Research Database (CGRD), an electronic healthcare delivery system derived from a group of Chang Gung Memorial Hospitals (CGMH) in geographically distinct parts of Taiwan. Briefly, CGMH is the largest healthcare delivery system in Taiwan, providing approximately 10–$12\%$ of the healthcare services of the Taiwan National Health Insurance (NHI) program [46]. The Taiwan NHI program is a compulsory, single-payer health insurance program that covers over $99\%$ of Taiwan’s population [47]. The CGRD, containing detailed individual patient-level EHR including diagnosis, prescription, and laboratory test results, has been validated in some disease populations [46,48]. We conducted this analysis using the CGRD data from January 2003 to December 2018. ## 4.2. Study Design and Study Cohort By using the EHR database, we first identified patients aged 30–89 years at the time of initiation of denosumab or raloxifene between 1 January 2003 and 31 December 2017 to be included in the study. To develop the new user cohort, only patients with ≥1-year records before treatment initiation were included in the study. To assess the risk of CVD, patients were excluded with history of stroke, myocardial infarction, heart failure, and cancer. Patients receiving procedures for cardiovascular diseases, including percutaneous coronary intervention (PCI) and coronary artery bypass surgery (CABG), were also identified before treatment initiation and excluded from analyses. Because raloxifene was used in women, male patients were excluded from the analysis. Operational definitions and codes for disease conditions and procedures are listed in Table S1. ## 4.3. Comparison Groups New users of denosumab were defined as patients who never taken raloxifene within one year before treatment initiation, and the earliest date of denosumab prescribed was defined as the index date. The same criteria were applied for new users of raloxifene without prior denosumab treatment. PSM was applied to balance differences in baseline demographic and clinical characteristics between the denosumab and raloxifene groups. The propensity score of initiating denosumab or raloxifene was estimated by logistic regression with above mentioned baseline characteristics, with a 1:1 ratio. ## 4.4. Outcomes and Follow-Up The outcomes of interest were composite of major CVD, including myocardial infarction, congestive heart failure, and ischemic stroke. These outcomes were assessed based on hospital discharge diagnoses. Individual outcomes of composite CVD were also analyzed. ## 4.5. Study Covariates Baseline variables considered in the analyses included patient demographics; comorbid conditions, such as the Charlson Comorbid Index [49]. The following study covariates were measured at baseline as potential confounders to be adjusted in the analyses: procedures of PCI and CABG; preexisting hyperlipidemia; and prior medication use: anti-thrombotic medications (anti-coagulants, anti-platelets), aspirin, glucose-lowering agents, lipid-lowering agents, antihypertensive agents, vitamin D, or calcium supplementations, as well as other osteoporosis therapy besides study drugs. ## 4.6. Statistical Analysis Data were summarized as mean ± standard deviation (SD) for continuous variables and n (%) for categorical variables in the study cohort. PSM was employed to balance the differences in baseline demographic and clinical characteristics between the denosumab and raloxifene groups. New users of denosumab and raloxifene were matched at a 1:1 ratio using the greedy algorithm of the PSM method. The covariate balance between the denosumab and raloxifene groups was measured using the standardized mean difference (SDM), and an SDM of <0.1 was considered as no meaningful difference [50]. We used the Cox proportional model to estimate the aHR of the study outcomes between the new users of denosumab and raloxifene. Time to CVD endpoint was used using Kaplan–Meier analysis with log-rank tests. Cox proportional hazard regression was performed for composite incident CVD events and individual cardiovascular events. Importantly, to assess the heterogeneous effects of denosumab (versus raloxifene) with different baseline characteristics, stratified analyses were performed in the matched cohorts by age < 65 years (vs. age ≥ 65 years) and baseline eGFR groups (≥60, 30–59.9, and <30 mL/min/1.73 m2). A two-tailed test (p value < 0.05) was considered statistically significant. All statistical analyses were performed using SAS 9.4 (SAS Institute, Cary, NC, USA). ## 5. Conclusions In conclusion, denosumab is superior to raloxifene in lowering risks of all-cause mortality and certain ischemic strokes in osteoporotic women. Taken into the concept of organ protection in the pharmacological treatment of osteoporosis, denosumab could be a better choice than raloxifene. ## References 1. Lane N.E.. Epidemiology, etiology, and diagnosis of osteoporosis. Am. J. Obstet. Gynecol. (2006) 194 S3-S11. DOI: 10.1016/j.ajog.2005.08.047 2. Shen Y., Huang X., Wu J., Lin X., Zhou X., Zhu Z., Pan X., Xu J., Qiao J., Zhang T.. The Global Burden of Osteoporosis, Low Bone Mass, and Its Related Fracture in 204 Countries and Territories, 1990-2019. Front. Endocrinol. (2022) 13 882241. DOI: 10.3389/fendo.2022.882241 3. Lakatta E.G., Levy D.. Arterial and cardiac aging: Major shareholders in cardiovascular disease enterprises: Part I: Aging arteries: A "set up" for vascular disease. Circulation (2003) 107 139-146. DOI: 10.1161/01.CIR.0000048892.83521.58 4. Rajsic S., Gothe H., Borba H.H., Sroczynski G., Vujicic J., Toell T., Siebert U.. Economic burden of stroke: A systematic review on post-stroke care. Eur. J. Health Econ. (2019) 20 107-134. DOI: 10.1007/s10198-018-0984-0 5. Srivastava M., Deal C.. Osteoporosis in elderly: Prevention and treatment. Clin. Geriatr. Med. (2002) 18 529-555. DOI: 10.1016/S0749-0690(02)00022-8 6. Lello S., Capozzi A., Scambia G.. Osteoporosis and cardiovascular disease: An update. Gynecol. Endocrinol. (2015) 31 590-594. DOI: 10.3109/09513590.2015.1041908 7. Kim H.S., Nam S.T., Mun S.H., Lee S.K., Kim H.W., Park Y.H., Kim B., Won K.J., Kim H.R., Park Y.M.. DJ-1 controls bone homeostasis through the regulation of osteoclast differentiation. Nat. Commun. (2017) 8 1519. DOI: 10.1038/s41467-017-01527-y 8. Harper E., Forde H., Davenport C., Rochfort K.D., Smith D., Cummins P.M.. Vascular calcification in type-2 diabetes and cardiovascular disease: Integrative roles for OPG, RANKL and TRAIL. Vascul. Pharmacol. (2016) 82 30-40. DOI: 10.1016/j.vph.2016.02.003 9. Cummings S.R., San Martin J., McClung M.R., Siris E.S., Eastell R., Reid I.R., Delmas P., Zoog H.B., Austin M., Wang A.. Denosumab for prevention of fractures in postmenopausal women with osteoporosis. N. Engl. J. Med. (2009) 361 756-765. DOI: 10.1056/NEJMoa0809493 10. Morony S., Tintut Y., Zhang Z., Cattley R.C., Van G., Dwyer D., Stolina M., Kostenuik P.J., Demer L.L.. Osteoprotegerin inhibits vascular calcification without affecting atherosclerosis in ldlr(-/-) mice. Circulation (2008) 117 411-420. DOI: 10.1161/CIRCULATIONAHA.107.707380 11. Helas S., Goettsch C., Schoppet M., Zeitz U., Hempel U., Morawietz H., Kostenuik P.J., Erben R.G., Hofbauer L.C.. Inhibition of receptor activator of NF-kappaB ligand by denosumab attenuates vascular calcium deposition in mice. Am. J. Pathol. (2009) 175 473-478. DOI: 10.2353/ajpath.2009.080957 12. Recker R., Lappe J., Davies K., Heaney R.. Characterization of perimenopausal bone loss: A prospective study. J. Bone Miner. Res. (2000) 15 1965-1973. DOI: 10.1359/jbmr.2000.15.10.1965 13. Akhter M.P., Lappe J.M., Davies K.M., Recker R.R.. Transmenopausal changes in the trabecular bone structure. Bone (2007) 41 111-116. DOI: 10.1016/j.bone.2007.03.019 14. D'Amelio P., Grimaldi A., Di Bella S., Brianza S.Z.M., Cristofaro M.A., Tamone C., Giribaldi G., Ulliers D., Pescarmona G.P., Isaia G.. Estrogen deficiency increases osteoclastogenesis up-regulating T cells activity: A key mechanism in osteoporosis. Bone (2008) 43 92-100. DOI: 10.1016/j.bone.2008.02.017 15. D'Amelio P., Isaia G.C.. The use of raloxifene in osteoporosis treatment. Expert. Opin. Pharmacother. (2013) 14 949-956. DOI: 10.1517/14656566.2013.782002 16. Ettinger B., Black D.M., Mitlak B.H., Knickerbocker R.K., Nickelsen T., Genant H.K., Christiansen C., Delmas P.D., Zanchetta J.R., Stakkestad J.. Reduction of vertebral fracture risk in postmenopausal women with osteoporosis treated with raloxifene: Results from a 3-year randomized clinical trial. Multiple Outcomes of Raloxifene Evaluation (MORE) Investigators. JAMA (1999) 282 637-645. DOI: 10.1001/jama.282.7.637 17. Barrett-Connor E., Grady D., Sashegyi A., Anderson P.W., Cox D.A., Hoszowski K., Rautaharju P., Harper K.D., Investigators M.. Raloxifene and cardiovascular events in osteoporotic postmenopausal women: Four-year results from the MORE (Multiple Outcomes of Raloxifene Evaluation) randomized trial. JAMA (2002) 287 847-857. DOI: 10.1001/jama.287.7.847 18. Hsu T.W., Hsu C.N., Wang S.W., Huang C.C., Li L.C.. Comparison of the Effects of Denosumab and Alendronate on Cardiovascular and Renal Outcomes in Osteoporotic Patients. J. Clin. Med. (2019) 8. DOI: 10.3390/jcm8070932 19. Chen C.L., Chen N.C., Wu F.Z., Wu M.T.. Impact of denosumab on cardiovascular calcification in patients with secondary hyperparathyroidism undergoing dialysis: A pilot study. Osteoporos. Int. (2020) 31 1507-1516. DOI: 10.1007/s00198-020-05391-3 20. Saitta A., Morabito N., Frisina N., Cucinotte D., Corrado F., D'Anna R., Altavilla D., Squadrito G., Minutoli L., Arcoraci V.. Cardiovascular effects of raloxifene hydrochloride. Cardiovasc. Drug. Rev. (2001) 19 57-74. DOI: 10.1111/j.1527-3466.2001.tb00183.x 21. Mijatovic V., van der Mooren M.J., Kenemans P., de Valk-de Roo G.W., Netelenbos C.. Raloxifene lowers serum lipoprotein(A) in healthy postmenopausal women: A randomized, double-blind, placebo-controlled comparison with conjugated equine estrogens. Menopause (1999) 6 134-137. DOI: 10.1097/00042192-199906020-00010 22. Huo S., Shi W., Ma H., Yan D., Luo P., Guo J., Li C., Lin J., Zhang C., Li S.. Alleviation of Inflammation and Oxidative Stress in Pressure Overload-Induced Cardiac Remodeling and Heart Failure via IL-6/STAT3 Inhibition by Raloxifene. Oxid. Med. Cell. Longev. (2021) 2021 6699054. DOI: 10.1155/2021/6699054 23. Shi H., Santos H.O., de Souza I.G.O., Hoilat G.J., Martins C.E.C., Varkaneh H.K., Alkhwildi J.A., Hejji A.T., Almuqayyid F., Abu-Zaid A.. The Effect of Raloxifene Treatment on Lipid Profile in Elderly Individuals: A Systematic Review and Meta-analysis of Randomized Clinical Trials. Clin. Ther. (2021) 43 297-317. DOI: 10.1016/j.clinthera.2021.07.017 24. Yang F., Li N., Gaman M.A., Wang N.. Raloxifene has favorable effects on the lipid profile in women explaining its beneficial effect on cardiovascular risk: A meta-analysis of randomized controlled trials. Pharmacol. Res. (2021) 166 105512. DOI: 10.1016/j.phrs.2021.105512 25. Shoback D., Rosen C.J., Black D.M., Cheung A.M., Murad M.H., Eastell R.. Pharmacological Management of Osteoporosis in Postmenopausal Women: An Endocrine Society Guideline Update. J. Clin. Endocrinol. Metab. (2020) 105 587-594. DOI: 10.1210/clinem/dgaa048 26. Kanis J.A., Cooper C., Rizzoli R., Reginster J.Y.. European guidance for the diagnosis and management of osteoporosis in postmenopausal women. Osteoporos. Int. (2019) 30 3-44. DOI: 10.1007/s00198-018-4704-5 27. Management of Osteoporosis in Postmenopausal Women: The Position Statement of The North American Menopause Society'' Editorial, P. Management of osteoporosis in postmenopausal women: The 2021 position statement of The North American Menopause Society. Menopause (2021) 28 973-997. DOI: 10.1097/GME.0000000000001831 28. Arceo-Mendoza R.M., Camacho P.M.. Postmenopausal Osteoporosis: Latest Guidelines. Endocrinol. Metab. Clin. North Am. (2021) 50 167-178. DOI: 10.1016/j.ecl.2021.03.009 29. Fraser S.D., Taal M.W.. Multimorbidity in people with chronic kidney disease: Implications for outcomes and treatment. Curr. Opin. Nephrol. Hypertens. (2016) 25 465-472. DOI: 10.1097/MNH.0000000000000270 30. Bowling C.B., Plantinga L., Phillips L.S., McClellan W., Echt K., Chumbler N., McGwin G., Vandenberg A., Allman R.M., Johnson T.M.. Association of Multimorbidity with Mortality and Healthcare Utilization in Chronic Kidney Disease. J. Am. Geriatr. Soc. (2017) 65 704-711. DOI: 10.1111/jgs.14662 31. Lee W.C., Lee Y.T., Li L.C., Ng H.Y., Kuo W.H., Lin P.T., Liao Y.C., Chiou T.T., Lee C.T.. The Number of Comorbidities Predicts Renal Outcomes in Patients with Stage 3(-)5 Chronic Kidney Disease. J. Clin. Med. (2018) 7. DOI: 10.3390/jcm7120493 32. Cicek M., Buckley J., Pearson-Stuttard J., Gregg E.W.. Characterizing Multimorbidity from Type 2 Diabetes: Insights from Clustering Approaches. Endocrinol. Metab. Clin. North Am. (2021) 50 531-558. DOI: 10.1016/j.ecl.2021.05.012 33. Lin K.M., Lu C.L., Hung K.C., Wu P.C., Pan C.F., Wu C.J., Syu R.S., Chen J.S., Hsiao P.J., Lu K.C.. The Paradoxical Role of Uric Acid in Osteoporosis. Nutrients (2019) 11. DOI: 10.3390/nu11092111 34. Barrett-Connor E., Mosca L., Collins P., Geiger M.J., Grady D., Kornitzer M., McNabb M.A., Wenger N.K., Raloxifene Use for The Heart Trial I.. Effects of raloxifene on cardiovascular events and breast cancer in postmenopausal women. N. Engl. J. Med. (2006) 355 125-137. DOI: 10.1056/NEJMoa062462 35. Dayspring T., Qu Y., Keech C.. Effects of raloxifene on lipid and lipoprotein levels in postmenopausal osteoporotic women with and without hypertriglyceridemia. Metabolism (2006) 55 972-979. DOI: 10.1016/j.metabol.2006.03.005 36. Hofbauer L.C., Brueck C.C., Shanahan C.M., Schoppet M., Dobnig H.. Vascular calcification and osteoporosis--from clinical observation towards molecular understanding. Osteoporos. Int. (2007) 18 251-259. DOI: 10.1007/s00198-006-0282-z 37. Samelson E.J., Miller P.D., Christiansen C., Daizadeh N.S., Grazette L., Anthony M.S., Egbuna O., Wang A., Siddhanti S.R., Cheung A.M.. RANKL inhibition with denosumab does not influence 3-year progression of aortic calcification or incidence of adverse cardiovascular events in postmenopausal women with osteoporosis and high cardiovascular risk. J. Bone. Miner. Res. (2014) 29 450-457. DOI: 10.1002/jbmr.2043 38. Suzuki S., Suzuki M., Hanafusa N., Tsuchiya K., Nitta K.. Denosumab Recovers Aortic Arch Calcification During Long-Term Hemodialysis. Kidney. Int. Rep. (2021) 6 605-612. DOI: 10.1016/j.ekir.2020.12.002 39. Eriksen E.F., Lyles K.W., Colon-Emeric C.S., Pieper C.F., Magaziner J.S., Adachi J.D., Hyldstrup L., Recknor C., Nordsletten L., Lavecchia C.. Antifracture efficacy and reduction of mortality in relation to timing of the first dose of zoledronic acid after hip fracture. J. Bone Miner. Res. (2009) 24 1308-1313. DOI: 10.1359/jbmr.090209 40. Wu P.H., Lin M.Y., Huang T.H., Lee T.C., Lin S.Y., Chen C.H., Kuo M.C., Chiu Y.W., Chang J.M., Hwang S.J.. Kidney Function Change and All-Cause Mortality in Denosumab Users with and without Chronic Kidney Disease. J. Pers. Med. (2022) 12. DOI: 10.3390/jpm12020185 41. Ono T., Hayashi M., Sasaki F., Nakashima T.. RANKL biology: Bone metabolism, the immune system, and beyond. Inflamm. Regen. (2020) 40 2. DOI: 10.1186/s41232-019-0111-3 42. Bando J.K., Gilfillan S., Song C., McDonald K.G., Huang S.C., Newberry R.D., Kobayashi Y., Allan D.S.J., Carlyle J.R., Cella M.. The Tumor Necrosis Factor Superfamily Member RANKL Suppresses Effector Cytokine Production in Group 3 Innate Lymphoid Cells. Immunity (2018) 48 1208-1219.e1204. DOI: 10.1016/j.immuni.2018.04.012 43. Guerrini M.M., Okamoto K., Komatsu N., Sawa S., Danks L., Penninger J.M., Nakashima T., Takayanagi H.. Inhibition of the TNF Family Cytokine RANKL Prevents Autoimmune Inflammation in the Central Nervous System. Immunity (2015) 43 1174-1185. DOI: 10.1016/j.immuni.2015.10.017 44. Loser K., Mehling A., Loeser S., Apelt J., Kuhn A., Grabbe S., Schwarz T., Penninger J.M., Beissert S.. Epidermal RANKL controls regulatory T-cell numbers via activation of dendritic cells. Nat. Med. (2006) 12 1372-1379. DOI: 10.1038/nm1518 45. Hanada R., Leibbrandt A., Hanada T., Kitaoka S., Furuyashiki T., Fujihara H., Trichereau J., Paolino M., Qadri F., Plehm R.. Central control of fever and female body temperature by RANKL/RANK. Nature (2009) 462 505-509. DOI: 10.1038/nature08596 46. Shao S.C., Chan Y.Y., Kao Yang Y.H., Lin S.J., Hung M.J., Chien R.N., Lai C.C., Lai E.C.. The Chang Gung Research Database-A multi-institutional electronic medical records database for real-world epidemiological studies in Taiwan. Pharmacoepidemiol. Drug Saf. (2019) 28 593-600. DOI: 10.1002/pds.4713 47. Cheng T.M.. Taiwan's new national health insurance program: Genesis and experience so far. Health Aff. (2003) 22 61-76. DOI: 10.1377/hlthaff.22.3.61 48. Huang Y.T., Chen Y.J., Chang S.H., Kuo C.F., Chen M.H.. Discharge status validation of the Chang Gung Research database in Taiwan. Biomed. J. (2021) 45 907-913. DOI: 10.1016/j.bj.2021.12.006 49. Sundararajan V., Henderson T., Perry C., Muggivan A., Quan H., Ghali W.A.. New ICD-10 version of the Charlson comorbidity index predicted in-hospital mortality. J. Clin. Epidemiol. (2004) 57 1288-1294. DOI: 10.1016/j.jclinepi.2004.03.012 50. Austin P.C.. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun. Stat. Simul. Comput. (2009) 38 1228-1234. DOI: 10.1080/03610910902859574
--- title: 'Associations with Blood Lead and Urinary Cadmium Concentrations in Relation to Mortality in the US Population: A Causal Survival Analysis with G-Computation' authors: - Nasser Laouali - Tarik Benmarhnia - Bruce P. Lanphear - Youssef Oulhote journal: Toxics year: 2023 pmcid: PMC9966985 doi: 10.3390/toxics11020133 license: CC BY 4.0 --- # Associations with Blood Lead and Urinary Cadmium Concentrations in Relation to Mortality in the US Population: A Causal Survival Analysis with G-Computation ## Abstract Using the parametric g-formula, we estimated the 27-year risk of all-cause and specific causes of mortality under different potential interventions for blood lead (BLLs) and urinary cadmium (UCd) levels. We used data on 14,311 adults aged ≥20 years enrolled in the NHANES-III between 1988 and 1994 and followed up through 31 Dec 31 2015. Time and cause of death were determined from the National Death Index records. We used the parametric g-formula with pooled logistic regression models to estimate the relative and absolute risk of all-cause, cardiovascular, and cancer mortality under different potential threshold interventions for BLLs and UCd concentrations. Median follow-up was 22.5 years. A total of 5167 ($36\%$) participants died by the end of the study, including 1550 from cardiovascular diseases and 1135 from cancer. Increases in BLLs and creatinine-corrected UCd levels from the 5th to the 95th percentiles were associated with risk differences of $4.17\%$ (1.54 to 8.77) and $6.22\%$ (4.51 to 12.00) for all-cause mortality, $1.52\%$ (0.09 to 3.74) and $1.06\%$ (−0.57 to 3.50) for cardiovascular disease mortality, and $1.32\%$ (−0.09 to 3.67) and $0.64\%$ (−0.98 to 2.80) for cancer mortality, respectively. Interventions to reduce historical exposures to lead and cadmium may have prevented premature deaths, especially from cardiovascular disease. ## 1. Introduction Cardiovascular disease and cancer are the two leading causes of death, accounting for 17.9 million and 9 million deaths worldwide, respectively, in 2016 [1]. While there has been a marked decline in cardiovascular disease mortality rates, cancer mortality is increasing and is even the leading cause of death in many countries [2]. In 2021, it was estimated that 608,570 Americans would die from cancer, corresponding to more than 1600 deaths per day [3]. These two causes of death, while seemingly different, share common risk factors, including being overweight or obese [4,5], not being physically active [6,7], having a poor diet [8], and exposure to environmental chemicals such as lead and cadmium [9]. Lead and cadmium are among the most common environmental and occupational pollutants derived from natural resources or as a byproduct of industries such as mining. In the last decades, lead and cadmium exposures have declined sharply in the United States [9,10]; however, in 2019, the US Agency for Toxic Substances and Disease Registry ranked lead and cadmium second and seventh, respectively, on the National Priorities List of substances that pose the most significant potential threat to human health [11]. While epidemiological studies have consistently reported that exposures to lead and cadmium above the safe standards are associated with increased risks of cardiovascular, cancer, and all-cause mortality [12,13], studies have also reported that chronic low-concentration lead exposure (1–5 μg/dL in the blood) may increase the risk of premature death, especially from cardiovascular diseases [14,15]. In contrast, there are conflicted results for cancer mortality risk at this range of exposure [15]. Results on urinary cadmium concentrations and mortality were reviewed by Larsson et al. in a recent meta-analysis, which highlighted the limited number of prospective studies ($$n = 9$$ studies) and conflicting results, concluding that there is a need for further large prospective studies [16]. Some limitations persist when analyzing the prospective associations of lead and cadmium with mortality outcomes. First, the non-adjustment for dietary intake (one of the main sources of exposure) and predictors of cancer and cardiovascular disease mortality. Second, the non-consideration of mutual adjustment between these two metals that could confound each other. The two metals share common sources and may also share common mechanistic pathways, such as inflammation, oxidative stress, and recently, epigenetic changes [17]. Third, all previous prospective studies used a Cox proportional hazard model to estimate a single averaged hazard ratio, although the hazard ratio may change over the study’s follow-up. As a result, the conclusions from the study may critically depend on the duration of the follow-up [18]. Finally, previous studies focused on the consequences of exposures to these metals, but it is also important to model the population level impact of potential policies or interventions aimed at reducing exposure levels for informing public health. The hazard ratio alone does not provide this information since the clinical significance depends on the baseline rate [19]. Thus, it is necessary to address all of these limitations and consider the use of models that allow us to overcome the proportional hazards assumption and evaluate the impact of potential interventions on these exposures. Therefore, we first aimed to extend the follow-up of previous studies in the US National Health and Nutrition Examination Survey (NHANES-III). Secondly, we aimed to use a pooled logistic regression model for the discrete-time hazard within the parametric g-formula to address previous limitations while flexibly simulating the expected all-cause and specific causes of mortality distributions for hypothetical interventions related to lead and cadmium exposures in the US population. ## 2.1. Design and Participants The US National Health and Nutrition Examination Survey (NHANES) data, which were collected, stored, and analyzed by the Centers for Disease Control and Prevention (CDC), were used for this study. The NHANES is an ongoing survey conducted by the CDC that uses a representative sample of non-institutionalized civilians in the US, selected using a complex, multistage, stratified, clustered probability design. Information on participants was collected by interviews and in personal physical examinations by the CDC. The interview includes background information such as socio-demographic, dietary, and health-related questions. The examination component consists of medical and physiological measurements, as well as laboratory tests. The National Center for Health Statistics Ethics Review Board approved all NHANES protocols, and all survey participants completed a consent form. The detailed protocol on NHANES methodology and data collection is available at https://www.cdc.gov/nchs/nhanes/index.htm (accessed on 12 March 2021). For this study, adults aged ≥20 years enrolled in the NHANES-III between 1988 and 1994 with data on blood lead and urinary cadmium concentrations ($$n = 16$$,040) were included. Exposure and covariate data from NHANES-III were then linked to the National Death Index mortality data. ## 2.2. Mortality Data A full description of the mortality linkage method is available from the National Center for Health Statistics (NCHS) [20]. Briefly, the de-identified and anonymized data of the NHANES III participants were linked to National Death Index mortality files based on 12 identifiers for each participant (e.g., Social Security number, sex, and date of birth) with a probabilistic matching algorithm to determine mortality status. The NCHS public-use linked mortality file provides mortality follow-up data from the date of NHANES III survey participation until 31 December 2015 (1988–2015). Participants with no matched death record at this date were assumed to be alive during the entire follow-up period. In a validation study using mortality-linked data from the first NHANES study (NHANES-I; 1971–75), $96\%$ of deceased participants and $99\%$ of those still alive were classified correctly [21]. The underlying causes of death were recorded in the public-use linked mortality file using the following ICD-10 codes: cardiovascular diseases including heart diseases (I00–I09, I11, I13, I20–I51) and cerebrovascular diseases (I60–I69) and malignant neoplasms (C00-C97). From the 16,040 participants with baseline data, 1729 had missing data on mortality and other covariates. The final study sample included 14,311 participants. ## 2.3. Measurements of Blood Lead and Urinary Cadmium Blood and urine samples were collected by the CDC during the medical examination. The laboratory methods for processing these samples are described in detail elsewhere [22]. Briefly, the blood and urine specimens were frozen (−30 and −20 °C, respectively), stored, and shipped for analysis to the Division of Laboratory Sciences, National Center for Environmental Health (Centers for Disease Control and Prevention in Atlanta, GA, USA). Lead (µg/dL) concentration was measured in whole blood using inductively coupled plasma mass spectrometry. Urinary cadmium was measured by graphite furnace atomic absorption with Zeeman background correction using the CDC method [22] with the modification proposed by Pruszkowska et al. [ 23]. Specimens were analyzed in duplicate and the average of the two measurements was reported. The detection limits were 1.0 μg/dL (0.048 μmol/L) and 0.03 μg/L for blood lead and urinary cadmium, respectively. For study participants who had concentrations of blood lead below the level of detection ($$n = 1217$$; $8.5\%$), values were imputed using LOD/√2 [0.7 μg/dL]. Urinary creatinine measured using the Jaffe reaction with a Beckman Synchron AS/ASTRA Clinical Analyzer (Beckman Instruments, Inc., Brea, CA, USA) was used to account for urine dilution. ## 2.4. Covariates Baseline covariates were collected when individuals participated in a household interview and demographic information—including sex (male/female), age (continuous; years), race-ethnicity (non-Hispanic whites, non-Hispanic blacks, Mexican Americans, and others), poverty-to-income ratio (categorized in tertiles), the number of years of education attended and completed (continuous; years), area of residence (metro and non-metro counties), and smoking status (current, former, and never). Information on body-mass index ([BMI] continuous; kg/m [2]), physical activity (none, 1 to 14 times, 15 or more times; per month), and overall dietary quality indexes (continuous) was obtained during the medical examination. Dietary intake was collected using a 24-h dietary recall. We derived the diet quality indexes, as measured by the Healthy Eating Index 2015 (HEI-2015) [24] and the adapted dietary inflammatory index [25], from the daily intakes of foods/beverages, energy, and nutrients based on the 24-h dietary recall. A complete description of the development of these scores is described in Supplementary Methods. ## 2.5. Statistical Analysis Complete data on exposures, covariates, and mortality were available for 14,311 participants. We log-transformed (base 2) blood lead and urinary cadmium concentrations to reduce the influence of outliers, and descriptive and bivariate analyses are reported as geometric means and geometric standard errors (SE) by population characteristics. We used the parametric g-computation to estimate the risk ratio (RR) and risk differences (RD) of all-cause and specific causes of mortality under hypothetical interventions. In 1986, Robins introduced g-methods, a class of causal inference techniques that allows building an outcome prediction model based on observed quantities, and then predicts potential outcomes under potential hypothetical intervention [26,27]. In recent years, there have been substantial advances in the application of this method, which has been used, for instance, to evaluate hypothetical interventions for sources of lead exposure on BLLs [28]. The parametric g-formula is a generalization of the standardization method that enables flexible simulation and estimation of survival curves to visualize the time-specific effect estimates of any form of hypothetical intervention. A more detailed discussion of this method is presented elsewhere [29,30,31]. Briefly, we first estimated the expected mortality for all observations using confounders and metal exposure levels as predictors in a pooled logistic regression. The discrete-time hazards of all-cause and specific causes of mortality for each 2-fold increase in the baseline metals concentrations (log2-transformed to reduce skewness) were then estimated. Then, we used the model fit to predict every participant’s mortality while manipulating the exposure by setting it to [1] “high level” metal concentrations for every participant and [2] “low level” metal concentrations for every participant. Finally, we estimated the average difference between the expected mortality. We used the non-parametric bootstrap method ($M = 200$) to calculate the confidence intervals around the estimate (the 2.5th and 97.5th percentiles as the lower and upper confidence interval limits, respectively). There are no known safety levels for the blood and urinary concentrations of these metals in adults. Therefore, we chose the interventions listed below based on previous epidemiologic analyses, as discussed elsewhere [14]. We compared the estimated risk of mortality under the following interventions: [1] all participants were assigned a high concentration (e.g., 95th percentile values of the metal distributions) with the estimated risk of mortality under the intervention. [ 2] All participants were assigned a low concentration (e.g., 5th percentile values of the metal distributions). This approach assumed a linear association between metal concentration and death. To check this assumption, we used multivariable restricted cubic splines with three knots placed at the 5th, 50th, and 95th percentiles of each metal concentration distribution to provide a graphical presentation [32]. Splines allowed us to test whether there was any departure from linearity. Finally, we also considered interventions comparing quartile groups of lead and cadmium concentrations. We categorized metal concentrations into quartiles and estimated the discrete-time hazards of all-cause and specific causes of mortality for each quartile with the first quartile group, the lowest metal concentration, as the reference category. We then compared the estimated risk of mortality under the following interventions: [1] all participants belonged to the bottom quartile group with the estimated risk of mortality under the intervention. [ 2] All participants belonged to the low quartile group. Models were adjusted for age, sex, ethnicity, poverty index, education level, area of residence, smoking status, BMI, physical activity, diet quality evaluated by the healthy eating index, and metal concentrations (mutual adjustment). The selection of potential confounders was done a priori. We also included product terms between the metal concentrations and time in all models to account for the time-varying risk. All analyses were weighted by the provided sample weights to account for the unequal probabilities of inclusion and response rates. Additionally, we investigated age (<50 years and ≥50 years) and sex-specific estimates in stratified analyses since previous studies reported potential effect modifications of the associations between metal concentrations and mortality by sex and age [12,14]. We also applied the Wald test [33] to assess the difference in associations of metal exposures with risk of mortality between subgroups of sex and age. A p-value < 0.1 was considered to be statistically significant. ## 2.6. Sensitivity Analyses We ran sensitivity analyses using unweighted models that did not account for unequal probabilities of inclusion and response rates in the NHANES survey because the weighted method was inefficient due to the large variability in assigned weights [34]. Unweighted analysis yields correct estimates when models are adjusted for the auxiliary variables used to define the weights (i.e., age, sex, and ethnicity) [34]. Finally, as smoking is a source of lead and cadmium, the main analysis was adjusted for exposure to tobacco smoke measured by cotinine level and pack-years of cumulative active smoking to further account for residual confounding by smoking. Statistical analyses were performed using R version 4.0.4 and Statistical Analysis System software version 9.4 (SAS Institute, Cary, NC, USA). ## 3. Results A total of 14,311 participants were included (mean age 48.0 ± 18.1 years) for this analysis. The blood lead concentrations ranged from 0.70 to 56.0 μg/dL (0.034 to 2.70 μmol/L), with a geometric mean (GM) of 2.97 μg/dL (geometric standard error [GSE]= 1.01). BLLs were higher in older participants, males, current and former smokers, those who reported drinking alcohol more than four time per month, and those who were in the categories of lower healthy eating index and low poverty-to-income ratio (Table 1). The urinary creatinine-standardized cadmium concentrations ranged from 0.002 to 23.35 μg/g, with a GM of 0.36 μg/g (GSE = 1.01). Participants who had the highest concentrations of urinary cadmium were older and more likely to be female, current and former smokers, and to not practice physical activity. There were no major differences among other participant characteristics (Table 1). ## 3.1. Blood Lead and Urinary Cadmium Concentrations and All-Cause and Cause-Specific Mortality During a median follow-up of 22.5 years (IQR 16.3–24.7), 5167 ($36\%$) participants died with 1550 ($30\%$) and 1135 ($22\%$) deaths attributable to cardiovascular disease and cancer, respectively. When comparing a potential intervention assigning all participants to 5th percentile values (0.70 μg/dL and 0.04 μg/g for blood lead and urinary cadmium levels; respectively) to an intervention assigning all participants to 95th percentile values (9.70 μg/dL and 1.63 μg/g for blood lead and urinary cadmium levels; respectively), we observed $138\%$ ($95\%$ CI, 14 to 196) and $126\%$ ($95\%$ CI, 77 to 383) increases in the risk of all-cause mortality at 27 years of follow-up for blood lead and urinary cadmium, respectively. On the absolute scale, we observed $4.17\%$ ($95\%$ CI, 1.54 to 8.77) and $6.22\%$ ($95\%$ CI, 4.51 to 12.00) increases in all-cause mortality associated with blood lead and urinary cadmium, respectively (Figure 1, Table 2). For cause-specific mortality, comparing the 5th to the 95th percentile assignments, blood lead and urinary cadmium showed a $109\%$ ($95\%$ CI, 4 to 268) and $37\%$ ($95\%$ CI, −0.20 to 239) increased risk of cardiovascular disease mortality at 27 years of follow-up, respectively (Figure 1, Table 2). The estimated 27-year cancer mortality risk after intervention to set all participants to the 95th percentile compared to the 5th percentile was increased by $287\%$ ($95\%$ CI, 12 to 691) and $144\%$ ($95\%$ CI, −0.34 to 281) for blood lead and urinary cadmium, respectively (Figure 1, Table 2). We did not observe any departure from linearity in the associations between metal concentrations and mortality when using smoothing splines (Supplemental Material; Figure S1). When we considered interventions comparing quartile groups of blood lead and urinary cadmium levels, the estimated RD of 27-year all-cause mortality after intervening to set all participants to the fourth quartile of metal concentrations compared to setting all participants to the first quartile of metal concentrations showed a percentage point increase of $5.52\%$ ($95\%$ CI, 1.25 to 18.25) and $10.48\%$ ($95\%$ CI, 5.15 to 27.99) for blood lead and urinary cadmium, respectively (Figure 2, Table 3). The corresponding increases in the 27-year RD for cardiovascular disease mortality were $5.18\%$ ($95\%$ CI, 0.47 to 48.61) and $6.31\%$ ($95\%$ CI, 3.76 to 50.38) for blood lead and urinary cadmium, respectively. For cancer mortality, the 27-year RD increased $4.17\%$ ($95\%$ CI, −0.28 to 26.00) and $0.13\%$ ($95\%$ CI, −1.57 to 48.35) for blood lead and urinary cadmium, respectively (Figure 2, Table 3). ## 3.2. Analyses Stratified by Sex and Age We found that the associations between the blood lead concentration and all-cause (p-heterogeneity = 0.021) and cardiovascular disease (p-heterogeneity <0.001) mortality were more pronounced for older (≥50 years) compared to younger (<50 years) participants, while for cancer mortality, the association was similar for younger and older participants (p-heterogeneity = 0.429) (Supplemental Figure S2 and Table S1). The associations between the urinary cadmium concentration and all-cause (p-heterogeneity = 0.020) and cardiovascular disease (p-heterogeneity = 0.093) mortality were more pronounced for older (≥50 years) compared to younger (<50 years) participants, while for cancer mortality, the association was similar for younger and older participants (p-heterogeneity = 0.525) (Supplemental Figure S2 and Table S1). Analyses stratified by sex showed that the associations between the urinary cadmium concentration and all-cause (p-heterogeneity = 0.098), cardiovascular disease (p-heterogeneity = 0.032), and cancer (p-heterogeneity = 0.102) mortality were more pronounced in men. There was no difference in the mortality risks associated with blood lead concentration and all-cause (p-heterogeneity = 0.622), cardiovascular disease (p-heterogeneity = 0.202), and cancer (p-heterogeneity = 0.206) mortality (Supplemental Figure S3 and Table S2). ## 3.3. Sensitivity Analyses When the models were unweighted, we observed the same pattern of associations for all-cause and cancer mortality both for lead and cadmium exposures (Supplemental Table S3). However, the association between blood lead concentration and cardiovascular disease mortality was more pronounced with an RR of 2.09 ($95\%$ CI, 1.04 to 3.68) at 27 years of follow-up compared to the model without the survey weights (RR = 1.02 (0.52 to 2.80), which corresponded to a percentage point increase of $0.69\%$ ($95\%$ CI, −2.72 to 3.85) (Supplemental Table S3). The model with the adjustment for more precise tobacco smoking measurement showed the same pattern of associations (Supplemental Table S4). ## 4. Discussion Using a large, nationally representative sample of US adults, we found that in those with high concentrations of blood lead and urinary cadmium, there were excess deaths from all causes (corresponding to 6,460,092 and 9,635,916 deaths, respectively, when considering NHANES weights), cardiovascular disease (corresponding to 2,354,758 and 1,642,134 deaths, respectively), and cancer (corresponding to 2,044,921 and 991,477 deaths, respectively) after 27 years. The associations were more pronounced for older participants, except for cancer mortality associated with blood lead levels. In addition, all of these associations were more pronounced in men than in women for urinary cadmium concentrations. Our findings are in line with previous population-based studies showing that exposures to lead and cadmium were associated with increased risks of cardiovascular, cancer, and all-cause mortality [12,13,14,15,16]. A recent study showed that environmental declines in lead and cadmium exposures were associated with a $32\%$ reduction in cardiovascular disease mortality [35]. Two studies reported that exposure to chronic, low concentrations of lead was associated with premature death, especially from cardiovascular disease [14,15]. In contrast, there was no association with cancer mortality risk [15]. Several epidemiological studies have prospectively examined the associations of cadmium exposure in relation to the risk of all-cause [12,36,37,38,39,40,41], cancer [37,42,43,44], and cardiovascular disease mortality [12,37,38,39,42]. Most studies have reported a positive association between urinary cadmium concentrations and mortality, except one study for cardiovascular disease mortality [12], two studies for cancer mortality [37,42] and two studies for all-cause mortality [12,36]. In comparison to other studies, our study explored the potential effect on mortality for different theoretical interventions of exposure levels of these metals. Environmental exposure to lead occurs continuously over a lifetime and lead is retained in the body for decades. Blood lead is an established biomarker of recent exposure, although it also shows a small component with a half-life of 5–20 years that reflects endogenous exposure from bone lead redistribution [45]. Several mechanisms have been proposed for the role of lead in cardiovascular events. Lead exposure can result in oxidative stress, inflammation, and diminished endothelium relaxation, and it promotes the development of atherosclerosis and thrombosis [46]. In addition, lead is a well-known risk factor for hypertension and has been associated with peripheral arterial disease, electrocardiographic abnormalities, left-ventricular hypertrophy, alteration of cardiac conduction, cardiac disease, and increased mortality due to cardiovascular disease [45,47]. Regarding cancer toxicity, the mechanisms by which lead may lead to tumor development is unclear. However, lead has been defined as a “probable human carcinogen” by the International Agency for Research on Cancer [48]. It has been proposed that lead can facilitate the process of carcinogenesis by inhibiting DNA synthesis and repair and by interacting with binding proteins, thus hindering tumor suppressor proteins [49]. Lead may also affect carcinogenesis through mechanisms involving oxidative damage, induction of apoptosis, and altered signaling pathways [50]. While blood cadmium tends to reflect recent exposures, urinary cadmium reflects kidney cadmium contents and, with a half-life of 15–30 years, is an established biomarker of cumulative body burden. The cardiovascular toxic effects of cadmium exposure have been well described. Experimental evidence supports a role for cadmium in atherosclerosis, including increased inflammation [51] and endothelial oxidative stress [52,53]. Results from epidemiological studies show that high cadmium exposure is associated with hypertension [54], growth of atherosclerotic plaques [55,56] and cardiovascular disease [57,58,59]. As mentioned recently by Lamas et al. [ 60], strong evidence supports that it is time to recognize metal contaminants in the evaluation, treatment, and prevention of cardiovascular diseases. Evidence for cancer toxicity stems from various mechanisms. Cadmium may promote carcinogenesis through induction of oxidative stress [61,62], suppression of DNA, and changes in DNA methylation and apoptosis repair [62,63,64]. Another alternative hypothesis is through estrogenic activities [65]. This study used the NHANES III dataset, a large, national survey of which the findings are generalizable to the adult, non-institutionalized, U.S. population. There are many strengths to this study, including its large sample-size and random sampling, the mutual adjustment between lead and cadmium metals, the adjustment for dietary intake (one of the main sources of exposure) and, most importantly, we used a method that was not based on proportional hazard assumptions and the graphic representation of the risk at each follow-up time allowed us to show that the risk varied over the duration of the follow-up. For example, the risk differences for cardiovascular disease mortality associated with blood lead and urinary cadmium concentrations were +$1.52\%$ and +$1.06\%$ at the end of the follow-up and +$0.47\%$ and +$0.27\%$ at the mid-follow-up (13 year), respectively (Supplemental Tables S5 and S6). Nevertheless, there are important limitations to note. The key limitation is that we relied on blood concentrations for lead; therefore, we did not account for cumulative chronic or long-term lead exposure. Urine measurements for lead are better indicators of cumulative exposure and would have strengthened the results. Furthermore, covariable data were only available at baseline. Thus, exposure-confounder relationships were not well defined temporally. Another limitation is that we relied on death certificates for the underlying cause of death, but they are imperfect [66]. Most importantly, although the main potential confounding factors were accounted for, there could be other cardiovascular disease and cancer risk factors and unmeasured confounding from other toxicants, especially in occupational settings, which may have influenced our findings. Finally, because internal dose metrics cannot correctly define the complete history of exposure and duration, the timing that correlates most strongly with the observed health effects are typically unknown or highly uncertain. This study focused only on hypothetical interventions related to lead and cadmium exposures to simulate what would have been the benefits of historical interventions. We recommend that future directions in environmental health research explore other interventions based on, for example, dietary and lifestyle factors, which may be complementary to pollutant-based interventions. In conclusion, our findings suggest that blood lead and urinary cadmium are associated with all-cause, cardiovascular disease, and cancer mortality. Despite the continuous decrease in lead and cadmium exposures in the U.S. population, we confirmed the previously reported associations and showed that several deaths could have been prevented under interventions to reduce the blood lead concentration from 9.70 to 0.70 μg/dL and the creatinine-corrected urinary cadmium concentration from 1.63 to 0.04 μg/g. Interventions to reduce historical exposures to these metals would be more effective (on the absolute scale) among individuals ≥50 years of age as well as in men. ## References 1. NCD Mortality and Morbidity 2. Sung H., Ferlay J., Siegel R.L., Laversanne M., Soerjomataram I., Jemal A., Bray F.. Global cancer statistics 2020: GLOBOCAN estimates of incidence and mortality worldwide for 36 cancers in 185 countries. CA Cancer J. Clin. (2021.0) 71 209-249. DOI: 10.3322/caac.21660 3. Siegel R.L., Miller K.D., Fuchs H.E., Jemal A.. Cancer Statistics, 2021. CA Cancer J. Clin. (2021.0) 71 7-33. DOI: 10.3322/caac.21654 4. Parra-Soto S., Cowley E.S., Rezende L.F.M., Ferreccio C., Mathers J.C., Pell J.P., Ho F.K., Celis-Morales C.. Associations of six adiposity-related markers with incidence and mortality from 24 cancers-findings from the UK Biobank prospective cohort study. BMC Med. (2021.0) 19. DOI: 10.1186/s12916-020-01848-8 5. Zhang Y.B., Pan X.F., Chen J., Cao A., Xia L., Zhang Y., Wang J., Li H., Liu G., Pan A.. Combined lifestyle factors, all-cause mortality and cardiovascular disease: A systematic review and meta-analysis of prospective cohort studies. J. Epidemiol. Community Health (2021.0) 75 92-99. DOI: 10.1136/jech-2020-214050 6. Ekblom-Bak E., Halldin M., Vikstrom M., Stenling A., Gigante B., de Faire U., Leander K., Hellenius M.L.. Physical activity attenuates cardiovascular risk and mortality in men and women with and without the metabolic syndrome—A 20-year follow-up of a population-based cohort of 60-year-olds. Eur. J. Prev. Cardiol. (2020.0) 28 1376-1385. DOI: 10.1177/2047487320916596 7. Rock C.L., Thomson C., Gansler T., Gapstur S.M., McCullough M.L., Patel A.V., Andrews K.S., Bandera E.V., Spees C.K., Robien K.. American Cancer Society guideline for diet and physical activity for cancer prevention. CA Cancer J. Clin. (2020.0) 70 245-271. DOI: 10.3322/caac.21591 8. Naghshi S., Sadeghi O., Willett W.C., Esmaillzadeh A.. Dietary intake of total, animal, and plant proteins and risk of all cause, cardiovascular, and cancer mortality: Systematic review and dose-response meta-analysis of prospective cohort studies. BMJ (2020.0) 370 m2412. DOI: 10.1136/bmj.m2412 9. Duan W., Xu C., Liu Q., Xu J., Weng Z., Zhang X., Basnet T.B., Dahal M., Gu A.. Levels of a mixture of heavy metals in blood and urine and all-cause, cardiovascular disease and cancer mortality: A population-based cohort study. Environ. Pollut. (2020.0) 263 114630. DOI: 10.1016/j.envpol.2020.114630 10. Dignam T., Kaufmann R.B., LeStourgeon L., Brown M.J.. Control of Lead Sources in the United States, 1970–2017: Public Health Progress and Current Challenges to Eliminating Lead Exposure. J. Public Health Manag. Pract. (2019.0) 25 S13-S22. DOI: 10.1097/PHH.0000000000000889 11. Agency for Toxic Substances and Disease Registry: Substance Priority List 12. Menke A., Muntner P., Silbergeld E.K., Platz E.A., Guallar E.. Cadmium levels in urine and mortality among U.S. adults. Environ. Health Perspect. (2009.0) 117 190-196. DOI: 10.1289/ehp.11236 13. Schober S.E., Mirel L.B., Graubard B.I., Brody D.J., Flegal K.M.. Blood lead levels and death from all causes, cardiovascular disease, and cancer: Results from the NHANES III mortality study. Environ. Health Perspect. (2006.0) 114 1538-1541. DOI: 10.1289/ehp.9123 14. Lanphear B.P., Rauch S., Auinger P., Allen R.W., Hornung R.W.. Low-level lead exposure and mortality in US adults: A population-based cohort study. Lancet Public Health (2018.0) 3 e177-e184. DOI: 10.1016/S2468-2667(18)30025-2 15. Menke A., Muntner P., Batuman V., Silbergeld E.K., Guallar E.. Blood lead below 0.48 micromol/L (10 microg/dL) and mortality among US adults. Circulation (2006.0) 114 1388-1394. DOI: 10.1161/CIRCULATIONAHA.106.628321 16. Larsson S.C., Wolk A.. Urinary cadmium and mortality from all causes, cancer and cardiovascular disease in the general population: Systematic review and meta-analysis of cohort studies. Int. J. Epidemiol. (2016.0) 45 782-791. DOI: 10.1093/ije/dyv086 17. Park S.S., Skaar D.A., Jirtle R.L., Hoyo C.. Epigenetics, obesity and early-life cadmium or lead exposure. Epigenomics (2017.0) 9 57-75. DOI: 10.2217/epi-2016-0047 18. Hernan M.A.. The hazards of hazard ratios. Epidemiology (2010.0) 21 13-15. DOI: 10.1097/EDE.0b013e3181c1ea43 19. Coory M., Lamb K.E., Sorich M.. Risk-difference curves can be used to communicate time-dependent effects of adjuvant therapies for early stage cancer. J. Clin. Epidemiol. (2014.0) 67 966-972. DOI: 10.1016/j.jclinepi.2014.03.006 20. 20. (NCHS)—National Center for Health Statistics NCHS Data Linked to NDI Mortality FilesNCHSHyattsville, MD, USA2016. NCHS Data Linked to NDI Mortality Files (2016.0) 21. 21. National Center for Health Statistics NHANES I Epidemiologic Follow-Up Survey (NHEFS): Calibration Sample for NDI Matching MethodologyNCHSWashington, DC, USA2009. NHANES I Epidemiologic Follow-Up Survey (NHEFS): Calibration Sample for NDI Matching Methodology (2009.0) 22. Gunter E.W., Lewis B.G., Koncikowski S.M.. Laboratory Procedures Used for the Third National Health and Nutrition Examination Survey (NHANES III), 1988–1994 (1996.0) 23. Pruszkowska E., Carnrick G.R., Slavin W.. Direct determination of cadmium in urine with use of a stabilized temperature platform furnace and Zeeman background correction. Clin. Chem. (1983.0) 29 477-480. DOI: 10.1093/clinchem/29.3.477 24. Krebs-Smith S.M., Pannucci T.E., Subar A.F., Kirkpatrick S.I., Lerman J.L., Tooze J.A., Wilson M.M., Reedy J.. Update of the Healthy Eating Index: HEI-2015. J. Acad. Nutr. Diet. (2018.0) 118 1591-1602. DOI: 10.1016/j.jand.2018.05.021 25. Laouali N., Mancini F.R., Hajji-Louati M., El Fatouhi D., Balkau B., Boutron-Ruault M.C., Bonnet F., Fagherazzi G.. Dietary inflammatory index and type 2 diabetes risk in a prospective cohort of 70,991 women followed for 20 years: The mediating role of BMI. Diabetologia (2019.0) 62 2222-2232. DOI: 10.1007/s00125-019-04972-0 26. Robins J.. A new approach to causal inference in mortality studies with a sustained exposure period—Application to control of the healthy worker survivor effect. Math. Model. (1986.0) 17 1393-1512. DOI: 10.1016/0270-0255(86)90088-6 27. Hernan M.A., Robins J.M.. Estimating causal effects from epidemiological data. J. Epidemiol. Community Health (2006.0) 60 578-586. DOI: 10.1136/jech.2004.029496 28. Oulhote Y., Le Bot B., Poupon J., Lucas J.P., Mandin C., Etchevers A., Zmirou-Navier D., Glorennec P.. Identification of sources of lead exposure in French children by lead isotope analysis: A cross-sectional study. Environ. Health (2011.0) 10 75. DOI: 10.1186/1476-069X-10-75 29. Breskin A., Edmonds A., Cole S.R., Westreich D., Cocohoba J., Cohen M.H., Kassaye S.G., Metsch L.R., Sharma A., Williams M.S.. G-computation for policy-relevant effects of interventions on time-to-event outcomes. Int. J. Epidemiol. (2021.0) 49 2021-2029. DOI: 10.1093/ije/dyaa156 30. Keil A.P., Edwards J.K., Richardson D.B., Naimi A.I., Cole S.R.. The parametric g-formula for time-to-event data: Intuition and a worked example. Epidemiology (2014.0) 25 889-897. DOI: 10.1097/EDE.0000000000000160 31. Ahern J., Hubbard A., Galea S.. Estimating the effects of potential public health interventions on population disease burden: A step-by-step illustration of causal inference methods. Am. J. Epidemiol. (2009.0) 169 1140-1147. DOI: 10.1093/aje/kwp015 32. Herndon J.E., Harrell F.E.. The restricted cubic spline as baseline hazard in the proportional hazards model with step function time-dependent covariables. Stat. Med. (1995.0) 14 2119-2129. DOI: 10.1002/sim.4780141906 33. Kaufman J.S., MacLehose R.F.. Which of these things is not like the others?. Cancer (2013.0) 119 4216-4222. DOI: 10.1002/cncr.28359 34. Korn E.L., Graubard B.I.. Epidemiologic studies utilizing surveys: Accounting for the sampling design. Am. J. Public Health (1991.0) 81 1166-1173. DOI: 10.2105/AJPH.81.9.1166 35. Ruiz-Hernandez A., Navas-Acien A., Pastor-Barriuso R., Crainiceanu C.M., Redon J., Guallar E., Tellez-Plaza M.. Declining exposures to lead and cadmium contribute to explaining the reduction of cardiovascular mortality in the US population, 1988–2004. Int. J. Epidemiol. (2017.0) 46 1903-1912. DOI: 10.1093/ije/dyx176 36. Nakagawa H., Nishijo M., Morikawa Y., Miura K., Tawara K., Kuriwaki J., Kido T., Ikawa A., Kobayashi E., Nogawa K.. Urinary cadmium and mortality among inhabitants of a cadmium-polluted area in Japan. Environ. Res. (2006.0) 100 323-329. DOI: 10.1016/j.envres.2005.08.014 37. Nawrot T.S., Van Hecke E., Thijs L., Richart T., Kuznetsova T., Jin Y., Vangronsveld J., Roels H.A., Staessen J.A.. Cadmium-related mortality and long-term secular trends in the cadmium body burden of an environmentally exposed population. Environ. Health Perspect. (2008.0) 116 1620-1628. DOI: 10.1289/ehp.11667 38. Tellez-Plaza M., Navas-Acien A., Menke A., Crainiceanu C.M., Pastor-Barriuso R., Guallar E.. Cadmium exposure and all-cause and cardiovascular mortality in the U.S. general population. Environ. Health Perspect. (2012.0) 120 1017-1022. DOI: 10.1289/ehp.1104352 39. Tellez-Plaza M., Guallar E., Howard B.V., Umans J.G., Francesconi K.A., Goessler W., Silbergeld E.K., Devereux R.B., Navas-Acien A.. Cadmium exposure and incident cardiovascular disease. Epidemiology (2013.0) 24 421-429. DOI: 10.1097/EDE.0b013e31828b0631 40. Suwazono Y., Nogawa K., Morikawa Y., Nishijo M., Kobayashi E., Kido T., Nakagawa H., Nogawa K.. Impact of urinary cadmium on mortality in the Japanese general population in cadmium non-polluted areas. Int. J. Hyg. Environ. Health (2014.0) 217 807-812. DOI: 10.1016/j.ijheh.2014.05.001 41. Suwazono Y., Nogawa K., Morikawa Y., Nishijo M., Kobayashi E., Kido T., Nakagawa H., Nogawa K.. All-cause mortality increased by environmental cadmium exposure in the Japanese general population in cadmium non-polluted areas. J. Appl. Toxicol. (2015.0) 35 817-823. DOI: 10.1002/jat.3077 42. Li Q., Nishijo M., Nakagawa H., Morikawa Y., Sakurai M., Nakamura K., Kido T., Nogawa K., Dai M.. Relationship between urinary cadmium and mortality in habitants of a cadmium-polluted area: A 22-year follow-up study in Japan. Chin. Med. J. (2011.0) 124 3504-3509. PMID: 22340168 43. Lin Y.S., Caffrey J.L., Lin J.W., Bayliss D., Faramawi M.F., Bateson T.F., Sonawane B.. Increased risk of cancer mortality associated with cadmium exposures in older Americans with low zinc intake. J. Toxicol. Environ. Health A (2013.0) 76 1-15. DOI: 10.1080/15287394.2012.722185 44. Garcia-Esquinas E., Pollan M., Tellez-Plaza M., Francesconi K.A., Goessler W., Guallar E., Umans J.G., Yeh J., Best L.G., Navas-Acien A.. Cadmium exposure and cancer mortality in a prospective cohort: The strong heart study. Environ. Health Perspect. (2014.0) 122 363-370. DOI: 10.1289/ehp.1306587 45. Toxicological Profile for Lead 46. Vaziri N.D.. Mechanisms of lead-induced hypertension and cardiovascular disease. Am. J. Physiol. Heart Circ. Physiol. (2008.0) 295 H454-H465. DOI: 10.1152/ajpheart.00158.2008 47. Navas-Acien A., Guallar E., Silbergeld E.K., Rothenberg S.J.. Lead exposure and cardiovascular disease—A systematic review. Environ. Health Perspect. (2007.0) 115 472-482. DOI: 10.1289/ehp.9785 48. Inorganic and organic lead compounds. IARC Monogr. Eval. Carcinog. Risks Hum. (2006.0) 87 1-471. PMID: 17191367 49. Caffo M., Caruso G., Fata G.L., Barresi V., Visalli M., Venza M., Venza I.. Heavy metals and epigenetic alterations in brain tumors. Curr. Genom. (2014.0) 15 457-463. DOI: 10.2174/138920291506150106151847 50. Silbergeld E.K.. Facilitative mechanisms of lead as a carcinogen. Mutat. Res. (2003.0) 533 121-133. DOI: 10.1016/j.mrfmmm.2003.07.010 51. Knoflach M., Messner B., Shen Y.H., Frotschnig S., Liu G., Pfaller K., Wang X., Matosevic B., Willeit J., Kiechl S.. Non-toxic cadmium concentrations induce vascular inflammation and promote atherosclerosis. Circ. J. (2011.0) 75 2491-2495. DOI: 10.1253/circj.CJ-11-0196 52. Almenara C.C., Broseghini-Filho G.B., Vescovi M.V., Angeli J.K., Faria Tde O., Stefanon I., Vassallo D.V., Padilha A.S.. Chronic cadmium treatment promotes oxidative stress and endothelial damage in isolated rat aorta. PLoS ONE (2013.0) 8. DOI: 10.1371/journal.pone.0068418 53. Angeli J.K., Cruz Pereira C.A., de Oliveira Faria T., Stefanon I., Padilha A.S., Vassallo D.V.. Cadmium exposure induces vascular injury due to endothelial oxidative stress: The role of local angiotensin II and COX-2. Free. Radic. Biol. Med. (2013.0) 65 838-848. DOI: 10.1016/j.freeradbiomed.2013.08.167 54. Caciari T., Sancini A., Fioravanti M., Capozzella A., Casale T., Montuori L., Fiaschetti M., Schifano M.P., Andreozzi G., Nardone N.. Cadmium and hypertension in exposed workers: A meta-analysis. Int. J. Occup. Med. Environ. Health (2013.0) 26 440-456. DOI: 10.2478/s13382-013-0111-5 55. Fagerberg B., Barregard L., Sallsten G., Forsgard N., Ostling G., Persson M., Borne Y., Engstrom G., Hedblad B.. Cadmium exposure and atherosclerotic carotid plaques--results from the Malmo diet and Cancer study. Environ. Res. (2015.0) 136 67-74. DOI: 10.1016/j.envres.2014.11.004 56. Fagerberg B., Bergstrom G., Boren J., Barregard L.. Cadmium exposure is accompanied by increased prevalence and future growth of atherosclerotic plaques in 64-year-old women. J. Intern. Med. (2012.0) 272 601-610. DOI: 10.1111/j.1365-2796.2012.02578.x 57. Julin B., Wolk A., Thomas L.D., Akesson A.. Exposure to cadmium from food and risk of cardiovascular disease in men: A population-based prospective cohort study. Eur. J. Epidemiol. (2013.0) 28 837-840. DOI: 10.1007/s10654-013-9841-8 58. Obeng-Gyasi E.. Chronic cadmium exposure and cardiovascular disease in adults. J. Environ. Sci. Health A (2020.0) 55 726-729. DOI: 10.1080/10934529.2020.1737459 59. Tellez-Plaza M., Jones M.R., Dominguez-Lucas A., Guallar E., Navas-Acien A.. Cadmium exposure and clinical cardiovascular disease: A systematic review. Curr. Atheroscler. Rep. (2013.0) 15 356. DOI: 10.1007/s11883-013-0356-2 60. Lamas G.A., Ujueta F., Navas-Acien A.. Lead and Cadmium as Cardiovascular Risk Factors: The Burden of Proof Has Been Met. J. Am. Heart Assoc. (2021.0) 10 e018692. DOI: 10.1161/JAHA.120.018692 61. Hart B.A., Lee C.H., Shukla G.S., Shukla A., Osier M., Eneman J.D., Chiu J.F.. Characterization of cadmium-induced apoptosis in rat lung epithelial cells: Evidence for the participation of oxidant stress. Toxicology (1999.0) 133 43-58. DOI: 10.1016/S0300-483X(99)00013-X 62. Rani A., Kumar A., Lal A., Pant M.. Cellular mechanisms of cadmium-induced toxicity: A review. Int. J. Environ. Health Res. (2014.0) 24 378-399. DOI: 10.1080/09603123.2013.835032 63. Asara Y., Marchal J.A., Carrasco E., Boulaiz H., Solinas G., Bandiera P., Garcia M.A., Farace C., Montella A., Madeddu R.. Cadmium modifies the cell cycle and apoptotic profiles of human breast cancer cells treated with 5-fluorouracil. Int. J. Mol. Sci. (2013.0) 14 16600-16616. DOI: 10.3390/ijms140816600 64. Pierron F., Baillon L., Sow M., Gotreau S., Gonzalez P.. Effect of low-dose cadmium exposure on DNA methylation in the endangered European eel. Environ. Sci. Technol. (2014.0) 48 797-803. DOI: 10.1021/es4048347 65. Stoica A., Katzenellenbogen B.S., Martin M.B.. Activation of estrogen receptor-alpha by the heavy metal cadmium. Mol. Endocrinol. (2000.0) 14 545-553. DOI: 10.1210/mend.14.4.0441 66. McGivern L., Shulman L., Carney J.K., Shapiro S., Bundock E.. Death Certification Errors and the Effect on Mortality Statistics. Public Health Rep. (2017.0) 132 669-675. DOI: 10.1177/0033354917736514 67. van Woudenbergh G.J., Theofylaktopoulou D., Kuijsten A., Ferreira I., van Greevenbroek M.M., van der Kallen C.J.. Adapted dietary inflammatory index and its association with a summary score for low-grade inflammation and markers of glucose metabolism: The Cohort study on Diabetes and Atherosclerosis Maastricht (CODAM) and the Hoorn study. Am. J. Clin. Nutr. (2013.0) 98 1533-1542. DOI: 10.3945/ajcn.112.056333 68. Shivappa N., Steck S.E., Hurley T.G., Hussey J.R., Hebert J.R.. Designing and developing a literature-derived, population-based dietary inflammatory index. Public Health Nutr. (2014.0) 17 1689-1696. DOI: 10.1017/S1368980013002115 69. Cavicchia P.P., Steck S.E., Hurley T.G., Hussey J.R., Ma Y., Ockene I.S., Hébert J.R.. A new dietary inflammatory index predicts interval changes in serum high-sensitivity C-reactive protein. J. Nutr. (2009.0) 139 2365-2372. DOI: 10.3945/jn.109.114025 70. Willett W.C., Howe G.R., Kushi L.H.. Adjustment for total energy intake in epidemiologic studies. Am. J. Clin. Nutr. (1997.0) 65 9S-1228. DOI: 10.1093/ajcn/65.4.1220S 71. Panizza C.E., Shvetsov Y.B., Harmon B.E., Wilkens L.R., Le Marchand L., Haiman C., Reedy J., Boushey C.J.. Testing the Predictive Validity of the Healthy Eating Index-2015 in the Multiethnic Cohort: Is the Score Associated with a Reduced Risk of All-Cause and Cause-Specific Mortality?. Nutrients (2018.0) 10. DOI: 10.3390/nu10040452 72. Reedy J., Lerman J.L., Krebs-Smith S.M., Kirkpatrick S.I., Pannucci T.E., Wilson M.M., Subar A.F., Kahle L.L., Tooze J.A.. Evaluation of the Healthy Eating Index-2015. J. Acad. Nutr. Diet. (2018.0) 118 1622-1633. DOI: 10.1016/j.jand.2018.05.019
--- title: The Relationship between Mental and Physical Minor Health Complaints and the Intake of Dietary Nutrients authors: - Hiroyo Kagami-Katsuyama - Maremi Sato-Ueshima - Kouji Satoh - Yuko Tousen - Hidemi Takimoto - Mari Maeda-Yamamoto - Jun Nishihira journal: Nutrients year: 2023 pmcid: PMC9966989 doi: 10.3390/nu15040865 license: CC BY 4.0 --- # The Relationship between Mental and Physical Minor Health Complaints and the Intake of Dietary Nutrients ## Abstract Presenteeism is a problem that needs to be solved urgently, both for individual workers and for society overall. In this report, we propose the concept of MHC, which refers to mild mental and physical complaints subjectively perceived by individuals that are not caused by illness. We also planned to examine what kind of physical and mental disorder MHC is and whether food is effective as a method of self-care for MHC. First, we conducted “the comprehensive survey to establish an integrated database of food, gut microbiome, and health information” (the “Sukoyaka Health Survey”) and obtained data on psychosomatic disorders and intakes of dietary nutrients. As a result, through factor analysis and item response theory analysis, we found the following specific examples of MHC: lack of vigor, irritability, fatigue, and somatic complaints. In addition, analysis of the relationship between these four MHC levels and the intake dietary nutrients indicated that they are closely related and that MHC levels can be improved by consuming sufficient amounts of multiple nutrients. ## 1. Introduction Recently, the problems of lowered work productivity (presenteeism) and economic loss have emerged, caused by mental and physical disorders such as sleep debts and excessive stress. Economic losses due to presenteeism in Japan as a whole have been calculated by the WHO-HPQ, a presenteeism scale, to be JPY 19.3 trillion per year, with the average annual loss per worker amounting to several hundred thousand yen [1]. On the other hand, it has become evident that workers with a high level of well-being without psychosomatic disorders have a high level of productivity and creativity [2]. Therefore, improving individual productivity, reducing economic loss, improving mental and physical disorders, and forming a healthy, long-lived, and energetic society is an urgent problem. Given this situation, we first considered that there are two types of disorders that cause presenteeism: psychosomatic disorders caused by illness and minor psychosomatic disorders not caused by illness. Thus, in this report, we refer to psychosomatic disorders that are not related to illness as minor health complaints (MHCs), which are subjectively felt by the individual. It is not hard to imagine that MHCs can result in lower productivity; therefore, it is important to find ways to improve MHCs not only for an individual but for society as a whole. MHCs are not a disease; thus, many people hope to improve their complaints through self-care. Self-care can be conducted in several ways such as diet changes, exercise, bathing, and moderate sun exposure. In particular, it has long been empirically accepted that eating and health are closely related. For example, there are some phrases that indicate or partially indicate this, such as “An apple a day keeps the doctor away” (an English proverb), “Drink morning tea even if you come back from far away”, and “A persimmon turns red, a doctor turns blue” (proverbs in Japanese). People today strongly view food as an important way to maintain good health. The concept of functional foods has taken root in various countries, including Japan, and each country has established systems around functional foods [3]. Thus, according to reports by international research organizations, the functional food market continues to grow consistently [4]. Furthermore, there are many reports on how daily diet relates to health in people with illnesses [5,6] as well as randomized controlled trial studies examining how functional foods relate to health in healthy people with minor mental and physical complaints [7,8]. On the other hand, there have not been many studies on the relationship between daily diet and health among healthy people with mild physical and mental disorders. In this study, we examined whether MHCs can be improved through daily diet and micronutrient intake changes. ## 2.1. Study Design and Population We performed this study using data obtained from “the comprehensive survey to establish an integrated database of food, gut microbiome, and health information” (the “Sukoyaka Health Survey”). The “Sukoyaka Health Survey” was conducted on Japanese men and women aged 20 years or older and younger than 80 years except for “a patient with serious cerebrovascular disease, heart disease, liver disease, kidney disease, gastrointestinal disease, or infection requiring notification”. ( COVID-19 infection was included in the list of infections requiring notification.) In the “Sukoyaka Health Survey”, we conducted two surveys on the same subjects: one in summer and one in winter. The “Sukoyaka Health Survey” was conducted in fiscal 2019 and 2020 as part of the Strategic Innovation Creation Program (SIP) project. In this study, we analyzed data obtained from subjects who gave written informed consent in fiscal years 2019 and 2020 at Hokkaido Information University. ## 2.2. Measurement of the Level of Psychosomatic Disorders The “Brief Job Stress Questionnaire (BJSQ)” is a 29-item questionnaire querying psychological and physical stress reactions. The group prepared this questionnaire through the Stress Measurement Study Group of the “Study Group on the Prevention of Work-related Diseases” of the Ministry of Health, Labour, and Welfare [9]. ## 2.3. Estimation of Dietary Nutrient Intakes (Nutritional Survey) Subjects completed a food record questionnaire (breakfast, lunch, dinner, and snack) on any given day, weighing themselves on a kitchen scale as appropriate. Additionally, we took dietary photographs whenever possible. The supervising dietitian checked the content of the food record questionnaire and the dietary photographs, asked the subjects about insufficient information in calculating nutrition, and calculated the daily nutritional intake from the food record questionnaire. Calculations followed the Japanese Food Standard Component Table 2015 (Seventh Revised Edition) [10]. ## 2.4. Analysis of the Relationship between the MHC Levels and the Intake of Dietary Nutrients We analyzed the relationship between the MHC levels (lack of vigor, irritability, fatigue, and somatic symptoms) and dietary nutrient intake calculated from the Nutritional Survey. The dietary nutrient intakes were compared between relatively high-MHC-level subjects and relatively low-level subjects. Relatively higher and lower MHCs were above and below the median for each complaint (lack of vigor, irritability, fatigue, and somatic symptoms). ## 2.5. Statistical Analysis We performed exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) using R software packages [11,12,13]. In addition, the response properties of stress response measures were calculated using item response theory (IRT) [14,15]. Unidimensional structures of each item were confirmed, and two-parameter logistic (2PL) IRT analysis was performed. As for EFA, CFA, and IRT, the answers to the BJSQ were converted to 0-0-1-1 in this order and in four steps of 1-2-3-4. The exceptions were the answers to Q01 (I have been very active), Q02 (I have been full of energy), and Q03 (I have been lively). We set 1-2-3-4 as 1-1-0-0 and unified the orientation of the point. Thus, the points of the negative answers tended relatively higher in all questions. The response curves were demonstrated by the use of averaged numbers of mass of each item. Except for the EFA, CFA, and IRT, the scoring for each question on the BJSQ followed that previously reported [9]. Mann–Whitney U tests were used for intergroup comparisons. A p-value of <0.05 was considered significant. For analysis, SPSS ver. 25 (IBM Japan, Tokyo, Japan) was used. ## 2.6. Ethics The “Sukoyaka Health Survey” was conducted following the ethical principles based on the Declaration of Helsinki (revised by the World Medical Association Fortareza General Assembly in October 2013) and in compliance with the Ethical Guidelines for Medical Research for Persons (revised by the Ministry of Education, Culture, Sports, Science and Technology and the Ministry of Health, Labour, and Welfare on 28 February 2017). We obtained written informed consent from all subjects. The Bioethics Committee of Hokkaido Information University reviewed and approved the feasibility of clinical trials and the ethical and scientific validity (approval date: 22 April 2019; approval number: 2019-04). ## 3.1. Baseline Characteristics Figure 1 shows a detailed flowchart of the population selected for this study. First, according to the questionnaire’s response, 887 subjects who completed this study were classified as occupational and unemployed. Moreover, we excluded 75 subjects who answered “no occupation or no answer” from all the analyses. In addition, 18 subjects with high psychological stress reactions throughout the year were excluded based on the “Manual for Implementation of Stress-Check System” [9]. Ultimately, we included 794 participants in the analysis as the “working subject population”. Table 1 shows demographic information of the analyzed population (794 persons: 2019 fiscal year: 520 persons; 2020 fiscal year: 274 persons). The gender composition was 3:7 for male-to-female participants. The majority of the age group was from 40 to 60 years old. ## 3.2. Characterization of the Analysis Population Using the BJSQ Figure 2 presents the results of six psychosomatic responses (lack of vigor, irritability, fatigue, anxiety, depressed mood, and somatic symptoms) caused by stress assessed by the BJSQ. The method for determining the degree of psychosomatic disorders followed the “Manual for Implementation of Stress-Check System” of the Ministry of Health, Labour, and Welfare [9]. Consequently, comparisons between summer (gray) and winter (white) showed no inter-seasonal differences in the distributed frequencies of subjects, both in males and females, in all psychosomatic responses. In the subsequent analysis, we describe notable results of the combined analysis of summer and winter data. First, about $40\%$ of the subjects were judged “C: normal” in all six psychosomatic responses. Regarding lack of vigor, approximately $20\%$ of subjects in both sexes were confirmed to have decreased vigor (“A: mild, B: slightly mild”) (Figure 2a). Irritability and fatigue were observed in approximately $15\%$ of subjects, with “D: slightly severe” and “E: severe” in approximately $3\%$ (Figure 2b,c); anxiety and depressed mood were observed in approximately $10\%$ of subjects, with “D: slightly severe” and “E: severe” in approximately $3\%$ (Figure 2d,e). Finally, for somatic symptoms, unlike the other five items, both men and women showed about $80\%$ of the total in both “B: slightly mild, C: normal” (Figure 2f). Additionally, “D: slightly severe, E: severe” was combined, totaling approximately $13\%$. Next, we conducted an EFA on summer data using the 29 items of the BJSQ psychological and physical stress reactions in the “working subject population”. First, we examined the sample validity of Kaiser–Meyer–Olkin to confirm that factorial analyses were feasible with the present data. Given the KMO level of 0.86, we confirmed that the sample size could be satisfactorily subjected to factorial analyses. When we then examined the optimal number of factors, a five-factor solution was concluded to be valid from a screen plot showing the decay situation of eigenvalues. A five-factor solution was also supported in parallel analyses and the MAP criteria. We adopted a five-factor solution (ML1–ML5) from the above findings. Then, the EFA method and oblique rotation method (Oblimin rotation method) were carried out and analyzed using the maximum likelihood approach. Table 2 shows the oblique solution pattern matrix of the factor analysis. Compared with the published report (x), ML1 consisted of the two published scales (anxiety and depressed mood), and two questions were from one published scale (somatic symptoms). ML2 consisted of one published scale (lack of vigor), ML3 consisted of one published scale (irritability), ML4 consisted of one published scale (fatigue), and ML5 consisted of the questions excluding the two questions classified as ML1 from the one published scale (somatic symptoms). Next, we performed a CFA on the winter data. This confirmed the validity of the above five scales: ML1 consisted mainly of items about anxiety and depressed mood; ML2 consisted of items about lack of vigor; ML3 consisted of items about irritability; ML4 consisted of items about fatigue; and ML5 consisted of items about somatic symptoms. As for assessing the goodness-of-fit, the values of the comparative fit index (CFI), Tucker–Lewis index (TLI), and standardized root mean square residual (SRMR) were 0.889, 0.878, and 0.056, respectively; accordingly, we concluded that the assumed five-factor scale was valid. ## 3.3. Evaluation of MHCs Using the BJSQ In addition, we attempted to distinguish relatively mild psychosomatic disorders from less mild psychosomatic disorders. Based on the factor analysis results, we judged there to be uni-dimensionality in all the items and used the five factors derived in this study. Figure 3 shows the average for each summary of each factor, and item–response curves are shown. There were five factors: ML1–ML5. The x-axis reflects the psychosomatic disorder level (θ) as a latent trait. The findings showed that the ML1–ML4 had high slopes and discriminatory power. On the other hand, the slope of the ML5 was very gradual, confirming that it appeared at a broad psychosomatic disorder level. The psychosomatic disorder level shown here is the threshold. This property indicates the projected point to the x-axis at the corresponding response rate of 0.5. Looking at each threshold of the five factors, the lack of vigor appeared in the early stage, with the highest threshold on the left. Next, fatigue, irritability, and finally anxiety and depression mood followed. Accordingly, we defined lack of vigor, irritability, fatigue, and somatic symptoms, which appeared in the stage of the mild psychosomatic disorders as MHC. In subsequent analyses, the degree of each MHC was quantified by reference to Table 2: lack of vigor in Q1–Q3, irritability in Q4–Q6, fatigue in Q7–Q9, and somatic symptoms in Q19–26 and 28. ## 3.4. Relationship between MHC and the Intake of Dietary Nutrients We investigated the relationship between four items of MHC and thirty-nine items pertaining to the intake of dietary nutrients (sodium, potassium, calcium, magnesium, phosphorus, iron, zinc, copper, vitamin A, retinol, beta-cryptoxanthin, beta-carotene equivalents, vitamin D, vitamin E, vitamin K, vitamin B1, vitamin B2, niacin, vitamin B6, vitamin B12, folic acid, pantothenic acid, vitamin C, saturated fatty acids, monounsaturated fatty acids, polyunsaturated fatty acids, cholesterol, total dietary fiber, soluble dietary fiber, insoluble dietary fiber, n-3 fatty acids, n-6 fatty acids, triacylglycerol equivalents, manganese, iodine, selenium, chromium, molybdenum, and biotin). First, as for lack of vigor, irritability, fatigue, and somatic symptoms, we divided the population into a subpopulation below the median and a subpopulation above the median based on MHC level. ( For the median, anxiety was 7 points, irritability was 6 points, fatigue was 6 points, and somatic symptoms was 15 points.) Then, as for lack of vigor, irritability, fatigue, and somatic symptoms, we conducted inter-subpopulation comparisons and confirmed whether significant differences were present or not. Table 3 shows seventeen nutrients with significant differences shared by three or more items of MHCs. We also evaluated the relationship between caloric intake and MHC, but no significant relationship was found. Next, we set the cut-off values for these seventeen nutrients. The cut-offs were set so that all four items of MHC were above “the low subpopulation” in Table 3 and below “the high subpopulation” in Table 3. For each subject, we counted how many of the seventeen nutrients shown in Table 3 were at or above the cut-off. First, we classified subjects according to the number ($\frac{0}{1}$–$\frac{4}{5}$–$\frac{8}{9}$–$\frac{12}{13}$–17) of nutrient intake levels above the cut-offs. Hereafter, we refer to the subgroups as the 0N group, 1–4N group, 5–8N group, 9–12N group, and 13–17N group, respectively. Figure 4 shows the mean MHC levels for each classified group. For every four items of MHC, we examined whether there were significant differences between the 1–4N group, 5–8N group, 9–12N group, and 13–17N groups, using the 0N groups as a control. The findings showed significant differences between the 5–8N, 9–12N, and 13–17N groups in lack of vigor; between the 9–12N and 13–17N groups in irritability; between the 9–12N and 13–17N groups in fatigue; and within the 13–17N group in somatic symptoms. ## 4. Discussion Tosen et al. reported that stress, sleep quality, and comprehensive health questionnaire items are required to assess mild psychosomatic disorder [16]. Using the “Brief Job Stress Questionnaire” (BJSQ), which is also presented in Tosen’s report [16], this study examined minor health complaints (MHCs). The BJSQ can be used to measure and evaluate psychosomatic disorder conveniently; in addition, it is a questionnaire with high validity [9]. First, we characterized the present population using the BJSQ. In Figure 2, the subject frequencies of the six psychosomatic responses are shown as: (a) lack of vigor, (b) irritability, (c) fatigue, (d) anxiety, (e) depressed mood, and (f) somatic symptoms. Approximately $40\%$ of the subjects were judged normal (C) in all six psychosomatic responses, showing a similar tendency to that reported previously [17]. On the other hand, the frequency of subjects with a lack of vigor, irritability, and fatigue was approximately 5–$7\%$ lower than that reported previously, and the frequency of subjects with high anxiety and depressed mood was approximately half of that reported previously [17]. In addition, the frequency of subjects with high somatic symptoms was 6–$9\%$ lower in females and 2–$4\%$ lower in males compared with that previously reported [17]. The distribution frequency of “psychosomatic disorder “ in this working subject population compared with that reported previously [17] shows that the population was slightly less stressed. Therefore, we decided to use factor analyses to investigate whether such populations could be used as controls to capture psychosomatic disorders on a scale similar to that previously reported. Thus, we conducted factor analysis to examine whether it is reasonable to classify psychosomatic disorder into six categories even in the group with relatively mild psychosomatic disorders. The results showed that the classification performed in this report using factor analysis had two minor points that differed from the classification method for psychosomatic disorders used previously [9,17]. First, the anxiety and depressed moods, which had previously been evaluated as separate reaction scores, were separated into a single group, anxiety + depression. In the second point, two questions, Q27 (no appetite) and Q29 (not sleeping well), which were initially grouped into somatic complaints, were integrated into the same group with a new ML1 (anxiety + depressed mood). Therefore, somatic complaints (physical complaints) were reduced to nine questions (ML5) (Table 2). In addition, ML2, ML3, and ML4 could be separated into three: lack of vigor, irritability, and fatigue, as in the previous report [9,17]. Despite these slight differences, we can conclude that the results of the present factor analysis are generally consistent with the scales presented in previous reports [9,17]. In subsequent analyses, however, we adopted ML1–ML5 as a classification of psychosomatic disorders that would be more suited to the present population. Then, IRT was employed to analyze what kind of psychosomatic disorders appeared according to the degree of psychosomatic disorder. Figure 3 shows the item–response curves for each factor scale. To restate, there were five factors used here: ML1, consisting mainly of items about anxiety and depressed mood; ML2, consisting of items about lack of vigor; ML3, consisting of items about irritability; ML4, consisting of items about fatigue; and ML5, consisting of items about somatic symptoms. Examining the respective thresholds, among the five factors, “lack of vigor” exhibited the leftmost threshold and emerged in the early psychosomatic disorder level. Next, “fatigue” and “irritability” appeared, and “anxiety and depressed mood” appears the latest. “ Lack of vigor” also emerged at lower psychosomatic disorder levels, indicating that a “lack of vigor” is already present when other factors reach the threshold. The results suggest that “lack of vigor” is a complaint recognized at a relatively low level of psychosomatic disorder, followed by “irritability” and “fatigue” and finally “anxiety and depression” at the highest level of psychosomatic disorder. The slope for “somatic complaints” was very gradual, indicating that they appeared at a wide range of levels of psychosomatic disorder. These results were consistent with those in previous reports [9]. Previous reports have noted that the most notable symptom to look for when observing more serious stress problems is depression and that appropriately addressing anxiety and depression is extremely important in mental health practice [9]. Therefore, in this study, with the exception of these complaints, we defined MHCs as a lack of vigor, irritability, fatigue, and somatic symptoms that may appear in the relatively mild stage of psychosomatic disorders. The degree of MHC was measured by Q1–Q3, Q4–Q6, Q7–Q9, and Q19–26, 27, and 29, as shown in Table 2, for lack of vigor, irritability, fatigue, and somatic symptoms. The degree of MHC could be measured by Q1–Q3; irritability could be measured by Q4–Q6; fatigue could be measured by Q7–Q9; and somatic symptoms could be measured by Q19–26, 27, and 29, as shown in Table 2. The relationships between individual scores on the four MHC items and nutrient intakes from the diet were examined. The results showed that the 17 nutrients listed in Table 3 were associated with almost all four MHC items. Furthermore, as shown in Figure 4, for each of these 17 components, a relatively high intake of multiple components was associated with improvement in the four MHC items. The following is a brief summary and discussion of the 17 nutrients of note that may be related to MHCs. First, folic acid has been reported to improve depressive symptoms [18]. Although depressive symptoms are not part of MHCs, folic acid may also have effects on MHC with milder levels of psychosomatic disorders. In addition, dietary fiber, especially insoluble fiber, improves the intestinal environment [19], and improvement of the intestinal environment has been reported to lead to improved health [20]. Dietary fiber may also improve defecation [21]; some reports suggest that improved defecation may lead to an improved quality of life [22,23]. It is understood that minerals and vitamins are necessary for maintaining a healthy mind and body; however, it is interesting that vitamin B6, which promotes the biosynthesis of neurotransmitters (dopamine, serotonin, GABA, etc.) [ 24], and carotene, which is reported to relieve mental stress [25], were among the 17 nutrients in this study. Finally, although we have shown that MHC and dietary nutrient intake are closely related, the results of the analyses in this report cannot demonstrate a causal relationship. However, as noted above, some of the nutrients that may be important in improving MHC in this study have already been reported to contribute directly or indirectly to improving physical and mental health problems. Therefore, it is reasonable to assume that nutrient intakes affect MHC levels. Notably, this report suggests that taking relatively high amounts of several nutrients may be more conducive to improving MHC levels. This would imply that daily dietary nutrients are crucial for preventing MHCs in the healthy individuals included in this analysis. However, a limitation of this study is that we did not consider the different nutrient intake requirements of different individuals. Several nutrients in this study did not have appropriate recommended intakes in the general standards of Japan. In addition, a detailed assessment of physical activity and expiratory gas analysis is mandatory to calculate calorie requirements; however, we did not include these two factors in the study. Thus, it is advisable to assess that nutrient intake requirements differ between individuals with different levels of physical activity and body size. Based on these facts, it will be necessary to make an overall judgment about how the amount of nutrients we should take to improve MHC levels by referring to the threshold levels established in this study and the generally recommended intake. ## 5. Conclusions In this report, we proposed the idea of “minor health complaints” (MHCs), which refers to psychosomatic disorders that are not related to illness and are subjectively felt by the individual. Specifically, we treated four physical and mental complaints (decreased vitality, irritability, fatigue, and somatic complaints) as MHCs. Furthermore, we have shown that MHC and dietary nutrient intake are closely related, and taking relatively high amounts of several nutrients may be more conducive to improving MHC levels. ## References 1. Suga M., Araki T., Suzuki T., Miyaki K., Nakayama T.. [Presenteeism-a pilot study towards the integration of medicine and business administration] Presenteeism-igaku to keieigaku no yuugo ni muketa Pilot Study. Res. Work. Competency (2017.0) 5 25-42 2. Proto E.. Are happy workers more productive?. IZA World Labor (2016.0) 315. DOI: 10.15185/izawol.315 3. Domínguez Díaz L., Fernández-Ruiz V., Cámara M.. An international regulatory review of food health-related claims in functional food products labeling. J. Funct. Foods (2020.0) 68 103896. DOI: 10.1016/j.jff.2020.103896 4. Functional Foods Market Size Worth $586.1 Billion by 2030 5. Martín-Peláez S., Fito M., Castaner O.. Mediterranean Diet Effects on Type 2 Diabetes Prevention, Disease Progression, and Related Mechanisms. A Review. Nutrients (2020.0) 12. DOI: 10.3390/nu12082236 6. Knight E., Geetha T., Burnett D., Babu J.R.. The Role of Diet and Dietary Patterns in Parkinson’s Disease. Nutrients (2022.0) 14. DOI: 10.3390/nu14214472 7. Maeda-Yamamoto M., Honmou O., Sasaki M., Haseda A., Kagami-Katsuyama H., Shoji T., Namioka A., Namioka T., Magota H., Oka S.. The Impact of Purple-Flesh Potato (. Nutrients (2022.0) 14. DOI: 10.3390/nu14122446 8. Shimizu K., Fukuda M., Yamamoto H.. Effect of Repeated Intake of Imidazole Dipeptides-containing Drink on Healthy People with Feeling of Fatigue from Daily Activities—The Results of 207 Volunteers Enrolled in the 1st Recruitment. Jpn. Pharmacol. Ther. (2009.0) 37 255-263 9. Kato M.. Final development of the Brief Job Stress Questionnaire mainly used for assessment of the individuals. The Ministry of Labour Sponsored Grant for the Prevention of Work-Related Illness (2000.0) 10. Ministry of Education, Culture, Sports, Science and Technology: Standard Tables of Food Composition in Japan—2015—(Seventh Revised Edition) 11. 11. R Core Team R: A Language and Environment for Statistical ComputingR Core TeamVienna, Austria2021Available online: https://www.R-project.org/(accessed on 6 February 2023). R: A Language and Environment for Statistical Computing (2021.0) 12. Revelle W.. psych: Procedures for Psychological, Psychometric, and Personality Research; Evanston, Illinois. (2021.0) 13. Rosseel Y.. Lavaan: An R package for structural equation modeling. J. Stat. Softw. (2012.0) 48 1-36. DOI: 10.18637/jss.v048.i02 14. Rizopoulos D.. ltm: An R Package for Latent Variable Modeling and Item Response Analysis. J. Stat. Softw. (2006.0) 17 1-25. DOI: 10.18637/jss.v017.i05 15. Partchev I., Maris G.. irtoys: A Collection of Functions Related to Item Response Theory (IRT). (2017.0) 16. Tousen Y., Mizushima R., Koitani N., Kurotani K., Nishihira J., Maeda-Yamamoto M., Takimoto H.. A literature review on the relationship between questionnaires for assessing minor health complaints and health status indices among Japanese people. Jpn. J. Public Health (2022.0) 69 368-382 17. Manual for Implementation of Stress-Check System based on Industrial Safety and Health Act. (2022.0) 18. Coppen A., Bailey J.. Enhancement of the antidepressant action of fluoxetine by folic acid: A randomised, placebo controlled trial. J. Affect. Disord. (2000.0) 60 121-130. DOI: 10.1016/S0165-0327(00)00153-1 19. Zeng H., Lazarova D.L., Bordonaro M.. Mechanisms linking dietary fiber, gut microbiota and colon cancer prevention. World J. Gastrointest Oncol. (2014.0) 6 41-51. DOI: 10.4251/wjgo.v6.i2.41 20. Skonieczna-Żydecka K., Marlicz W., Misera A., Koulaouzidis A., Łoniewski I.. Microbiome—The Missing Link in the Gut-Brain Axis: Focus on Its Role in Gastrointestinal and Mental Health. J. Clin. Med. (2018.0) 7. DOI: 10.3390/jcm7120521 21. Nakaji S., Tokunaga S., Sakamoto J., Todate M., Shimoyama T., Umeda T., Sugawara K.. Relationship between lifestyle factors and defecation in a Japanese population. Eur. J. Nutr. (2002.0) 41 244-248. DOI: 10.1007/s00394-002-0380-4 22. Kinoshita Y., Shoji S., Hayashi T., Okumura H.. A cross-sectional analysis of the health-related quality of life and work productivity in Japanese subjects with self-reported chronic constipation using the National Health and Wellness Survey 2017. J. Jpn. Soc. Gastroenterol. (2020.0) 117 504-513 23. Kyo H., Hasebe H., Ishihara K., Ito M., Uebaba K., Suzuki N.. Effects of the Intake of Fruits Granola on Defecation and Quality of Life. Jpn. J. Complement. Altern. Med. (2017.0) 14 23-26 24. Sato K.. Why is vitamin B6 effective in alleviating the symptoms of autism?. Med. Hypotheses (2018.0) 115 103-106. DOI: 10.1016/j.mehy.2018.04.007 25. Iijima Y., Soeno N., Inomata M., Shioiri T., Saito R., Kimoto K., Tomabechi K., Mita R., Inoue K., Ikegami S.. Influence of β-carotene and Vitamin C on Physiologic and Psychologic Responses to Stress. Jpn. J. Nutr. (1995.0) 53 93-102. DOI: 10.5264/eiyogakuzashi.53.93
--- title: 'Evaluation of the Effectiveness of the Policy of Holding the Second Dose of Vaccination: Lessons from the Outbreak in Ho Chi Minh City' authors: - Vu Thi Thu Trang - Le Van Truong - Truong Van Dat - Randa Elsheikh - Nguyen Tuan Anh - Dang Xuan Thang - Vo Viet Thang - Abdelrahman M. Makram - Nguyen Tien Huy journal: Vaccines year: 2023 pmcid: PMC9967005 doi: 10.3390/vaccines11020293 license: CC BY 4.0 --- # Evaluation of the Effectiveness of the Policy of Holding the Second Dose of Vaccination: Lessons from the Outbreak in Ho Chi Minh City ## Abstract ### Abstract The coronavirus disease 2019 (COVID-19) pandemic has caused a lot of ethical controversy in the equal provision of healthcare, including vaccination. Therefore, our study was designed to assess the impact of Ho Chi Minh City’s policy to hold the second dose of the COVID-19 vaccine. Using a cross-sectional study design to assess low saturation of peripheral oxygen (SPO2) risk based on vaccination status, we included patients who were confirmed to have SARS-CoV-2 and were treated at home. The stepwise method was used to determine participants’ low SPO2 risk-related factors. The average age of the 2836 respondents was 46.43 ± 17.33 (years). Research results have shown that seven factors are related to the low SPO2 status of participants, including age, sneezing, shortness of breath, coughing, and fainting as COVID-19 symptoms, the number of people living with COVID-19, and a history of lung disease. A statistically significant ($$p \leq 0.032$$) finding in this study was that fully vaccinated patients had a $6\%$ lower risk of low SPO2 compared to the first dose less than 21 days group. This result was similar in the vaccine holder group ($p \leq 0.001$). Holding the second dose of the COVID-19 vaccine is associated with a lower SPO2 risk than that of fully vaccinated patients. Therefore, this approach should be considered by governments as it could bring a greater benefit to the community. ## 1. Introduction Coronavirus disease 2019 (COVID-19) is a respiratory disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2). The disease was first identified in December 2019 in Wuhan, China [1]. It then spread rapidly on a global scale and, as of June 2022, there were nearly 600 million confirmed cases and over 6 million deaths [2]. To overcome the global burden of this devastating pandemic, rapid transmission control actions were widely implemented [3] and vaccines were developed with international collaborations [4]. Most studies suggest that vaccines are still effective against circulating variants, and possibly against severe disease and death [5]. However, how long vaccine-induced immunity lasts and how transmissibility has been affected by the vaccine are still unanswered questions [5,6]. The situation in low- and middle-income countries (LMICs) may be somewhat different from their counterparts (middle- and higher-income countries) due to various reasons. These include different incremental cost-effectiveness ratios of various vaccine types in different situations [7], vaccine hesitancy or low acceptance rates [8], the usage of vaccines with lower efficacy [9], and the lower overall purchasing capacity for the vaccines [10]. In Vietnam, an LMIC, the fourth wave of the pandemic lasted for more than eight months with a total of over two million confirmed COVID-19 cases and 35,480 deaths nationwide, of which the largest contribution came from Ho Chi Minh City, with over 20,000 deaths. According to the report of the Vietnam National Steering Committee on COVID-19 prevention and control, in the fourth wave, almost $100\%$ of COVID-19 patients were infected with the Delta variant [11]. Infection control policies such as social distancing and mandatory mask-wearing did not seem to be effective at this stage [12]. An expected solution to the problem is vaccination; however, the Vietnamese government is facing a shortage of vaccine supplies and medical staff. The solution applied by the government was to prioritize vaccines based on people’s risk factors and to administer the vaccines through a center-based policy, where people receive vaccines at larger, more crowded healthcare facilities [12]. The centralized vaccination process, however, has resulted in the exposure of uninfected individuals. Ho Chi Minh City government has applied the measure of vaccination with the first dose widely throughout the city, starting with the elderly, people who have comorbidities, and medical staff [11]. The solution has proved effective when counting the number of people having access to the COVID-19 vaccine; specifically, by the end of September 2021 in District 5, Ho Chi Minh City, $98\%$ of people over 18 years old have received COVID-19 vaccines, of which $31\%$ have received two doses [13]. However, this approach has generated controversy throughout implementation regarding doubts about the evidence for the benefit of a single dose of vaccine [12,14]. Therefore, our study was designed to evaluate the effectiveness of administering only one dose on the saturation of peripheral oxygen (SPO2) index of future COVID-19 patients treated at home. ## 2.1. Study Design, Population, and Conduction The cross-sectional descriptive study which follows the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement (checklist available as Supplementary Table S1) [14] was conducted from 15 August to 15 November 2021, according to decision No. 980/UBND-VP issued on 14 August 2021 by the People’s Committee of District 5, Ho Chi Minh City, Vietnam. The main purpose of the study was to detect the severe condition of COVID-19 patients treated at home through a low SPO2 index (SPO2 ≤ $93\%$) [15], thus making a timely admission decision. All patients with confirmed COVID-19 being treated at home were provided a link by the medical staff to take part in a Zalo® group, a WhatsApp®-comparable software program, after which they were provided with a questionnaire study to assess their initial status. ## 2.2. Questionnaire Design and Survey Conduction The questionnaire consisted of two parts. The first part had ten questions about basic demographics. The second part contained 40 questions about the clinical status and the 23 symptoms of COVID-19 based on the WHO report, SPO2 index, highest and lowest body temperature in the last 24 h, and sneezing, which is a common finding in COVID-19 patients but not reported in the WHO list [15]. Our study divided the characteristics of COVID-19 vaccination into three groups, including the first dose less than 21 days group, which included patients who were infected with COVID-19 within 21 days following the administration of the first dose; the full vaccination group, which included patients with COVID-19 who have received the second dose at or after 21 days; and the delayed second dose group, which included COVID-19 patients who had had their first dose for more than 21 days but had not yet received their second dose. The time of 21 days after vaccination was established based on the evidence of stable immunity formation following the first and second doses of vaccine [16,17,18]. ## 2.3. Data Analysis We underwent a descriptive statistical analysis using T-test student, Chi-square, and ANOVA tests to compare demographic characteristics and clinical characteristics according to vaccination status and classification of SPO2 index (normal SPO2 and low SPO2). Multivariable linear regression analysis was used to evaluate factors related to the low SPO2 of patients vaccinated against COVID-19, using the stepwise AIC method on the MASS package to determine the optimal model. All analyses were performed on R language version 4.1.0. ## 2.4. Ethical Considerations Before the conduction of this study, data were collected from the home care program for COVID-19 patients at the People’s Committee of District 5, Ho Chi Minh City (Decision No. 980/UBND-VP dated 14 August 2021). The benefits (e.g., increased knowledge about the policy impact) and probable burdens (e.g., time burden and some feelings of discomfort) were explained to all patients before the distribution of the survey. It was emphasized that there are no direct benefits to the patients, including financial incentives. It was also emphasized that patients can withdraw at any time by quitting the survey page; however, once the questionnaire has been submitted, there was no way of deleting the collected data as personal data were not collected. To ensure further confidentiality, the IP address tracking was disabled to disallow any attempt at identifying the enrolled participants. All data will be stored for five years after the publication of this manuscript. Before moving on to the questionnaire page, electronic informed consent was obtained from all patients after reading all the details about the project. If the participant ticked “I consent to fill in the questionnaire”, they were redirected to the questionnaire page. Otherwise, a skip-logic function ended the survey. ## 3. Results Our program had a response rate of $38.5\%$ ($\frac{2548}{6616}$ patients), of which $56.3\%$ were female. The average age of participants was 46.43 ± 17.33 (years), and $7.3\%$ ($\frac{186}{2548}$) of the patients were in the full vaccination group. Table 1 shows that the rate of asymptomatic patients in the “first dose less than 21 days”, “holding the second dose”, and “full vaccination” groups tends to increase statistically as the time from injection and number of vaccinations increases, with rates of $25.1\%$, $28.8\%$, and $39.3\%$, respectively. The number of symptoms with which the patients typically presented revealed a similar trend, with the average number of symptoms being 3.27 ± 3.47 (points), 3.12 ± 3.42 (points), and 1.92 ± 2.58 (points), respectively. The study results also showed that the lowest and highest temperature and the lowest SPO2 recorded in the last 24 h of patients who were not previously vaccinated showed statistically significantly higher values than those of the vaccinated group. Furthermore, the number of presenting symptoms in the unvaccinated group was significantly higher than in the rest of the groups, with a positive rate of $\frac{17}{23}$ symptoms. Table 2 shows that the group of normal SPO2 patients has a statistically significantly lower mean age than the low SPO2 group. The average time from vaccination to confirmed COVID-19 in the low SPO2 group was 24.33 ± 12.05 (days), which was statistically significantly lower than that of the normal SPO2 group, with 27.52 ± 13.65 (days). According to the medical record, the low SPO2 group had a statistically significantly higher rate of underlying disease ($63.0\%$) than the normal SPO2 group, including hypertension, cardiovascular disease, diabetes, other lung diseases, dementia, and kidney disease. Figure 1 shows that the total number of symptoms corresponding to patients taking the first dose less than 21 days, holding the second dose, and fully vaccinated reached the highest value on the first day after a confirmed infection with SARS-CoV-2 and tended to decrease and maintain a steady rate over the next ten days. Figure 2 shows that patients in the first dose less than 21 days group had an average SPO2 value of 95.4 ± $4.3\%$, the index of holding the second dose group was 96.8 ± $3.0\%$, and the full vaccination group had an index of 97.4 ± $1.6\%$. The group of patients with a first dose from 21 days to 50 days and a full vaccine from 21 days to 65 days showed a stable SPO2 index above $97\%$. Table 3 shows that a ten-year increase in patient age was associated with a $2\%$ ($95\%$ CI: 1–$3\%$, $p \leq 0.001$) increased risk of low SPO2. Each increase in cohabitation was associated with a $1\%$ ($95\%$ CI: 0–$2\%$, $$p \leq 0.025$$) increase in the risk of low SPO2. Patients with a history of lung diseases were associated with a $35\%$ ($95\%$ CI: 21–$51\%$, $p \leq 0.001$) increased risk of low SPO2. The female gender was associated with a $5\%$ ($95\%$ CI: 2–$7\%$, $$p \leq 0.002$$) reduction in the risk of low SPO2. Positive symptoms including shortness of breath, sneezing, fainting, and cough were significantly associated with an increased risk of low SPO2 of $30\%$, $10\%$, $8\%$, and $5\%$, respectively. Patients infected with SARS-CoV-2, including both the first dose of less than 21 days and full vaccination groups, were statistically significantly associated with a $6\%$ reduction in the risk of low SPO2. The above model recorded a good forecast of the low SPO2 situation, expressed through the area under the curve (AUC) reaching $86.4\%$. Supplementary Table S2 presents the same models along with a comparison between males and females with respect to the other covariates. ## 4. Discussion Our study aimed to determine the factors associated with low SPO2 in Vietnamese post-vaccination COVID-19 patients treated at home to evaluate the effects of delaying the second dose or administering only one dose of vaccination. Through a multivariable logistic regression model, nine factors that affected COVID-19 patients’ SPO2 index were identified. A 10-year increase in age, developing sneezing, shortness of breath, coughing, and fainting as COVID-19 symptoms, co-living with an additional one person above the reported average, and having a history of lung disease were all associated with a higher risk of developing low SPO2. Contrastingly, COVID-19 patients belonging to the group of stable immunity (full vaccination group or holding second dose group) and female gender were found to be linked to a lower risk of developing low SPO2. Furthermore, vaccination was associated with an increased proportion of asymptomatic patients, with $15.3\%$, $25.1\%$, $28.8\%$, and $39.3\%$ asymptomatic rates in patients who did not receive any dose, first dose less than 21 days, holding the second dose, and full vaccination, respectively. In 2021, Moghadas et al. evaluated the optimal time for the administration of the second dose of the Moderna and Pfizer-BioNTech vaccines. They found a significant reduction in infection, hospitalization, and death rates when the second dose was deferred 12–15 weeks from the first dose less than 21 days [20]. Similarly, Silva et al. studied the ideal delay between COVID-19 dose administration and its effect on ICU admission rates. As in our study, they reported that a minimum of a 4-week delay in second dose administration is expected to decrease ICU admission rates, as it gives time to the first dose less than 21 days to achieve a higher efficacy [21]. An Oxford study carried out to assess the reactogenicity and the immunogenicity following the delay in the administration of the ChAdOx1 nCoV-19 included participants with 8–12, 15–25, and 44–45-week intervals between the doses. Antibody levels measured 6 months (median = 3738, IQR = 1824–6625) after the administration of the second dose were found to be significantly higher ($p \leq 0.001$) than those with a 15–25-week interval between the first and second doses (median = 1860, IQR = 917–4992), showing that the delay in the administration of the second dose of AstraZeneca vaccine increases its efficacy [22]. Regarding the factors associated with worse COVID-19 outcomes, a systematic review was conducted to determine the factors associated with higher mortality risk following SARS-CoV-2 infection. Consistent with our study, old age, male gender, and previous lung diseases had a positive correlation with COVID-19 mortality. A prolonged inflammatory response following the weakening of immunity and the excessive release of type 2 cytokines was found to be the reason behind worse outcomes in elderly patients [23]. Moreover, hypoxia in COVID-19 patients is associated with viral lung injury, with alternating regions of hyperventilation and hypoventilation [24]. Therefore, sneezing influences airflow and ventilation pressure change [25], which may be responsible for the more common occurrence of low SPO2 in this group of patients. Our findings suggest that several benefits can be obtained from holding the second dose of vaccination, which are comparable to the benefits seen in the full vaccination group. Achieving the full effectiveness of the first dose in less than 21 days was shown to be associated with higher asymptomatic rates, and less severe outcomes. A longer interval between vaccination and COVID-19 infection was further associated with improved SPO2 rates, which indicates that a second dose delay could contribute to lowering ICU admissions and mortality rates. However, our findings should be cautiously interpreted on a global scale, due to various reasons including the higher use rate of AstraZeneca and the lower BMI when compared to higher-income countries, as well as the statistically significant differences between the two groups, detected in Table 2. Although, as far as we know, this is the only study investigating the delay in the administration of COVID-19 vaccines in Vietnam, better results could have been obtained by investigating individual vaccines rather than unspecified vaccines as was carried out by Payne et al. in the UK [26]. As for the analysis of individual COVID-19 vaccines, a study in Canada by Hall et al. reported that extending the interval between the two doses of the Pfizer-BioNTech vaccine from 3–6 weeks to 8–16 weeks leads to a better antibody response in female healthcare workers [27,28]. The same results were reported earlier for the AstraZeneca vaccine for both the second and the booster dose [22]. However, we were not able to do so due to the limited variability of sample sizes for each vaccine type. Another limitation of this study is the study design, which did not allow for a follow-up. Moreover, we were not able to assess more detrimental clinical features and outcomes, choosing to analyze only the SPO2. Lastly, it is important to consider the findings of this study in the context of wider public health in Vietnam, where nearly a quarter of the population is hesitant to receive the vaccine or offer it to their children [29,30]. Moreover, people with medical or allergic history are more likely to decline vaccination, further contributing to the health inequalities gap related to vaccination [29]. ## 5. Conclusions The study shows that holding the second dose of vaccination against COVID-19 is as effective as obtaining a full vaccination in terms of increasing the asymptomatic rate and reducing the rate of low SPO2. Moreover, the SPO2 rate was different in the unvaccinated patients and the group receiving the first dose of the vaccine less than 21 days following SARS-CoV-2 infection. Our study provides further evidence for policymakers about how vaccine distribution ensures maximum protection for the community in the face of limited vaccine supply. However, the study also suggests the need for policies to limit the process of cross-contamination during vaccination, helping to improve the effectiveness of patient protection and avoid outbreaks related to an infection at the vaccination site. Although it has not been given sufficient attention yet, sneezing was present in $21.4\%$ of our sample and was associated with a $10\%$ increased risk of low SPO2 together with an increased risk of disease transmission. ## References 1. Li Q., Guan X., Wu P., Wang X., Zhou L., Tong Y., Ren R., Leung K.S.M., Lau E.H.Y., Wong J.Y.. Early Transmission Dynamics in Wuhan, China, of Novel Coronavirus–Infected Pneumonia. N. Engl. J. Med. (2020.0) 382 1199-1207. DOI: 10.1056/nejmoa2001316 2. COVID Live-Coronavirus Statistics-Worldometer 3. Rios P., Radhakrishnan A., Williams C., Ramkissoon N., Pham B., Cormack G.V., Grossman M.R., Muller M.P., Straus S.E., Tricco A.C.. Preventing the transmission of COVID-19 and other coronaviruses in older adults aged 60 years and above living in long-term care: A rapid review. Syst. Rev. (2020.0) 9 1-8. DOI: 10.1186/s13643-020-01486-4 4. Ball P.. The lightning-fast quest for COVID vaccines—And what it means for other diseases. Nature (2021.0) 589 16-18. DOI: 10.1038/d41586-020-03626-1 5. Mistry P., Barmania F., Mellet J., Peta K., Strydom A., Viljoen I.M., James W., Gordon S., Pepper M.S.. SARS-CoV-2 Variants, Vaccines, and Host Immunity. Front Immunol. (2021.0) 12 809244. DOI: 10.3389/fimmu.2021.809244 6. Fathizadeh H., Afshar S., Masoudi M.R., Gholizadeh M.P., Asgharzadeh M., Ganbarov K., Köse S., Yousefi M., Kafil H.S.. SARS-CoV-2 (Covid-19) vaccines structure, mechanisms and effectiveness: A review. Int. J. Biol. Macromol. (2021.0) 188 740-750. DOI: 10.1016/j.ijbiomac.2021.08.076 7. Siedner M.J., Alba C., Fitzmaurice K.P., Gilbert R.F., Scott J.A., Shebl F.M., Ciaranello A., Reddy K.P., Freedberg K.A.. Cost-effectiveness of Coronavirus Disease 2019 Vaccination in Low-and Middle-Income Countries. J. Infect. Dis. (2022.0) 226 1887-1896. DOI: 10.1093/infdis/jiac243 8. Solís Arce J.S., Warren S.S., Meriggi N.F., Scacco A., McMurry N., Voors M., Syunyaev G., Malik A.A., Aboutajdine S., Adeojo O.. COVID-19 vaccine acceptance and hesitancy in low- and middle-income countries. Nat. Med. (2021.0) 27 1385-1394. DOI: 10.1038/s41591-021-01454-y 9. Nachman R., Levine R., Jaffe E.. Synthesis of actin by cultured guinea pig megakaryocytes: Complex formation with fibrin. Biochim. Biophys. Acta Gen. Subjects (1978.0) 543 91-105. DOI: 10.1016/0304-4165(78)90457-9 10. Pasquale S., Gregorio G.L., Caterina A., Francesco C., Beatrice P.M., Vincenzo P., Caterina P.M.. COVID-19 in Low- and Middle-Income Countries (LMICs): A Narrative Review from Prevention to Vaccination Strategy. Vaccines (2021.0) 9. DOI: 10.3390/vaccines9121477 11. National Committee for the Prevention and Control of COVID-19 in Vietnam: Results of Two Years of Implementation of COVID-19 Prevention and Control 12. Tough R.. Ho Chi Minh City during the fourth wave of COVID-19 in Vietnam. City Soc. (2021.0) 33. DOI: 10.1111/ciso.12413 13. Ho Chi Minh City Media Center 14. Thoi P.T.. Ho Chi Minh City- the front line against COVID-19 in Vietnam. City Soc. (2020.0) 32. DOI: 10.1111/ciso.12284 15. 15. World Health Organization (WHO) Coronavirus Disease (COVID-19)WHOGeneva, Switzerland2020Available online: https://www.who.int/health-topics/coronavirus(accessed on 20 November 2022). Coronavirus Disease (COVID-19) (2020.0) 16. Stephen B.. Thacker Centers for Disease Control and Prevention Library: COVID-19 Science Update released: 27 October 2020 Edition 60 17. Stephen B.. Thacker Centers for Disease Control and Prevention Library: COVID-19 Science Update released: 30 July 2021 Edition 100 18. Wei J., Stoesser N., Matthews P.C., Ayoubkhani D., Studley R., Bell I., Bell J.I., Newton J.N., Farrar J., Diamond I.. Antibody responses to SARS-CoV-2 vaccines in 45,965 adults from the general population of the United Kingdom. Nat. Microbiol. (2021.0) 6 1140-1149. DOI: 10.1038/s41564-021-00947-3 19. World Health Organization—Western Pacific Region: The Asia-Pacific Perspective: Redefining Obesity and Its Treatment. (2000) 20. Moghadas S.M., Vilches T.N., Zhang K., Nourbakhsh S., Sah P., Fitzpatrick M.C., Galvani A.P.. Evaluation of COVID-19 vaccination strategies with a delayed second dose. PLoS Biol. (2021.0) 19. DOI: 10.1371/journal.pbio.3001211 21. Silva P.J.S., Sagastizábal C., Nonato L.G., Struchiner C.J., Pereira T.. Optimized delay of the second COVID-19 vaccine dose reduces ICU admissions. Proc. Natl. Acad. Sci. USA (2021.0) 118 e2104640118. DOI: 10.1073/pnas.2104640118 22. Flaxman A., Marchevsky N.G., Jenkin D., Aboagye J., Aley P.K., Angus B., Belij-Rammerstorfer S., Bibi S., Bittaye M., Cappuccini F.. Reactogenicity and immunogenicity after a late second dose or a third dose of ChAdOx1 nCoV-19 in the UK: A substudy of two randomised controlled trials (COV001 and COV002). Lancet (2021.0) 398 981-990. DOI: 10.1016/s0140-6736(21)01699-8 23. Parohan M., Yaghoubi S., Seraji A., Javanbakht M.H., Sarraf P., Djalali M.. Risk factors for mortality in patients with Coronavirus disease 2019 (COVID-19) infection: A systematic review and meta-analysis of observational studies. Aging Male (2020.0) 23 1416-1424. DOI: 10.1080/13685538.2020.1774748 24. Serrano R., Corbella X., Rello J.. Management of hypoxemia in SARS-CoV-2 infection: Lessons learned from one year of experience, with a special focus on silent hypoxemia. J. Intensiv. Med. (2021.0) 1 26-30. DOI: 10.1016/j.jointm.2021.02.001 25. Beni H.M., Hassani K., Khorramymehr S.. Study of the sneezing effects on the real human upper airway using fluid–structure interaction method. J. Braz. Soc. Mech. Sci. Eng. (2019.0) 41 181. DOI: 10.1007/s40430-019-1677-z 26. Payne R.P., Longet S., Austin J.A., Skelly D.T., Dejnirattisai W., Adele S., Meardon N., Faustini S., Al-Taei S., Moore S.C.. Immunogenicity of standard and extended dosing intervals of BNT162b2 mRNA vaccine. Cell (2021.0) 184 5699-5714.e11. DOI: 10.1016/j.cell.2021.10.011 27. Hall V.G., Ferreira V.H., Wood H., Ierullo M., Majchrzak-Kita B., Manguiat K., Robinson A., Kulasingam V., Humar A., Kumar D.. Delayed-interval BNT162b2 mRNA COVID-19 vaccination enhances humoral immunity and induces robust T cell responses. Nat. Immunol. (2022.0) 23 380-385. DOI: 10.1038/s41590-021-01126-6 28. Martinez D.R., Ooi E.E.. A potential silver lining of delaying the second dose. Nat. Immunol. (2022.0) 23 349-351. DOI: 10.1038/s41590-022-01143-z 29. Van Nhu H., Tuyet-Hanh T.T., Quang L.N., Linh T.N.Q., Tien T.Q.. COVID-19 Vaccine Acceptance in Vietnam: An Online Cross-Sectional Study. Asia Pac. J. Public Health (2021.0) 34 131-133. DOI: 10.1177/10105395211053732 30. Huynh G., Nguyen H.T.N., Van Tran K., Le An P., Tran T.D.. Determinants of COVID-19 vaccine hesitancy among parents in Ho Chi Minh City, Vietnam. Postgrad. Med. (2022.0) 134 303-308. DOI: 10.1080/00325481.2022.2044142
--- title: Dietary Supplementation of Cedryl Acetate Ameliorates Adiposity and Improves Glucose Homeostasis in High-Fat Diet-Fed Mice authors: - Jingya Guo - Mengjie Li - Yuhan Zhao - Seong-Gook Kang - Kunlun Huang - Tao Tong journal: Nutrients year: 2023 pmcid: PMC9967006 doi: 10.3390/nu15040980 license: CC BY 4.0 --- # Dietary Supplementation of Cedryl Acetate Ameliorates Adiposity and Improves Glucose Homeostasis in High-Fat Diet-Fed Mice ## Abstract Cedryl acetate (CA), also called acetyl cedrene, is approved by the FDA as a flavoring or adjuvant to be added to foods. In this study, we aimed to investigate the preventive benefits of CA on obesity and obesity-related metabolic syndrome caused by a high-fat diet (HFD). Three groups of C57BL/6J mice (ten-week-old) were fed Chow, an HFD, or an HFD with CA supplementation (100 mg/kg) for 19 weeks. We observed that CA supplementation significantly reduced weight gain induced by an HFD, decreased the weight of the visceral fat pads, and prevented adipocyte hypertrophy in mice. Moreover, mice in the CA group showed significant improvements in hepatic lipid accumulation, glucose intolerance, insulin resistance, and gluconeogenesis compared with the mice in the HFD group. Since 16S rRNA analysis revealed that the gut microbiota in the CA and HFD groups were of similar compositions at the phylum and family levels, CA may have limited effects on gut microbiota in HFD-fed mice. The beneficial effects on the metabolic parameters of CA were reflected by CA’s regulation of metabolism-related gene expression in the liver (including Pepck, G6Pase, and Fbp1) and the epididymal white adipose tissues (including PPARγ, C/EBPα, FABP4, FAS, Cytc, PGC-1α, PRDM16, Cidea, and COX4) of the mice. In summary, a potent preventive effect of CA on HFD-induced obesity and related metabolic syndrome was highlighted by our results, and CA could be a promising dietary component for obesity intervention. ## 1. Introduction Obesity refers to one of the multifactor-induced chronic metabolic diseases and is characterized by an increase in adipocytes (size and number) and an abnormal elevation of body fat percentage [1]. In 2020, the World Health Organization (WHO) reported that between 1975 and 2016, the global prevalence of obesity nearly tripled; in excess of 1.9 billion people aged ≥18 were overweight, while more than 650 million adults were obese [2]. Furthermore, obesity is an essential danger for many chronic diseases, such as nonalcoholic steatohepatitis, hypertension, muscle atrophy, and type 2 diabetes mellitus, which place an enormous burden on society [3,4]. Therefore, obesity has become not only an epidemic worldwide but also a major public health concern. During the past few decades, searching for new substances from foods, food ingredients, and natural products that are capable of preventing or treating obesity and related metabolic syndromes has been a popular research issue [5]. The gut microbiota is a complicated microbial community dwelling in the gastrointestinal tract and develops a close symbiotic relationship with its host [6,7]. To date, the gut microbiota has been recognized as being related to various physiological and pathological processes in the human body, which include preventing pathogen colonization, fermenting indigestible food components, synthesizing some vitamins, and contributing to the maturation of the immune system [8]. Notably, mounting evidence supports a connection between the dysbiosis of microbial ecology and metabolic diseases [9,10,11], particularly obesity [12,13]. Bäckhed et al. transplanted the microbiota of normally grown mice to germ-free mice, and the latter had a $60\%$ increase in body fat content with reduced food intake, which provided pioneering evidence linking the gut microbiota with the progression of obesity [14]. In addition, certain natural products that have been reported to prevent or treat obesity are also capable of modifying gut microbiota, such as eugenol [15], Korean red ginseng [16], and pistachio [17]. Cedryl acetate (CA, Figure 1), also called acetyl cedrene, is a tricyclic sesquiterpene that naturally exists in the essential oils of Platycladus orientalis [18] and Schinus molle L. (false pepper) [19]. Since it has stable, long-lasting, and intense woody fragrances, CA is commonly used in producing cosmetics, soaps, perfumes, and other products with fragrances. Aside from the above roles of CA, it is worth noting that it has been approved by the U.S. Food and Drug Administration (FDA) for use as a flavoring or adjuvant in foods [20], and it is permitted as a spice for food according to China’s National Food Safety Standard (GB 29938-2020). At present, CA has been reported to have α-glucosidase inhibitory activity [21] and antifungal [22] activity. Our previous studies have revealed that oral administration of two analogs of CA, i.e., α-cedrene [23] and methyl cedryl ether [24], can prevent or reverse HFD-induced obesity and abnormal metabolic aberrations in rodents. Nevertheless, whether the consumption of CA affects obesity and related metabolic syndromes has not yet been elucidated. Herein, we intended to assess the protective potential of dietary CA supplementation against high-fat diet (HFD)-induced obesity and explore whether the gut microbiota has a pronounced role in modifying the beneficial effects of CA against host adiposity. ## 2.1. Animals, Diets, and Design Twenty-four 7-week-old C57BL/6J male mice were acquired from Vital River (Beijing, China). The animals were accommodated in the Animal Center (SYXK (Jing) 2020-0052 at 22 ± 2 °C, 40–$70\%$ relative humidity with a 12 h light to 12 h dark cycle). Mice were allocated into three groups ($$n = 8$$), i.e., Chow, HFD, and CA groups following three weeks of acclimatization. The standard Chow diet was provided by HFK Bioscience (Beijing, China, Supplementary Table S1), while the HFD (containing $40\%$ fat) and HFD with supplementation of $0.1\%$ CA (w/w) were obtained by following the previous description by Kim et al. [ 25]. The formulations of the HFD and HFD with CA supplementation are presented in Table 1. During the 19-week experiment, all the mice were fed their respective diets accompanied by weekly weighing. Access to food and water was freely available to all the mice. At the end of the experiment, fresh stool samples were collected and stored immediately at −80 °C. Blood samples were obtained and stored in serum at −80 °C. The mice were then sacrificed by cervical dislocation. The perirenal, retroperitoneal, mesenteric, and epididymal white adipose tissue (eWAT) of the mice were taken out, weighed, and stored at −80 °C after immediate freezing in liquid nitrogen. ## 2.2. Dosage Information The dose of CA (analytical reagent grade, Aladdin, Shanghai, China) for the mice in the study was 100 mg/kg of body weight, which is equivalent to 8.11 mg/kg of body weight in humans. The conversion calculations were estimated using the body surface area (BSA) normalization methodology as reported formerly by Reagan-Shaw et al. [ 26]. ## 2.3. Oral Glucose Tolerance Test (OGTT) and Fasting Blood Glucose (FBG) FBG testing was conducted on the mice with morning fasting (6 h, between 8:00 a.m. and 2:00 p.m.). The concentration of blood glucose was tested by a glucometer with blood glucose test strips, which were both obtained from Accu-Chek (Basel, Switzerland). On the following day, fasting was performed as described above; then, the OGTT was performed after gavaging the mice with D-(+)-glucose (2 g/kg body weight concentration, Sigma, St. Louis, MO, USA). The blood glucose concentration was measured before administration (recorded as 0 min), and the blood glucose concentrations were tested at 15, 30, 60, 90, and 120 min after administration; the area under the curve (AUC) was also calculated. ## 2.4. Insulin Tolerance Test (ITT) Mice fasted as described in 2.3; ITT tests were performed after the mice were injected with insulin intraperitoneally. The concentration of insulin, purchased from Novo Nordisk (Copenhagen, Denmark), was 0.75 U/kg of body weight. The blood glucose measurement time was the same as described in 2.3. ## 2.5. Pyruvate Tolerance Test (PTT) The PTT was performed by intraperitoneal injection of sodium pyruvate after 16 h of fasting (7:00 p.m. to 11:00 a.m.) in the mice. The concentration of sodium pyruvate, purchased from Blotopped (Beijing, China), was 1 g/kg of body weight. The blood glucose measurement time was the same as that stated in 2.3. ## 2.6. Serum Biochemical Analysis Biochemical parameter kits were obtained from Bio-Technology and Science (Beijing, China). The concentrations of all biochemical parameters in our study were measured with a Thermo Fisher Indiko analyzer (Waltham, MA, USA). ## 2.7. Histopathological Analysis Tissues from the liver and eWAT were necropsy-dissected and fixed in a $4\%$ paraformaldehyde solution for at least 24 h. After dehydrating with ethanol and embedding the tissue in paraffin, the staining of tissues with a thickness of 4 to 5 μm was performed by hematoxylin and eosin (H&E). Observation of histopathological changes was performed under a Leica DM750 microscope (Nussloch, Germany). The size of the adipocytes was calculated by Image J (version 1.53). ## 2.8. Gut Microbiota Analysis DNA was isolated from the fecal matter produced by the mice. A Thermo Fisher Nano-300 microspectrophotometer (Waltham, MA, USA) was applied to measure the purity of the DNA. PCR was performed to assemble the 16S rRNA gene (V3–V4 region), followed by sequencing in the San Diego Illumina Nova-PE250 (San Diego, CA, USA) at the Novogene Bioinformatics Institute. Sequence analysis was performed by QIIME (version 1.9.1., Flagstaff, AZ, USA), and the same operational classification units (OTUs) were assigned to series with ≥$97\%$ identity. Classification annotation was performed by the Ribosome Database Project (RDP) version 2.2 classifier. α-Diversity was analyzed using PAST3 (Olso, Norway). Non-metric multidimensional scaling (NMDS) and the unweighted pair group method with arithmetic mean (UPGMA) were calculated with the Novogene cloud platform (Beijing, China) on the basis of the weighted UniFrac distance. Following the linear discriminant analysis (LDA), the LDA effect size (LEfSe) analysis was conducted to identify the differential bacterial taxa from the level of the phylum to genus with two filters ($p \leq 0.05$ and LDA score > 4) (http://huttenhower.sph.harvard.edu/galaxy/) (accessed on 11 February 2022). ## 2.9. Real-Time Quantitative PCR (RT-qPCR) RNA was extracted by Blotopped TRIzol reagent (Beijing, China), and then an RNA reverse transcription procedure was performed (TIANGEN, Beijing, China). A Bio-Rad CFX96 real-time PCR system (Hercules, CA, USA) was used to perform RT-qPCR analysis, accompanied by TIANGEN SYBER Green Supermix (Beijing, China). The mRNA expression was normalized using β-actin expression. The primer sequences are listed in Table 2. ## 2.10. Statistical Analysis For all the data, the mean ± SEM was expressed and plotted as a graph with GraphPad Prism (version 9.0). One-way ANOVA was used to assess differences between groups and was considered statistically significant at $p \leq 0.05.$ ## 3.1. CA Has a Significant Preventive Effect against HFD-Induced Body Weight Gain in Mice After 19 weeks of HFD feeding, significantly higher body weight and weight gain were observed in the mice in the HFD group compared with those in the Chow group, demonstrating that 19 weeks of HFD feeding was successful in inducing obesity. On the contrary, CA was significant in preventing HFD-induced obesity, which was evidenced by a decrease in the cumulative body weight gain in the mice (Table 3). In addition, the HFD and CA groups were comparable in food intake (Table 3); thus, the preventative effect of CA on HFD-induced obesity may not be attributed to appetite suppression. ## 3.2. CA Dramatically Decreases Visceral Fat Pad Weight, Attenuates Adipocyte Hypertrophy, and Improves Serum Lipid Profile in Mice Consistent with the preventive effects of CA against body weight gain, the CA group had a significantly lower visceral fat-pad weight than the HFD group. ( Figure 2A). Meanwhile, CA remarkably reduced adipocyte size in the epididymal white adipose tissue (eWAT) compared with that in the HFD group (2035 µm2 vs. 4275 µm2), as visualized by H&E-stained and quantified sections (Figure 2B–D). Serum chemistry parameters related to blood lipid levels were also measured, and CA supplementation produced a significant decrease in serum LDL-C, HDL-C, and TC (Table 4). ## 3.3. CA Improves Glucose Intolerance and Insulin Resistance in Mice Common metabolic diseases associated with obesity also include impaired glucose metabolism; thus, whether CA could improve glucose metabolism was evaluated. In our study, compared with that in the HFD group, FBG decreased significantly in the CA group (Figure 3A). After glucose administration, the glucose levels at 60, 90, and 120 min were significantly lower in the CA group than in the HFD group (Figure 3B). The AUC0–120min was calculated, and it indicated that glucose intolerance was improved with the supplementation of CA, as demonstrated by a significant decrease in the CA group in comparison with the HFD group (Figure 3C). Regarding insulin tolerance, after insulin injection, an evident decrease in plasma glucose concentrations at 15, 30, 60, and 90 min occurred in the CA group compared with those in the HFD group (Figure 3D). The AUC0–120min was also remarkably reduced by CA supplementation, revealing that CA could improve insulin resistance (Figure 3E). In the OGTT and ITT tests, it was noteworthy that the CA group’s AUC levels were comparable to those of the Chow group. ## 3.4. CA Inhibits Hepatic Gluconeogenesis in HFD-Fed Mice PTT was performed, and the HFD group exhibited an increase in gluconeogenesis compared with the Chow group. By contrast, in comparison with the HFD group, CA supplementation significantly inhibited the elevation of blood glucose levels at 30, 60, 90, and 120 min, and the AUC0–120 min was significantly decreased (Figure 4A,B). Collectively, during the PTT test, it was indicated that CA could significantly inhibit gluconeogenesis in HFD-fed mice. To further explore the mechanisms by which CA impacts glucose production, we measured the expression of genes participating in gluconeogenesis, a primary determinant of hepatic glucose production. As assessed by RT-qPCR, of the three key enzymes involved in gluconeogenesis—Pepck, G6Pase, and Fbp1—all were downregulated after CA supplementation, especially Pepck and G6Pase (Figure 4C–E). ## 3.5. CA Protectes against Hepatic Lipid Accumulation in Mice Obesity is commonly companioned by hepatic lipid accumulation, and we further examined whether CA could prevent hepatic lipid accumulation in mice. When compared with mice of the HFD group, liver weight was significantly reduced by CA supplementation (Figure 5A). In addition, CA showed a clear reduction in lipid accumulation and an orderly arrangement of hepatocytes, while mice in the HFD group showed a significant increase in neutral lipid droplets and an irregular arrangement of hepatocytes (Figure 5B). In addition, supplementation with CA significantly decreased the serum levels of ALT and AST, two markers of hepatotoxicity (Table 4). ## 3.6. Limited Effects of CA on the Modifying Gut Microbiota in HFD-Fed Mice By sequencing the 16S rRNA gene, we measured the influence of CA supplementation on the structure and composition of the gut microbiota in mice. The indexes of α-diversity were analyzed at the OTU level and included ACE, Chao1, Simpson, and Shannon, with the former two indexes representing community richness and the latter two indexes representing community diversity. For the above-mentioned indexes, no considerable difference was found in the CA group compared with the HFD group (Figure 6A–D). Regarding β-diversity, the microbial communities of the CA group clustered together with the HFD group, while they were separated in the Chow group (Figure 6E). In addition, we calculated the composition of the gut microbiota in mice. Our results revealed that a similar composition at the phylum and family levels of the gut microbiota was observed in the CA and HFD groups, while the composition was significantly altered in the HFD group compared with the Chow group (Table 5). In addition, cluster analysis according to the UPGMA protocol indicated that no segregation existed between the CA and HFD groups, while clear segregation existed between the Chow and HFD groups (Figure 7A). Regarding the family level, the gut microbiota compositions were similar in the CA and the HFD groups, and the 17 most abundant taxa at the family level showed no significant difference between them (Figure 7B). An LEfSe analysis was also carried out, and only LDA scores over 4 were marked. The results revealed that dramatical taxonomical changes existed in the mice in the HFD group compared with those in the Chow group (Supplementary Figure S1A,B). Nevertheless, CA supplementation had a limited effect on modifying the gut microbiota, as revealed by the LEfSe analysis (Supplementary Figure S2A,B). ## 3.7. CA Alters the Expression of Metabolic Genes in eWAT Considering that the WAT has an essential position in energy metabolism, we measured the mRNA expression of adipogenesis-associated genes in the eWAT in order to further evaluate the mechanism through which CA decreased the accumulation of lipid deposits. The C/EBPα and PPAR-γ are the major regulators of early adipogenesis but are also required to maintain the differentiated state of mature adipocytes. Our results confirmed that CA supplementation induced a marked decrease in C/EBPα and PPAR-γ expression compared with those in the HFD group (Figure 8A,B). We also observed a decrease in the expression of the main lipid synthesis molecules in the CA group, including FABP4 and FAS (Figure 8C,D), which could respond to PPAR-γ and C/EBPα. This was consistent with our observation that CA reduced adipocyte hypertrophy (Figure 2). In addition, the results revealed that CA supplementation could also upregulate genes associated with the thermogenesis of eWAT, including PGC-1α, PRDM16, Cidea, Cytc, and COX4 (Figure 8E–I). ## 4. Discussion In the present research, compared with mice in the HFD group, a decrease of approximately $30\%$ in final body weight was observed in CA-treated mice (100 mg/kg) after 19 weeks; meanwhile, mice in the CA group had comparable body weights to those in the Chow group, indicating that CA has the potential to potently prevent obesity. Importantly, CA’s weight loss effect appeared to be unrelated to toxicity. During the 19-week study, neither death nor clinical signs of adverse treatment-related effects were observed. CA was reported to have a median acute lethal dose of 44,750 mg/kg in rats [27]. According to China’s acute toxicity dose classification standards (GB 15193.3-2014), it has been known to be practically non-toxic. Herein, no hepatotoxicity (Table 4), abnormalities, pathological histological changes, or specific injury manifestations related to the subjects were found on their anatomy. Nonetheless, comprehensive safety studies are needed before the pharmaceutical use of CA. Excessive consumption of an HFD has doubtlessly contributed to the prevalence of obesity [28]. A multitude of studies has recently indicated that modifications in the gut microbiota due to HFDs are related to the obesity epidemic [29,30]. In the gut microbiota, the main dominant phyla are Bacteroides and Firmicutes [31], and an increase in Firmicutes and a decrease in Bacteroidetes (i.e., increased F/B ratio) are generally accompanied by HFD consumption [32], which is similar to our results (Table 5). Nevertheless, contrary to this, a number of studies have observed that this parameter was not altered to any degree and have even found that obese animals have a decreased F/B ratio [33,34,35]. On the one hand, these differences may be due to several factors that could affect the gut microbiota, in terms of the genetic background of the host, age, sex, and time of gut transport [36,37]. On the other hand, a point that is easily ignored but indeed exists is that different methodologies could also affect the results of the gut microbiota composition. Allali et al. reported a difference in the determination of microbial diversity and species richness between sequencing platforms and library preparation protocols [38]. Regarding the changes in the gut microbiota between the CA and the HFD groups, no significant changes in α-diversity or β-diversity were observed, and the gut microbiota compositions were similar in the two groups. In detail, compared with mice fed with HFD, CA-treated mice exhibited no significant change in the relative abundance of the nine most abundant phylum-level taxa, and they did not show significant changes in the 17 most abundant family-level taxa. These results indicated that CA may ameliorate HFD-induced obesity independently of the gut microbiota. Similarly, some dietary components and natural products exert beneficial effects without altering the gut microbiota. Trans-resveratrol reduced both weight gain and serum insulin levels in mice while scarcely modifying the profile of the gut bacteria [39]. Metformin reduces adiposity and/or low inflammation per se to improve metabolic syndrome instead of interacting directly with the gut microbiota [40]. Nevertheless, certain plant components have been reported to have beneficial effects associated with the gut microbial composition in obesity models, such as xanthohumol derivatives [41,42] and epigallocatechin-3-gallate [43]. In summary, our results initially suggest that CA may have a limited effect on modifying the gut microbiota, and the effects of different natural products on the gut microbiota are probably related to the properties of the substances themselves. Excessive accumulation of WAT is a defining characteristic of obesity, and the remodeling process of WAT includes the proliferation of adipocytes (defined as adipogenesis) [44]. The adipogenesis process is critically regulated by various signaling molecules and several key adipose transcription factors, particularly PPARγ and C/EBPα, the activation of which is essential for adipocyte differentiation [45]. In addition, they contribute to encoding lipid synthesis-related molecules, including FABP4 and FAS, which could promote lipogenic binding and lipid storage [46]. Our results showed that the PPARγ, C/EBPα, FABP4, and FAS genes were significantly downregulated in CA-treated mice (Figure 8), suggesting that CA may have the ability to reduce lipid accumulation and further alleviate obesity by restraining adipogenesis and lipid synthesis. In addition, CA enhanced the expression of PGC-1α, Cidea, PRDM16, COX4, and Cytc in the eWAT after CA supplementation (Figure 8). These genes participate in numerous biological functions, and their increased expression plays a role in promoting thermogenesis [47]. Altogether, these data suggest that CA could primarily downregulate genes involved in adipogenesis and lipid synthesis and upregulate genes related to thermogenesis; thus, the effect of CA on lowering lipid levels and alleviating obesity and its related metabolic syndrome was probably associated with the alteration of these metabolism-related genes. Obesity is often followed by hepatic lipid accumulation resulting from an increased supply of lipids to the liver as lipids from adipose tissue spill over into the circulation [48,49]. In the hepatocytes, such an excessive supply contributes to the aggregation of lipid droplets. In addition, obesity leads to systemic insulin resistance, one of the fundamental aspects of the etiology of type 2 diabetes [50]. According to our results, CA supplementation reduced obesity, alleviated hepatic lipid accumulation, and inhibited hepatic gluconeogenesis in mice fed an HFD. Moderate weight loss was reported to reverse the accumulation of hepatic lipids and insulin resistance and normalize hepatic glucose production by reducing gluconeogenesis [51]. Hence, in CA-treated mice, the amelioration in hepatic lipid accumulation and insulin resistance observed may be a secondary event following a reduction in adiposity. There is a need to evaluate the mechanism by which CA improves hepatic lipid accumulation and insulin resistance in the future. ## 5. Conclusions Overall, our research indicated that CA significantly reduced body weight gain, decreased the weight of visceral fat pads, and prevented adipocyte hypertrophy in HFD-fed mice. The HFD-induced hepatic lipid accumulation and impaired glucose metabolism were also ameliorated in mice in the CA group. The above beneficial effects of CA on obesity and obesity-related metabolic syndrome may be independent of the gut microbiota and rather associated with its regulation of metabolism-related gene expression. In a word, CA has the potential to be a prospective dietary component for obesity prevention. ## References 1. Crowley V.E.F., Yeo G.S.H., O’Rahilly S.. Obesity Therapy: Altering the Energy Intake-and-Expenditure Balance Sheet. Nat. Rev. Drug Discov. (2002.0) 1 276-286. DOI: 10.1038/nrd770 2. Obesity and Overweight 3. Goedeke L., Perry R.J., Shulman G.I.. Emerging Pharmacological Targets for the Treatment of Nonalcoholic Fatty Liver Disease, Insulin Resistance, and Type 2 Diabetes. Annu. Rev. Pharmacol. Toxicol. (2019.0) 59 65-87. DOI: 10.1146/annurev-pharmtox-010716-104727 4. Jung U.J., Choi M.S.. Obesity and Its Metabolic Complications: The Role of Adipokines and the Relationship between Obesity, Inflammation, Insulin Resistance, Dyslipidemia and Nonalcoholic Fatty Liver Disease. Int. J. Mol. Sci. (2014.0) 15 6184-6223. DOI: 10.3390/ijms15046184 5. Sirtori C.R., Pavanello C., Calabresi L., Ruscica M.. Nutraceutical Approaches to Metabolic Syndrome. Ann. Med. (2017.0) 49 678-697. DOI: 10.1080/07853890.2017.1366042 6. Zhao Y., Li M., Wang Y., Geng R., Fang J., Liu Q., Kang S.G., Zeng W.C., Huang K., Tong T.. Understanding the Mechanism Underlying the Anti-Diabetic Effect of Dietary Component: A Focus on Gut Microbiota. Crit. Rev. Food Sci. Nutr. (2022.0) 62 1-21. DOI: 10.1080/10408398.2022.2045895 7. Magne F., Gotteland M., Gauthier L., Zazueta A., Pesoa S., Navarrete P., Balamurugan R.. The Firmicutes/Bacteroidetes Ratio: A Relevant Marker of Gut Dysbiosis in Obese Patients?. Nutrients (2020.0) 12. DOI: 10.3390/nu12051474 8. Jandhyala S.M., Talukdar R., Subramanyam C., Vuyyuru H., Sasikala M., Nageshwar Reddy D.. Role of the Normal Gut Microbiota. World J. Gastroenterol. (2015.0) 21 8787-8803. DOI: 10.3748/wjg.v21.i29.8787 9. Zheng S., Wang Y., Fang J., Geng R., Li M., Zhao Y., Kang S.G., Huang K., Tong T.. Oleuropein Ameliorates Advanced Stage of Type 2 Diabetes in Db/Db Mice by Regulating Gut Microbiota. Nutrients (2021.0) 13. DOI: 10.3390/nu13072131 10. Fan Y., Pedersen O.. Gut Microbiota in Human Metabolic Health and Disease. Nat. Rev. Microbiol. (2021.0) 19 55-71. DOI: 10.1038/s41579-020-0433-9 11. Hills R.D., Pontefract B.A., Mishcon H.R., Black C.A., Sutton S.C., Theberge C.R.. Gut Microbiome: Profound Implications for Diet and Disease. Nutrients (2019.0) 11. DOI: 10.3390/nu11071613 12. Abenavoli L., Scarpellini E., Colica C., Boccuto L., Salehi B., Sharifi-Rad J., Aiello V., Romano B., De Lorenzo A., Izzo A.A.. Gut Microbiota and Obesity: A Role for Probiotics. Nutrients (2019.0) 11. DOI: 10.3390/nu11112690 13. Torres-Fuentes C., Schellekens H., Dinan T.G., Cryan J.F.. The Microbiota-Gut-Brain Axis in Obesity. Lancet Gastroenterol. (2017.0) 2 747-756. DOI: 10.1016/S2468-1253(17)30147-4 14. Bäckhed F., Ding H., Wang T., Hooper L.V., Koh G.Y., Nagy A., Semenkovich C.F., Gordon J.I.. The Gut Microbiota as an Environmental Factor That Regulates Fat Storage. Proc. Natl. Acad. Sci. USA (2004.0) 101 15718-15723. DOI: 10.1073/pnas.0407076101 15. Li M., Zhao Y., Wang Y., Geng R., Fang J., Kang S.G., Huang K., Tong T.. Eugenol, a Major Component of Clove Oil, Attenuates Adiposity, and Modulates Gut Microbiota in High-Fat Diet-Fed Mice. Mol. Nutr. Food Res. (2022.0) 66 e2200387. DOI: 10.1002/mnfr.202200387 16. Lee S.Y., Yuk H.G., Ko S.G., Cho S.G., Moon G.S.. Gut Microbiome Prolongs an Inhibitory Effect of Korean Red Ginseng on High-Fat-Diet-Induced Mouse Obesity. Nutrients (2021.0) 13. DOI: 10.3390/nu13030926 17. Terzo S., Mule F., Caldara G.F., Baldassano S., Puleio R., Vitale M., Cassata G., Ferrantelli V., Amato A.. Pistachio Consumption Alleviates Inflammation and Improves Gut Microbiota Composition in Mice Fed a High-Fat Diet. Int. J. Mol. Sci. (2020.0) 21. DOI: 10.3390/ijms21010365 18. Lei H., Wang Y., Liang F., Su W., Feng Y., Guo X., Wang N.. Composition and Variability of Essential Oils of. Biochem. Syst. Ecol. (2010.0) 38 1000-1006. DOI: 10.1016/j.bse.2010.09.018 19. D’Addabbo T., Argentieri M.P., Laquale S., Candido V., Avato P.. Relationship between Chemical Composition and Nematicidal Activity of Different Essential Oils. Plants (2020.0) 9. DOI: 10.3390/plants9111546 20. Cedryl Acetate 21. Sultan S., Choudhary M.I., Khan S.N., Fatima U., Atif M., Ali R.A., Rahman A.U., Fatmi M.Q.. Fungal Transformation of Cedryl Acetate and A-Glucosidase Inhibition Assay, Quantum Mechanical Calculations and Molecular Docking Studies of Its Metabolites. Eur. J. Med. Chem. (2013.0) 62 764-770. DOI: 10.1016/j.ejmech.2013.01.036 22. Wang S., Liu F., Zhou Q., Xu S., Huang S., Luo J.. Preparation of Cedrol and Cedryl Acetate and Their Antifungal Activities. J. For. Res. (2021.0) 6 74-80 23. Tong T., Yu R., Park T.. A-Cedrene Protects Rodents from High-Fat Diet-Induced Adiposity Via Adenylyl Cyclase 3. Int. J. Obes. (2019.0) 43 202-216. DOI: 10.1038/s41366-018-0176-0 24. Li M., Kang S.-G., Huang K., Tong T.. Dietary Supplementation of Methyl Cedryl Ether Ameliorates Adiposity in High-Fat Diet-Fed Mice. Nutrients (2023.0) 15. DOI: 10.3390/nu15030788 25. Kim Y.J., Park T.. Genes Are Differentially Expressed in the Epididymal Fat of Rats Rendered Obese by a High-Fat Diet. Nutr. Res. (2008.0) 28 414-422. DOI: 10.1016/j.nutres.2008.03.015 26. Reagan-Shaw S., Nihal M., Ahmad N.. Dose Translation from Animal to Human Studies Revisited. FASEB J. (2007.0) 22 659-661. DOI: 10.1096/fj.07-9574LSF 27. Jenner P.M., Hagan E.C., Taylor J.M., Cook E.L., Fitzhugh O.G.. Food Flavourings and Compounds of Related Structure I. Acute Oral Toxicity. Food Cosmet. Toxicol. (1964.0) 2 327-343. DOI: 10.1016/S0015-6264(64)80192-9 28. Murphy E.A., Velazquez K.T., Herbert K.M.. Influence of High-Fat Diet on Gut Microbiota: A Driving Force for Chronic Disease Risk. Curr. Opin. Clin. Nutr. Metab. Care (2015.0) 18 515-520. DOI: 10.1097/MCO.0000000000000209 29. Amabebe E., Robert F.O., Agbalalah T., Orubu E.S.F.. Microbial Dysbiosis-Induced Obesity: Role of Gut Microbiota in Homoeostasis of Energy Metabolism. Br. J. Nutr. (2020.0) 123 1127-1137. DOI: 10.1017/S0007114520000380 30. Moossavi S., Bishehsari F.. Microbes: Possible Link between Modern Lifestyle Transition and the Rise of Metabolic Syndrome. Obes. Rev. (2019.0) 20 407-419. DOI: 10.1111/obr.12784 31. Bruce-Keller A.J., Salbaum J.M., Luo M., Blanchard E.t., Taylor C.M., Welsh D.A., Berthoud H.R.. Obese-Type Gut Microbiota Induce Neurobehavioral Changes in the Absence of Obesity. Biol. Psychiatry (2015.0) 77 607-615. DOI: 10.1016/j.biopsych.2014.07.012 32. Stojanov S., Berlec A., Štrukelj B.. The Influence of Probiotics on the Firmicutes/Bacteroidetes Ratio in the Treatment of Obesity and Inflammatory Bowel Disease. Microorganisms (2020.0) 8. DOI: 10.3390/microorganisms8111715 33. Weitkunat K., Stuhlmann C., Postel A., Rumberger S., Fankhanel M., Woting A., Petzke K.J., Gohlke S., Schulz T.J., Blaut M.. Short-Chain Fatty Acids and Inulin, but Not Guar Gum, Prevent Diet-Induced Obesity and Insulin Resistance through Differential Mechanisms in Mice. Sci. Rep. (2017.0) 7 6109. DOI: 10.1038/s41598-017-06447-x 34. Cao W., Chin Y., Chen X., Mi Y., Xue C., Wang Y., Tang Q.. The Role of Gut Microbiota in the Resistance to Obesity in Mice Fed a High Fat Diet. Int. J. Food Sci. Nutr. (2020.0) 71 453-463. DOI: 10.1080/09637486.2019.1686608 35. Tung Y.C., Chang W.T., Li S., Wu J.C., Badmeav V., Ho C.T., Pan M.H.. Citrus Peel Extracts Attenuated Obesity and Modulated Gut Microbiota in Mice with High-Fat Diet-Induced Obesity. Food Funct. (2018.0) 9 3363-3373. DOI: 10.1039/C7FO02066J 36. Li Y., Chen T., Li Y., Tang Y., Huang Z.. Gut Microbiota Are Associated with Sex and Age of Host: Evidence from Semi-Provisioned Rhesus Macaques in Southwest Guangxi, China. Ecol. Evol. (2021.0) 11 8096-8122. DOI: 10.1002/ece3.7643 37. Valeri F., Endres K.. How Biological Sex of the Host Shapes Its Gut Microbiota. Front. Neuroendocrinol. (2021.0) 61 100912. DOI: 10.1016/j.yfrne.2021.100912 38. Allali I., Arnold J.W., Roach J., Cadenas M.B., Butz N., Hassan H.M., Koci M., Ballou A., Mendoza M., Ali R.. A Comparison of Sequencing Platforms and Bioinformatics Pipelines for Compositional Analysis of the Gut Microbiome. BMC Microbiol. (2017.0) 17. DOI: 10.1186/s12866-017-1101-8 39. Etxeberria U., Arias N., Boqué N., Macarulla M.T., Portillo M.P., Martínez J.A., Milagro F.I.. Reshaping Faecal Gut Microbiota Composition by the Intake of Trans-Resveratrol and Quercetin in High-Fat Sucrose Diet-Fed Rats. J. Nutr. Biochem. (2015.0) 26 651-660. DOI: 10.1016/j.jnutbio.2015.01.002 40. Adeshirlarijaney A., Zou J., Tran H.Q., Chassaing B., Gewirtz A.T.. Amelioration of Metabolic Syndrome by Metformin Associates with Reduced Indices of Low-Grade Inflammation Independently of the Gut Microbiota. Am. J. Physiol. Endocrinol. Metab. (2019.0) 317 E1121-E1130. DOI: 10.1152/ajpendo.00245.2019 41. Miranda C.L., Johnson L.A., de Montgolfier O., Elias V.D., Ullrich L.S., Hay J.J., Paraiso I.L., Choi J., Reed R.L., Revel J.S.. Non-Estrogenic Xanthohumol Derivatives Mitigate Insulin Resistance and Cognitive Impairment in High-Fat Diet-Induced Obese Mice. Sci. Rep. (2018.0) 8 613. DOI: 10.1038/s41598-017-18992-6 42. Zhang Y., Bobe G., Revel J.S., Rodrigues R.R., Sharpton T.J., Fantacone M.L., Raslan K., Miranda C.L., Lowry M.B., Blakemore P.R.. Improvements in Metabolic Syndrome by Xanthohumol Derivatives Are Linked to Altered Gut Microbiota and Bile Acid Metabolism. Mol. Nutr. Food Res. (2020.0) 64 e1900789. DOI: 10.1002/mnfr.201900789 43. Ushiroda C., Naito Y., Takagi T., Uchiyama K., Mizushima K., Higashimura Y., Yasukawa Z., Okubo T., Inoue R., Honda A.. Green Tea Polyphenol (Epigallocatechin-3-Gallate) Improves Gut Dysbiosis and Serum Bile Acids Dysregulation in High-Fat Diet-Fed Mice. J. Clin. Biochem. Nutr. (2019.0) 65 34-46. DOI: 10.3164/jcbn.18-116 44. Haider N., Larose L.. Harnessing Adipogenesis to Prevent Obesity. Adipocyte (2019.0) 8 98-104. DOI: 10.1080/21623945.2019.1583037 45. Jakab J., Miskic B., Miksic S., Juranic B., Cosic V., Schwarz D., Vcev A.. Adipogenesis as a Potential Anti-Obesity Target: A Review of Pharmacological Treatment and Natural Products. Diabetes Metab. Syndr. Obes. (2021.0) 14 67-83. DOI: 10.2147/DMSO.S281186 46. Tang J.J., Li J.G., Qi W., Qiu W.W., Li P.S., Li B.L., Song B.L.. Inhibition of Srebp by a Small Molecule, Betulin, Improves Hyperlipidemia and Insulin Resistance and Reduces Atherosclerotic Plaques. Cell Metab. (2011.0) 13 44-56. DOI: 10.1016/j.cmet.2010.12.004 47. Waki H., Park K.W., Mitro N., Pei L., Damoiseaux R., Wilpitz D.C., Reue K., Saez E., Tontonoz P.. The Small Molecule Harmine Is an Antidiabetic Cell-Type-Specific Regulator of Ppargamma Expression. Cell Metab. (2007.0) 5 357-370. DOI: 10.1016/j.cmet.2007.03.010 48. Canfora E.E., Jocken J.W., Blaak E.E.. Short-Chain Fatty Acids in Control of Body Weight and Insulin Sensitivity. Nat. Rev. Endocrinol. (2015.0) 11 577-591. DOI: 10.1038/nrendo.2015.128 49. Tong T., Ryu S.E., Min Y., de March C.A., Bushdid C., Golebiowski J., Moon C., Park T.. Olfactory Receptor 10j5 Responding to A-Cedrene Regulates Hepatic Steatosis Via the Camp–Pka Pathway. Sci. Rep. (2017.0) 7 9471. DOI: 10.1038/s41598-017-10379-x 50. Kahn B.B., Flier J.S.. Obesity and Insulin Resistance. J. Clin. Investig. (2000.0) 106 473-481. DOI: 10.1172/JCI10842 51. Petersen K.F., Dufour S., Befroy D., Lehrke M., Hendler R.E., Shulman G.I.. Reversal of Nonalcoholic Hepatic Steatosis, Hepatic Insulin Resistance, and Hyperglycemia by Moderate Weight Reduction in Patients with Type 2 Diabetes. Diabetes (2005.0) 54 603-608. DOI: 10.2337/diabetes.54.3.603
--- title: Whole-Transcriptome Analysis Highlights Adenylyl Cyclase Toxins-Derived Modulation of NF-κB and ERK1/2 Pathways in Macrophages authors: - Taoran Zhao - Ruihua Li - Mengyin Qian - Meirong Wang - Xiaozheng Zhang - Yuhan Wang - Xinghui Zhao - Jun Xie journal: Toxins year: 2023 pmcid: PMC9967024 doi: 10.3390/toxins15020139 license: CC BY 4.0 --- # Whole-Transcriptome Analysis Highlights Adenylyl Cyclase Toxins-Derived Modulation of NF-κB and ERK1/2 Pathways in Macrophages ## Abstract Edema toxin (ET), one of the main toxic factors of Bacillus anthracis (B. anthracis), is a kind of potent adenylate cyclase (AC). B. anthracis has adapted to resist macrophage microbicidal mechanisms in part by secreting ET. To date, there is limited information on the pathogenic mechanisms used by ET to manipulate macrophage function, especially at the transcriptome level. We used RNA sequencing to study transcriptional changes in RAW264.7 cells treated with ET. We aimed to identify molecular events associated with the establishment of infection and followed changes in cellular proteins. Our results indicate that ET inhibited TNF-α expression in the RAW264.7 mouse macrophage cell line by activating the cAMP/PKA pathway. ET challenge of macrophages induced a differential expression of genes that participate in multiple macrophage effector functions such as cytokine production, cell adhesion, and the inflammatory response. Furthermore, ET influenced the expression of components of the ERK$\frac{1}{2}$, as well as the NF-αB signaling pathways. We also showed that ET treatments inhibit the phosphorylation of the ERK$\frac{1}{2}$ protein. ET also attenuated NF-αB subunit p65 phosphorylation and transcriptional activity of NF-αB via the cAMP/PKA pathway in macrophages. Since the observed modulatory effects were characteristic only of the bacterial exotoxin ET, we propose this may be a mechanism used by B. anthracis to manipulate macrophages and establish systemic infection. ## 1. Introduction Bacillus anthracis (B. anthracis) is the pathogen that causes anthrax, an acute rapidly progressing infectious disease that affects both humans and animals. The virulence factors of B. anthracis primarily include capsule and anthrax toxins [1]. Anthrax toxins have three components: the protective antigen (PA), the lethal factor (LF), and the edema factor (EF) [2]. LF and PA together constitute the lethal toxin (LT), which cleaves mitogen-activated protein kinase kinases (MAPKK) 1–4, 6, and 7 to inactivate the associated pathway. The edema toxin (ET) is composed of EF and PA, and EF is a calmodulin-dependent adenylate cyclase (AC). The entry of EF into host cells is mediated by the interaction between PA and host cell surface receptors; once EF enters the host cell, intracellular cyclic adenosine monophosphate (cAMP) levels increase dramatically [3]. Several studies have indicated that ET helps the early dissemination of B. anthracis within the host by altering the antimicrobial function of macrophages. ET markedly modified the patterns of bacterial dissemination, by leading to apparent direct dissemination to the spleen and by provoking lymphoid cell apoptosis [4]. Macrophages are critical for the early host defense response to B. anthracis. Previous studies have shown that mice in which macrophages were depleted were killed more rapidly by B. anthracis than untreated mice [5,6]. Interestingly, mice lacking the myeloid-specific toxin receptor were completely resistant to B. anthracis infection, while wild-type mice were highly sensitive [7]. However, the mechanisms underlying how ET affects the function of macrophages are not well studied. cAMP, an intracellular second messenger, regulates cellular functions by its interactions with effector molecules, protein kinase A (PKA), or exchange proteins directly activated by cAMP (Epac) [8,9]. In the innate immune system, the elevation of the level of cAMP within phagocytes (including monocytes, macrophages, and neutrophils) could modulate three key effector functions of these cells: generation of inflammatory mediators (e.g., cytokine, chemokine, and lipids), phagocytosis, and intracellular killing of ingested pathogens [10]. In mammalian cells, cAMP can be synthesized by endogenous AC and degraded by phosphodiesterase (PDE) [10]. To date, there are 10 known isoforms of AC [11] and 11 distinct PDE gene families [12]. Moreover, both ACs and PDEs are differentially expressed in various cell types and are localized in different spatial compartments within the cell. As a result, cAMP signaling is under precise spatiotemporal control [13]. As an exogenous AC for the mammalian host, ET interferes with the physiological homeostasis of intracellular cAMP and down-regulates the defense function of macrophages. EF(H351A) is an EF mutant with decreased AC activity characterized by the substitution of histidine (H) at position 351 (H351) by alanine (A) [14]. A previous study suggested that EF(H351A) represents a potential anthrax toxin decoy because it retains PA-binding ability but has significantly weaker activity [15]. However, our study found that ET(H351A) (composed of EF(H351A) and PA) can still slightly increase intracellular cAMP levels and leads to systemic toxicity in a mouse model [16]. Whether ET(H351A) could regulate macrophage function is still unknown. Few sequencing studies have examined transcriptome changes in ET-challenged macrophages [17]. Further, no studies have measured changes in the macrophage transcriptome associated with ET(H351A) treatment. Herein, we used the RAW264.7 cell line, which is a monocyte-derived macrophage cell line, as an in vitro model to characterize global changes in gene expression in macrophages treated with ET or ET(H351A). Whole-transcriptome analysis by RNA-based next-generation sequencing (RNA-seq) shows that challenge by both ET and ET(H351A) alters the macrophage transcriptome by inducing significant changes in the expression of genes involved in various innate immune effector functions. One of the findings of our RNA-seq screen was that ET and ET(H351A) challenge influenced the expression of components in both the extracellular signal-regulated kinases 1 (ERK1) and ERK2 as well as the nuclear factor kappa B (NF-κB) signaling pathways. Further experimental verification showed a reduction in phosphorylation on the effector protein mediated the ET and ET(H351A) inhibition of ERK1 and ERK2 and NF-κB signaling pathways. ## 2.1. ET and ET(H351A) Inhibited TNF-α Expression by Activating cAMP/PKA Pathway TNF-α is a dominant factor in the macrophage response to bacterial pathogens; its level was reported to be restricted by ET [18]. After being treated with 100 ng/mL ET, the intracellular levels of cAMP in RAW264.7 cells increased over 150-fold, while treatment with 100 ng/mL ET(H531A) only induced a 3-fold increase in the intracellular cAMP levels (Figure 1A). TNF-α secretion from macrophages markedly increased upon stimulation by bacteria or pathogen-associated molecular patterns (PAMPs) (i.e., LPS, Figure 1B). This effect could be dramatically reversed by the addition of ET but not ET(H531A) (Figure 1B). Co-incubation of different concentrations of 8-Bromo-cAMP also inhibited the induction of TNF-α secretion by LPS in a dose-dependent manner (Figure 1B). With regard to the effects on transcription, LPS-induced TNF-α mRNA expression was suppressed by ET and 8-Bromo-cAMP, as well as by ET(H351A) (Figure 1C). However, in the absence of LPS, ET(H351A) and 8-Bromo-cAMP showed weaker inhibitory effects on the transcription of TNF-α (Figure 1D). Together, these results indicated that ET inhibits LPS-induced TNF-α expression in macrophages by elevating intracellular cAMP levels. We next investigated whether ET inhibits TNF-α expression in macrophages by activating PKA, a main downstream target of intracellular cAMP [9]. H89 is a widely used PKA inhibitor. In the presence of LPS, both ET-mediated and cAMP-mediated inhibition of TNF-α secretion was reversed by H89 (Figure 1E). Furthermore, we transfected the luciferase reporter gene for the TNF promoter into RAW264.7 cells to determine the effects of ET and H89 on TNF promoter activation. ET alone could inhibit TNF promoter activation, while H89 alone did not have a significant effect (Figure 1F). When H89 was applied to macrophages with ET, ET-induced inhibition of the TNF promoter was reversed (Figure 1F). Thus, activation of PKA is involved in the ET-mediated decrease in TNF-α secretion. ## 2.2. ET and ET(H351A) Induced Global Changes in Gene Expression of Macrophages To assess the changes in gene expression after ET challenge, whole-transcriptome analysis was performed by RNA-based next-generation sequencing (RNA-seq) using RAW264.7 mouse macrophages, which were challenged with PA, ET(H351A), or ET. Figure 2A shows the sample correlation/clustering study of gene expression profiles, which clearly showed a distinct pattern of the total RNA of ET-treated samples vs. ET(H351A)-treated samples. The two samples clustered for each experimental condition, showing that sample variability was not a major contributor to our data set. Interestingly, the transcriptional profile of ET-stimulated macrophages clearly separated from the control group (PA treated), while ET(H351A)-treated samples clustered together with PA-treated samples (Figure 2A). Next, differential expression analysis was carried out between each treatment condition using the limma method, where the standard of p-value cutoff ≤ 0.05 and the log fold change \|log2FC\| ≥ 1 was utilized to compile a list of differentially expressed genes (DEGs) for further analyses. The degree of DEGs was determined by the treatments and was plotted based on whether genes were up- or down-regulated compared to the PA group (Figure 2B). Stimulation of macrophages with ET induced changes in 4094 genes: 2046 ($49.98\%$) and 2048 ($50.02\%$) genes were up- and down-regulated, respectively. It should be noted that ET(H351A) stimulation induced 1107 DEGs, which was fewer than the number of DEGs obtained following ET treatment (Figure 2C). The most significant DEG for each treatment was identified upon inspection of Figure 2D. Of these, Ptchd1 and Scara3 were down-regulated while Thbs1 and Rab44 were up-regulated in both ET- and ET(H351A)-treated samples compared to PA-treated samples. The Ptchd1 gene encodes a protein involved in synaptic transmission, whose deficiency induces a neurodevelopmental disorder [19,20]. The Scara3 gene encodes a macrophage scavenger receptor-like protein and was reported to protect cells from oxidative stress-induced cell damage by removing oxidizing molecules or harmful products of oxidation [21,22]. The Thbs1 gene encodes Thrombospondin-1, associated with platelet activation and wound healing [23]. Rab44 gene levels are commonly decreased in macrophages during differentiation from their precursor cells; however, short-term treatment with IFN and LPS could elevate the level of Rab44 in macrophages [24]. In addition, there are 693 common DEGs among all DEGs induced by ET or ET(H351A) treatment. Hierarchical clustering analysis showed that these common DEGs were classified into three clusters (Figure 2E). Cluster 1 included approximately half of the common DEGs that exhibited up-regulated expression in the ET or ET(H351A) treatment groups relative to the PA treatment groups. However, the DEGs in cluster 1 presented higher levels of increase in the ET treatment groups than those in the ET(H351A) groups. By contrast, the rest of the DEGs in clusters 2 and 3 displayed lower levels of down-regulation in the ET treatment groups than those in the ET(H351A) groups. ## 2.3. ET and ET(H351A) Influenced Macrophage Biological Processes To better understand and classify the biological implications of DEGs induced by ET or ET(H351A) stimulation in macrophages, the enrichment of DEGs in the gene ontology (GO) category of biological process was analyzed using the ClusterProfiler tool. Overall, 307 and 77 significant (adjusted p-value < 0.01) biological processes were identified for ET and ET(H351A) treatment, respectively (Supplemental Tables S1 and S2). To analyze the relationship between the enriched terms, as shown in Supplemental Figures S1 and S2, the most significant GO terms were structured in the form of a directed acyclic graph (DAG) to represent a network of complex correlation of ‘child’ and ‘parent’. The more ‘child’ a GO term is, the more the term is related to a specific biological process. Figure 3A reports that the most ‘child’ terms in the ET(H351A) treatment group included cell chemotaxis, leukocyte cell–cell adhesion, inflammatory response, positive regulation of cytokine production, positive regulation of peptidyl-tyrosine phosphorylation, regulation of cell activation, and response to bacterium and T cell activation. The most ‘child’ terms in the ET treatment group included those in the positive regulation of apoptotic processes, positive regulation of cell adhesion, regulation of the mitogen-activated protein kinase (MAPK) cascade, regulation of ERK1 and ERK2 cascades, and rRNA processing. It is worth noting that the DEGs involving rRNA processing were all down-regulated by ET, while DEGs involving the other four biological processes were consistently up- and down-regulated (Figure 3B). The enrichment of common DEGs between the ET or ET(H351A) treatment groups was also analyzed. Figure 3C reports that the most significantly enriched biological processes were the regulation of the ERK1 and ERK2 cascade, regulation of epithelial cell proliferation, positive regulation of cytokine production, cellular response to biotic stimulus, and cell chemotaxis. Most of these processes are related to the innate immune response of the macrophage. To further analyze the similarities and differences between ET and ET(H351A) treatments in affecting the macrophage biological processes, the count of DEGs induced by each treatment and the corresponding p-value of each biological process above were plotted (Figure 3D). In the ET(H351A) treatment group, regardless of rRNA processing, the remaining 14 biological processes were all significantly enriched (adjusted p-value < 0.01) with 20 to 60 DEGs in each process. However, in the ET treatment group, 12 of 15 biological processes were significantly enriched, with many more DEGs in each process (70 to 160 DEGs). From the enrichment results of biological processes in the GO analysis, DEGs induced by ET or ET(H351A) treatment were significantly enriched in the regulation of the ERK1 and ERK2 cascade (ID: GO 0070372) (Figure 3C,D). A total of 83 and 29 DEGs in the ET and ET(H351A) treatment groups, respectively, were enriched in this cascade and the FPKMs of each DEG are shown in a heat map (Figure 4A, Supplemental Figures S3 and S4). Several cytokine or cytokine-related genes were involved in the regulation of the ERK1 and ERK2 cascade, including CCL2 and IL-6. However, the key elements ERK1 and ERK2 were not affected by ET or ET(H351A) treatment at the mRNA level. ERK$\frac{1}{2}$ can be activated and phosphorylated under LPS stimulation. PD0325901 is a potent ERK$\frac{1}{2}$ phosphorylation inhibitor. The total and phosphorylated ERK$\frac{1}{2}$ levels were evaluated by Western blotting. Densitometry analysis of Western blot bands showed that ET stimulation down-regulated both total and phosphorylated ERK$\frac{1}{2}$ levels, while ET(H351A) treatment alone showed a down-regulation of total ERK$\frac{1}{2}$ (Figure 4B–D). ## 2.4. ET and ET(H351A) Modulated Cytokine Pathways and Signaling Pathways Next, we identified pathways relevant to the challenge with ET or ET(H351A). Using ClusterProfiler, the list of DEGs was mapped onto predefined pathways from the Kyoto Encyclopedia of Genes and Genomes (KEGG) database. We limited our analysis to highly significant pathways with a $p \leq 0.01$, which resulted in 25 and 17 pathways for ET and ET(H351A) treatment, respectively (Table A1). The ET(H351A)/ET-macrophage transcriptome reinforced the pathogenic potential of AC toxins by the number of significant pathways linked to pathogens that subvert immune cells (malaria, Epstein–Barr virus, legionellosis, herpesvirus, African trypanosomiasis). Meanwhile, the type I diabetes mellitus pathway was significantly affected in the ET(H351A) treatment group, and the pathways of cancer, rheumatoid arthritis, and inflammatory bowel disease in both the ET(H351A) and ET treatment groups. Macrophages, and especially the intracellular cAMP levels, continue to be linked to major diseases like those listed above [25], and these results suggest that regulation of intracellular cAMP levels could either play a role in the pathogenesis of these diseases or may represent a potential treatment approach. Moreover, several pathways involved in cell proliferation (ribosome biogenesis in eukaryotes, aminoacyl-tRNA biosynthesis, pyrimidine metabolism, one carbon pool by folate) were just enriched in the ET-treated group but not the ET(H351A)-treated group. Cytokine secretion is an important means for macrophages to inhibit pathogen invasion. In both the ET(H351A) and ET treatment groups, the cytokine–cytokine receptor interaction pathway enriched most DEGs by KEGG analysis (Figure 5A,B). In the RNAseq data, expression of the cytokine genes TNF and CCL2 was significantly down-regulated, and IL-6 was up-regulated in both the ET(H351A) and ET treatment groups (Figure 6A). We then investigated the expression of these cytokines at the protein level in response to different treatments. LPS-stimulated macrophages produced higher levels of TNF-α and CCL2, along with a lower level of IL-6 (Figure 6B). Consistent with RNAseq results, macrophages co-treated with ET and LPS synthesized lower levels of the pro-inflammatory cytokines TNF-α and CCL2, but higher levels of IL-6 (Figure 6B). The effects of ET(H351A) were similar to those of ET but much weaker. ET(H351A) limited LPS-induced TNF-α and CCL2 production but increased IL-6 secretion (Figure 6B). Interestingly, ET treatment also induced a significant increase in IL-10 secretion in macrophages with LPS stimuli (Figure 6B). However, IL-10 gene expression did not show any significant difference in the RNAseq data. Furthermore, the NF-κB signaling pathway was significantly affected in both the ET(H351A) and ET treatment groups (Figure 5A,B), with 13 and 34 DEGs enriched in each group, respectively. The enriched DEGs are shown in a heatmap (Figure 6C, Supplemental Figures S5 and S6) to further clarify the component regulated by ET(H351A) or ET. The NF-κB signaling pathway, and two cytokine genes Ccl4 and TNF were both down-regulated by ET(H351A) and ET treatments, but the cytokine gene IL-1b was up-regulated by ET(H351A) and ET, while two NF-κB signaling module component genes Relb and Iκbκb were up-regulated by ET, but not by ET(H351A). ## 2.5. ET Down-Regulated NF-κB Transcription Activity and p65 Phosphorylation To verify the regulation of the NF-κB pathway in response to ET or ET(H351A) stimulation, the luciferase reporter vector for NF-κB binding sites was transfected into RAW264.7 cells to reflect the binding potential of NF-κB to its target genes. The binding activity of NF-κB was significantly activated by LPS stimulation, while BAY 11-7082 (a specific NF-κB inhibitor) could abolish the activation due to LPS (Figure 7A). Interestingly, this activation effect of LPS could also be partially inhibited by ET or ET(H351A) (relative to the suppression effect of BAY 11-7082) (Figure 7A). Furthermore, in the absence of LPS, ET, but not ET(H351A), also partially suppressed the binding activity of NF-κB (Figure 7B). We next investigated whether ET inhibited NF-κB signaling by activating PKA, a main downstream target of intracellular cAMP. When H89 was applied to macrophages with ET, the ET-induced NF-κB binding inhibition was reversed (Figure 7C). Together, this data indicated that ET inhibits NF-κB signaling by activating the cAMP/PKA pathway. To determine whether the observed inhibition of NF-κB signaling was due to the increase in the levels of the inhibitor of kappa Bα (IκBα), the expression level of IκBα was evaluated by Western blotting. Modest differences in IκBα expression between cells treated with ET or ET(H351A) and untreated cells were identified (Figure 7D). The abundance of p65, the effector subunit of NF-κB, was further measured in RAW264.7 cells. Densitometry analysis of the Western blot bands for total p65 showed that stimulation with ET or ET(H351A) had no effect on p65 abundance (Figure 7D). However, treatment with both ET and BAY heavily impaired the phosphorylation of a key amino acid residue of p65 (S536) (Figure 7E). These data show that ET partially modulates the activation of the NF-κB signaling pathway by selectively interfering with the phosphorylation of p65. ## 3. Discussion Macrophages, which are a dynamic and heterogeneous cell type, are well known as an important component of innate host antibacterial immunity [26]. Macrophages are roughly classified into two groups: classically activated macrophages (M1) and alternatively activated macrophages (M2). The M1 phenotype macrophages produce high levels of pro-inflammatory cytokines, including TNF-α, CCL2, IL-1b, and IL-6, to kill microorganisms and increase the Th1 immune response [26,27]. By contrast, M2 phenotype macrophages are characterized by the low production of the pro-inflammatory cytokines and high production of the anti-inflammatory cytokine IL-10 [26,27]. Several bacterial pathogens have evolved strategies to interfere with macrophage activation and to modulate host responses [28]. For example, Mycobacterium tuberculosis induces the polarization of macrophages toward the M1 phenotype during the early stages of infection [29,30], but polarizes the macrophages to the M2 phenotype at a later stage of infection via the virulence factor early secretory antigenic target ESAT-6 [31,32]. Coxiella burnetii, an obligate intracellular bacterium, survives in macrophages by stimulating an atypical M2 activation program [33]. Interestingly, certain bacteria have evolved to hijack the host cAMP axis by increasing the intracellular cAMP production of the host cell [34]. For instance, the pertussis toxin and CyaA of Bordetella pertussis, and the cholera toxin of Vibrio cholera, have both been reported to inhibit the host defense functions of myeloid phagocytes [34]. In this study, we demonstrated that ET and ET(H531A) treatment increased macrophage intracellular cAMP concentration and reduced macrophage TNF-α expression. Furthermore, TNF-α expression was negatively related to intracellular cAMP in a dose-dependent manner. ET inhibited TNF-α expression through the cAMP/PKA pathway. Furthermore, ET-treated macrophages produced higher anti-inflammatory cytokine IL-10 levels. Although a switch from M1 to M2 macrophage polarization occurs under various physiological and pathological conditions, TNF-α has been identified as a major anti-M2 factor [27]. In fact, many studies have shown an inverse relationship between the degree of TNF-α signaling and the number of M2 macrophages [27]. It is possible that in sepsis, TNF-α production could precede macrophage expansion [35], and monocytes could be induced into a pro-inflammation M1 phenotype by TNF-α stimulation [36]. Convincingly, the complete knockout of TNF led mice to increase the expression of M2-linked genes and M2 macrophage expansion [37]. Therefore, the suppressive effect of ET on TNF-α production might be the key to the induction of the M2 macrophage phenotype. However, IL-6, a cytokine typically associated with M1 polarization, was increased in ET-treated macrophages. This is consistent with previous studies on the relationship between macrophage polarization and cAMP signaling [38] or the Q fever pathogen C. burnetii [34]. In these studies, both stimuli inhibit TNF-α while inducing IL-6 in macrophages. Studies have indicated that IL-6 may inhibit the IFN-γ response during M. tuberculosis infection [39] and reduce the Th1 response in Yersinia enterocolitica-infected mice [40]. IL-6 and TGF-β1 act together to induce IL-10 production in T cells [41]. Thus, it is tempting to speculate that IL-6 may contribute to the immune modulatory role of macrophages. According to the review by Abbas et al., M2 macrophages can be further divided into several subsets [42], and ET or other cAMP agonists may induce macrophages into atypical M2 subsets. In 2006, Comer et al. performed a chip analysis on ET-treated mouse macrophages, and the results showed that ET treatment for 3 h and 6 h changed the expression levels of 71 and 259 genes, respectively [17]. Although these differentially expressed genes have shown that ET has strong and extensive cellular activity, relative to the overall transcriptome, these genes may not reflect the full impact of ET. In our study, ET treatment for 8 h induced 4094 DEGs, reflecting a widespread influence of ET on macrophages. Even ET(H531A), once thought to be a non-toxic mutant, induced 1107 DEGs in 8 h. This indicates that the use of ET(H531A) for the treatment of anthrax infection should be carefully re-considered. cAMP not only has a comprehensive immune-cell regulatory function but also participates in the activity and development of the nervous system [43]. The cAMP/PKA signaling pathway is critical for long-lasting synaptic and memory formation [44]. Disruption of this pathway by certain toxins could result in neurodevelopmental damage [45,46]. Ptchd1 is among the genes most negatively regulated by ET or ET(H351A) treatment. Ptchd1 encodes a transmembrane protein, whose mutation or deficiency is involved in neurodevelopmental disorders [19,20]. Therefore, inhibition of the Ptchd1 signal may be a mechanism of the cAMP/PKA pathway that contributes to neurodevelopmental defects. In eukaryotes, cAMP synthesis is canonically triggered via G protein-coupled receptor (GPCR)-mediated activation of endogenous transmembrane ACs. The functional diversity of cAMP signaling is tightly regulated by intracellular cAMP gradients and microdomains [47]. The destruction of cAMP compartmentation in normal cells increases cell proliferation and induces cell transformation [48]. ET, as an exogenous AC, intensively elevates the cAMP level in macrophages independent of GPCR. The cAMP molecules induced by ET are very likely distributed randomly within the cell. That may be the reason why ET treatment affected several cell proliferation-related pathways. An interesting observation from the RNAseq analysis was that ET and ET(H351A) modulated the MAPK signaling pathway (especially the ERK$\frac{1}{2}$ cascade) and the NF-κB signaling pathway. The innate immune response provides the first line of defense after infection, using a limited number of germline-encoded pattern recognition receptors (PRRs) to recognize the PAMPs of invariant pathogens [49,50]. Macrophages express several classes of PRRs, including Toll-like receptors (TLRs), RIG-I-like receptors (RLRs), NOD-like receptors (NLRs), and C-type lectin receptors (CLRs). Although receptor-proximal signaling mechanisms vary, all of these PRRs activate MAPK and NF-κB pathways, which are crucial for generating immune responses [49,50]. In the present study, the KEGG analysis of DEGs in macrophages induced by ET or ET(H351A) treatment suggested the enrichment of the NF-κB signaling pathway. NF-κB signaling is a master regulator of immunological transcriptional programs, including the inflammatory response to pathogens by innate immune cells [51]. Interactions between cAMP and NF-κB cascades have been described in various cell types, including, among others, diverse leukocyte subsets, fibroblasts, epithelial and endothelial cells, smooth muscle cells, and brain cells. Some studies have reported cell-type-specific effects of cAMP. For instance, cAMP inhibited NF-κB in 3T3 fibroblasts, whereas it induced NF-κB in brown adipocytes [52]. The NF-κB signaling module consists of five NF-κB monomers (RelA/p65, RelB, cRel, NF-κB 1 p50, and NF-κB p52), which can act as homo- or heterodimers, and five inhibitory proteins (IκB α, β, ε, γ, and δ) that make up the IκB protein family. The inactivated NF-κB proteins are sequestered in the cytoplasm through physical interaction with IκB proteins. Upon bacterial infection, the PAMPs of pathogens activate NF-κB signaling via the activation of the inhibitor kappa B kinase (IKK) trimeric complex. Once activated by phosphorylation, IKK further phosphorylates the IκB, which leads to the degradation of IκB and the release of NF-κB from the NF-κB/IκB complex, which allows NF-κB to activate the transcriptional activity of its target genes [52]. The p50/p65 combination is the most abundant and ubiquitously expressed NF-κB heterodimer. In the present study, ET treatment decreased the phosphorylation of the p65 subunit of the NF-κB transcription factor. This impairment may have led to the failure of this heterodimer to enter the nucleus. Several sites in the human and murine TNF promoters are designated as κB motifs, and these motifs are involved in the NF-κB -mediated regulation of TNF-α transcription [53]. Thus, ET interfered with the binding of NF-κB to its target genes (including TNF-α). There are two main types of intracellular cAMP transducers: cAMP-dependent PKA and the guanine exchange proteins that are directly activated by cAMP (EPAC-1 and EPAC-2). Of these, PKA is considered the main effector of cAMP in interacting with NF-κB [52]. The specific PKA inhibitor H89 reversed the ET-induced inhibition of NF-κB, suggesting the involvement of the cAMP/PKA pathway. The S536 residue of p65 can be phosphorylated by several kinases, including IKKα, IKKβ, Akt, TANK-binding kinase1 (TBK1), IKKε, and so on [54]. Among these kinases, the activity of non-canonical IκB kinases TBK1 and IKKε has been reported to be inhibited by cAMP increasement and PKA activation [55]. This may be the possible mechanism by which ET inhibits NF-κB transcription activity. MAKP includes four subsets: ERK1 and ERK2 (P44MAPK and P42MAPK, respectively); stress-activated protein kinases (SAPKs/JNKs); p38 kinase; and ERK5. The ERK1 and ERK2 pathways have been shown to have important roles in macrophages, regulating cytokine production via both transcriptional and post-transcriptional mechanisms. This study has shown that ET and ET(H351A) diminished the phosphorylation of ERK1 and ERK 2 as well as down-regulated the total protein levels of ERK1 and ERK 2. Activation of ERK1 and ERK2 signaling by all TLRs in primary macrophages is mediated by the MAP3K TPL2 [56]. In unstimulated cells, TPL2 forms a complex with the NF-κB subunit precursor protein p105, which inhibits the kinase activity of TPL2 [57,58]. TLR stimulation activates the IKK complex, which phosphorylates p105, inducing its K48-linked ubiquitylation and proteasome-mediated proteolysis [28]. After its release from p105-mediated inhibition, TPL2 can then phosphorylate MAPK kinase 1 (MKK1) and MKK2 upstream of ERK1 and ERK2. IKK2 also directly phosphorylates TPL2 at Ser400, which is a crucial regulatory residue in its carboxyl terminus that is required for LPS to induce ERK activation in macrophages [59,60]. This crosstalk between the ERK1 and ERK2 pathways and the NF-κB pathway may explain why these signaling pathways are always regulated simultaneously in macrophages. ## 4. Conclusions In this study, we show that both ET(H351A) and ET induce significant changes in the macrophage transcriptome. In silico analysis demonstrated that the biological processes involved were the regulation of the ERK1 and ERK2 cascade, regulation of epithelial cell proliferation, positive regulation of cytokine production, cellular response to biotic stimulus, and cell chemotaxis. Moreover, ET(H351A) and ET modulated both the cytokine-related pathways and NF-κB signaling pathways. Further experimental verification suggested that the inhibition of ERK1 and ERK2 of phosphorylation, as well as p65, may be the main targets of ET(H351A)- and ET-mediated modulation of the ERK1 and ERK2 as well as NF-κB signaling pathways. Our study provides novel insight into how the AC toxin helps pathogens evade host defense mechanisms, and may serve as a framework for further studies of B. anthracis infection prevention and treatment. ## 5.1. Toxins The PA, EF, and EF(H351A) proteins used in this investigation were expressed in Escherichia coli and were purified as previously described [16,61]. EF(H351A) is a variant of EF with a mutation of histidine (H) 351 into alanine (A) that leads to a mostly diminished but not eliminated AC activity [16]. The treatment of 100 ng/mL ET in this study suggests the combination of 100 ng/mL EF with 200 ng/mL PA, just as 100 ng/mL ET(H351A) indicates the co-treatment of 100 ng/mL EF(H351A) with 200 ng/mL PA. ## 5.2. Cell Culture The monocyte-derived mouse macrophage cell line RAW264.7 was obtained from the American Type Culture Collection and cultured in minimum essential medium (MEM) supplemented with $10\%$ fetal bovine serum, penicillin (100 U/mL), streptomycin (100 μg/mL), and glutamine (2 mM) at 37 °C under $5\%$ CO2. Before stimulation, RAW264.7 cells were seeded in 6-well plates at a density of 4 × 105 cells/well and cultured overnight. ## 5.3. Intracellular cAMP Measurement RAW264.7 cells pretreated with 200 ng/mL PA, 100 ng/ mL ET(H351A), or 100 ng/mL ET for 8 h were lysed using 0.1 M HCI. Total intracellular cAMP levels were assayed using the Monoclonal Anti-cAMP Antibody Based Direct cAMP ELISA Kit (Neweast Bioscience, Wuhan, China), following the manufacturer’s instructions. ## 5.4. Cytokine Production Supernatants were collected from all RAW264.7 cells in all pretreatment groups (10 ng/mL LPS (Sigma-Aldrich, Steinheim, Germany) plus 200 ng/mL PA, 100 ng/mL ET(H351A), or 100 ng/mL ET) and the control group. The levels of tumor necrosis factor-α (ΤNF-α), C-C motif chemokine 2 (CCL2), Interleukin-6 (IL-6), and IL-10 in the supernatants were determined using the cytometric bead array (CBA) mouse inflammation kits (BD Biosciences, San Jose, USA), following the manufacturer’s instructions. ## 5.5. Gene Expression Analysis with Quantitative Reverse Transcriptase-PCR Total RNA was extracted using RNeasy Plus Mini Kits (Qiagen, Duesseldorf, Germany) and reversed transcribed to cDNA using QuantiTect Reverse Transcription Kits (Qiagen, Duesseldorf, Germany), following the manufacturer’s instructions. A LightCycler (ABI Prism 7000) and an SYBR RT-PCR kit (Takara, Tokyo, Japan) were used for quantitative reverse transcriptase-PCR (qRT-PCR) analysis. TNF was amplified using the specific primer pair 5′-GGTCTGGGCCATAGAACTGA-3′ and 5′-CAGCCTCTTCTCATTCCTGC-3′, while β-actin was amplified using the specific primer pair 5′-ATGGAGGGGAATACAGCCC-3′ and 5′-TTCTTTGCAGCTCCTTCGTT-3′. The expression of TNF in each sample was normalized to β-actin expression. ## 5.6. Luciferase Reporter Gene Expression Assay RAW264.7 cells (2 × 105 cells/well) were seeded in 24-well plates. After 12 h, cells were co-transfected with 0.25 μg of Renilla-expressing plasmids (pRL-SV40-C; Beyotime Biotechnology, Shanghai, China) and either 1 μg of NF-κB binding site reporter plasmids (pNF-κB-TA-Luc; Beyotime Biotechnology, Shanghai, China) or 1 μg of TNF-α promoter reporter plasmids (pTNF-α-promoter-Luc; Beyotime Biotechnology, Shanghai, China) using TurboFect (Invitrogen, Carlsbad, USA). At 4 h post-transfection, cells were treated with PBS, 100 ng/mL ET(H351A), 100 ng/mL ET, 10 μM BAY 11-7082 (BAY), and 40 μM H89, in the presence or absence of 10 ng/mL LPS or 10 ng/mL TNF-α. Luciferase activity levels were determined using the Dual-Luciferase reporter assay system (Promega, Madison, WI, USA) using a microplate luminometer (GLOMAX96; Promega, Madison, WI, USA), following the manufacturer’s instructions. Firefly luciferase activity was normalized against Renilla luciferase activity. ## 5.7. RNA Isolation, Sequencing, and Analysis RAW264.7 cells were seeded in 10-cm dishes (107 cells/dish) and cultured in the incubator overnight. The next morning, cells were stimulated with PA 200 ng/mL, ET(H351A) 100 ng/mL, or ET 100 ng/mL for 8 h. After stimulation, the culture medium in each plate was discarded and the cells were harvested using Trizol (Life Technologies, Carlsbad, CA, USA) and stored at −80 °C until RNA extraction. Cells suspended in Trizol were transported on dry ice to Zhongkejingyun Bio-Information Technology Co., Ltd., (Beijing, China), where total RNA was isolated and RNA quality control was conducted. After passing the quality inspection, the rRNA was removed by hybridization capture based on the structure and sequence characteristics of the rRNA, and the remaining RNA samples were used for reverse transcription and library construction. The Illumina HiSeqTM4000/MisseqTM/X-Ten high-throughput sequencing platform was used to sequence the cDNA library and the raw sequencing data were analyzed with FastQC using Cutadapt to remove joints and Trimmomatic to remove low-quality bases and reads at both ends. The clean data were aligned to the Mus musculus reference genome assembly (GRC39.fa) using Hisat2, generating alignment files in BAM format. The number of fragments that overlap each Entrez gene was summarized using featureCounts, differential expression analysis between each challenge (ET(H351A) 100 ng/mL or ET 100 ng/mL), and the control condition (PA 200 ng/mL) was performed using the limma software package. A q-value cutoff ≤ 0.05 with an absolute \|log2FC\| ≥ 1 was used to determine differential expression. ## 5.8. Western Blotting The stimulated RAW264.7 cells were suspended in radioimmunoprecipitation assay (RIPA) buffer containing protease inhibitors (protease inhibitor cocktail tablets; Roche Applied Sciences, Mannheim, Germany) and phosphatase inhibitors (TransGen Biotech, Beijing, China) for 20 min. The proteins in the cell lysates were separated by electrophoresis, and the separated proteins were transferred to PVDF membranes (Millipore, MI, USA). After an overnight incubation in Tris-buffered saline supplemented with $0.2\%$ Tween 20 (TBST) and $5\%$ nonfat dry milk, membranes were incubated with antibodies against IκBα (1:500; Santa Cruz Biotechnology), p65 (1:500; Santa Cruz Biotechnology), p65S-phosphor-S536 (1:500, Santa Cruz Biotechnology), p-ERK$\frac{1}{2}$ (1:500, Cell Signaling Technology), ERK$\frac{1}{2}$ (1:1000, Cell Signaling Technology), or β-actin (1:1000, Abcam) for 2 h at room temperature. The membranes were washed with TBST and then incubated with the peroxidase-conjugated secondary antibodies (1:5000; Abcam) at room temperature for 1 h. ImmobilonTM Western Chemiluminescent HRP Substrate (Millipore) and the Western blotting imager (Clinx Scinence Instruments Co., Shanghai, China) were used to determine the protein expression. ## 5.9. Statistical Analysis GraphPad Prism 7.0 was used for statistical analyses. Significant differences between the means of the experimental and control groups were identified with Student’s t-test or with one-way ANOVA analysis. We considered $p \leq 0.05$ as statistically significant. ## References 1. Moayeri M., Leppla S.H., Vrentas C., Pomerantsev A.P., Liu S.. Anthrax Pathogenesis. Annu. Rev. Microbiol. (2015) 69 185-208. DOI: 10.1146/annurev-micro-091014-104523 2. Liu S., Moayeri M., Leppla S.H.. Anthrax lethal and edema toxins in anthrax pathogenesis. Trends Microbiol. (2014) 22 317-325. DOI: 10.1016/j.tim.2014.02.012 3. Drum C.L., Yan S.Z., Bard J., Shen Y.Q., Lu D., Soelaiman S., Grabarek Z., Bohm A., Tang W.J.. Structural basis for the activation of anthrax adenylyl cyclase exotoxin by calmodulin. Nature (2002) 415 396-402. DOI: 10.1038/415396a 4. Dumetz F., Jouvion G., Khun H., Glomski I.J., Corre J.P., Rougeaux C., Tang W.J., Mock M., Huerre M., Goossens P.L.. Noninvasive imaging technologies reveal edema toxin as a key virulence factor in anthrax. Am. J. Pathol. (2011) 178 2523-2535. DOI: 10.1016/j.ajpath.2011.02.027 5. Cote C.K., Rea K.M., Norris S.L., van Rooijen N., Welkos S.L.. The use of a model of in vivo macrophage depletion to study the role of macrophages during infection with Bacillus anthracis spores. Microb. Pathog. (2004) 37 169-175. DOI: 10.1016/j.micpath.2004.06.013 6. Cote C.K., Van Rooijen N., Welkos S.L.. Roles of macrophages and neutrophils in the early host response to Bacillus anthracis spores in a mouse model of infection. Infect. Immun. (2006) 74 469-480. DOI: 10.1128/IAI.74.1.469-480.2006 7. Liu S., Miller-Randolph S., Crown D., Moayeri M., Sastalla I., Okugawa S., Leppla S.H.. Anthrax toxin targeting of myeloid cells through the CMG2 receptor is essential for establishment of Bacillus anthracis infections in mice. Cell Host Microbe (2010) 8 455-462. DOI: 10.1016/j.chom.2010.10.004 8. Chin K.V., Yang W.L., Ravatn R., Kita T., Reitman E., Vettori D., Cvijic M.E., Shin M., Iacono L.. Reinventing the wheel of cyclic AMP: Novel mechanisms of cAMP signaling. Ann. N.Y. Acad. Sci. (2002) 968 49-64. DOI: 10.1111/j.1749-6632.2002.tb04326.x 9. Kopperud R., Krakstad C., Selheim F., Døskeland S.O.. cAMP effector mechanisms. Novel twists for an ‘old’ signaling system. FEBS Lett. (2003) 546 121-126. DOI: 10.1016/S0014-5793(03)00563-5 10. Serezani C.H., Ballinger M.N., Aronoff D.M., Peters-Golden M.. Cyclic AMP: Master regulator of innate immune cell function. Am. J. Respir. Cell Mol. Biol. (2008) 39 127-132. DOI: 10.1165/rcmb.2008-0091TR 11. Sunahara R.K., Dessauer C.W., Gilman A.G.. Complexity and diversity of mammalian adenylyl cyclases. Annu. Rev. Pharmacol. Toxicol. (1996) 36 461-480. DOI: 10.1146/annurev.pa.36.040196.002333 12. Lin C.S., Xin Z.C., Lin G., Lue T.F.. Phosphodiesterases as therapeutic targets. Urology (2003) 61 685-691. DOI: 10.1016/S0090-4295(02)02439-1 13. Bock A., Annibale P., Konrad C., Hannawacker A., Anton S.E., Maiellaro I., Zabel U., Sivaramakrishnan S., Falcke M., Lohse M.J.. Optical Mapping of cAMP Signaling at the Nanometer Scale. Cell (2020) 182 1519-1530. DOI: 10.1016/j.cell.2020.07.035 14. Shen Y., Zhukovskaya N.L., Guo Q., Florián J., Tang W.J.. Calcium-independent calmodulin binding and two-metal-ion catalytic mechanism of anthrax edema factor. EMBO J. (2005) 24 929-941. DOI: 10.1038/sj.emboj.7600574 15. Gupta M., Alam S., Bhatnagar R.. Catalytically inactive anthrax toxin(s) are potential prophylactic agents. Vaccine (2007) 25 8410-8419. DOI: 10.1016/j.vaccine.2007.09.063 16. Zhao T., Zhao X., Liu J., Meng Y., Feng Y., Fang T., Zhang J., Yang X., Li J., Xu J.. Diminished but Not Abolished Effect of Two His351 Mutants of Anthrax Edema Factor in a Murine Model. Toxins (2016) 8. DOI: 10.3390/toxins8020035 17. Comer J.E., Galindo C.L., Zhang F., Wenglikowski A.M., Bush K.L., Garner H.R., Peterson J.W., Chopra A.K.. Murine macrophage transcriptional and functional responses to Bacillus anthracis edema toxin. Microb. Pathog. (2006) 41 96-110. DOI: 10.1016/j.micpath.2006.05.001 18. Hoover D.L., Friedlander A.M., Rogers L.C., Yoon I.K., Warren R.L., Cross A.S.. Anthrax edema toxin differentially regulates lipopolysaccharide-induced monocyte production of tumor necrosis factor alpha and interleukin-6 by increasing intracellular cyclic AMP. Infect. Immun. (1994) 62 4432-4439. DOI: 10.1128/iai.62.10.4432-4439.1994 19. Ung D.C., Iacono G., Meziane H., Blanchard E., Papon M.A., Selten M., van Rhijn J.R., Montjean R., Rucci J., Martin S.. Ptchd1 deficiency induces excitatory synaptic and cognitive dysfunctions in mouse. Mol. Psychiatry (2018) 23 1356-1367. DOI: 10.1038/mp.2017.39 20. Tora D., Gomez A.M., Michaud J.F., Yam P.T., Charron F., Scheiffele P.. Cellular Functions of the Autism Risk Factor PTCHD1 in Mice. J. Neurosci. (2017) 37 11993-12005. DOI: 10.1523/JNEUROSCI.1393-17.2017 21. Han H.J., Tokino T., Nakamura Y.. CSR, a scavenger receptor-like protein with a protective role against cellular damage causedby UV irradiation and oxidative stress. Hum. Mol. Genet. (1998) 7 1039-1046. DOI: 10.1093/hmg/7.6.1039 22. Bock A.J., Nymoen D.A., Brenne K., Kaern J., Davidson B.. SCARA3 mRNA is overexpressed in ovarian carcinoma compared with breast carcinoma effusions. Hum. Pathol. (2012) 43 669-674. DOI: 10.1016/j.humpath.2011.06.003 23. Chen H., Herndon M.E., Lawler J.. The cell biology of thrombospondin-1. Matrix Biol. J. Int. Soc. Matrix Biol. (2000) 19 597-614. DOI: 10.1016/S0945-053X(00)00107-4 24. Tokuhisa M., Kadowaki T., Ogawa K., Yamaguchi Y., Kido M.A., Gao W., Umeda M., Tsukuba T.. Expression and localisation of Rab44 in immune-related cells change during cell differentiation and stimulation. Sci. Rep. (2020) 10 10728. DOI: 10.1038/s41598-020-67638-7 25. Burg A.R., Tse H.M.. Redox-Sensitive Innate Immune Pathways During Macrophage Activation in Type 1 Diabetes. Antioxid. Redox Signal. (2018) 29 1373-1398. DOI: 10.1089/ars.2017.7243 26. Murray P.J.. Macrophage Polarization. Annu. Rev. Physiol. (2017) 79 541-566. DOI: 10.1146/annurev-physiol-022516-034339 27. Martinez F.O., Gordon S.. The M1 and M2 paradigm of macrophage activation: Time for reassessment. F1000prime Rep. (2014) 6 13. DOI: 10.12703/P6-13 28. Wang N., Liang H., Zen K.. Molecular mechanisms that influence the macrophage m1-m2 polarization balance. Front. Immunol. (2014) 5 614. DOI: 10.3389/fimmu.2014.00614 29. Ehrt S., Schnappinger D., Bekiranov S., Drenkow J., Shi S., Gingeras T.R., Gaasterland T., Schoolnik G., Nathan C.. Reprogramming of the macrophage transcriptome in response to interferon-gamma and Mycobacterium tuberculosis: Signaling roles of nitric oxide synthase-2 and phagocyte oxidase. J. Exp. Med. (2001) 194 1123-1140. DOI: 10.1084/jem.194.8.1123 30. Chacon-Salinas R., Serafin-Lopez J., Ramos-Payan R., Mendez-Aragon P., Hernandez-Pando R., Van Soolingen D., Flores-Romo L., Estrada-Parra S., Estrada-Garcia I.. Differential pattern of cytokine expression by macrophages infected in vitro with different Mycobacterium tuberculosis genotypes. Clin. Exp. Immunol. (2005) 140 443-449. DOI: 10.1111/j.1365-2249.2005.02797.x 31. Refai A., Gritli S., Barbouche M.R., Essafi M.. Mycobacterium tuberculosis Virulent Factor ESAT-6 Drives Macrophage Differentiation Toward the Pro-inflammatory M1 Phenotype and Subsequently Switches It to the Anti-inflammatory M2 Phenotype. Front. Cell. Infect. Microbiol. (2018) 8 327. DOI: 10.3389/fcimb.2018.00327 32. Raju B., Hoshino Y., Belitskaya-Levy I., Dawson R., Ress S., Gold J.A., Condos R., Pine R., Brown S., Nolan A.. Gene expression profiles of bronchoalveolar cells in pulmonary TB. Tuberculosis (2008) 88 39-51. DOI: 10.1016/j.tube.2007.07.003 33. Benoit M., Barbarat B., Bernard A., Olive D., Mege J.L.. Coxiella burnetii, the agent of Q fever, stimulates an atypical M2 activation program in human macrophages. Eur. J. Immunol. (2008) 38 1065-1070. DOI: 10.1002/eji.200738067 34. Raker V.K., Becker C., Steinbrink K.. The cAMP Pathway as Therapeutic Target in Autoimmune and Inflammatory Diseases. Front. Immunol. (2016) 7 123. DOI: 10.3389/fimmu.2016.00123 35. Blumenstein M., Boekstegers P., Fraunberger P., Andreesen R., Ziegler-Heitbrock H.W., Fingerle-Rowson G.. Cytokine production precedes the expansion of CD14+CD16+ monocytes in human sepsis: A case report of a patient with self-induced septicemia. Shock (1997) 8 73-75. DOI: 10.1097/00024382-199707000-00012 36. Skinner N.A., MacIsaac C.M., Hamilton J.A., Visvanathan K.. Regulation of Toll-like receptor (TLR)2 and TLR4 on CD14dimCD16+ monocytes in response to sepsis-related antigens. Clin. Exp. Immunol. (2005) 141 270-278. DOI: 10.1111/j.1365-2249.2005.02839.x 37. Schleicher U., Paduch K., Debus A., Obermeyer S., Konig T., Kling J.C., Ribechini E., Dudziak D., Mougiakakos D., Murray P.J.. TNF-Mediated Restriction of Arginase 1 Expression in Myeloid Cells Triggers Type 2 NO Synthase Activity at the Site of Infection. Cell Rep. (2016) 15 1062-1075. DOI: 10.1016/j.celrep.2016.04.001 38. Sciaraffia E., Riccomi A., Lindstedt R., Gesa V., Cirelli E., Patrizio M., De Magistris M.T., Vendetti S.. Human monocytes respond to extracellular cAMP through A2A and A2B adenosine receptors. J. Leukoc. Biol. (2014) 96 113-122. DOI: 10.1189/jlb.3A0513-302RR 39. Nagabhushanam V., Solache A., Ting L.M., Escaron C.J., Zhang J.Y., Ernst J.D.. Innate inhibition of adaptive immunity: Mycobacterium tuberculosis-induced IL-6 inhibits macrophage responses to IFN-gamma. J. Immunol. (2003) 171 4750-4757. DOI: 10.4049/jimmunol.171.9.4750 40. Dube P.H., Handley S.A., Lewis J., Miller V.L.. Protective role of interleukin-6 during Yersinia enterocolitica infection is mediated through the modulation of inflammatory cytokines. Infect. Immun. (2004) 72 3561-3570. DOI: 10.1128/IAI.72.6.3561-3570.2004 41. Jankovic D., Trinchieri G.. IL-10 or not IL-10: That is the question. Nat. Immunol. (2007) 8 1281-1283. DOI: 10.1038/ni1207-1281 42. Shapouri-Moghaddam A., Mohammadian S., Vazini H., Taghadosi M., Esmaeili S.A., Mardani F., Seifi B., Mohammadi A., Afshari J.T., Sahebkar A.. Macrophage plasticity, polarization, and function in health and disease. J. Cell. Physiol. (2018) 233 6425-6440. DOI: 10.1002/jcp.26429 43. Beavo J.A., Brunton L.L.. Cyclic nucleotide research-still expanding after half a century. Nat. Rev. Mol. Cell Biol. (2002) 3 710-718. DOI: 10.1038/nrm911 44. Pittenger C., Huang Y.Y., Paletzki R.F., Bourtchouladze R., Scanlin H., Vronskaya S., Kandel E.R.. Reversible inhibition of CREB/ATF transcription factors in region CA1 of the dorsal hippocampus disrupts hippocampus-dependent spatial memory. Neuron (2002) 34 447-462. DOI: 10.1016/S0896-6273(02)00684-0 45. Qian B., Zen Z., Zheng Z., Wang C., Song J.. A preliminary study on the mechanism of the neurosteroid-mediated ionotropic receptor dysfunction in neurodevelopmental toxicity induced by decabromodiphenyl ether. Ecotoxicol. Environ. Saf. (2021) 217 112198. DOI: 10.1016/j.ecoenv.2021.112198 46. Liu J., Liu B., Yuan P., Cheng L., Sun H., Gui J., Pan Y., Huang D., Chen H., Jiang L.. Role of PKA/CREB/BDNF signaling in PM2.5-induced neurodevelopmental damage to the hippocampal neurons of rats. Ecotoxicol. Environ. Saf. (2021) 214 112005. DOI: 10.1016/j.ecoenv.2021.112005 47. Zaccolo M., Pozzan T.. Discrete microdomains with high concentration of cAMP in stimulated rat neonatal cardiac myocytes. Science (2002) 295 1711-1715. DOI: 10.1126/science.1069982 48. Zhang J.Z., Lu T.W., Stolerman L.M., Tenner B., Yang J.R., Zhang J.F., Falcke M., Rangamani P., Taylor S.S., Mehta S.. Phase Separation of a PKA Regulatory Subunit Controls cAMP Compartmentation and Oncogenic Signaling. Cell (2020) 182 1531-1544. DOI: 10.1016/j.cell.2020.07.043 49. Newton K., Dixit V.M.. Signaling in innate immunity and inflammation. Cold Spring Harb. Perspect. Biol. (2012) 4 a006049. DOI: 10.1101/cshperspect.a006049 50. Kawai T., Akira S.. The role of pattern-recognition receptors in innate immunity: Update on Toll-like receptors. Nat. Immunol. (2010) 11 373-384. DOI: 10.1038/ni.1863 51. Liu T., Zhang L., Joo D., Sun S.C.. NF-kappaB signaling in inflammation. Signal Transduct. Target. Ther. (2017) 2 1-9 52. Gerlo S., Kooijman R., Beck I.M., Kolmus K., Spooren A., Haegeman G.. Cyclic AMP: A selective modulator of NF-kappaB action. Cell. Mol. Life Sci. (2011) 68 3823-3841. DOI: 10.1007/s00018-011-0757-8 53. Diamant G., Dikstein R.. Transcriptional control by NF-kappaB: Elongation in focus. Biochim. Biophys. Acta (2013) 1829 937-945. DOI: 10.1016/j.bbagrm.2013.04.007 54. Viatour P., Merville M.P., Bours V., Chariot A.. Phosphorylation of NF-kappaB and IkappaB proteins: Implications in cancer and inflammation. Trends. Biochem. Sci. (2005) 30 43-52. DOI: 10.1016/j.tibs.2004.11.009 55. Xu J., Jia Y.F., Tapadar S., Weaver J.D., Raji I.O., Pithadia D.J., Javeed N., Garcia A.J., Choi D.S., Matveyenko A.V.. Inhibition of TBK1/IKKepsilon Promotes Regeneration of Pancreatic beta-cells. Sci. Rep. (2018) 8 15587. DOI: 10.1038/s41598-018-33875-0 56. Gantke T., Sriskantharajah S., Sadowski M., Ley S.C.. IκB kinase regulation of the TPL-2/ERK MAPK pathway. Immunol. Rev. (2012) 246 168-182. DOI: 10.1111/j.1600-065X.2012.01104.x 57. Waterfield M.R., Zhang M., Norman L.P., Sun S.-C.. NF-κB1/p105 Regulates Lipopolysaccharide-Stimulated MAP Kinase Signaling by Governing the Stability and Function of the Tpl2 Kinase. Mol. Cell (2003) 11 685-694. DOI: 10.1016/S1097-2765(03)00070-4 58. Beinke S., Deka J., Lang V., Belich M.P., Walker P.A., Howell S., Smerdon S.J., Gamblin S.J., Ley S.C.. NF-kappaB1 p105 negatively regulates TPL-2 MEK kinase activity. Mol. Cell. Biol. (2003) 23 4739-4752. DOI: 10.1128/MCB.23.14.4739-4752.2003 59. Robinson M.J., Beinke S., Kouroumalis A., Tsichlis P.N., Ley S.C.. Phosphorylation of TPL-2 on serine 400 is essential for lipopolysaccharide activation of extracellular signal-regulated kinase in macrophages. Mol. Cell. Biol. (2007) 27 7355-7364. DOI: 10.1128/MCB.00301-07 60. Roget K., Ben-Addi A., Mambole-Dema A., Gantke T., Yang H.T., Janzen J., Morrice N., Abbott D., Ley S.C.. IkappaB kinase 2 regulates TPL-2 activation of extracellular signal-regulated kinases 1 and 2 by direct phosphorylation of TPL-2 serine 400. Mol. Cell. Biol. (2012) 32 4684-4690. PMID: 22988300 61. Xu J.J., Dong D.Y., Song X.H., Ge M., Li G.L., Fu L., Zhuang H.L., Chen W.. Expression, purification and characterization of the recombinant anthrax protective antigen. Sheng Wu Gong Cheng Xue Bao = Chin. J. Biotechnol. (2004) 20 652-655
--- title: 'Electrocardiographic Pathological Findings Caused by the SARS-CoV-2 Virus Infection: Evidence from a Retrospective Multicenter International Cohort Longitudinal Pilot Study of 548 Subjects' authors: - Nicola Susca - Antonio Giovanni Solimando - Paola Borrelli - Donatello Marziliano - Francesco Monitillo - Pasquale Raimondo - Domenico Vestito - Agostino Lopizzo - Gaetano Brindicci - Mohammad Abumayyaleh - Ibrahim El-Battrawy - Annalisa Saracino - Salvatore Grasso - Natale Daniele Brunetti - Vito Racanelli - Francesco Santoro journal: Journal of Cardiovascular Development and Disease year: 2023 pmcid: PMC9967030 doi: 10.3390/jcdd10020058 license: CC BY 4.0 --- # Electrocardiographic Pathological Findings Caused by the SARS-CoV-2 Virus Infection: Evidence from a Retrospective Multicenter International Cohort Longitudinal Pilot Study of 548 Subjects ## Abstract COVID-19 has threatened the capability of receiving and allocating patients in emergency departments (EDs) all over the world. This is a retrospective cohort study to explore the role of a simple procedure like an ECG to screen for the severity of COVID-19 on admission to the ED. For this study, 548 consecutive patients were enrolled in a multicenter international registry and stratified upon ECG on admission with a simple distinction between normal vs. abnormal rhythm. Among patients in the abnormal ECG group were those with heart rates higher than 100 beats per minute and/or atrial fibrillation. Survival in patients with normal ECG rhythm was deemed below $75\%$ after 58 days and then stabilized, while survival in patients with abnormal ECG rhythm was deemed below $75\%$ after 11 days and below $50\%$ after 21 days. A multivariate analysis including abnormal rhythm, gender, age, diabetes, obesity, respiratory failure during hospitalization, heart failure during hospitalization, and abnormal rhythm was an independent predictor of death (HR 7.20 $95\%$ CI 3.63–14.28, $p \leq 0.01$). This finding, if confirmed in large prospective studies, is promising for identifying a cheap and simple procedure for patients in need of a closer look. ## 1. Introduction SARS-CoV-2 can determine infections with very different clinical characteristics and severity, ranging from asymptomatic to acute respiratory distress syndrome (ARDS). In addition to the lung, the organ most affected by the virus, many other organs are adversely affected due to the widespread expression of its target receptors ACE and TMPRSS2 [1]. Emerging evidence suggests an important cardiac involvement, with consequences on patient outcome [2,3]. Among hospitalized patients with COVID-19, a variable percentage of patients, about 20–$30\%$, can have electrophysiological disturbances [4]. Considering the mechanisms underlying the arrhythmias in COVID-19, both direct and indirect effects of SARS-CoV-2 infection can be recognized. In the first group, we can recognize direct cardiotoxicity, dysregulation of the RAAS, endothelial damage and thromboinflammation, immune-dysregulation-induced cytokine storm, and demand-supply mismatch [5], with the pivotal potential mechanisms being hypoxia, myocarditis, abnormal host immune response, myocardial ischemia, myocardial strain, electrolyte derangements, metabolic and endocrine implications [6], intravascular volume imbalances and drug side effects [5]. Dysautonomia is a common feature as well, especially in the form of postural orthostatic tachycardia syndrome (POTS) [7]. The various proposed mechanisms proposed to be responsible for POTS differ between the acute setting and the chronic one. First, the viral infection leads to a hyper-responsiveness of the immune system, with autoimmune inflammation and immunosuppression. Next, in the chronic phase, both immune and non-immune mechanisms can interact in driving the POTS. On top of that, hypovolemia, hydration state, and other factors co-intervene [7]. In this study, we aimed to: [1] Evaluate whether patients with SARS-CoV-2 infection present with a different rate of ECG abnormalities, suggesting cardiac involvement and [2] assess the prognostic implications of ECG in patients hospitalized because of COVID-19, as well as construct a multivariable model based on these findings. We, therefore, assessed whether the ECG anomalies could be linked to the action of the virus and whether this affected the patients’ mortality. ## 2. Materials and Methods A total of 548 consecutive patients admitted with a COVID-19 diagnosis were enrolled in this retrospective longitudinal cohort study. Patients were enrolled from March 2020 to December 2020 from 4 European hospitals: the infectious diseases unit and intensive care unit of Hospital-University Polyclinic of Bari, Italy; the Department of Infectious disease, Vittorio Emanuele II Hospital, Bisceglie, Italy; the Department of Infectious Disease, San Carlo Hospital, Potenza, Italy; and the First Department of Medicine, Faculty of Medicine, University Medical Centre Mannheim, Germany. An ECG was performed for each at admission, and several anamnestic, biochemical, and other parameters were collected. SARS-CoV-2 testing was based on the protocol released by the local and Italian authorities and as previously described [8]. In detail, laboratory confirmation of SARS-CoV-2 was defined by a positive result on a real-time reverse transcriptase-polymerase chain reaction (RT-PCR) assay performed on nasopharyngeal swabs or lower respiratory tract aspirates. Chest X-ray and, when needed, thoracic computerized tomography (CT) scan were performed to confirm the diagnosis. Patients were further divided into [1] COVID-19 and normal rhythm detected by ECG group (sinus rhythm with heart rate between 60–100 beats per minute) and [2] COVID-19 and abnormal ECG, including all subjects with any rhythm abnormality detected by ECG testing. Among patients within the abnormal ECG group were included those with heart rates higher than 100 beats per minute and/or atrial fibrillation. This study was approved by the institutional ethics board, which waived the need for informed consent, and was performed in accordance with the ethical standards laid down in the 1964 Declaration of Helsinki and its later amendments. The study considered all adults (aged ≥ 18 years) with laboratory-confirmed SARS-CoV-2 infection and admitted them as in-patients in both intensive and non-intensive care units. Clinical data were collected from electronic health records, including age, sex, smoking habit, triage vital signs and presenting symptoms, comorbidities, current medications, laboratory test results (pre-defined disease-specific panel), and duration and outcome of follow-up. ## 2.1. ECG Analysis and Definitions Twelve-lead standard ECGs were recorded on admission with a CARDIOLINE® HD+ ECG machine. The ECGs were retrieved by our dedicated institutional ECG storage server and independently analyzed by two cardiologists (FS, FM). The ECGs were analyzed before proceeding with an assessment of clinical outcome, which was therefore unknown to the ECG readers. ## 2.2. Statistical Analysis Descriptive analysis was carried out using mean and standard deviation or median and interquartile range (IQR) for the quantitative variables and percentages values for the qualitative ones. Normality distribution for quantitative variables was assessed by the Shapiro–Wilk test. Univariate comparisons were investigated between groups (ECG with normal and abnormal rhythm) and explicative variables using the Pearson chi-square test or the Fisher’s exact test for categorical data, the Student’s t-test for independent data, or non-parametric Wilcoxon rank-sum test when appropriate for continuous data. Survival analysis was performed by applying the Kaplan–Meier estimator and log-rank test for equality of survivor functions. The association with clinical features was analyzed with the Cox model of proportional hazards (hazard ratio (HR) and $95\%$ CI), and the applicability assumption was evaluated by the Schoenfeld test. Statistical significance was taken at the <0.05 level. All analyses were performed using STATA software v15.1 (StataCorp, College Station, TX, USA). ## 3.1. Descriptive Characteristics and Comparisons Five hundred forty-eight adult patients were included. There were 327 ($59.7\%$) females and 221 ($40.7\%$) males, with a mean age of 61.8 ± 16.9 years (range 18–98). Table 1 shows clinical, laboratory, and socio-demographic characteristics for all patients and stratified by ECG status (normal and abnormal rhythm). We found statistically significant differences between ECG status and outcome (χ2 = 52.66, df = 1, $p \leq 0.001$) and heart failure during hospitalization (χ2 = 13.24, df = 1, $p \leq 0.001$). ## 3.2. Survival Analysis Survival in patients with normal ECG rhythm was deemed below $75\%$ after 58 days and then stabilized, while survival in patients with abnormal ECG rhythm was deemed below $75\%$ after 11 days and below $50\%$ after 21 (Table 2). At any time, median survival was always higher in subjects with a normal ECG rhythm than in those with an abnormal rhythm. This result was deemed statistically significant in both uni- and multivariate analyses (Figure 1 and Table 3). Hazard ratio and corresponding $95\%$ CIs were determined in univariate analyses through the Cox model for overall survival to evaluate relationships between ECG tracings and overall survival, showing a statistically significant HR both at univariate (HR = 8.92, $95\%$CI 4.60–17.26, $p \leq 0.001$) and multivariate analysis (HR = 7.20, $95\%$CI 3.63–14.28, $p \leq 0.001$). Furthermore, a higher death rate was observed in patients affected by heart failure (HR = 3.98, $95\%$CI 2.10–7.53, $p \leq 0.001$) in univariate and multivariate analyses (HR = 2.74, $95\%$CI 1.33–5.72, $p \leq 0.001$). No significant association was appreciated between sex, diabetes, obesity and lung failure, and overall survival. ## 4. Discussion A major pathological step of SARS-CoV-2 is the trigger function exerted in developing cardiac arrhythmias. Despite the caveats existing in the given rhythm disturbance, Guo et al. defined malignant rhythm disturbance as sustained ventricular tachycardia persisting for more than 30 s, with hemodynamic instability or ventricular fibrillation [9]. The authors uncovered troponin as a potential prognostic discriminator associated with rhythm disturbances in this scenario. The underlying mechanisms of arrhythmias in COVID-19 can be comprised within the clinical consequence of the acute myocardial injury, electrical instability associated with QT elongation, e.g., associated with hypokalemia, hypomagnesemia, and drug use. In addition, a key role is also played by the direct electrophysiological effect of cytokines on the myocardium, with a prolongation in the duration of the action potential and consequent demodulation of the calcium and potassium [10] channels. There is often a cytokine hyperactivation of the sympathetic system centrally and peripherally, as well as electrolyte abnormalities resulting from an alteration of renal function [11]. Despite these data, shreds of evidence regarding rapid and effective point-of-care ECG monitoring potentially stratifying COVID-19 based on simpler rhythm classification are scanty. While acknowledging this study’s limitations due to the lack of statistical power and the need for prospective trials to validate our hypothesis-generating report, to our knowledge, this is the first report highlighting the potential role of a basic ECG-oriented screening of patients at risk of a worse prognosis. This observation holds the potential to implement the already validated prognosticators’ tools [8,12] with a potential impact on outpatient settings besides hospital-admitted subjects, also in the era of effective outpatient therapies [13,14,15]. The value of this study lies in demonstrating the strong association with an unfavorable outcome of a simple and easily achievable parameter such as heart rhythm analysis. A fair number of studies have deepened the prognostic role of ECG in COVID-19 patients, finding significant combinations of various parameters with increased mortality or critical illness: signs of previous myocardial infarction [16], acute change in the ST tract and T wave [16,17], left bundle branch block [18], intraventricular block [19], premature atrial beats [19], right bundle branch block [19], right ventricular strain [20], fragmented QRS [21], heart rate variability [22], poor R wave progression [23], and lengthening of QTc interval and subsequent development of life-threatening arrhythmias [3]. COVID-19 increases the risk of myocarditis, but it appears that direct myocardial involvement in SARS-CoV-2 infection is relatively rare [24] compared to extensive evidence of ECG alterations or increased heart damage enzymes. Similarly, SARS-CoV-2 increases the risk of acute myocardial infarction [25], but a presentation with a classic ST tract elevation remains rare [19]. In summary, heart damage is strongly associated with COVID-19 mortality [26]. With this work, we propose the simple evaluation of the rhythm at ECG (altered rhythm vs. normal rhythm) as an important piece for the prognostic stratification of the patient at admission to the emergency department, potentially being considered as a standing alone factor associated with increased mortality. After adequate validation as part of prospective studies, heart rhythm assessment could be used, along with other simple clinical, instrumental, and laboratory evaluations (such as the presence of dyspnea, chest ultrasound, and arterial blood gas analysis) as a screening tool in territorial structures and in a non-hospital and outpatient setting to direct the patient to a higher intensity treatment path. Moreover, based on resourceful in silico predictors available [27], a novel therapeutical landscape can be better tailored based on the given cardiological profiling. ## 5. Conclusions Collectively, even though our aim is far from considering this study as a turning point in COVID-19 early management, its significance can be acknowledged as it could integrate and improve already existing prognostic indicators by recognizing pathological cardiac rhythm. An early and simple assessment of patients can be instrumental in providing the most appropriate and accurate allocation, especially in the outpatient setting, namely, in general practice. ## References 1. Lukassen S., Chua R.L., Trefzer T., Kahn N.C., Schneider M.A., Muley T., Winter H., Meister M., Veith C., Boots A.W.. SARS-CoV-2 Receptor ACE2 and TMPRSS2 Are Primarily Expressed in Bronchial Transient Secretory Cells. EMBO J. (2020) 39 e015114. DOI: 10.15252/embj.20105114 2. Evans P.C., Rainger G.E., Mason J.C., Guzik T.J., Osto E., Stamataki Z., Neil D., Hoefer I.E., Fragiadaki M., Waltenberger J.. Endothelial Dysfunction in COVID-19: A Position Paper of the ESC Working Group for Atherosclerosis and Vascular Biology, and the ESC Council of Basic Cardiovascular Science. Cardiovasc. Res. (2020) 116 2177-2184. DOI: 10.1093/cvr/cvaa230 3. Santoro F., Monitillo F., Raimondo P., Lopizzo A., Brindicci G., Gilio M., Musaico F., Mazzola M., Vestito D., Benedetto R.D.. QTc Interval Prolongation and Life-Threatening Arrhythmias During Hospitalization in Patients with Coronavirus Disease 2019 (COVID-19): Results from a Multicenter Prospective Registry. Clin. Infect. Dis. (2021) 73 e4031-e4038. DOI: 10.1093/cid/ciaa1578 4. Zareini B., Rajan D., El-Sheikh M., Jensen M.H., Højbjerg Lassen M.C., Skaarup K., Hansen M.L., Biering-Sørensen T., Jabbari R., Kirk O.. Cardiac Arrhythmias in Patients Hospitalized with COVID-19: The ACOVID Study. Heart Rhythm O2 (2021) 2 304-308. DOI: 10.1016/j.hroo.2021.03.008 5. Bovolenta P., Liem R.K.H., Mason C.A.. Glial Filament Protein Expression in Astroglia in the Mouse Visual Pathway. Dev. Brain Res. (1987) 33 113-126. DOI: 10.1016/0165-3806(87)90181-7 6. Lisco G., De Tullio A., Stragapede A., Solimando A.G., Albanese F., Capobianco M., Giagulli V.A., Guastamacchia E., De Pergola G., Vacca A.. COVID-19 and the Endocrine System: A Comprehensive Review on the Theme. J. Clin. Med. (2021) 10. DOI: 10.3390/jcm10132920 7. Blitshteyn S., Whitelaw S.. Postural Orthostatic Tachycardia Syndrome (POTS) and Other Autonomic Disorders after COVID-19 Infection: A Case Series of 20 Patients. Immunol. Res. (2021) 69 205-211. DOI: 10.1007/s12026-021-09185-5 8. Solimando A.G., Susca N., Borrelli P., Prete M., Lauletta G., Pappagallo F., Buono R., Inglese G., Forina B.M., Bochicchio D.. Short-Term Variations in Neutrophil-to-Lymphocyte and Urea-to-Creatinine Ratios Anticipate Intensive Care Unit Admission of COVID-19 Patients in the Emergency Department. Front. Med. (2021) 7 625176. DOI: 10.3389/fmed.2020.625176 9. Guo T., Fan Y., Chen M., Wu X., Zhang L., He T., Wang H., Wan J., Wang X., Lu Z.. Cardiovascular Implications of Fatal Outcomes of Patients with Coronavirus Disease 2019 (COVID-19). JAMA Cardiol. (2020) 5 811. DOI: 10.1001/jamacardio.2020.1017 10. Berlansky S., Sallinger M., Grabmayr H., Humer C., Bernhard A., Fahrner M., Frischauf I.. Calcium Signals during SARS-CoV-2 Infection: Assessing the Potential of Emerging Therapies. Cells (2022) 11. DOI: 10.3390/cells11020253 11. Jain U.. Effect of COVID-19 on the Organs. Cureus (2020) 12 e9540. DOI: 10.7759/cureus.9540 12. Liao D., Zhou F., Luo L., Xu M., Wang H., Xia J., Gao Y., Cai L., Wang Z., Yin P.. Haematological Characteristics and Risk Factors in the Classification and Prognosis Evaluation of COVID-19: A Retrospective Cohort Study. Lancet Haematol. (2020) 7 e671-e678. DOI: 10.1016/S2352-3026(20)30217-9 13. Gupta A., Gonzalez-Rojas Y., Juarez E., Crespo Casal M., Moya J., Falci D.R., Sarkis E., Solis J., Zheng H., Scott N.. Early Treatment for Covid-19 with SARS-CoV-2 Neutralizing Antibody Sotrovimab. N. Engl. J. Med. (2021) 385 1941-1950. DOI: 10.1056/NEJMoa2107934 14. Bavaro D., Diella L., Solimando A., Cicco S., Buonamico E., Stasi C., Ciannarella M., Marrone M., Carpagnano F., Resta O.. Bamlanivimab and Etesevimab Administered in an Outpatient Setting for SARS-CoV-2 Infection. Pathog. Glob. Health (2022) 116 297-304. DOI: 10.1080/20477724.2021.2024030 15. Hammond J., Leister-Tebbe H., Gardner A., Abreu P., Bao W., Wisemandle W., Baniecki M., Hendrick V.M., Damle B., Simón-Campos A.. Oral Nirmatrelvir for High-Risk, Nonhospitalized Adults with Covid-19. N. Engl. J. Med. (2022) 386 1397-1408. DOI: 10.1056/NEJMoa2118542 16. Rosén J., Noreland M., Stattin K., Lipcsey M., Frithiof R., Malinovschi A., Hultström M.. ECG Pathology and Its Association with Death in Critically Ill COVID-19 Patients, a Cohort Study. PLoS ONE (2021) 16. DOI: 10.1371/journal.pone.0261315 17. Yang D., Li J., Gao P., Chen T., Cheng Z., Cheng K., Deng H., Fang Q., Yi C., Fan H.. The Prognostic Significance of Electrocardiography Findings in Patients with Coronavirus Disease 2019: A Retrospective Study. Clin. Cardiol. (2021) 44 963-970. DOI: 10.1002/clc.23628 18. De Carvalho H., Leonard-Pons L., Segard J., Goffinet N., Javaudin F., Martinage A., Cattin G., Tiberghien S., Therasse D., Trotignon M.. Electrocardiographic Abnormalities in COVID-19 Patients Visiting the Emergency Department: A Multicenter Retrospective Study. BMC Emerg. Med. (2021) 21. DOI: 10.1186/s12873-021-00539-8 19. Mccullough S.A., Goyal P., Krishnan U., Choi J.J., Safford M.M., Okin P.M.. Electrocardiographic Findings in Coronavirus Disease-19: Insights on Mortality and Underlying Myocardial Processes. J. Card. Fail. (2020) 26 626-632. DOI: 10.1016/j.cardfail.2020.06.005 20. Raad M., Gorgis S., Dabbagh M., Chehab O., Parikh S., Singh G.. Right Heart Strain on Presenting 12-Lead Electrocardiogram Predicts Critical Illness in COVID-19. JACC Clin. Electrophysiol. (2021) 7 485-493. DOI: 10.1016/j.jacep.2020.09.013 21. Özdemir İ.H., Özlek B., Özen M.B., Gündüz R., Çetin N., Özlek E., Yıldız B.S., Tıkız H.. Fragmented QRS is a marker of mortality in patients with severe COVID-19: A retrospective observational study. Anatol. J. Cardiol. (2021) 25 811-820. DOI: 10.5152/AnatolJCardiol.2021.62 22. Mol M.B.A., Strous M.T.A., van Osch F.H.M., Vogelaar F.J., Barten D.G., Farchi M., Foudraine N.A., Gidron Y.. Heart-Rate-Variability (HRV), Predicts Outcomes in COVID-19. PLoS ONE (2021) 16. DOI: 10.1371/journal.pone.0258841 23. Santoro F., Núñez-Gil I.J., Vitale E., Viana-Llamas M.C., Romero R., Maroun Eid C., Feltes Guzman G., Becerra-Muñoz V.M., Fernández Rozas I., Uribarri A.. Aspirin Therapy on Prophylactic Anticoagulation for Patients Hospitalized with COVID-19: A Propensity Score-Matched Cohort Analysis of the HOPE-COVID-19 Registry. J. Am. Heart Assoc. (2022) 11 e024530. DOI: 10.1161/JAHA.121.024530 24. Patone M., Mei X.W., Handunnetthi L., Dixon S., Zaccardi F., Shankar-Hari M., Watkinson P., Khunti K., Harnden A., Coupland C.A.C.. Risks of Myocarditis, Pericarditis, and Cardiac Arrhythmias Associated with COVID-19 Vaccination or SARS-CoV-2 Infection. Nat. Med. (2022) 28 410-422. DOI: 10.1038/s41591-021-01630-0 25. Katsoularis I., Fonseca-Rodríguez O., Farrington P., Lindmark K., Fors Connolly A.-M.. Risk of Acute Myocardial Infarction and Ischaemic Stroke Following COVID-19 in Sweden: A Self-Controlled Case Series and Matched Cohort Study. Lancet (2021) 398 599-607. DOI: 10.1016/S0140-6736(21)00896-5 26. Weng H., Yang F., Zhang L., Jin H., Liu S., Fan F., Liu Z., Zheng X., Yang H., Li Y.. Joint Predictive Value of CTnI and NT-ProBNP on Mortality in Patients with Coronavirus Disease 2019: A Retrospective Research in Wuhan, China. J. Transl. Intern. Med. (2021) 9 177-184. DOI: 10.2478/jtim-2021-0034 27. Karami H., Derakhshani A., Ghasemigol M., Fereidouni M., Miri-Moghaddam E., Baradaran B., Tabrizi N., Najafi S., Solimando A., Marsh L.. Weighted Gene Co-Expression Network Analysis Combined with Machine Learning Validation to Identify Key Modules and Hub Genes Associated with SARS-CoV-2 Infection. J. Clin. Med. (2021) 10. DOI: 10.3390/jcm10163567
--- title: Development and Comparative In Vitro and In Vivo Study of BNN27 Mucoadhesive Liposomes and Nanoemulsions for Nose-to-Brain Delivery authors: - Maria Kannavou - Kanelina Karali - Theodora Katsila - Eleni Siapi - Antonia Marazioti - Pavlos Klepetsanis - Theodora Calogeropoulou - Ioannis Charalampopoulos - Sophia G. Antimisiaris journal: Pharmaceutics year: 2023 pmcid: PMC9967044 doi: 10.3390/pharmaceutics15020419 license: CC BY 4.0 --- # Development and Comparative In Vitro and In Vivo Study of BNN27 Mucoadhesive Liposomes and Nanoemulsions for Nose-to-Brain Delivery ## Abstract Intranasal administration offers an alternative and promising approach for direct nose-to-brain delivery. Herein, we developed two chitosan (CHT)-coated (and uncoated) nanoformulations of BNN27 (a synthetic C-17-spiro-dehydroepiandrosterone analogue), liposomes (LIPs), and nanoemulsions (NEs), and compared their properties and brain disposition (in vitro and in vivo). LIPs were formulated by thin film hydration and coated with CHT by dropwise addition. BNN27-loaded NEs (BNEs) were developed by spontaneous emulsification and optimized for stability and mucoadhesive properties. Mucoadhesive properties were evaluated by mucin adherence. Negatively charged CHT-coated LIPs (with $0.1\%$ CHT/lipid) demonstrated the highest coating efficiency and mucoadhesion. BNEs containing $10\%$ w/w Capmul-MCM and $0.3\%$ w/w CHT demonstrated the optimal properties. Transport of LIP or NE-associated rhodamine-lipid across the blood–brain barrier (in vitro) was significantly higher for NEs compared to LIPs, and the CHT coating demonstrated a negative effect on transport. However, the CHT-coated BNEs demonstrated higher and faster in vivo brain disposition following intranasal administration compared to CHT-LIPs. For both BNEs and LIPs, CHT-coating resulted in the increased (in vivo) brain disposition of BNN27. Current results prove that CHT-coated NEs consisting of compatible nasal administration ingredients succeeded in to delivering more BNN27 to the brain (and faster) compared to the CHT-coated LIPs. ## 1. Introduction Dehydroepiandrosterone (DHEA) is one of the most abundant neuroactive steroids. It is synthesized in the human adrenal cortex as well as in the brain by neurons and glia [1]. It is a multifaceted agent, interacting with steroid and neurotransmitter receptors, and acts as an endogenous precursor for the biosynthesis of androgens, estrogens, and their metabolites [2]. It is known that DHEA increases the effects of the excitatory neurotransmitter glutamate, decreases the inhibitory neurotransmitter γ-aminobutyric acid (GABA), and stimulates the release of acetylcholine (Ach) in the hippocampus. Interestingly, all of the above-mentioned neurotransmitters have altered levels in patients with depression and stress-related disorders, suggesting the potential use of neurosteroids as therapeutic tools [1]. However DHEA’s metabolism to estrogens and androgens results in serious side-effects that limit its applicability as a long-term therapeutic agent despite its proven neuroprotective activity. A new family of small synthetic C17-spiro DHEA derivatives named microneurotrophins have been synthesized [3] to mimic endogenous neurotrophin effects through the activation of specific receptors. BNN27 [3β,21-dihydroxy-17R,20-epoxy-5-pregnene], a BBB-permeable C17-spiroepoxy DHEA derivative, with agonistic activity against the neurotrophin receptors ΤrkA and p75NTR, exhibits anti-apoptotic properties in vitro and in vivo [3,4,5]. Its main advantage compared to DHEA is the lack of hormone-receptor activation while preserving its anti-apoptotic and neurotrophic properties [6], therefore, it can be considered as a novel lead compound for developing non-toxic, BBB-permeable, neurotrophin receptor agonists and antagonists for therapeutic applications in neurodegenerative diseases, brain trauma, and neuropathic pain [7]. In order to enhance the highly interesting profile of BNN27 as a CNS therapeutic, nose-to-brain delivery was considered herein as a brain targeting strategy [8,9,10]. In most cases, treatments for CNS disorders are administered parenterally, reducing drug effectiveness and potency. Even if the lipophilicity of the drug does not impede its accessibility to the brain through the circulation, systemic clearance significantly reduces the drug bioavailability [8]. In addition, plasma protein binding delays delivery to the brain through circulation, and the peripheral side effects of systemic administration routes have triggered the hunt for alternative routes that deliver drugs directly to the brain. The intranasal route has emerged as a powerful strategy to circumvent the BBB [9,10], by delivering drugs directly to the brain through the nasal cavity via olfactory neurons. This pathway is associated with enhanced safety, increased patient compliance, remarkable ease of administration, rapid onset of action as well as minimized systemic exposure [9]. The design of suitable nose-to-brain formulations with enhanced mucoadhesive properties, large surface contact area and penetration enhancement, is necessary for successful intranasal drug delivery [10]. Herein, we prepared two different formulation types, LIPs and nanoemulsions with the optimal physicochemical and mucoadhesive properties, and compared them for their potential to deliver BNN27 to the brain following nasal administration. To the best of our knowledge, this is the first time that such mucoadhesive formulations have been compared for their brain disposition after intranasal delivery. ## 2. Materials and Methods 1,2-Distearoyl-sn-glycerol-3-phosphatidylcholine (PC) and 1,2-distearoyl-sn-glycero-3-phospho-(19-rac-glycerol) (sodium salt) (PG), and Lissamine Rhodamine B phosphatidylethanolamine or Rhodamine-lipid (RHO), were purchased from Avanti Polar Lipids (Alabaster, AL). Capmul MCM was received as a gift sample from Abitec Corporation Limited (Columbus, OH, USA). Labrafac Lipophile WL 1349, Labrafac PG, and Transcutol HP were received as gift samples from Gattefosse (Lyon, France). Tween 80 was purchased from Fisher BioReagents and Tween 20 from BioChemica UK Ltd. Carbopol 974P was kindly provided by Chemix SA (Athens, Greece). Cholesterol, mucin from porcine stomach Type III bound sialic acid 0.5–$1.5\%$ (partially purified powder), low molecular weight chitosan (LMW-CHT, with a molecular weight of 50–190 kDa and 75–$85\%$ deacetylated) and medium molecular weight chitosan (MMW-CHT, with a molecular weight between 190–310 kDa and 75–$85\%$ deacetylated), and all other excipients for the nanoemulsion preformulation studies were purchased from Sigma-Aldrich or Merck. BNN27 was kindly provided by Bionature Ltd. For the quantification of BNN27, an enzymatic method—a cholesterol kit purchased from Biotechnological applications LTD (Athens, Greece)—was used. All other chemicals were of analytical quality and were purchased from Sigma-Aldrich or Merck (Darmstadt, Germany). ## 2.1. Quantification of BNN27 Concentration in Formulations For the measurement of BNN27 loading in the various formulations during formulation development, as a routine everyday quantification method, the CO/PAP enzymatic method [11,12], which is used for the measurement of blood cholesterol (Cholesterol Determination Kit, Biotechnological Applications), was applied (due to the structural similarity of BNN27 with cholesterol). For this, appropriate BNN27 calibration curves were constructed by preparing BNN27 standard solutions of known concentrations ranging between 0 and 150 ppm in ethanol. A sample of 100 µL from each solution was mixed with 100 µL of PBS (or empty LIPs, in order to see if the presence of LIP components disturbed the BNN27 measurement) and 1 mL of the Cholesterol Measurement Kit reagent was added. After mixing and 15 min incubation at 37 °C, the optical density at 510 nm was measured by a Shimatzu UV-1205 spectrophotometer. The measurement of BNN27 in ΝΕs or solutions of NE ingredients is not possible with this method due to interactions between the NE ingredients and reagent. Thereby, the BNN27 extraction was applied for the measurement of BNN27 in NEs or in NE ingredients (oils, surfactants and co-surfactants). For this, 1 mL of BNN27-containing NEs or solutions was vigorously mixed (by vortex) with 2 mL chloroform for (at least) 2 min. Then, after complete separation of the two phases, the aqueous phase was removed and the organic was evaporated. The drug was re-suspended in 200 μL of ethanol and mixed with 1 mL of reagent. After incubation at 37 °C for 15 min, the sample OD-510 nm was measured. The BNN27 content was calculated from a calibration curve that was conducted by the same method (applied in each case) using known amounts of BNN27 mixed with empty LIPs, blank NEs, or NE components. ## 2.2.1. Preparation of BNN27-Loaded LIPs MLV LIPs composed of PC or PC/PG at 9:1 (mol/mol) and loaded with BNN27 were prepared by the thin-film hydration method [13,14]. For this, the lipid(s), together with the drug, was dissolved in the CHCl3/MeOH (2:1 v/v) mixture in a round bottom flask and a thin lipid film was formed by rotor-evaporation of the organic solvents. The thin lipid film was hydrated with PBS, pH 7.40. Different amounts of BNN27 were included in the lipid solution in order to identify the maximum loading conditions. After the formation of multilamellar vesicles, their size was reduced by probe sonication (Sonics & Materials) in order to produce SUV LIPs. For this, 3–5 min of sonication with a tapered microtip was applied at $35\%$ intensity until a clear dispersion was produced. Following sonication, the SUV LIPs were incubated at room temperature for 1 h in order to anneal any structural defects. Separation of LIPs from the non-encapsulated BNN27 was achieved by centrifugation (15,000 rpm for 30 min) and supernatant filtration through 0.45 μm pore filters, in order to separate the LIP dispersions from any non-incorporated and precipitated BNN27. LIPs were stored at 4 °C before use. ## 2.2.2. Preparation of Chitosan-Coated LIPs For the LIP coating, two types of CHT were used: low molecular weight (LMW) and medium molecular weight (MMW). CHT solutions of different concentrations were prepared in isotonic acetate buffer (pH = 4.40). Then, 0.5 mL of the LIP dispersion (in PBS) was mixed with the dropwise addition of an equal volume of the appropriate CHT solution to give CHT/lipid (w/w) ratios of 0.1, 0.3, or 0.5, under continuous stirring for 1 h at room temperature [13], followed by overnight incubation at 4 °C [15]. Then, CHT-coated LIPs were harvested from the reaction mixture by centrifugation at 15,000 rpm for 15 min. The supernatant (non-adhered CHT) was removed and the pellet, consisting of CHT-coated LIPs, was re-suspended in PBS. For coating efficiency determination, the phospholipid content of the samples was measured by the Stewart assay and compared to the total phospholipid (before centrifugation), as previously reported [13,16]. The coating efficiency (%) was calculated from the equation:Coating Efficiency (CE) % = LPRE (lipid-in-precipitate)/LTOT (Total Lipid) × 100[1] ## 2.2.3. Physicochemical Properties of LIPs The size distribution (mean hydrodynamic diameter and polydispersity index) and ζ-potential of the LIP dispersions were measured by dynamic light scattering (DLS) and laser Doppler electrophoresis (LDE), respectively, on a Nano-ZS (Nanoseries, Malvern Instruments), which measures the mass distribution of the particle size as well as the electrophoretic mobility of the dispersed particles. Measurements were made at 25 °C with a fixed angle of 173°. Sizes quoted are the z-average mean (dz) for the liposomal hydrodynamic diameter (nm). Calculation of ζ-potential (mV) was carried out by the instrument from electrophoretic mobility, which was measured in small volume disposable zeta cells and converted to zeta potential by in-built software that applies the Helmholtz–Smoluchowski equation. For measurements, samples were diluted to have a 0.4 mg/mL lipid concentration. ## 2.3. Preparation of BNN27-Loaded Nanoemulsion In order to formulate BNN27-loaded nanoemulsions (BNEs), we followed the previously reported methodologies for the optimal formulations of nanoemulsions intended for nose-to-brain delivery of various drugs such as risperidone [17,18], rivastigmine [19], and quetiapine fumarate [20]. The selection of ingredients was based on the BNN27 solubility measurements, ternary phase diagrams (to select the optimal surfactant/co-surfactant ratio in the surfactant-co-surfactant mixture (Smix)), and the physicochemical properties/stability of the formulated NEs. ## 2.3.1. Solubility of BNN27 in Potential NE Ingredients In order to select the optimal materials, BNN27 solubility studies were carried out. The solubility of BNN27 was measured in three different oils that are commonly used in NEs [18,19,20,21], Capmul MCM, Labrafac Lipophile WL 1349, and Labrafac PG. Additionally three surfactants, Tween 20, Tween 80, and Cremophor EL RH40 as well as three co-surfactants, PEG 400, Transcutol HP, and a Transcutol HP/propylene glycol (1:1) mixture, were studied. For solubility determination, BNN27 was added as a solid powder in the different liquids or mixtures until a noticeable amount of the powder was not solubilized, and stirred for 24 h on a mechanical rocking shaker (Kisher Biotech) at room temperature (25.0 ± 2.3 °C). Afterward, only the samples that still had visible undissolved solid powder were further processed by centrifugation (10,000× g for 10 min) and the supernatants were diluted. Finally, BNN27 solubilized in the liquids was measured by the method described in detail above (Section 2.1). ## 2.3.2. Optimization of NE Formulation After the selection of the best oil and co-surfactant (based of BNN27 solubility), and due to the fact that two of the tested surfactants, Tween 20 and Tween 80, demonstrated similar (very high) ability to solubilize BNN27, both surfactants were studied with the scope to finally select the one that conferred NEs with the maximum stability. Since it is known from the relevant literature that the best surfactant/co-surfactant ratio in the case when Tween 80 is used (as surfactant) and Transcutol HP/propylene glycol (1:1) (as co-surfactant) is 4:2 (w/w) [18], we initially constructed similar ternary diagrams to find out whether the same optimal ratio also applies when Tween 20 is used as a surfactant. Additionally, since it has been reported that in NEs containing Tween 80 together with Transcutol HP/propylene glycol mixtures, the NE globule size is the lowest when the percent of Smix is $44\%$ (w/w), and does not further decrease when increasing the Smix content [17,20], a number of NEs containing Tween 20 as the surfactant instead of Tween 80 and different Smix amounts were formulated and evaluated for the effect of the Smix amount on the corresponding NE’s globule size and transparency. Finally, the globule sizes of the NEs constructed using Capmul MCM as the oil phase at $8\%$ or $10\%$ (w/w), and Tween 20 or Tween 80 as the surfactant (always mixed with Transcutol HP/propylene glycol (1:1) (as the co-surfactant) with $44\%$ (w/w) Smix content), were compared in order to identify which of the two surfactants, Tween 20 or Tween 80, produced NEs with the lowest globule size and highest stability. ## 2.3.3. Preparation of BNN27-Loaded NEs After the identification of the optimal NE composition, BNN27-loaded NEs were formulated by the spontaneous emulsification (titration) method. For this, a saturated solution of BNN27 in Capmul MCM was prepared by adding 40 mg/mL BNN27 in the oil phase and applying magnetic stirring. Then, in the BNN27 Capmul MCM phase, the Smix (containing Tween 80 as surfactant) was added until a clear mixture was produced. Finally, H2O was added dropwise and stirred to produce clear NEs of BNN27. If all of the BNN27 is loaded with a $10\%$ w/w oil phase, the NEs should contain 4 mg of BNN27 per mL of NE. Mucoadhesive, chitosan (CHT) or carbopol (CAR) coated BNN27-loaded NEs (BNEs) were also prepared. For this, concentrated BNEs (using the minimum volume of the external phase) were initially prepared, and then mixed with the required volume of CHT or CAR aqueous solution to attain a final CHT or CAR concentration of $0.3\%$ w/w. After the addition of CHT or CAR, the BNEs were allowed to homogenize by continuous stirring for 1 h. ## 2.3.4. Physicochemical Properties and Stability of NEs The quality and stability of the various NE or BNE formulations constructed were evaluated by dilution tests, centrifugation tests, measurements of pH, transmittance, globule size distribution, and ζ-potential. The dilution test was performed by diluting 1 mL of NEs to 100 mL with d.d. H2O, and applying optical observation of the NEs for clarity/turbidity. For the centrifugation test, NEs were centrifuged at 300× g for 15 min and examined by visual observation if they remained monophasic or if phase separation occurred. The pH of the NEs was measured in 5 mL samples placed in a 10 mL beaker by a pH meter (Consort P902). The transmittance percent (%T) of NEs at 650 nm was measured using a UV–VIS spectrophotometer. Finally, the globule size distribution and ζ-potential of the NEs were measured with the same methods described for LIPs (§ 2.2). For measurement, the NE samples were diluted 20 times in d.d. H2O. All measurements were performed in triplicate. The stability of the NEs (coated and non-coated) was evaluated by applying all of the methods above-mentioned at various time points (1, 7, 14, 21, 30, 60 d) during storage (for up to 2M) at room temperature as well as at 4 °C. ## 2.4. Mucoadhesive Properties The adsorption of mucin on the surface of LIPs or NEs was used as a method to assess the mucoadhesive properties of the BNN27-loaded LIPs and NEs [13]. For LIPs, 1 mL of mucin aqueous solution (0.5 mg/mL) was mixed (vortexed) with an equal volume of each LIP dispersion (lipid concentration at 2 mg/mL) at room temperature and the dispersions were centrifuged at 15,000 rpm for 30 min. Free mucin was measured in the supernatant. PC and PC/PG (negatively charged) LIPs were studied, before and after coating. The same protocol was used for the measurement of the mucoadhesive properties of NEs, after the NEs were diluted in order to eliminate any turbidity that would affect the measurements. For measurement of the free mucin in the supernatants, the Bradford colorimetric method was used [22]. Samples (and standard solutions) were incubated for 20 min at 37 °C after the addition of the Bradford reagent, and then absorbance at 595 nm was measured (Shimatzu UV-1205 spectrophotometer). A mucin calibration curve was prepared by measuring the mucin standard solutions, and the mucin content of each sample was calculated from the calibration curve. Finally, the amount of mucin adsorbed on the samples was calculated as the difference between the total and free mucin. ## 2.5.1. Cytotoxicity Assessment Immortalized human brain microvascular endothelial cells (hCMEC/D3) as well as human embryonic kidney cells (HEK) were used. HEK cells were grown in a high glucose DMEM medium supplemented with $10\%$ FBS and $1\%$ antibiotic-antimycotic solution (Invitrogen, Carlsbad, CA, USA). The cells were cultured at 37 °C, $5\%$ CO2/saturated humidity. The medium was changed every 2–3 days. hCMEC/D3 cells (passage 25–35) were obtained under license from the Institut National de la Sante et de la RechercheMedicale, INSERM, Paris, France and grown in EndoGRO medium (Merck, Darmstadt, DE) supplemented with 10 mM HEPES, 1 ng/mL basic FGF (bFGF), 1.4 μM hydrocortisone, 5 μg/mL ascorbic acid, penicillin-streptomycin, chemically defined lipid concentrate, and $5\%$ ultralow IgG FBS. All cultureware was coated with 0.1 mg/mL rat tail collagen type I (BD Biosciences, Franklin Lakes, NJ, USA). The cytotoxicity of the liposomal and nanoemulsion samples toward the hCMEC/D3 and HEK cells was evaluated with the MTT assay. Briefly, 25,000 cells were seeded in collagen pre-coated 24-well plates and after overnight incubation, the medium was replaced with the amount of each sample required to confer 1 μM BNN27, and incubated at 37 °C and in $5\%$ CO2 for 48 h. After completion of the cell/vesicle incubations, MTT solution was added in all samples and after 2 h (for HEK cells) or 4 h (for hCMEC/D3 cells), acidified isopropanol was used to dissolve the formazan crystals that were formed. Viable cells (%) were calculated based on the equation: (A620 sample-A620 background)/(A620 control-A620 background) × 100, where the A620 control is the OD-620 nm of untreated cells, and the A620 background is the OD-620 nm of MTT without cells. ## 2.5.2. Cell-Monolayer Permeation Studies For the monolayer studies, hCMEC/D3 cells were seeded on Transwell filters (polycarbonate six-well, pore size 0.4 μm; Millipore Merck, Darmstadt, DE) pre-coated with type I collagen, at 5 × 104 cells/cm2. The detailed procedure followed, and the tests that were carried out to verify that the monolayer produced was intact (transendothelial electrical resistance measurements (TEER) and Lucifer yellow (LY) permeability calculation) are described in detail elsewhere [14,23]. The permeability of the samples was determined after the preparation of RHO-labeled nanoformulations. After adding the samples to the top of the permeation filter (0.2 μM RHO per filter), the transport of RHO was calculated by the fluorescence intensity measurements ($\frac{540}{585}$) of the samples taken from the basolateral portion at selected time points (10, 30, 60, 90, and 120 min). For the extraction of RHO-lipid from the formulations, the Folch method was applied, as previously reported [24]. ## 2.6. Transmission Electron Microscopy (TEM) LIPs (0.5–1 mg/mL) were re-suspended in 10 mM HEPES (to eliminate potential artifacts from phosphate salts) while NEs were diluted in water. Then, all types of samples were negatively stained with $1\%$ phosphotungstic acid in dH2O (freshly prepared), washed three times with dH2O, drained with the tip of a tissue paper, and observed at 100,000 eV with a JEM-2100 (Jeol, Tokyo, Japan) transmission electron microscope (TEM) [25]. ## 2.7. In Vivo Studies For the in vivo study, C57BL/6J 8-week old mice were utilized. Animals were housed and maintained in a 12-h light/dark cycle and fed ad libitum. All procedures were performed according to the European Union policy (Directive $\frac{86}{609}$/EEC) (carried out in compliance with Greek Government guidelines) and institutionally approved protocols (Veterinary Directorate of Prefecture of Heraklion (Crete) and FORTH ethics committee (License number: EL91-BIOexp-02)). Mice were anesthetized using an intraperitoneal injection of a ketamine (Nerketan 10, 100 mg/mL) and xylazine (Xylapan, 20 mg/mL) cocktail. Once the hind-limp was lost, the mice were fixed in the supine position, 25 µL of each formulation was administered to each mouse in 60 s intervals, via 1–2 µL dose alternatively into each nostril [26]. Animals were sacrificed 60 and 120 min after the completion of the administration. Brains were quickly dissected out of the cranium; excess blood was wiped off, and the brain was snap frozen until further processing. After weighing, brain samples were homogenized in 300 µL ice-cold distilled water/methanol solution ($\frac{25}{75}$ v/v) and sonicated for 20 min at 4 °C. Next, three volumes of ice-cold acetonitrile were added, followed by sonication for 10 min and centrifugation at 14,000× g/15 min/4 °C. After an extra addition of 100 µL ice-cold acetonitrile to the supernatant, a 10 min-sonication and centrifugation at 14,000× g/15 min/4 °C were carried out. The supernatant was vacuum-dried at a SpeedVac without heating. Prior to analysis, samples were stored at −80 °C. To quantify the BNN27 levels, the detection limit of the enzymatic method used for BNN27 quantification in the formulations (>2 ppm) was not low enough; therefore, a liquid chromatography mass spectrometry (LC-MSn) method was developed using deuteriated pregnenolone (pregnenolone 17,21,21,21-D4) as the internal standard (70 ng/mL). The analysis was performed on an LTQ-Orbitrap Velos mass spectrometer (MS) (Thermo Fisher Scientific, Bremen, Germany) connected to an Accela ultra-high-performance LC (UHPLC) system. An Acquity UPLC BEH C18 VanGuard pre-column (130 Å, 1.7 µm, 2.1 mm × 100 mm) coupled to an Acquity UPLC BEH C18 column (130 Å, 1.7 µm, 2.1 mm × 5 mm) was used. Quality control samples were prepared at three concentrations (low, medium, high) to monitor the instruments’ performance and chromatographic integrity over time. Monitoring occurred in positive ion mode. The standard curve concentration range was 1–2000 ng/mL ($Y = 0.000546519$ + 0.000897604∙X; R2 = 0.9673; W:1/x). The injection volume was set at 5 µL, and the mobile phase flow rate was set at 0.2 mL/min. Mobile phase solvents were A ($95\%$ H2O, $5\%$ methanol, $0.1\%$ formic acid) and B (methanol, $0.1\%$ formic acid). The eluting gradient program was the following: 0–0.1 min ($40\%$ A, $60\%$ B), 0.1–0.5 min ($20\%$ A, $80\%$ B), 0.6–6.5 min ($5\%$ A, $95\%$ B), 6.51–8.0 min ($40\%$ A, $60\%$ B). Data were processed with Xcalibur software (version 2.1, Thermo Scientific, Waltham, MA, USA) and data analysis was conducted using the R programming language. In order to calculate the BNN27 levels in brain tissue, expressed as administered dose percent (ID%), the following formula was applied: ID%/g brain = (Xbrain/X in dose) × 100; where: Xbrain = BNN27 (mg) per g of weighted brain tissue, and X in dose = BNN27 (mg) in 25 µL solution for intranasal administration. ## 2.8. Statistical Analysis All results were expressed as mean ± S.D from at least three independent experiments. Most data were analyzed by using one way ANOVA followed by the Bonferroni post hoc test. $p \leq 0.05$ was considered statistically significant for all comparisons. When more factors were compared, two-way ANOVA was performed. The significance of comparisons is presented in the graphs as: * $p \leq 0.05$, $p \leq 0.01$, * $p \leq 0.001$, **** $p \leq 0.0001.$ ## 3.1. BNN-Loaded LIPs and Chitosan-Coated LIPs All calibration curves constructed both in the absence and presence of empty LIPs were linear; proving that the quantification method applied resulted in the accurate determination of BNN27 in LIPs (see Supplementary Materials Figure S1). ## 3.1.1. BNN Loading in LIPs The BNN27 content of each LIP type was calculated from the appropriate calibration curve, after mixing equal volumes of the liposomal sample and pure ethanol, and applying the method described above. The results of studies carried out to optimize BNN27 loading in LIPs can be seen in Figure 1a. As seen, BNN27 loading could not confer a D/L (mol/mol) ratio higher than 0.056 that was realized when using an initial D/L ratio equal to 0.1 (mol/mol). Furthermore the loading efficiency was not significantly modified by using different initial D/L ratios, between 0.05 and 0.167. Concerning the lipid membrane composition effect on BNN27 loading in LIPs, as seen from the results in Figure 1b, the addition of negative charged lipids (PG) resulted in a significant decrease in BNN27 incorporation into the LIPs. It is well-known that the composition of LIP lipid membranes determines the lipophilic drug partitioning/incorporation in LIPs, since these drugs are incorporated in the lipid membrane. Regardless of the decrease in BNN27 loading, the addition of the negative charge is very important for liposomal surface modification such as coatings with chitosan, as demonstrated previously and verified by the results presented in the following section. From the Figure 1b results, the nanosize of both types (lipid membrane compositions) of BNN27 loaded LIPs was confirmed as well as their narrow size distribution. As anticipated, the addition of PG in the LIP membrane conferred a significant increase or the vesicle negative zeta potential. ## 3.1.2. Coating of BNN27-Loaded LIPs with Chitosan The coating efficiency of chitosan-LIPs (CHT-Lip), together with the physicochemical properties of the various LIP types prepared, is presented in Table 1. As demonstrated, the vesicle coating efficiency was substantially higher in the case of PC/PG LIPs compared to PC LIPs, irrespective of the amount/type of CHT used. This is in agreement with a previous report that the CHT coating of LIPs cannot be further increased above the 0.1 w/w chitosan/lipid ratio [13]. Indeed, when the medium molecular weight (MMW) chitosan was used, vesicles with the lowest size (1147 ± 86 nm) and polydispersity index (PDI = 0.447) were formed at 0.1 w/w CHT/lipid ratio. Higher CHT amounts resulted in the formation of liposomal clusters with large sizes and high PDI values. Furthermore, the vesicle ζ-potential value, which is used as a measure of the vesicles’ mucoadhesive capacity, was not significantly increased when higher than 0.1 CHT/lipid ratios were used (Table 1); thereby we continued our experiments with LIPs that were coated using 0.1 w/w CHT/lipid. Uncharged PC vesicles had much lower CHT coating efficiencies (Table 1) compared to the negatively charged PC/PG LIPs. It was previously explained that in addition to the electrostatic interactions between positively charged chitosan and negatively charged LIPs, the coating process is regulated by a number of other mechanisms such as the formation of hydrogen bonds between the hydrogen of the polysaccharide and the nitrogen groups of the polar head of PC [27]. Nevertheless, differences between PC LIPs that were coated with LMW chitosan and MMW chitosan (concerning their size distribution and zeta-potential) were not significant, most probably due to the very low amount of CHT adhered on these non-charged vesicles. In contrast, the molecular weight of CHT had a more obvious effect when negatively charged LIPs were studied. Indeed, the variations in the mean hydrodynamic diameter of the LIPs and ζ-potential values, between vesicles that were coated with LMW CHT and MMW CHT, were statistically significant ($p \leq 0.0001$). The coating of vesicles with MMW chitosan resulted in increased liposomal size and ζ-potential values approx. by 1.5 times compared to the coating with LMW CHT, but most importantly, the vesicle zeta-potential values were similarly increased (Table 1). For the latter reason, MMW CHT coated LIPs were used in all of the following studies. Thereby, the LIPs used in the following studies were PC/PG/BNN27 CHT-LIPs with MMW CHT. ## 3.2.1. BNN27 Solubility Studies—Selection of NE Ingredients The solubility of practically insoluble (in water) BNN27 in selected potential ingredients of NEs is reported in Figure 2. As seen, between the oils tested (Figure 2a), the highest solubility was measured in Capmul MCM, so this oil was selected as the oil phase for NEs. Concerning the solubility of BNN27 in surfactants (Figure 2b), the highest solubility was measured in Tween 20 and Tween 80; therefore, we selected to study both surfactants in order to select the surfactant that acquired the optimal NE properties. Finally, between the potential co-surfactants tested (Figure 2c), very high BNN27 solubilities were found in both the Transcutol and Transcutol/propylene glycol (1:1) mixture. In fact, we selected to use the Transcutol/PG mixture, since it has been previously reported to impart high stability and lower globule size (compared to other co-surfactants) to the NEs [20]. ## 3.2.2. Selection of Optimal Surfactant and Composition for NE The type of NE formed depends on the properties of the oil, surfactant, and co-surfactant. Many surfactants cannot lower the oil–water interfacial tension sufficiently to form NEs, so co-surfactant addition is necessary. Co-surfactants additionally ensure that the interfacial film is flexible enough to deform readily around each droplet as their intercalation between the primary surfactant molecules decreases both the polar head group interactions [18]. In this study, Tween 20 or Tween 80 was selected as potential surfactants and Transcutol/PG (1:1) as the co-surfactant system (Smix). Previously, by the construction of ternary phase diagrams, it was demonstrated that when Tween 80 was used a surfactant with the same oil phase (Capmul MCM) and co-surfactant system (Smix), the optimal surfactant/co-surfactant (Smix) ratio was 4:2 [18]. Herein, we sought to confirm if the same ratio also applied when Tween 20 was used as the surfactant. Ternary phase diagrams were constructed by varying the Tween 20/Smix ratios to be: 1:2, 2:2, 3:2, and 4:2 (see Supplementary Materials Figure S2), for the selection of the ingredient amounts as required in order to avoid metastable formulations. From the ternary diagrams, it was confirmed that the 4:2 surfactant/Smix ratio was also optimal for Tween 20. Therefore, the optimal NE formulations that were selected for further experiments consisted by Capmul:surfactant + Smix:water at $\frac{8}{44}$/48 and $\frac{10}{44}$/46 ratios, where Smix consists of (Tween 80 or Tween 20):Transcutol:propylene glycol at a $\frac{4}{1}$/1 ratio. After confirming that the 4:2 ratio (for surfactant/co-surfactants) was also optimal when Tween 20 was used as the surfactant, several NEs using $8\%$ Capmul MCM as the oil, and increasing concentrations of the surfactant (S) + co-surfactants (Smix) between 24 and up to $44\%$, were prepared in order to select the composition that acquired the optimal NE properties. As seen in Figure 3, composition B6 had the best properties with regard to the globule size distribution and transmittance (%). Thereby, we decided to use $44\%$ concentration of S + Smix. Comparison of various NE formulations with Tween 80 or Tween 20 as the surfactant was then realized (after preparing various formulations) for the identification of the best surfactant. The results of their physicochemical properties and short-term physical stability study are presented in Figure 4. As seen in Figure 4a, the NEs with Tween 80 had a significantly lower globule size for all oil concentrations tested, although their transmittance (%) was not affected by the different surfactants. Furthermore, Tween 80-containing NEs demonstrated higher stability (Figure 4c) during a preliminary 8 day at 25 °C stability study compared to the corresponding Tween 20-containing-NEs (Figure 4b), where a significant increase ($p \leq 0.001$) in globule size was observed after 8 day of storage at room temperature. In both types of NEs (with Tween 80 and Tween 20), when the oil concentration was increased so did the NE globule size. The last experiment showed that Tween 80-containing NEs had better properties and stability, and thereby Tween 80 was used as the surfactant in the following studies. For the latter decision, it was also taken into consideration that Tween 80 belongs to the class of non-ionic surfactants and is widely used since it is less toxic compared to ionic surfactants, and additionally, it is less affected by pH and ionic strength [19]. ## 3.2.3. Properties of BNN27-Loaded NEs After the selection of the ingredients and the optimal S + Smix percent, BNN27-loaded NEs (BNEs) were formulated using the method described above. Two types of BNN27-loaded NEs were prepared: one with an oil phase of $8\%$ (w/w) and the other of $10\%$ (w/w), which were evaluated for their properties and stability. In general, a NE exhibits the characteristics of its external phase. There are several techniques for identifying the type of emulsion. Dilution studies are based on the fact that emulsions are soluble only in the liquid that forms their continuous phase. When diluted with water, no change was observed in the BNEs’ droplet size and clarity, indicating that the BNEs are oil-in-water emulsions. Additionally, neither phase separation nor creaming was observed after centrifugation of the NEs, suggesting the stability of the systems. The physicochemical properties and quantitative test data of the formulated BNEs are shown in Figure 5a. The pH of all BNEs is between 4.05 and 5.73, which is within the previously considered normal pH range of nasal fluid (3.5–6.4) [28]. However, it has recently been suggested that the pH of the nasal cavity is restricted in the range of 5.5–6.5 [29,30]. Thereby, the CAR-BNEs may most probably cause irritation following instillation on the nasal mucosa, while the CHT-coated ones may be considered as marginal. This is an issue that should be further explored in future studies. The high degree of transparency of the non-coated BNEs verified that clear dispersions were formulated, while the CHT or CAR coated NEs had lower transmittance percentages due to the contribution of these components to turbidity. The low PDI of the non-coated BNEs indicates that they are a monodispersed system (Figure 5a). All types of BNEs were also tested for drug content and found to demonstrate high BNN27 loading, ranging between approx. $90\%$ and $99\%$ of the amount of BNN27 used for their preparation (Figure 5a). In the stability studies, the BNEs exhibited no precipitation of drug, creaming, phase separation, or flocculation on visual observation, and were found to be stable after centrifugation. When stored at 25 °C as well as at 4 °C (Figure 5b–e), only negligible changes in the quantitative parameters of the BNEs containing the $10\%$-oil phase were observed after 2 months of storage (Figure 5d,e). In contrast, the BNEs with the $8\%$-oil phase exhibited significant increases in their mean globule size (Figure 5b,c), especially during storage at 25 °C. Similar stability studies were also carried out for the CHT- and CAR-coated BNEs (see Supplementary Materials Figure S3), where it was demonstrated that the BNEs with the $10\%$-oil phase were more stable for both types of coated BNEs compared with the corresponding BNE types with the $8\%$-oil phase. For this reason, BNN27-loaded NEs with $10\%$ (w/w)-oil content (Capmul MCM) were used for the next in vitro and in vivo studies. The higher amount of oil phase in the BNEs would also provide the capability to load higher amounts of BNN27. Furthermore, it was previously reported the Capmul improved the NE-loaded drug permeation through the BBB barrier [21]. Additionally, the drug content of all BNE types studied for their stability was measured at all time points, and no significant decrease was demonstrated in the BNN27 content (or else no drug leakage occurred) in any of the tested BNE types (see Supplementary Materials Figure S4). ## 3.3. Cytotoxicity Evaluation The cytotoxicity of the prepared NEs toward the hCMEC/D3 and HEK-294 cells was evaluated by the MTT method. As seen in Figure 6, no toxicity was observed in any of the cell lines studied after 48 h of incubation with BNEs. ## 3.4. TEM Morphology Transmission electron microscopy was performed to complete the characterization of the mucoadhesive nanoformulations. in Figure 7a, a polymeric chitosan membrane formed around the liposomal vesicles can be observed. It is interesting to highlight the different LIP mean-diameters observed by TEM and measured by DLS (which measures the hydrodynamic diameter), the last being more than two times smaller. For the BNEs, the TEM micrographs confirmed the differences in size of CHT-BMNEs with different percentages of the oil phase, as seen in Figure 7b,c. The latter differences are in agreement with the DLS measurements of the corresponding NE-types (Figure 5a). ## 3.5. Mucoadhesive Properties Mucoadhesive properties were calculated as the percentage of mucin attached to LIPs or NEs. As seen in Figure 8, uncoated vesicles and NEs showed a small nonspecific mucoadhesion. As anticipated, PC/PG CHT-LIP is characterized by a strong positive surface charge (Table 1), which is why they exhibited significantly higher mucoadhesive properties compared to the PC CHT-LIPs. The uncoated PC/PG-LIPs were negatively charged and exhibited low mucoadhesive properties. The adhesion process is rather complex and several theories have been proposed to explain the adhesion of polymeric materials [31,32,33,34]. Diffusion is one of the main theories proposed to describe mucoadhesion, which includes the action of polymer-chain entanglement [35]. The diffusion theory states that inter-penetration of the polymer and mucin chains may lead to prolonged mucosal and mechanical adhesion. According to the diffusion theory, we expect that since the MW of MMW-CHT is nearly 10 times greater than that of LMW-CHT, the contribution of the natural entanglement to the adhesion between MMW-CHT and mucin should be much stronger than that between LMW-CHT and mucin. Indeed, as seen in Figure 8a, MMW-CHT conferred significantly higher mucoadhesive properties to the PC/PG coated LIPs compared to LMW-CHT. High molecular weight chitosan (HMW) was not used in our studies because it has been previously shown to confer lower mucoadhesive properties than MMW, a result attributed to the higher probability of the very long chains of HMW-CHT to bend, leading to fewer available amino groups, and thus providing a lower positive charge for interaction [36]. Concerning the mucoadhesive properties of the coated BNEs, as seen in Figure 8b, both polymers, CHT and CAR, resulted in a significant increase in the mucoadhesive properties of non-coated BNEs, however, the mucoadhesive properties of the CHT-coated BNEs were about two times higher than those of the corresponding CAR-coated BNEs. Indeed, it has also been previously observed that the mucosal capacity of CHT is significantly higher than the relevant properties of CAR [37,38]. Finally, the oil content (percent) of the BNEs did not seem to have any effect on the mucoadhesive properties of the BNEs, at least for BNEs with $8\%$ or $10\%$ (w/w) oil. ## 3.6. In Vitro and In Vivo Studies The permeability of the formulation-incorporated RHO across a cellular model of the BBB as well as the brain disposition of BNN27 following the intranasal administration of BNN-loaded formulation to mice were evaluated in the last part of the current study, in order to compare the various types of formulations. It should be clarified at this point that the hCMEC/D3 monolayer is considered as a model of “intact” BBB (with tight connections between the cells), while the nasal CNS barrier is proposed to be more ‘‘leaky’’ because of continuous neuron turnover [29,39]. Nevertheless, we thought that it would be interesting to conduct comparative studies of the two formulation types under identical conditions in the hCMEC/D3 model and also after intranasal administration (in vivo). ## 3.6.1. In Vitro Permeability across hCMEC/D3 Monolayer For the permeability study, formulations incorporating 1 mol% RHO as a lipophilic drug model were used. LIP (PC/PG) and CHT-coated LIP (coated with MMW CHT) as well as NE and CHT-NE with $10\%$ Capmul were prepared; the physicochemical properties of the formulations are presented in Figure 9a. During monolayer formation, the TEER of the monolayers was measured and was found to gradually increase from 31.5 Ω × cm2 (at day 3) to 111.1 Ω × cm2 (after simvastatin treatment), in agreement with previous results [23,24]. Lucifer yellow (LY, a marker for paracellular transport) was added in all of the monolayer experiments (wells) and its permeability was measured as a method to identify any toxicity of the samples toward the monolayer integrity, which would thus impair the accuracy of the transport results. As seen in Figure 9a, LY permeability was practically un-influenced by the samples, ranging between 1.06 × 10 3 and 1.24 × 10 3 cm/min for all samples, being in good agreement with the previously reported values [23,24]. In preliminary cytotoxicity studies, it was confirmed that all the formulations used were non-cytotoxic toward the hCMEC/D3 cells following 2 h incubation with the cells at a final RHO-lipid concentration of 0.2 µM (see Supplementary Materials Figure S5), as also seen for the BNN27-incorporating formulations at a 1 µM concentration after 48 h (Figure 6). However, the substantial difference in the cytotoxicity of the two formulation types is a matter that needs to be highlighted. Indeed, RHO-incorporating LIPs (non-coated and coated) did not confer any cytotoxicity toward the hCMEC/D3 cells at 5 µM concentration (of RHO) after 48 h (Figure S5a), whereas the RHO-incorporated NEs diminished the cells even at 10 times a lower concentration (of RHO) and only 2 h incubation (Figure S5b). Concerning the comparison between LIPs and NEs, interestingly, the translocation of NE-associated RHO and the corresponding permeability values were approx. three times higher for the corresponding values of LIPs. This is a very interesting finding, especially since such differences between LIPs and NEs have not been previously reported (to the best of our knowledge). Concerning the effect of coating, the transport of the coated-nanoformulations (Figure 9b) as well as the corresponding permeability of the formulation-associated RHO (Figure 9c) were found to be lower than the corresponding values of the control (non-coated) nanoformulations, for both formulation types (LIPs and NEs) (Figure 9b), although in some cases, the differences noted were not statistically significant. This fact can be attributed to the adhesion of the CHT LIPs or NE droplets to the monolayer, resulting in delayed permeability. Free RHO (micellar solution) did not show any difference compared to LIP-RHO (not shown), while compared to the NEs, its permeability was significantly lower, which is in good agreement with the previously reported results [40]. ## 3.6.2. BNN27 In Vivo Intranasal Delivery Studies The calibration curve used for the calculation of the BNN27 concentration in the brain samples is presented in Figure S6 (see Supplementary Materials). As seen in Figure 10, the brain disposition of BNN27 after intranasal administration of BNN27 nanoformulations to C57BL/6J mice showed that CHT-BNEs achieved the highest levels of BNN27 in the brain compared to all of the other formulations used. Specifically, the BNN27 levels in the brain 1 h after nasal administration of CHT-BNE were more than 3.5 times higher than the corresponding concentrations after the administration of LIPs (coated or non-coated) or non-coated BNEs. At 2 h post-administration, the differences in brain disposition were lower, since all other formulations demonstrated increased brain concentrations of BNN27 at this time point (compared to the 1 h time point), except for CHT-BNE, but the BNN27 concentration after administration of CHT-BNEs was still more than 2-fold higher compared to all other nanoformulations (Figure 10b). As demonstrated, 2 h post-administration, the non-coated BNEs achieved a higher BNN27 brain concentration compared to non-coated LIPs (although the difference was not statistically significant); however, nasal administration of CHT-coated nanoformulations resulted in higher brain concentrations of BNN27 in most cases, with the exception of LIPs at 1 h post-administration. Another very interesting observation from the data of Figure 10b is the very rapid brain disposition of CHT-BNE-associated-BNN27, which reached the highest BNN27 brain concentration already 1 h post-administration, in contrast to all of the other nanoformulations that demonstrated the highest BNN27 brain concentrations 2 h post-administration. It has been reported that the natural biodegradable polymer chitosan may enhance the penetration and absorption of drugs through the nasal mucosa and may also delay mucociliary clearance [41], thus increasing the absorption of drugs after intranasal delivery. Additional advantages of coating with CHT are its excellent biocompatibility as well as the fact that it is a well-tolerated polymer [42]. Several studies have confirmed a double role for CHT as a coating of NEs, proving that chitosan-coated NEs conferred the highest fluxes and nasal mucosa permeability compared to the corresponding uncoated NEs [10]. ## 4. Conclusions Herein, two types of nanoformulations, LIPs and NEs, were developed using biocompatible ingredients, optimized for drug loading, stability, and mucoadhesive properties, and evaluated as nanoformulations for the nose-to-brain delivery of BNN27. The effect of the CHT-coating was evaluated for both formulation types. The results showed that NEs (consisted of Capmul MCM, Tween 80, Transcutol, and propylene glycol) had significantly higher in vitro BBB-crossing capability compared to LIPs (consisted of PC and PG) (Figure 9); however, the CHT-coating resulted in a slight decrease in the BBB monolayer transport for both formulation types, perhaps due to the adhesion of CHT-coated vesicles/globules on the apical side of the monolayer, thus preventing their transport to the basal side. In agreement with the monolayer permeability results, BNEs demonstrated significantly higher brain disposition of BNN27 compared to the BNN27-loaded LIPs (Figure 10). This result may be attributed to the previously reported effect of Capmul MCM to enhance drug permeability across the nasal mucosa and thus the brain disposition of drugs [21,43]. Furthermore, in vivo, the CHT-coating conferred enhanced brain concentrations of BNN27 for both formulation types (NEs and LIPs). In the case of LIPs, the effect of the CHT-coating became evident only at 2 h post-administration, suggesting that prolonged retention of the coated LIPs at the site of administration (nasal mucosa) is probably implicated in the higher BNN27 brain amounts, as also proposed in other cases [44]. In fact, it has previously been reported that CHT increased NE permeability, which was attributed to its penetration enhancing properties by the transient opening of tight junctions [45]. Higher permeability of a CHT NE formulation was also suggested to explain the higher AUC and shorter Tmax of the NE-associated drug (in the brain) compared to i.v. or nasal administration of the drug solution [46]. Interestingly, CHT-BNE demonstrated very high brain disposition of BNN27 already at 1 h post-administration, which was not the case for the non-coated BNE. The latter result suggests that perhaps the presence of CHT on the much smaller globules of the BNE (compared to LIPs) has a more direct effect on the permeability and brain disposition of the drug, and not only on the retention of the drug on the nasal mucosa. The particle size of the formulations may also be important. Indeed, it has been suggested that particle size and not the CHT coating is the most important determining factor for NE brain disposition following nasal delivery, and that formulations with sizes around 100 nm demonstrate prolonged residence in the nasal cavity, whereas NEs of larger sizes have faster clearance. It was additionally proven by imaging methods that NEs with globule sizes of 100 nm were transported through the trigeminal and the olfactory nerves to the olfactory bulb, while larger NE transport through the nose-to-brain route was lower due to higher mucociliary clearance [47]. This last report could perhaps explain the significant difference between the CHT LIPs and CHT BNE regarding the BNN27 brain disposition. However, as highlighted above, it has to be emphasized that the hydrodynamic mean diameters of liposomes (especially the ones coated with CHT) measured by DLS were much larger than their actual sizes, as proven from the TEM micrographs (Figure 7a). Nevertheless, the CHT LIPs had significantly larger diameters compared to the CHT BNEs (Figure 7), suggesting that this may result in higher mucociliary clearance and reduced transport through the nose-to-brain route, according to previous findings [47]. A direct comparison of LIPs and NEs for nose-to-brain delivery of drugs has never been evaluated before; therefore, we cannot discuss the current results with respect to previous studies. We identified two studies reporting a direct comparison of LIPs and NEs for other drug delivery applications: one that compared the skin delivery of retinyl palmitate [48], and another in which the two formulation types were compared for the brain delivery of melatonin after i.v. injection [49]. In the first report, the cumulative amount of drug that permeated through human skin after 38 h was 6.67 ± 1.58 mg, and 4.36 ± 0.21 mg for NEs and LIPs, respectively, while the NE flux was significantly higher than that of the LIPs. In the same study, it was reported that the LIPs resulted in significantly higher skin retention compared to NEs, while the NEs disrupted the skin more [48]. In the second study, although the melatonin bioavailability was similar for LIPs and NEs, a NE formulation containing medium chain triglycerides as the oil phase resulted in a higher fraction of animals that reached the critical concentration of the drug in brain extracellular fluid (and also faster) compared to the LIPs [49]. In summary, the CHT-BNE formulation developed herein was demonstrated to confer faster and higher nose-to-brain delivery of BNN27 compared to CHT-LIPs. Such NEs could be considered as alternative systems for the brain delivery of lipophilic drugs following intranasal administration. Nevertheless, extended biocompatibility and toxicity studies are required to exclude any potential toxicity issues in relation to the cytotoxicity differences between the LIPs and NEs above-mentioned due to the high surfactant content of NEs. ## References 1. Kokras N., Dioli C., Paravatou R., Sotiropoulos M.G., Delis F., Antoniou K., Calogeropoulou T., Charalampopoulos I., Gravanis A., Dalla C.. Psychoactive Properties of BNN27, a Novel Neurosteroid Derivate, in Male and Female Rats. Psychopharmacology (2020) 237 2435-2449. DOI: 10.1007/s00213-020-05545-5 2. Charalampopoulos I., Remboutsika E., Margioris A.N., Gravanis A.. Neurosteroids as Modulators of Neurogenesis and Neuronal Survival. Trends Endocrinol. Metab. (2008) 19 300-307. DOI: 10.1016/j.tem.2008.07.004 3. Calogeropoulou T., Avlonitis N., Minas V., Alexi X., Pantzou A., Charalampopoulos I., Zervou M., Vergou V., Katsanou E.S., Lazaridis I.. Novel Dehydroepiandrosterone Derivatives with Antiapoptotic, Neuroprotective Activity. J. Med. Chem. (2009) 52 6569-6587. DOI: 10.1021/jm900468p 4. Bennett J.P., O’Brien L.C., Brohawn D.G.. Pharmacological Properties of Microneurotrophin Drugs Developed for Treatment of Amyotrophic Lateral Sclerosis. Biochem. Pharmacol. (2016) 117 68-77. DOI: 10.1016/j.bcp.2016.08.001 5. Lisa S., Iban- Arias R., Kokona D., Charalampopoulos I., Gravanis A., Thermos K.. Effects of Novel Synthetic Microneurotrophins in Diabetic Retinopathy. Springerplus (2015) 4 L25. DOI: 10.1186/2193-1801-4-S1-L25 6. Tsika C., Tzatzarakis M.N., Antimisiaris S.G., Tsoka P., Efstathopoulos P., Charalampopoulos I., Gravanis A., Tsilimbaris M.K.. Quantification of BNN27, a Novel Neuroprotective Spiroepoxy Dehydroepiandrosterone Derivative in the Blood and Retina of Rodents, after Single Intraperitoneal Administration. Pharmacol. Res. Perspect. (2021) 9 e00724. DOI: 10.1002/prp2.724 7. Gravanis A., Pediaditakis I., Charalampopoulos I.. Synthetic Microneurotrophins in Therapeutics of Neurodegeneration. Oncotarget (2017) 8 9005-9006. DOI: 10.18632/oncotarget.14667 8. Agrawal M., Saraf S., Saraf S., Antimisiaris S.G., Chougule M.B., Shoyele S.A., Alexander A.. Nose-to-Brain Drug Delivery: An Update on Clinical Challenges and Progress towards Approval of Anti-Alzheimer Drugs. J. Control. Release (2018) 281 139-177. DOI: 10.1016/j.jconrel.2018.05.011 9. Bourganis V., Kammona O., Alexopoulos A., Kiparissides C.. Recent Advances in Carrier Mediated Nose-to-Brain Delivery of Pharmaceutics. Eur. J. Pharm. Biopharm. (2018) 128 337-362. DOI: 10.1016/j.ejpb.2018.05.009 10. Bonferoni M.C., Rossi S., Sandri G., Ferrari F., Gavini E., Rassu G., Giunchedi P.. Nanoemulsions for “Nose-to-Brain” Drug Delivery. Pharmaceutics (2019) 11. DOI: 10.3390/pharmaceutics11020084 11. Richmond W.. Preparation and Properties of a Cholesterol Oxidase from Nocardia Sp. and Its Application to the Enzymatic Assay of Total Cholesterol in Serum. Clin. Chem. (1973) 19 1350-1356. DOI: 10.1093/clinchem/19.12.1350 12. Allain C.C., Poon L.S., Chan C.S., Richmond W., Fu P.C.. Enzymatic Determination of Total Serum Cholesterol. Clin. Chem. (1974) 20 470-475. DOI: 10.1093/clinchem/20.4.470 13. Zaru M., Manca M.L., Fadda A.M., Antimisiaris S.G.. Chitosan-Coated LIPs for Delivery to Lungs by Nebulisation. Colloids Surf. B Biointerfaces (2009) 71 88-95. DOI: 10.1016/j.colsurfb.2009.01.010 14. Markoutsa E., Papadia K., Giannou A.D., Spella M., Cagnotto A., Salmona M., Stathopoulos G.T., Antimisiaris S.G.. Mono and Dually Decorated NanoLIPs for Brain Targeting, in vitro and in vivo Studies. Pharm. Res. (2014) 31 1275-1289. DOI: 10.1007/s11095-013-1249-3 15. Filipovic J., Enjak I.J.. Mucoadhesive Chitosan-Coated LIPs: Characteristics. J. Microencapsul. (2001) 18 3-12. DOI: 10.1080/026520401750038557 16. Stewart J.C.. Colorimetric determination of phospholipids with ammonium ferrothiocyanate. Anal. Biochem. (1980) 104 10-14. DOI: 10.1016/0003-2697(80)90269-9 17. Kumar M., Misra A., Babbar A.K., Mishra A.K., Mishra P., Pathak K.. Intranasal nanoemulsion based brain targeting drug delivery system of risperidone. Int. J. Pharm. (2008) 358 285-291. DOI: 10.1016/j.ijpharm.2008.03.029 18. Kumar M., Pathak K., Misra A.. Formulation and Characterization of Nanoemulsion-Based Drug Delivery System of Risperidone. Drug Dev. Ind. Pharm. (2009) 35 387-395. DOI: 10.1080/03639040802363704 19. Haider M.F., Khan S., Gaba B., Alam T., Baboota S., Ali J., Ali A.. Optimization of rivastigmine nanoemulsion for enhanced brain delivery: In-vivo and toxicity evaluation. J. Mol. Liq. (2018) 255 384-396. DOI: 10.1016/j.molliq.2018.01.123 20. Boche M., Pokharkar V.. Quetiapine Nanoemulsion for Intranasal Drug Delivery: Evaluation of Brain-Targeting Efficiency. AAPS PharmSciTech (2017) 18 686-696. DOI: 10.1208/s12249-016-0552-9 21. Mallick A., Gupta A., Hussain A., Aparajay P., Singh S., Singh S.K., Dev A.. Intranasal delivery of gabapentin loaded optimized nanoemulsion for augmented permeation. J. Drug Deliv. Sci. Technol. (2020) 56 101606. DOI: 10.1016/j.jddst.2020.101606 22. Bradford M.M.. A Rapid and Sensitive Method for the Quantitation of Microgram Quantities of Protein Utilizing the Principle of Protein-Dye Binding. Anal. Biochem. (1976) 72 248-254. DOI: 10.1016/0003-2697(76)90527-3 23. Poller B., Gutmann H., Krähenbühl S., Weksler B., Romero I., Couraud P.O., Tuffin G., Drewe J., Huwyler J.. The Human Brain Endothelial Cell Line HCMEC/D3 as a Human Blood-Brain Barrier Model for Drug Transport Studies. J. Neurochem. (2008) 107 1358-1368. DOI: 10.1111/j.1471-4159.2008.05730.x 24. Markoutsa E., Papadia K., Clemente C., Flores O., Antimisiaris S.G.. Anti-Aβ-MAb and dually decorated nanoLIPs: Effect of Aβ1-42 peptides on interaction with hCMEC/D3 cells. Eur. J. Pharm. Biopharm. (2012) 81 49-56. DOI: 10.1016/j.ejpb.2012.02.006 25. Franken L.E., Boekema E.J., Stuart M.C.A.. Transmission Electron Microscopy as a Tool for the Characterization of Soft Materials: Application and Interpretation. Adv. Sci. (2017) 4 1600476. DOI: 10.1002/advs.201600476 26. Kanazawa T., Fukuda M., Suzuki N., Suzuki T.. Novel Methods for Intranasal Administration Under Inhalation Anesthesia to Evaluate Nose-to-Brain Drug Delivery. J. Vis. Exp. (2018) 141 e58485. DOI: 10.3791/58485 27. Perugini P., Genta I., Pavanetto F., Conti B., Scalia S., Baruffini A.. Study on Glycolic Acid Delivery by LIPs and Microspheres. Int. J. Pharm. (2000) 196 51-61. DOI: 10.1016/S0378-5173(99)00439-1 28. Illum L Transport of Drugs from the Nasal Cavity to the Central Nervous System. Eur. J. Pharm. Sci. (2000) 11 1-18. DOI: 10.1016/S0928-0987(00)00087-7 29. Battaglia L., Panciani P.P., Muntoni E., Capucchio M.T., Biasibetti E., De Bonis P., Mioletti S., Fontanella M., Swaminathan S.. Lipid nanoparticles for intranasal administration: Application to nose-to-brain delivery. Expert Opin. Drug Deliv. (2018) 15 369-378. DOI: 10.1080/17425247.2018.1429401 30. Zha S., Wong K.L., All A.H.. Intranasal Delivery of Functionalized Polymeric Nanomaterials to the Brain. Adv. Health Mater. (2022) 11 e2102610. DOI: 10.1002/adhm.202102610 31. Campion R.P.. The Influence of Structure on Autohesion (Self-Tack) and Other Forms of Diffusion into Polymers. J. Adhes. (1975) 7 1-23. DOI: 10.1080/00218467508078895 32. Deryaguin B.V., Toporow Y.P., Mueler V.M.. On the Relationship between the Electrostatic and Molecular Component of the Adhesion of Elastic Particles to a Solid Surface. J. Colloid Interface Sci. (1997) 58 528-533. DOI: 10.1016/0021-9797(77)90162-X 33. Kaelble D.H.. A Surface Energy Analysis of Bioadhesion. Polymer (1977) 18 475-482. DOI: 10.1016/0032-3861(77)90164-1 34. Helfand E., Tagami Y.. Theory of the Interface between Immiscible Polymers. J. Chem. Phys. (1972) 57 1821-1823. DOI: 10.1063/1.1678491 35. Thongborisute J., Takeuchi H.. Evaluation of Mucoadhesiveness of Polymers by BIACORE Method and Mucin-Particle Method. Int. J. Pharm. (2008) 354 204-209. DOI: 10.1016/j.ijpharm.2007.12.001 36. Sun Y., Cui F., Shi K., Wang J., Niu M., Ma R.. The Effect of Chitosan Molecular Weight on the Characteristics of Spray-Dried Methotrexate-Loaded Chitosan Microspheres for Nasal Administration. Drug Dev. Ind. Pharm. (2009) 35 379-386. DOI: 10.1080/03639040802395185 37. Karn P.R., Vanić Z., Pepić I., Škalko-Basnet N.. Mucoadhesive Liposomal Delivery Systems: The Choice of Coating Material. Drug Dev. Ind. Pharm. (2011) 37 482-488. DOI: 10.3109/03639045.2010.523425 38. Takeuchi H., Matsui Y., Yamamoto H., Kawashima Y.. Mucoadhesive Properties of Carbopol or Chitosan-Coated LIPs and Their Effectiveness in the Oral Administration of Calcitonin to Rats. J. Control. Release (2003) 86 235-242. DOI: 10.1016/S0168-3659(02)00411-X 39. Dhuria S.V., Hanson L.R., Frey W.H.. Intranasal delivery to the central nervous system: Mechanisms and experimental considerations. J. Pharm. Sci. (2010) 99 1654-1673. DOI: 10.1002/jps.21924 40. Kadakia E., Shah L., Amiji M.M.. Mathematical Modeling and Experimental Validation of Nanoemulsion-Based Drug Transport across Cellular Barriers. Pharm. Res. (2017) 34 1416-1427. DOI: 10.1007/s11095-017-2158-7 41. Rinaldi F., Oliva A., Sabatino M., Imbriano A., Hanieh P.N., Garzoli S., Mastroianni C.M., De Angelis M., Miele M.C., Arnaut M.. Antimicrobial Essential Oil Formulation: Chitosan Coated Nanoemulsions for Nose to Brain Delivery. Pharmaceutics (2020) 12. DOI: 10.3390/pharmaceutics12070678 42. Salem L.H., El-Feky G.S., Fahmy R.H., El Gazayerly O.N., Abdelbary A.. Coated Lipidic Nanoparticles as a New Strategy for Enhancing Nose-to-Brain Delivery of a Hydrophilic Drug Molecule. J. Pharm. Sci. (2020) 109 2237-2251. DOI: 10.1016/j.xphs.2020.04.007 43. Jojo G.M., Kuppusamy G., De A., Karri V.V.S.N.R.. Formulation and optimization of intranasal nanolipid carriers of pioglitazone for the repurposing in Alzheimer’s disease using Box-Behnken design. Drug Dev. Ind. Pharm. (2019) 45 1061-1072. DOI: 10.1080/03639045.2019.1593439 44. Formica M.L., Real D.A., Picchio M.L., Catlin E., Donnelly R.F., Paredes A.J.. On a highway to the brain: A review on nose-to-brain drug delivery using nanoparticles. Appl. Mater. Today (2022) 29 101631. DOI: 10.1016/j.apmt.2022.101631 45. Ahmad N., Ahmad R., Alam A., Samim M., Iqbal Z., Ahmad F.. Quantification and evaluation of thymoquinone loaded mucoadhesive nanoemulsion for treatment of cerebral ischemia. Int. J. Biol. Macromol. (2016) 88 320-332. DOI: 10.1016/j.ijbiomac.2016.03.019 46. Abdou E.M., Kandil S.M., El Miniawy H.M.. Brain targeting efficiency of antimigrain drug loaded mucoadhesive intranasal nanoemulsion. Int. J. Pharm. (2017) 529 667-677. DOI: 10.1016/j.ijpharm.2017.07.030 47. Ahmad E., Feng Y., Qi J., Fan W., Ma Y., He H., Xia F., Dong X., Zhao W., Lu Y.. Evidence ofnose-to-brain delivery of nanoemulsions: Cargoes but not vehicles. Nanoscale (2017) 9 1174-1183. DOI: 10.1039/C6NR07581A 48. Clares B., Calpena A.C., Parra A., Abrego G., Alvarado H., Fangueiro J.F., Souto E.B.. Nanoemulsions (NEs), LIPs (LPs) and solid lipid nanoparticles (SLNs) for retinyl palmitate: Effect on skin permeation. Int. J. Pharm. (2014) 473 591-598. DOI: 10.1016/j.ijpharm.2014.08.001 49. Molska A., Nyman A.K.G., Sofias A.M., Kristiansen K.A., Hak S., Widerøe M.. In vitro and in vivo evaluation of organic solvent-free injectable melatonin nanoformulations. Eur. J. Pharm. Biopharm. (2020) 152 248-256. DOI: 10.1016/j.ejpb.2020.05.003
--- title: 'Face Mask: As a Source or Protector of Human Exposure to Microplastics and Phthalate Plasticizers?' authors: - Jiong Cao - Yumeng Shi - Mengqi Yan - Hongkai Zhu - Shucong Chen - Ke Xu - Lei Wang - Hongwen Sun journal: Toxics year: 2023 pmcid: PMC9967050 doi: 10.3390/toxics11020087 license: CC BY 4.0 --- # Face Mask: As a Source or Protector of Human Exposure to Microplastics and Phthalate Plasticizers? ## Abstract Wearing masks has become the norm during the Coronavirus disease pandemic. Masks can reportedly interface with air pollutants and release microplastics and plastic additives such as phthalates. In this study, an experimental device was set up to simulate the impact of five kinds of masks (activated-carbon, N95, surgical, cotton, and fashion masks) on the risk of humans inhaling microplastics and phthalates during wearing. The residual concentrations of seven major phthalates ranged from 296 to 72,049 ng/g (median: 1242 ng/g), with the lowest and the highest concentrations detected in surgical (median: 367 ng/g) and fashion masks (median: 37,386 ng/g), respectively. During the whole inhalation simulation process, fragmented and 20–100 μm microplastics accounted for the largest, with a rapid release during the first six hours. After one day’s wearing, that of 6 h, while wearing different masks, 25–135 and 65–298 microplastics were inhaled indoors and outdoors, respectively. The total estimated daily intake of phthalates with indoor and outdoor conditions by inhalation and skin exposure ranged from 1.2 to 13 and 0.43 to 14 ng/kg bw/d, respectively. Overall, surgical masks yield a protective effect, while cotton and fashion masks increase human exposure to microplastics and phthalates both indoors and outdoors compared to no mask wearing. This study observed possible risks from common facemasks and provided suggestions to consumers for selecting suitable masks to reduce exposure risks from microplastics and phthalate acid. ## 1. Introduction Since the outbreak of Coronavirus disease (COVID-19), more than 500 million people have been infected, mainly by droplets and aerosol transmission [1,2]. The World Health Organization recommends mask-wearing as one of the best safeguard measures for breaking the transmission chain of the virus during the pandemic [3]. It has been reported that the attributable risk was about six times higher in non-mask wearers than that of mask wearers [4]. Accordingly, mask-wearing has become a norm, with 129 billion masks used every month in 2020 [5]. Although surgical masks, N95 respirators, and similar masks yield a more significant protective effect than other masks [6], they are often costly. As a result, activated carbon masks, cotton masks, and fashion masks have gained significant momentum. However, the wide use of masks has adversely affected the environment. Disposable masks are made of polymers such as polypropylene, polyurethane, polyacrylonitrile, polystyrene, polycarbonate, polyethylene, or polyester [7]. Many additives, including plasticizers, antioxidants, and flame retardants, are added to these materials in non-covalent forms to obtain products with greater performance [8]. Phthalate esters (PAEs), a common kind of plastic additives, have been detected in facemasks with a level up to 38 μg/g [9]. Used masks undergo oxidation or weathering into microplastics after discarding them into the environment when not properly disposed of, contributing to the ubiquitous occurrence of chemical additives and microplastics in nature [10,11]. Significantly, these emerging contaminants in masks affect the environment and are deleterious to humans during wearing. Once inhaled, part of microplastics keep away from the clearing mechanism of the respiratory tract and may lead to lung disease, such as asthma, chronic obstructive pulmonary disease, and even cancer, by producing radical oxygen species, inducing inflammation, damaging cellular structures, and blocking vessels [12,13]. PAEs, typical endocrine disrupters, can enter the body through dermal contact and inhalation while wearing masks and result in a series of endocrine disorders. Di (2-ethylhexyl) phthalate (DEHP), for example, can decrease female fertility [14], increase risk of allergic diseases and asthma in children [15], cause insulin resistance [16], and be associated with overweight and obesity [17]. Nonetheless, few studies have focused on the two pollutants’ extent of harm to human when wearing masks. Li et al. [ 18] found that masks released fibrous microplastics after 720 h but did not consider the indoor and outdoor conditions, and the interval time was excessively long, as the average wearing time of masks was less than 24 h. Although the estimated daily intake of PAEs based on the content reduction after degassing the mask has been calculated, about 2.0–20 ng/kg bw/day for adults [19,20], release characteristics of PAEs and various masks’ different effects on exposure were not concerned. Moreover, it is widely thought that masks represent a protector of human exposure to microplastics and phthalates. Overwhelming evidence substantiates that these pollutants are widely distributed in the air, with a definite risk of inhalation exposure when people do not wear masks [21,22]. Therefore, whether facemasks increase or decrease exposure to microplastics and phthalates remains uncertain. Herein, we detected PAEs in five types of masks, including activated-carbon masks, N95 masks, surgical masks, cotton masks, and fashion masks, and we developed simulated inhalation equipment based on the report of Li et al. [ 18], which was applied for collecting microplastics and phthalates during 24 h indoors and outdoors to further identify the effects of different facemasks while wearing. ## 2.1. Chemicals and Reagents Seven phthalate diester standards including dimethyl phthalate (DMP), diethyl phthalate (DEP), di-iso-butyl phthalate (DIBP), di-n-butyl phthalate (DNBP), butyl benzyl phthalate (BBZP), di(2-ethylhexyl) phthalate (DEHP), and di-n-octyl phthalate (DNOP) were purchased from Dr. Ehrenstorfer (Augsburg, Germany; purity ≥ $99.0\%$). Two isotope-labelled standards, namely d4-DNBP and d4-DEHP, were purchased from Sigma Aldrich (St. Louis, MO, USA; purity ≥ $96.7\%$). Details with regard to these target chemicals are listed in Table S1 in the Supporting Information (SI). Hexane and methanol used were of high-performance liquid chromatography (HPLC) grade and purchased from Anpel (Shanghai, China). The glass fiber filter (GFF; 0.45-μm pore size, Ø 25 mm) was purchased from Beihua (Beijing, China). ## 2.2. Facemask Collection A total of 11 brands of best-selling facemasks representing 5 main types (as shown in Figure S1), i.e., activated carbon (hereinafter referred to as “AC”; sample code: M1–2), N95 (M3–5), surgical (SU; M6–7), cotton (CO; M8–9) and fashion masks (FA; M10–11), were purchased from online retailers in October 2021 and then sealed in polyethylene bags for storage at 4 °C in the dark until exposure experiment or chemical analysis. Each category of facemask contained at least two items. All facemask samples were manufactured in China, and the unit price of each facemask varied from 0.12 to 10 CNY. The majority of the facemasks were made of non-woven and melt-blown fabrics except for cotton and fashion masks whose main bodies were cotton and polyurethane, respectively. A detailed list of samples analyzed in this study is provided in Table S2. ## 2.3. Experimental Approach The experimental system that had been reported by Li et al. [ 18] was adopted in the present study with minor modification. As shown in Figure 1, each experimental flow-path was composed of a mask sample, a suction flask with GFF filter, a rotermeter, and a vacuum pump. In order to simulate the human inhalation process, air was sucked through the testing mask continuously at a flow rate of 15 L/min [23]. A blank test collecting microplastics and phthalates without the mask fixed on top of the suction flask was also conducted at the same time to figure out the role facemasks played. This experiment was conducted indoors and outdoors, and no contamination control measures of sampling environments were applied to reflect a realistic situation of microplastics and phthalate diester inhalation (Figure 1). ## 2.4. PAEs Measurements Samples from several parts of the unused mask (without ear loops and melt nose strips) were cut into small pieces to provide a representative sample for extraction. A fraction of 0.05 g of mask pieces or laden GFF was placed in a 15 mL glass tube. The sample was then extracted with 5 mL of hexane, followed by fortifying 100 ng each of deuterated internal standards. After ultrasonication at 100 kHz for 30 min and centrifugation at 3000 rpm for 15 min, the solvent layer was transferred into another glass tube, and the extraction step was then repeated. Before instrument analysis, the combined extraction solvent was concentrated to 1 mL under a gentle nitrogen stream, centrifuged at 10,000 rpm for 3 min, and then spun into a gas chromatographic glass vial. Determination of target PAEs was conducted on an Agilent 7890A gas chromatography equipped with a 5977B mass spectrometry (GC-MS; Agilent Technologies, MA, USA) using an electron ion source and selected ion monitoring mode based on Cheng et al. ’s study [24]. A DB-5MS column (Agilent; 30 m × 0.25 mm × 0.25 μm) was used for chromatographic separation of target chemicals. Further instrument parameters and detailed information of mass spectrometry are listed in Tables S3 and S4. Chromatograms of standards and a real mask sample are shown in Figure S2. ## 2.5. Microplastic Detection The microplastics intercepted onto the GFF were observed and counted under a stereomicroscope (Sunny Optical Technology Co., Ltd., Ningbo, Zhejiang, China). To determine the exact membrane collecting microplastics, four types of membranes, including GFF, nylon filter, polytetrafluoroethylene filter, and mixed cellulose filter, which were commonly used for air sampling, were detected and compared the background of microplastic quantity. Except GFFs, 5–31 microplastics were observed on the latter three membranes (Table S5). At the same time, accounting for the non-plastic nature of GFF, which means low PAEs background concentration, GFF was finally chosen as the membrane for collecting microplastics and phthalates. ## 2.6. PAEs Exposure Assessment Based on the measured concentrations in GFFs and facemasks, human exposure to PAEs via inhalation and dermal contact due to mask-wearing were calculated using the following equations [25]:[1]EDIinh=m×TinhTbreBW×T where EDIinh is estimated inhalation daily intake (ng/kg bw/day); m is PAE amount in GFF after a 6 h process (ng) (method A: 4 h + 2 h mass on GFF; method B: 6 h mass on GFF); TinhTbre is the proportion of the time taken to inhale during a full breath, the value is $\frac{1}{3}$ (dimensionless) [26]; BW is the body weight of a person, 17.1 kg for a child, 60.6 kg for an adult [27,28]; and T is the exposure duration (1 day). [ 2]EDIder=MV×kp−l×Kssl−g×SA×TKcl−g×BW where EDIder is estimated dermal daily intake (ng/kg bw/day); M is PAE amount measured in the mask (ng); V is mask volume (estimated to be 52.0 cm3) [29,30]; kp−l is the rate of chemicals transferring from skin surface lipid to blood (cm/h) (shown in Table S6, [25]; and Kssl−g and Kcl−g are the partition coefficients (dimensionless) of PAEs between skin surface lipid and air, and between mask and air, respectively (Table S6, [31]). Kcl−g was omitted while calculating the EDIder of the air sample; SA is the contact surface area between skin and mask (estimated to be 166 cm2) [29]; and T is mask wearing time (assumed to be 6 h/day). ## 2.7. Quality Assurance and Quality Control Prior to use, glass tubes and GFFs were baked at 450 °C for 4 h, and 1.5 mL plastic centrifugal tubes were cleaned twice with methanol to remove the background PAEs. A blank sample was analyzed with every 10 samples. Among the target chemicals, DNBP, DEHP, and DNOP were detected in procedural blanks ranging from 1.94 to 7.63 ng/mL, and these values were subtracted from reported concentrations in the present study. Two mask samples were prepared with each batch to evaluate the repeatability of the analytical method. The coefficient of variation of target chemicals in different batches of pooled mask samples ranged from 6.7–$15.1\%$. The recoveries of the seven PAEs spiked into facemasks and GFFs were determined at two levels, i.e., 50 and 500 ng/g, and were in the range of 93–$126\%$ and 112–$128\%$ for masks, 51–$108\%$ and 86–$128\%$ for GFFs, respectively (Table S7). The limit of detection (LOD) and limit of quantification (LOQ) were defined as the concentration in matrix samples that generated signal-to-noise ratios of 3 and 10, respectively. The range of LODs was 0.002–3.41 ng/g for real samples (Table S7). A cotton lab coat and nitrile gloves were worn to avoid microplastic contamination during the experiment. Furthermore, GFFs used for sampling were cleaned twice with Milli-Q water, and no microplastic could be detected after this step. After sampling, the suction cup was cleaned with Milli-Q water to ensure that all microplastics were transferred onto the membrane, and GFFs were stored in clean membrane wares. The masks and suction cups were covered with aluminum foil when the equipment was not switched off. During microplastic detection, the microscope was covered by a nylon bag. ## 2.8. Statistical Analysis Qualitative software was used to integrate the peak area of the gas chromatogram of target chemicals. Values that were below LOD and between the LOD and LOQ were defined as zero and LOQ/√2, respectively, for statistical analysis. Statistical analyses, including Spearman’s correlation analysis, which gave correlation among 7 phthalates concentration in masks, ANOVA, which gave ∑7PAEs concentration difference among 5 types of masks, and a paired t-test analyzing the significant difference of PAEs collected in indoor and outdoor conditions, were performed using IBM SPSS Statistics 26.0. A p-value < 0.05 was statistically significant. ## 3.1. PAEs Residue Levels in Face Masks Among the seven targeted PAEs in the present study, DNBP and DEHP were the predominant plasticizers and detected in all mask samples, with mean concentrations of 264 and 6804 ng/g, respectively (Table 1). This is in agreement with their large production and extensive use in melt-blown fabric and non-woven fabric of masks [9]. In contrast, the detected frequencies of DIBP, DEP, DMP, and BBZP ranged from 18–$45\%$, and their mean concentrations were 805, 502, 558, and 51.2 ng/g, respectively. DNOP was not detected in any mask samples, which was consistent with the findings reported by Wang et al. [ 20]. The overall concentrations of the 7 PAEs (hereafter referred to as ∑7PAEs) in face masks ranged from 296 to 72,049 ng/g, with a median of 1401 ng/g. This result was consistent with those collected from China (median: 2050 ng/g), Europe (2890 ng/g), Japan (1459 ng/g), Korea (787 ng/g), and USA (1950 ng/g) [9,19]. In comparison to other skin-contact substances, ∑PAEs’ concentrations in facemasks were in the same order of magnitude with that of panty liners (∑9PAEs: 168–34,500 ng/g, median: 1830 ng/g) and pads (∑9PAEs: 205–11,200 ng/g, median: 362 ng/g) for females [32], infant clothes (∑6PAEs: 2290–51,900 ng/g, median: 4150 ng/g) [33], and children’s clothing (∑6PAEs: 1969–183,248 ng/g, median: 5579 ng/g) [25]. These results indicate that facemasks also represent an important source of PAE exposure in daily life and warrant further study. With regard to the facemask type, we found that the fashion mask had the highest ∑7PAEs concentrations (median: 37,386 ng/g), which was two orders of magnitude higher than that in the surgical mask (367 ng/g) and N95 mask (577 ng/g) (Table S8). The other two types of facemasks, i.e., activated-carbon and cotton mask, contained similar residual levels of ∑7PAEs, with median values ranging from 5000–5689 ng/g. The various ∑7PAEs levels of masks may be due to different production standards and processes. Nevertheless, there was no significant difference among five types of masks ($p \leq 0.05$). Similar to our study, Arribas et al. [ 34] detected organophosphate esters (OPEs), another important class of plasticizers, in masks and found no statistical difference in OPE levels in common facemasks. This result may be related to similarities in raw materials used and limited facemask sample size in the present study [9]. Considering individual PAEs, DEHP and DNBP yielded the most significant contributions, accounting for $47.3\%$ and $16.5\%$, respectively (Figure S3). It should be noted that DIBP yielded the largest contribution in cotton masks ($79.8\%$), which may be related to the special materials. Although not statistically significant, there was a positive correlation between DNBP and DEHP ($r = 0.52$, $p \leq 0.05$). This may be attributed to the multiple purposes for these two primary PAEs. As a representative low molecular weight PAE, DNBP is often used as a solvent to maintain the color and smell of materials, while DEHP is indicated to soften polyvinyl chloride materials [35]. Furthermore, Vimalkumar et al. [ 19] found that DEHP clustered with non-phthalate plasticizers such as dibutyl sebacate and bis (2-ethylhexyl) adipate in facemasks by principal component analysis. Phthalates are readily released from facemasks irresponsibly disposed of and believed to be a significant source of DEHP and other phthalates in the environment. An estimated 129 million masks (each weighing 4 g) were used every month worldwide during the COVID-19 pandemic in 2020 [5,29], suggesting that 5.2 × 105 tons of masks were consumed every month. Based on the median concentration of ∑7PAEs measured in this study and assuming that $44\%$ of the masks ended up as waste in the environment [36], 3415 tons of PAEs are calculated to have entered the environment. This is roughly equivalent to $0.04\%$ of the annual production of PAEs, and the disposal of these masks is threatening to be a huge problem for the environment. ## 3.2. Microplastic Quantity in Inhalation Measurements As shown in Figure 2, microplastics retained by GFFs under indoor or outdoor conditions were divided into fragments and fibers, and they were counted separately over time. After 24 h simulating the release experiment, the amount of microplastics (sum of fibers and fragments) was 62–487 items/GFF. However, Li et al. counted 4000–30,000 microplastics after 24 h process, which is two magnitudes higher than our finding [18]. We thought that the disparity might be related to the different experiment environment (an indoor condition with lots of textiles in Li et al.) and count rules of microplastics. The microplastics number ranked in the following order: blank > cotton > activated-carbon > fashion > N95 > surgical and cotton > blank > activated-carbon > fashion > surgical > N95 under indoor and outdoor environments, respectively (Tables S9–S11). The largest number of microplastics in the blank sample and less microplastic in the N95 sample under indoor conditions was the same as the discovery of Li et al. [ 18]. This finding indicated that most masks could reduce respiratory exposure to microplastics compared to not wearing them, except for cotton masks. In this respect, the surgical masks and N95 masks yielded the best performance in reducing the inhalation risk of microplastics, for use either indoors or outdoors. Furthermore, microplastics were significantly higher outdoors (62–254 items/GFF) than those of indoors (139–487 items/GFF) after the whole release process, which was different from previous studies [37] reporting microplastics’ pollution in air was severer in indoor conditions. The discrepancy could be explained by different indoor conditions: apartments contain many textiles contributing a mass of fibers [37], whereas a laboratory environment does not provide a large source of textile fibers. It should be noted that we failed to analyze the chemical composition of microplastics and compared the blank sample and mask samples because of the limit of instruments, so we did not discuss the source (masks themselves or air) of microplastics or compare the release ability of different masks. With regard to microplastic shape, fragmented items (68–$86\%$) were significantly more than fibers (14–$32\%$) in most mask samples except for cotton masks ($49\%$ vs. $50\%$) (Figure S4). The abundance of fibers at each point in time was the highest in cotton mask samples indoors and outdoors, reaching 119 and 245 items/GFF after 24 h release, which was 3.31 and 1.74 times higher in indoor and outdoor air samples, respectively (Table S9). Similarly, a study found that cotton masks released significantly more fibers [823] than surgical masks [85] when washing various types of masks in a washing machine [38]. This finding may be related to the loose structure of the cotton mask leading to a large amount of fiber released from the inner layer of this mask during inhalation [38,39]. In contrast, the abundance of fragmented microplastic in cotton masks was slightly lower than in indoor and outdoor air samples (Table S10), suggesting that cotton masks had a weak ability to protect against fragmented microplastics. Interestingly, the fiber amount detected in the activated carbon mask sample was more than the blank sample in the indoor environment, but the opposite findings were observed outdoors. This might be related to the larger amount of fiber in the air (Figure S4), as well as the stronger prevention and weaker release abilities of the activated carbon mask. ## 3.3. Microplastic Release Characteristics During the 24 h period, a biphasic release pattern was observed with rapid release in the first six hours, followed by slower release indoors and outdoors. Although it could not be precisely determined whether the microplastics collected came from the mask or the atmosphere, the variations in the quantity over time were similar to the release trend of the microplastics in the water environment of the mask. Similar equations (Equations [3]–[6]) can be used to fit the dynamic characteristics [40,41] (Figure 3). Elovich equation:[3]Qt=a+blnt Parabolic diffusion equation:[4]Qt=a + bt0.5 Power function equation:[5]Qt=btc Modified-Freundlich:[6]Qt=a + btc where t is the release time (h); Qt is the quantity of microplastics at the time t (item/GFF); a is the initial quantity of microplastics (item/GFF); and b and c is the rate constant. We analyzed relevant parameters for the four fitting equations (Table S12 for indoor conditions and Table S13 for outdoor conditions) and found that all R2 values were > 0.92, indicating the proper equations were chosen, and the modified Freundlich equation yielded the best fitting results (R2 > 0.98). A c value less than 1 in the power function equation and modified Freundlich equation suggested that the quantity of microplastics collected by GFFs decreased exponentially per unit time [42]. The b value in the four equations reflects the rate of increase in microplastics; a greater b was associated with a greater increase in rate [40]. In all samples, the cotton mask yielded the largest b value in both conditions, but the surgical mask and N95 mask showed the smallest b indoors and outdoors, respectively, similar to the quantity of microplastics we discussed before. This finding proved that the number of microplastics collected during release was closely related to the type of mask. This is inconsistent with the conclusion reached by Liang et al. [ 40], who reported no relationship between the release of microplastics in water and the type of mask. Moreover, Wu et al. [ 43] reported that the release of microplastics from surgical masks was greater than from ordinary and filtering facepiece masks. Indeed, during the process of this experiment, the mask not only released microplastics but also blocked microplastics from the air to various degrees. We analyzed the kinetic characteristics of microplastics, classified into five groups, 20–30, 30–100, 100–500, 500–1000, and >1000 μm, according to their size (Figure S5). As shown in Figure S6, in indoor and outdoor environments, microplastics with small sizes, such as 20–30 μm (indoor: 16–$63\%$; outdoor: 3–$29\%$) and 30–100 μm (indoor: 21–$46\%$; outdoor: 24–$60\%$), were the dominant groups during the inhalation experiment except for the cotton mask sample, while the 100–500 μm group was mostly collected by cotton mask. Interestingly, microplastics > 500 μm almost stopped increasing after 10 h since large-sized microplastics (mostly fibers) on the mask surface were released more easily and quickly at the beginning of the process [40], but in the subsequent 18 h, the large-sized microplastics in the air could be blocked by the mask, and fibers inside the mask are difficult to release due to their interweaving. ## 3.4. PAEs on GFFs and Associations with Microplastics We detected released PAEs on GFFs during the inhalation experiment and compared the indoor and outdoor mass of the ∑7PAEs (Figure S7). Interestingly, the mass of ∑7PAEs (mean: 999 ng vs. 411 ng, $p \leq 0.05$) was significantly higher indoors than outdoors. This result was consistent with the literature [44,45], indicating that indoor environments showed higher exposure sources than outdoors. After the whole inhalation experiment, the ∑7PAE mass of all GFF samples were as follows: blank > activated-carbon > fashion > N95 > cotton > surgical mask and fashion > cotton > blank > N95 > activated-carbon > surgical mask indoors and outdoors, respectively (Figure 4). This data indicated wearing masks could decrease human exposure to PAEs (i.e., acting as a “protector”) indoors, but in outdoor conditions, their effects differed and were affected by types. Interestingly, after further analysis of each compound’s average mass (as shown in Figure 4), we found that mask-wearing could increase the inhalation dose of DIBP (except for surgical masks outdoors) but reduce that of DNBP and DEHP, which can be attributed to their logKOA. DIBP had the lowest value and could easily enter the air and then be trapped by GFFs during “inhalation” [31]. Moreover, the mass of DEP in the fashion mask sample was higher than in indoor and outdoor air samples, which may be caused by the large amount of DEP (1421 ng/g) released from fashion masks during the inhalation experiment. Overall, masks can reduce the inhalation dose of ∑7PAEs but can release DIBP, while fashion masks represent an important source of DEP. During the “inhalation” process, the mass of ∑7PAEs increased gradually; the same equation as microplastics was not used considering the better linearity (R2 > 0.93) of PAEs except surgical and cotton mask samples outdoors, which was related to more complex flux outdoors. However, there was still a significant correlation between microplastics and PAEs (indoor: $r = 0.852$, $p \leq 0.01$; outdoor: $r = 0.620$, $p \leq 0.01$, as shown in Figure S8), and smaller microplastic size corresponded to a more severe PAE load (Table S14), as microplastics with smaller size could carry PAEs from processed materials because of the larger specific surface area [46]. ## 3.5. Exposure Estimation In recent years, several studies which analyzed the PAE exposure risk when wearing masks yielded comparable results despite using different methodologies. In the present study, the EDIinh of the ∑7PAEs ranged from 3.60 to 8.62 and from 1.47 to 7.49 ng/kg bw/d indoors and outdoors for children, and from 1.02 to 2.43 and from 0.493 to 2.11 ng/kg bw/d for adults, which was comparable with the EDI values (5.14 and 2.02 ng/kg bw/d for children and adults, respectively) calculated by Vimalkumar et al. [ 19]. Wearing fashion, cotton and activated-carbon masks indoors and outdoors increased the inhalation risk of PAEs, while surgical masks could reduce that of PAEs (Figure 5a,b). Furthermore, Wang et al. [ 20] calculated the EDIinhs caused by wearing surgical and N95 masks and found that the EDI of the former was almost ten-fold less than the latter, indicating the safer nature of the surgical mask. Taking dermal exposure into consideration, given PAEs’ high concentration in fashion masks (2724–72,049 ng/g), the EDIder of this mask reached 0.98 and 3.5 ng/kg bw/d for adults and children, respectively, while the level of N95 and surgical masks was smaller than 0.1 ng/kg bw/d (Figure S9b). However, a study reported that the dermal risk of 11 PAEs caused by wearing masks ranged from 3.71 to 639 ng/kg bw/d [9], which was significantly higher than in our study, while their PAE concentration (115–37,700 ng/g) in masks was comparable to our results (296–72,049 ng/g), which may be related to the different exposure scenarios and parameters selected for skin penetration. Xie et al. [ 9] multiplied the mass of PAEs in masks and the human body absorption rate ($20\%$) to calculate the dermal risk. Nonetheless, in our study, after considering the equilibrium of PAEs among mask, air, and skin surface and the rate of chemicals’ transfer from skin surface lipid to blood, the equivalent “absorption rate” was only 5.30 × 10−5–6.34 × 10−$3\%$ for individual phthalates, four to six orders of magnitude lower than the previous study, resulting in lower EDIder. After summing up the two exposure approaches (EDItotal), more types of masks could potentially be considerable sources of PAEs. Among the five types, fashion masks brought the most serious risk, followed by cotton and activated-carbon masks, while the surgical mask still played a protective role (Figure 5a). The EDItotal was 1.05–12.1 ng/kg bw/d indoors and 0.507–7.03 ng/kg bw/d outdoors, which was significantly lower than the EDI of dietary exposure (1.03–4.68 µg/kg bw/d) and reference dose (RfD) of a single PAE (RfD = 20–1000 µg/kg bw/d) [47,48,49,50,51,52,53], indicating human exposure to PAEs via masks may not pose a potential health risk. However, it should be borne in mind that fashion masks represent an important source of PAEs indoors and outdoors. Different from the complex and dependent risk assessments of PAEs, microplastics’ risk mainly appeared on one mask whether in indoor and outdoor conditions. Although nanoplastics smaller than 100 nm can easily enter the lungs due to their small size, fibers as long as 2475 μm have been documented in human lung tissues [54]. Based on the amount of 20–2475 μm microplastics trapped by the filter membrane (Figure 5b), after 6 h of wearing, 40–160 and 61–389 microplastics may be inhaled in indoor and outdoor environments, respectively, and cotton masks represent the most important source of microplastics. However, due to limitations in measurement instruments, this study failed to quantify nanoplastics. A previous study observed that an amount of approximately 109 of nanoplastic could be released when the mask was oscillated in a water environment for 4 h [55]. Although it differs from the breathing environment, the massive release of nanoplastics is still worthy of attention and needs further research and discussion. We further analyzed the difference between Method A and Method B while calculating the exposure dose of microplastics and PAEs (Figure 5a–d). Briefly speaking, the total exposure dose of PAEs of most masks and the inhaled microplastics of all masks increased. The EDItotal of fashion masks decreased from 3.66 to 3.41 ng/kg bw/d indoors and, in opposite fashion, from 3.14 to 1.98 ng/kg bw/d outdoors. As for different masks’ roles, activated-carbon, cotton, and fashion masks increased PAE exposure whether in indoor or outdoor conditions, but cotton masks were still the only mask increasing the inhalation amount of microplastics after changing the masks’ usage. In a word, different masks’ usage could result in different exposure doses; on one hand, it was suggested to change masks after 4 h to avoid viruses, but on the other hand, changing masks too often might increase some exposure to pollutants, microplastics at least. ## 4. Conclusions In this study, we discussed the role of five widely used masks on microplastic and phthalate plasticizer exposure with respect to environmental health. We observed considerable pollution of phthalates in widely used masks, with the lowest concentration in surgical masks, and the highest levels in fashion masks. Among seven phthalates, DNBP and DEHP were detected in each mask sample. During the inhalation experiment, microplastics increased faster in the first 6 h, and fragments and small-size microplastics accounted for the largest. For our main focus, surgical masks play a protective role, while wearing activated carbon, cotton, and fashion masks in indoor and outdoor environments increases human exposure to PAEs. Additionally, wearing cotton masks resulted in a larger inhalation risk than wearing no mask, while other types of masks acted as protectors against microplastics. However, we call for more studies and development based on some limitations. The limited number of masks was not enough to analyze the statistical significance of various masks ∑7PAE concentrations and to stand for exact exposure characteristics of various types of masks while wearing; secondly, our release experiment, which was conducted under laboratory conditions, only focused on the inhalation process, and exhalation was not concerned although inspiration–expiration ratio was considered to reduce deviation while calculating EDIinh; meanwhile, the simulation inhalation process, especially the air flow rate, was different from a real situation considering masks’ distinctive fluid resistance; finally, on account of the limit of instruments, we did not obtain the chemical composition of microplastics collected, and the fibers of the cotton mask sample were assumed to be microplastics, so there might be an overestimation of the microplastic exposure. Even so, our study corroborates that cotton and fashion masks are important sources of human exposure to microplastics and phthalates, highlighting that it is not recommended to wear these two masks unnecessarily in daily life to reduce exposure. In contrast, the surgical mask is a great choice against microplastics and phthalates, as well as the COVID-19 virus. ## References 1. 1. WHO Coronavirus Disease (COVID-19) Situation DashboardWorld Health OrganizationGeneva, SwitzerlandAvailable online: https://covid19.who.int/(accessed on 6 May 2022). Coronavirus Disease (COVID-19) Situation Dashboard 2. Zhang R.Y., Li Y.X., Zhang A.L., Wang Y., Molina M.J.. Identifying airborne transmission as the dominant route for the spread of COVID-19. Proc. Natl. Acad. Sci. USA (2020.0) 117 14857-14863. DOI: 10.1073/pnas.2009637117 3. Rational Use of Personal Protective Equipment for Coronavirus Disease (COVID-19) and Considerations during, Severe Shortages 4. Chu D.K., Akl E.A., Duda S., Solo K., Yaacoub S., Schunemann H.J.. Covid-Systematic Urgent Review: Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: A systematic review and meta-analysis. Lancet (2020.0) 395 1973-1987. DOI: 10.1016/S0140-6736(20)31142-9 5. Prata J.C., Silva A.L.P., Walker T.R., Duarte A.C., Rocha-Santos T.. COVID-19 Pandemic Repercussions on the Use and Management of Plastics. Environ. Sci. Technol. (2020.0) 54 7760-7765. DOI: 10.1021/acs.est.0c02178 6. Cheng Y.F., Ma N., Witt C., Rapp S., Wild P.S., Andreae M.O., Poschl U., Su H.. Face masks effectively limit the probability of SARS-CoV-2 transmission. Science (2021.0) 372 1439-1443. DOI: 10.1126/science.abg6296 7. Potluri P., Needham P., Scott R.A.. 6—Technical textiles for protection. Textiles for Protection (2005.0) 151-175 8. Benjamin S., Pradeep S., Sarath Josh M., Kumar S., Masai E.. A monograph on the remediation of hazardous phthalates. J. Hazard. Mater. (2015.0) 298 58-72. DOI: 10.1016/j.jhazmat.2015.05.004 9. Xie H.J., Han W.J., Xie Q., Xu T., Zhu M.H., Chen J.W.. Face mask-A potential source of phthalate exposure for human. J. Hazard. Mater. (2022.0) 422 126848. DOI: 10.1016/j.jhazmat.2021.126848 10. Fadare O.O., Okoffo E.D.. COVID-19 face masks: A potential source of microplastic fibers in the environment. Sci. Total Environ. (2020.0) 737 140279. DOI: 10.1016/j.scitotenv.2020.140279 11. Akhbarizadeh R., Dobaradaran S., Nabipour I., Tangestani M., Abedi D., Javanfekr F., Jeddi F., Zendehboodi A.. Abandoned COVID-19 personal protective equipment along the Bushehr shores, the Persian Gulf: An emerging source of secondary microplastics in coastlines. Mar. Pollut. Bull. (2021.0) 168 112386. DOI: 10.1016/j.marpolbul.2021.112386 12. Prata J.C.. Airborne microplastics: Consequences to human health?. Environ. Pollut. (2018.0) 234 115-126. DOI: 10.1016/j.envpol.2017.11.043 13. Dong C.-D., Chen C.-W., Chen Y.-C., Chen H.-H., Lee J.-S., Lin C.-H.. Polystyrene microplastic particles: In vitro pulmonary toxicity assessment. J. Hazard. Mater. (2020.0) 385 121575. DOI: 10.1016/j.jhazmat.2019.121575 14. Upson K., Sathyanarayana S., De Roos A.J., Thompson M.L., Scholes D., Dills R., Holt V.L.. Phthalates and risk of endometriosis. Environ. Res. (2013.0) 126 91-97. DOI: 10.1016/j.envres.2013.07.003 15. Braun J.M., Sathyanarayana S., Hauser R.. Phthalate exposure and children’s health. Curr. Opin. Pediatr. (2013.0) 25 247-254. DOI: 10.1097/MOP.0b013e32835e1eb6 16. Trasande L., Sathyanarayana S., Spanier A.J., Trachtman H., Attina T.M., Urbina E.M.. Urinary Phthalates are Associated with Higher Blood Pressure in Childhood. J. Pediatr. (2013.0) 163 747-753. DOI: 10.1016/j.jpeds.2013.03.072 17. Buser M.C., Murray H.E., Scinicariello F.. Age and sex differences in childhood and adulthood obesity association with phthalates: Analyses of NHANES 2007–2010. Int. J. Hydrog. Environ. Health (2014.0) 217 687-694. DOI: 10.1016/j.ijheh.2014.02.005 18. Li L., Zhao X.L., Li Z.Y., Song K.. COVID-19: Performance study of microplastic inhalation risk posed by wearing masks. J. Hazard. Mater. (2021.0) 411 124955. PMID: 33445045 19. Vimalkumar K., Zhu H.K., Kannan K.. Widespread occurrence of phthalate and non-phthalate plasticizers in single-use facemasks collected in the United States. Environ. Int. (2022.0) 158 106967. DOI: 10.1016/j.envint.2021.106967 20. Wang X.Y., Okoffo E.D., Banks A.P.W., Li Y., Thomas K.V., Rauert C., Aylward L.L., Mueller J.F.. Phthalate esters in face masks and associated inhalation exposure risk. J. Hazard. Mater. (2022.0) 423 127001. DOI: 10.1016/j.jhazmat.2021.127001 21. Liu K., Wang X.H., Fang T., Xu P., Zhu L.X., Li D.J.. Source and potential risk assessment of suspended atmospheric microplastics in Shanghai. Sci. Total Environ. (2019.0) 675 462-471. DOI: 10.1016/j.scitotenv.2019.04.110 22. Zhang L.B., Wang F.M., Ji Y.Q., Jiao J., Zou D.K., Liu L.L., Shan C.Y., Bai Z.P., Sun Z.R.. Phthalate esters (PAEs) in indoor PM10/PM2.5 and human exposure to PAEs via inhalation of indoor air in Tianjin, China. Atmos. Environ. (2014.0) 85 139-146. DOI: 10.1016/j.atmosenv.2013.11.068 23. Garcia G.J.M., Schroeter J.D., Kimbell J.S.. Olfactory deposition of inhaled nanoparticles in humans. Inhal. Toxicol. (2015.0) 27 394-403. DOI: 10.3109/08958378.2015.1066904 24. Cheng Z.P., Yao Y.M., Sun H.W.. Comparative uptake, translocation and subcellular distribution of phthalate esters and their primary monoester metabolites in Chinese cabbage. Sci. Total Environ. (2020.0) 742 140550. DOI: 10.1016/j.scitotenv.2020.140550 25. Tang Z.W., Chai M., Wang Y.W., Cheng J.L.. Phthalates in preschool children’s clothing manufactured in seven Asian countries: Occurrence, profiles and potential health risks. J. Hazard. Mater. (2020.0) 387 121681. DOI: 10.1016/j.jhazmat.2019.121681 26. Lessard M.R., Guerot E., Lorino H., Lemaire F., Brochard L.. Effects of pressure-controlled with different I:E ratios versus volume-controlled ventilation on respiratory mechanics, gas exchange, and hemodynamics in patients with adult respiratory distress syndrome. Anesthesiology (1994.0) 80 983-991. DOI: 10.1097/00000542-199405000-00006 27. Wang L.X., Zhao B., Liu C., Lin H., Zhang Y.P.. Analysis on exposure of indoor phthalic acid esters. Build. Sci. (2010.0) 26 73-80 28. 28. Ministry of Environmental Protection Exposure Factors Handbook of Chinese Population (Adults)China Environmental PressBeijing, China2013748. Exposure Factors Handbook of Chinese Population (Adults) (2013.0) 748 29. Liu R.Z., Mabury S.A.. Single-Use Face Masks as a Potential Source of Synthetic Antioxidants to the Environment. Environ. Sci. Technol. Lett. (2021.0) 8 651-655. DOI: 10.1021/acs.estlett.1c00422 30. Balazy A., Toivola M., Reponen T., Podgo R.A., Zimmer A., Grinshpun S.A.. Manikin-based performance evaluation of N95 filtering-facepiece respirators challenged with nanoparticles. Ann. Occup. Hydrog. (2006.0) 50 259-269 31. Saini A., Okeme J.O., Mark Parnis J., McQueen R.H., Diamond M.L.. From air to clothing: Characterizing the accumulation of semi-volatile organic compounds to fabrics in indoor environments. Indoor Air (2017.0) 27 631-641. DOI: 10.1111/ina.12328 32. Gao C.J., Kannan K.. Phthalates, bisphenols, parabens, and triclocarban in feminine hygiene products from the United States and their implications for human exposure. Environ. Int. (2020.0) 136 105465. DOI: 10.1016/j.envint.2020.105465 33. Li H.L., Ma W.L., Liu L.Y., Zhang Z., Sverko E., Zhang Z.F., Song W.W., Sun Y., Li Y.F.. Phthalates in infant cotton clothing: Occurrence and implications for human exposure. Sci. Total Environ. (2019.0) 683 109-115. DOI: 10.1016/j.scitotenv.2019.05.132 34. Arribas F.J., Moreno T., Bartroli R., Eljarrat E.. COVID-19 face masks: A new source of human and environmental exposure to organophosphate esters. Environ. Int. (2021.0) 154 106654. DOI: 10.1016/j.envint.2021.106654 35. Cao X.L.. Phthalate Esters in Foods: Sources, Occurrence, and Analytical Methods. Compr. Rev. Food Sci. Food Saf. (2010.0) 9 21-43. DOI: 10.1111/j.1541-4337.2009.00093.x 36. Tan M., Wang Y., Luo L., Hu J.. How the public used face masks in China during the coronavirus disease pandemic: A survey study. Int. J. Nurs. Stud. (2021.0) 115 103853. DOI: 10.1016/j.ijnurstu.2020.103853 37. Dris R., Gasperi J., Mirande C., Mandin C., Guerrouache M., Langlois V., Tassin B.. A first overview of textile fibers, including microplastics, in indoor and outdoor environments. Environ. Pollut. (2017.0) 221 453-458. DOI: 10.1016/j.envpol.2016.12.013 38. De Felice B., Antenucci S., Ortenzi M.A., Parolini M.. Laundering of face masks represents an additional source of synthetic and natural microfibers to aquatic ecosystems. Sci. Total Environ. (2022.0) 806 150495. DOI: 10.1016/j.scitotenv.2021.150495 39. Zambrano M.C., Pawlak J.J., Daystar J., Ankeny M., Cheng J.J., Venditti R.A.. Microfibers generated from the laundering of cotton, rayon and polyester based fabrics and their aquatic biodegradation. Mar. Pollut. Bull. (2019.0) 142 394-407. DOI: 10.1016/j.marpolbul.2019.02.062 40. Liang H., Ji Y., Ge W., Wu J., Song N.N., Yin Z.D., Chai C.. Release kinetics of microplastics from disposable face masks into the aqueous environment. Sci. Total Environ. (2022.0) 816 151650. DOI: 10.1016/j.scitotenv.2021.151650 41. Mallakpour S., Hatami M.. Fabrication and characterization of pH-sensitive bio-nanocomposite beads havening folic acid intercalated LDH and chitosan: Drug release and mechanism evaluation. Int. J. Biol. Macromol. (2019.0) 122 157-167. DOI: 10.1016/j.ijbiomac.2018.10.166 42. Arzhang F.G., Hasan T., Karim S.. Kinetics of arsenic release from naturally contaminated soils at half saturation moisture by various extractants. Environ. Res. (2021.0) 194 110658. PMID: 33359704 43. Wu P.F., Li J.P., Lu X., Tang Y.Y., Cai Z.W.. Release of tens of thousands of microfibers from discarded face masks under simulated environmental conditions. Sci. Total Environ. (2022.0) 806 150458. DOI: 10.1016/j.scitotenv.2021.150458 44. Wang X., Tao W., Xu Y., Feng J., Wang F.. Indoor phthalate concentration and exposure in residential and office buildings in Xi’an, China. Atmos. Environ. (2014.0) 87 146-152. DOI: 10.1016/j.atmosenv.2014.01.018 45. Net S., Sempere R., Delmont A., Paluselli A., Ouddane B.. Occurrence, Fate, Behavior and Ecotoxicological State of Phthalates in Different Environmental Matrices. Environ. Sci. Technol. (2015.0) 49 4019-4035. DOI: 10.1021/es505233b 46. Prüst M., Meijer J., Westerink R.H.S.. The plastic brain: Neurotoxicity of micro- and nanoplastics. Part Fibre Toxicol. (2020.0) 17 24. DOI: 10.1186/s12989-020-00358-y 47. Wang Y., Zhu H.K., Kannan K.. A Review of Biomonitoring of Phthalate Exposures. Toxics (2019.0) 7. DOI: 10.3390/toxics7020021 48. Tran T.M., Kannan K.. Occurrence of phthalate diesters in particulate and vapor phases in indoor air and implications for human exposure in Albany, New York, USA. Arch. Environ. Contam. Toxicol. (2015.0) 68 489-499. DOI: 10.1007/s00244-015-0140-0 49. 49. USEPA Human Health Noncarcinogen Fact Sheet for Di(2-ethylhexyl) Phthalate: (Human Health Noncarcinogen—Fish Ingestion Only)Indiana Department of Environmental ManagementIndianapolis, IN, USAAvailable online: https://www.epa.gov/gliclearinghouse/human-health-noncarcinogen-fact-sheet-di2-ethylhexyl-phthalate-human-health(accessed on 18 November 2022). Human Health Noncarcinogen Fact Sheet for Di(2-ethylhexyl) Phthalate: (Human Health Noncarcinogen—Fish Ingestion Only) 50. 50. USEPA Human Health Noncarcinogen Fact Sheet for Dibutyl phthalate: (Human Health Noncarcinogen—Fish Ingestion Only)Indiana Department of Environmental ManagementIndianapolis, IN, USAAvailable online: https://www.epa.gov/gliclearinghouse/human-health-noncarcinogen-fact-sheet-dibutyl-phthalate-human-health-noncarcinogen(accessed on 18 November 2022). Human Health Noncarcinogen Fact Sheet for Dibutyl phthalate: (Human Health Noncarcinogen—Fish Ingestion Only) 51. 51. USEPA Human Health Fact Sheet for Diethyl phthalate: (Human Health—Fish Ingestion Only)Ohio Environmental Protection AgencyChicago, IL, USAAvailable online: https://www.epa.gov/gliclearinghouse/human-health-fact-sheet-diethyl-phthalate-human-health-fish-ingestion-only-ohio(accessed on 18 November 2022). Human Health Fact Sheet for Diethyl phthalate: (Human Health—Fish Ingestion Only) 52. 52. USEPA Human Health—Water Ingestion Only Fact Sheet for Butyl Benzyl Phthalate: (Human Health—Water Ingestion Only)New York State Department of Environmental ConservationAlbany, NY, USAAvailable online: https://www.epa.gov/gliclearinghouse/human-health-water-ingestion-only-fact-sheet-butyl-benzyl-phthalate-human-health(accessed on 18 November 2022). Human Health—Water Ingestion Only Fact Sheet for Butyl Benzyl Phthalate: (Human Health—Water Ingestion Only) 53. Wang J., Chen G., Christie P., Zhang M., Luo Y., Teng Y.. Occurrence and risk assessment of phthalate esters (PAEs) in vegetables and soils of suburban plastic film greenhouses. Sci. Total Environ. (2015.0) 523 129-137. DOI: 10.1016/j.scitotenv.2015.02.101 54. Jenner L.C., Rotchell J.M., Bennett R.T., Cowen M., Tentzeris V., Sadofsky L.R.. Detection of microplastics in human lung tissue using μFTIR spectroscopy. Sci. Total Environ. (2022.0) 831 154907. DOI: 10.1016/j.scitotenv.2022.154907 55. Ma J., Chen F.Y., Xu H., Jiang H., Liu J.L., Li P., Chen C.C., Pan K.. Face masks as a source of nanoplastics and microplastics in the environment: Quantification, characterization, and potential for bioaccumulation. Environ. Pollut. (2021.0) 288 117748. DOI: 10.1016/j.envpol.2021.117748
--- title: Multiclass Mask Classification with a New Convolutional Neural Model and Its Real-Time Implementation authors: - Alexis Campos - Patricia Melin - Daniela Sánchez journal: Life year: 2023 pmcid: PMC9967054 doi: 10.3390/life13020368 license: CC BY 4.0 --- # Multiclass Mask Classification with a New Convolutional Neural Model and Its Real-Time Implementation ## Abstract The world has been greatly affected by the COVID-19 pandemic, causing people to remain isolated and decreasing the interaction between people. Accordingly, various measures have been taken to continue with a new normal way of life, which is why there is a need to implement the use of technologies and systems to decrease the spread of the virus. This research proposes a real-time system to identify the region of the face using preprocessing techniques and then classify the people who are using the mask, through a new convolutional neural network (CNN) model. The approach considers three different classes, assigning a different color to identify the corresponding class: green for persons using the mask correctly, yellow when used incorrectly, and red when people do not have a mask. This study validates that CNN models can be very effective in carrying out these types of tasks, identifying faces, and classifying them according to the class. The real-time system is developed using a Raspberry Pi 4, which can be used for the monitoring and alarm of humans who do not use the mask. This study mainly benefits society by decreasing the spread of the virus between people. The proposed model achieves $99.69\%$ accuracy with the MaskedFace-Net dataset, which is very good when compared to other works in the current literature. ## 1. Introduction The contingency of the COVID-19 virus has caused people around the world to use face masks as a measure to stop the spread of this disease and in turn, any other disease that can be air transmitted or by contact with other people. Some symptoms that could occur are dry cough, tiredness, fever, and headaches, among many other symptoms that have been occurring and changing throughout the pandemic [1]. Similarly, there are cases in which no symptoms occur. Therefore, it is imperative to stop the spread of infections and prevent humans from catching diseases. People who have been infected have a varied recovery time that could get worse depending on the person who is sick. On many occasions, it is determined that the sick person fulfills quarantine depending on the severity. Facial recognition is a method that has been growing in recent years and the industry has been revolutionizing it through artificial intelligence (AI) and machine learning (ML) techniques. People today relate heavily to facial recognition and using computer vision techniques and image processing makes it possible to solve this complex task. Face masks have become an item of daily use, and, even though the obligation to wear them has been diminished, there are many institutions where it is still a duty to use them; some of these places can range from public locations such as schools, universities to private locations, such as offices. A mask is placed incorrectly when in the human it is not covering the region of the mouth, nose, and chin. Companies and institutes should locate humans who are using the mask correctly or are not wearing it, in a specific area. A real-time system would solve this problem, allowing to identify people who wear a mask, who are using it incorrectly, or who are not wearing one at all, to be identified. The use of convolution neural networks (CNNs) facilitates the identification and classification of images, extracting the main features and locating patterns that can be observed. The objective of our work is to solve this problem through CNN and machine learning techniques to distinguish the appropriate use of face masks. This article proposes a real-time system using deep learning techniques and computer vision, to classify people into three classes, if they wear masks correctly, if they use masks incorrectly, or if they do not wear masks. The system is carried out using a video camera and a Raspberry Pi 4 combining the use of libraries such as OpenCV and TensorFlow, as well as Python as a programming language. The method will identify a person’s face and place a rectangular box labeled “Mask” if the person is appropriately using a mask, which occurs when a mask covers the nose, mouth, and chin; otherwise, if the person is only wearing a mask on their chin and mouth, the model will label the box as “Incorrect”, and if they are not wearing any mask at all or if it is only on their chin, then the label will be “NoMask”. Therefore, the proposed method will allow the virus transmission to slow down, potentially benefiting people’s health systems. The dataset used to test the effectiveness of the CNN architecture is MaskedFace-Net. This dataset is provided by Cabani [2] and Hammoudi [3], features around 137,016 images and is based on another Flick-dataset [4]. The MaskedFace-Net images have a size of 1024 × 1024 pixels, and images are internally classified into two subcategories named Correctly Masked Face Dataset (CMFD) and Incorrectly Masked Face Dataset (IMFD). The Raspberry Pi is a small, low-cost device that can be used as a computer and runs programs in Python [5]. The graphics processing unit (GPU) outputs and inputs all work on a circuit. GPIO Board pins are an important element that allows the RPi to access hardware programming to control the I/O device’s electronic circuitry and data processing. We can add a keyboard, power supply, monitor, and mouse that run on the Raspberry Pi through the HDMI connector. New models are available that can communicate with the Internet via Wi-Fi and Ethernet ports. The RPi can be used with the Raspbian operating system [6]. In summary, this study proposes a CNN model where computer vision, machine learning, and deep learning techniques are combined to achieve classification in the three classes (NoMask, Mask, IncorrectMask). In Figure 1 can be observed that the region of interest is identified by the input image, in this case, the face will be identified and when the image is identified, a CNN model will be used to determine the use of masks to classify it as one of the three categories. This article is structured as follows, in Section 2 there is information from different authors who investigated the use of face masks, comparing some of their results, followed by Section 3, which deals with the methodology implemented to create the convolutional neural network model, as well as the dataset and preprocessing. In Section 4 the results obtained from the model are presented as real-time examples. Section 5 shows the conclusions, as well as future work. ## 2. Background Artificial intelligence and deep learning (DL) are technologies that have been constantly growing and many applications around them have been developed and implemented in the industry, as they have been applied in various areas, such as pattern recognition, and image processing, among others. CNNs have proven to be quite efficient at solving pattern problems, some standard architectures such as Resnet [7], YOLO [8], and MobileNet [9] already have an integrated convolutional neural network model. Our research group has worked with convolutional neural networks with diverse goals, as we can find in [10], where a new CNN model in combination with image preprocessing and optimization algorithms was proposed for diabetic retinopathy classification. Additionally, in [11] they use a deep neural network model for guitar classification, including fuzzy edge detection to improve the accuracy. In [12] a new hybrid approach was proposed, using fuzzy logic integration in combination with a modular artificial neural network. In other works, such as in [13], they propose the use of the MobileNetV2 [14] neural network architecture in combination with single shot detector (SSD) [15] performing real-time prediction using libraries such as OpenCV with an alert system to detect people who use or do not use the mask with the use of a Raspberry Pi 4 achieving between $85\%$ and $95\%$ accuracy percentage. A model called Facemasknet is proposed in [16] where they achieve $98.6\%$ to identify people who are wearing the mask, those wearing improperly, and those who do not have masks. In [17], the authors use a deep learning model to classify images into whether they use masks or no masks, using a small database of 460 images for non-masks and 380 for face masks implementing MobileNetV2. Additionally, in [18], a CNN is utilized to classify humans who are using the mask correctly, incorrectly, or not wearing masks, with the help of the Flickr-Faces-HQ and MaskedFaceNet database, achieving $98.5\%$ accuracy. Haar Cascades is widely used to identify the region of the face as it [19] uses it, in turn with the MobileNetV2 architecture and with the Real Facemask dataset and MaskedFaceNet dataset reaches $99\%$ accuracy. Similarly, in [20], they use technologies for mask identification, analyzing in real time the category to which it belongs, classifying into two classes Mask and NoMask, adding methods to improve the dataset, and eliminating images with low light. Several works have managed to detect the use of face masks, and each of them uses a different method to classify the correct use of the masks. It is very common to utilize neural network architectures already tested, for example, MobileNet [21], YOLO [22,23,24], Inception [25] or Resnet [26,27,28] each one reaching different percentages of precision and different type of classification. Other authors [29,30,31], have managed to solve the same problem with their own convolutional neural network models. Table 1 compares different studies carried out by various authors, where the purpose was to classify the correct use of face masks, generally performing a type of multiclass classification, although many others are based on binary studies. Most of the authors’ works use Python as a programming language in conjunction with its libraries, such as Tensorflow, OpenCV, and Keras, achieving between $90\%$ and $99\%$ accuracy. Other works have mostly used CNN architectures such as YOLO, MobileNet, or Resnet, whereas we proposed a CNN with one less convolutional layer that is fast to train on the MaskedFaceNet dataset, as we can realize from Table 1, where they proposed different classification models. Some other related works use the MobileNet architecture to demonstrate that ROpenPose runs faster than a number of the current state-of-the-art models and performs detection similarly [35]. The authors in [36], in order to breakdown and reconstruct spherical iris signals and extract more robust geometric properties of the iris surface, suggest using a spherical–orthogonal–symmetric Haar wavelet. Additionally, in [37], they propose a multi-stage unsupervised stereo matching method based on the cascaded Siamese network. In [38], a two-stage multi-view stereo network is suggested for quick and precise depth estimation. ## 3. Methodology The proposed real-time system mainly helps to decrease the spread of any infection that is transmitted by air, by identifying people according to the use of face masks. This can be carried out through monitoring and using alarms according to the rules that each location has. This section will cover the proposed solution to solve this problem, as well as the proposed architecture to detect people according to the use of face masks and enclose the region of the face in a rectangle. Basically, the system monitors and identifies the use of masks, see Figure 2, where from our convolutional neural network model it classifies people with real-time images into three classes: Mask which is when the person has a mask and green mark, IncorrectMask when the person has the mask incorrectly identified by the yellow color, and NoMask which is when the human is not using any mask and is indicated by red color. The proposed system utilizes computer vision techniques and deep learning techniques to detect the region of the face. There will be an automatic indication, using a Raspberry Pi 4 and a camera, of the people who are using masks, who do not wear masks, or use masks incorrectly. ## 3.1. Convolutional Neural Network In order to handle challenging image-driven pattern recognition problems, CNNs are typically utilized, and their precise and straightforward architecture makes using artificial neural networks (ANNs) easier, as they are very effective in accessing the graphic properties of the image. In this case, only MaskedFace-Net images were used, see Figure 3, using the four classes provided by this dataset. Multiple experiments were performed, where each was run 30 times to obtain the average precision and standard deviation, and the best case was identified for each experiment. The MaskedFace-Net dataset is classified into four classes: correctly masked, uncovered chin, uncovered nose, and uncovered nose and mouth, where the last three are part of a higher class called incorrectly masked. There have been proposed three different models to identify the CNN architecture that offers the best results. The architecture of the proposed CNN model basically comprises two stages, where the learning stage contains four convolutional layers with ReLu as the activation function and max pooling applied between each layer, and in the classification stage, the class to which it belongs is identified according to the three proposed classes: Mask, NoMask, and IncorrectMask. Therefore, the proposed general architecture is shown in Figure 4. This model is designed to classify the correct use of the face masks, utilizing images of the database and applying preprocessing to each of them. Basically, we first find the main characteristics of the region of the face found in the image, then this model in the learning stage uses four convolutional layers applying max pooling between each of them, adding ReLu as activation function. Finally, in the classification stage, it will be assigned to the class that belongs, including Mask, NoMask, or IncorrectMask. We compared other convolutional neural networks models, and this one provided us with better results and great performance. As we can note in Figure 5, the global model used to implement our method shows all the convolutional layers, the max pooling as pooling operation and classification stage, ending with a dense layer classified into three classes. ## 3.2. Database In this work, three datasets are used to perform the training and testing of the CNN model, one for each class. To create the Mask class, we use the Correctly Masked Face Dataset, also to create IncorrectMask class we utilize the Incorrectly Masked Face Dataset. These two datasets are part of MaskedFaceNet, this dataset has around 137,016 images of faces with simulated face masks, and is based on another dataset, and as mentioned by the authors in [2,3], the images are completely free to use under a license, and the classification efficiency of face masks has been corroborated through test with CNN models. The third dataset is the Flickr-Faces-HQ dataset, which contains images originally 1024 × 1024 in size with a wide variety of people in terms of background, age, or ethnicity, and this dataset was used to create the NoMask class, an example of the content of this dataset can be found in Figure 6. In addition, the first 15,000 images in total were selected for training, testing, and validation; therefore, 5000 images were used for each class. The CNN training was divided into different percentages where $70\%$ was used for training, $20\%$ was used for testing and the remaining $10\%$ was used for validation. ## 3.3. Data Pre-Processing Image preprocessing using images from the MaskedFace-Net database is performed by classifying and tagging them into three different types of mask wear. To improve the percentage of accuracy, a face detection model well known as the Caffe model [39] was used. For the preprocessing of the image an existing background subtraction was used, in this processing algorithm a technique known as RGB mean [40] subtraction is used, see Figure 7. This face detection and the preprocessing algorithm are applied to all images of the dataset to facilitate the identification of the mask and the class to which it belongs. ## Caffe Model Caffe model is a pre-trained model for the face detection algorithm that uses deep learning and computer vision techniques. This model is very efficient with faces at different angles, it is written in C ++ and provides tools for Python and Matlab. The model for face detection is pre-trained with 300 × 300 images, an example of face detection is shown in Figure 8. The Caffe model uses the Resnet-10 architecture and is based on Single Shot Multibox Detector (SSD). It is efficient with rapid head movements and occlusion managing to identify the region of the face very well from different sides or angles even when wearing a face mask. ## 3.4. Classification Once the face detection was performed and preprocessing applied, the model that has been trained to identify the mask is used to classify the corresponding class, see Figure 9. This trained model recognizes the state of the face mask use according to the image obtained from a video camera, recognizing the face region, and was tested with black, gray, and blue face masks. Loading the model enable obtaining a good percentage of accuracy and acceptable results for real-time prediction, Python libraries are basically used for the creation of the model sequence and classification. ## 3.5. Raspberry Pi A Raspberry Pi 4 is a small device with a 4 GB ARM processor and HDMI inputs, USB, and microSD ports. The operating system is based on GNU/Linux and is called Raspberry Pi OS, which is a custom version of Debian. The proposed system uses it in combination with a camera to obtain the image in real-time, to monitor and identify the use of face masks. Our system basically performs face detection through the model loaded on the Raspberry Pi, in addition to identifying the real-time status of mask use. The circuit, in Figure 10, was used through a protoboard to perform light-on tests, according to the identified class. The Raspberry Pi sends a signal through its GPIO Board turning on the green LED in case the mask is placed correctly, yellow in case the mask is placed incorrectly, and red when the mask is not placed. The camera of the Raspberry Pi is placed at strategic points, according to the need to monitor in real-time and continuously detect people. ## 4. Results and Discussion The proposed model was tested in 30 experiments, in Table 2 it can be seen that the training that obtained the best results was number 18, in which $99.69\%$ accuracy and $2.15\%$ loss were obtained. These results compared to [18] are better on average, using the same datasets, in a similar way, Python libraries such as Keras, OpenCv, and Tensorflow were used. The average obtained through the 30 trainings is of $99.60\%$ with a standard deviation of $0.04\%$. Achieving a mean not so far from the best average obtained in experiment 18 and a very small standard deviation, so the separation between the average and the mean value is small. The confusion matrix of the training with the best results is observed in Figure 11, where the heat map is shown with the evaluated images of the training percentage. Of the 2991 images, we find that 2979 images were correctly classified in their respective classes, which in percentage it would be $99.60\%$, and is quite good according to the MaskedFace-Net dataset. We classify different parts of the MaskedFace-Net to corroborate the efficacy of the proposed model, where the first part uses the first 15,000 images of the dataset, while part 2 the next 15,000. In the same way, part 3 evaluates the subsequent images, and part 4 uses the last images of the dataset. As can be seen in Table 3, where the different parts of the dataset are evaluated, the best percentage for obvious reasons was obtained with Part 1, achieving $99.90\%$ accuracy. This is because that part of the classification model of the use of face masks was trained, and although the percentage of accuracy is not higher in the other parts, this is still quite similar to obtaining percentages of high precision and low loss. This was achieved with the images of the dataset with which our model was not trained. Multiple tests were performed using a video camera to obtain the input images of the model so that it can be evaluated in real time. In Figure 12 the result of the classification of an image when a human is wearing the mask is shown. In addition to classifying the class to which it belongs, at the same time it lights an LED according to the class, in this case the class is green color, because the mask is placed correctly. Through labeling, it is easier to detect humans who are not using the mask correctly, managing to detect if they belong to any of the three classes, which are Mask, NoMask, or IncorrectMask. In most cases, the results are quite satisfactory, managing to classify mask use correctly. Figure 13 shows the classification of the remaining two classes, as well as the LED, which shows the color corresponding to the class, the yellow color for the IncorrectMask class, and the red color for the NoMask class. The Raspberry Pi 4 sends the signals to turn on the corresponding LED thanks to the GPIO Board and the programming made in Python to activate the appropriate pin and load the model for face detection and the model trained for the multiclass classification of the correct use of masks. ## 5. Conclusions The proposed method is capable of classifying the correct use of multiclass face masks using a CNN model, in combination with computer vision. This is with the goal of avoiding transmitting the COVID-19 virus or any other virus that can be transmitted by air and this is achieved through real-time monitoring in strategic areas, by means of a Raspberry Pi 4, where it is possible to perform this identification to perform specific actions. The method will detect the face of a person and put a rectangular box labeled as Mask if the person is using a mask correctly, which happened when a mask is covering the nose, mouth, and chin. Otherwise, if the person is wearing the mask only covering the chin and mouth then the model will label it in a rectangular box as Incorrect, or if the person is not wearing any mask or if it is only on the chin then the label NoMask will show. The model manages to classify the use of masks In three classes: NoMask, Mask, and IncorrectMask, and through the GPIO Board of the Raspberry Pi, sends signals to light green, yellow, or red LED, respectively, obtaining an accuracy percentage of $99.69\%$, evaluated with the MaskedFaceNet dataset. The proposed system, therefore, will potentially help in decreasing the spread of the virus, helping people’s health systems. This solution could prevent restrictions from being breached in real time, improving the safety of people around us, and can be used in various areas such as schools, squares, and public or private places, among others. ## References 1. Syed A., Shirin E., Muhammed K., Syed Q.. Real-Time Face Mask Detection in Deep Learning using Convolution Neural Network. Proceedings of the 10th IEEE International Conference on Communication Systems and Network Technologies (CSNT) 2. Cabani A., Hammoudi K., Benhabiles H., Melkemi M.. MaskedFace-Net—A dataset of correctly/incorrectly masked face images in the context of COVID-19. Smart Health (2020.0) 19 100144. DOI: 10.1016/j.smhl.2020.100144 3. Hammoudi K., Cabani A., Benhabiles H., Melkemi M.. Validating the Correct Wearing of Protection Mask by Taking a Selfie: Design of a Mobile Application “CheckYourMask” to Limit the Spread of COVID-19. Comput. Model. Eng. Sci. (2020.0) 124 1049-1059 4. Karras T., Laine S., Aila T.. A Style-Based Generator Architecture for Generative Adversarial Networks. IEEE Trans. Pattern Anal. Mach. Intell. (2021.0) 43 4217-4228. DOI: 10.1109/TPAMI.2020.2970919 5. Wai Zhao C., Jegatheesan J., Chee Loon S.. Exploring IOT Application Using Raspberry Pi. Int. J. Comput. Netw. Appl. (2015.0) 2 27-34 6. Militante S., Dionisio N.. Real-Time Facemask Recognition with Alarm System using Deep Learning. Proceedings of the 2020 11th IEEE Control and System Graduate Research Colloquium (ICSGRC) 7. Targ S., Almeida D., Lyman K.. Resnet in Resnet: Generalizing Residual Architectures. Proceedings of the ICLRW 2016 8. Redmon J., Divvala S., Girshick R., Farhadi A.. You Only Look Once: Unified, Real-Time Object Detection. Proceedings of the 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 9. Khasoggi B., Ermatita E., Samsuryadi S.. Efficient mobilenet architecture as image recognition on mobile and embedded devices. Indones. J. Electr. Eng. Comput. Sci. (2019.0) 16 389-394. DOI: 10.11591/ijeecs.v16.i1.pp389-394 10. Cordero-Martínez R., Sánchez D., Melin P.. Hierarchical genetic optimization of convolutional neural models for diabetic retinopathy classification. Int. J. Hybrid Intell. Syst. (2022.0) 18 97-109. DOI: 10.3233/HIS-220004 11. Torres C., Gonzalez C.I., Martinez G.E.. Fuzzy Edge-Detection as a Preprocessing Layer in Deep Neural Networks for Guitar Classification. Sensors (2022.0) 22. DOI: 10.3390/s22155892 12. Varela-Santos S., Melin P.. A new modular neural network approach with fuzzy response integration for lung disease classification based on multiple objective feature optimization in chest X-ray images. Expert Syst. Appl. (2021.0) 168 114361. DOI: 10.1016/j.eswa.2020.114361 13. Yadav S.. Deep Learning based Safe Social Distancing and Face Mask Detection in Public Areas for COVID-19 Safety Guidelines Adherence. Int. J. Res. Appl. Sci. Eng. Technol. (2020.0) 8 1368-1375. DOI: 10.22214/ijraset.2020.30560 14. Sandler M., Howard A., Zhu M., Zhmoginov A., Chen L.-C.. MobileNetV2: Inverted Residuals and Linear Bottlenecks. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) 15. Araki R., Onishi T., Hirakawa T., Yamashita T., Fujiyoshi H.. MT-DSSD: Deconvolutional Single Shot Detector Using Multi Task Learning for Object Detection, Segmentation, and Grasping Detection. Proceedings of the 2020 IEEE International Conference on Robotics and Automation (ICRA) 16. Inamdar M., Mehendale N.. Real-Time Face Mask Identification Using Facemasknet Deep Learning Network. SSRN Electron. J. (2020.0) 3663305. DOI: 10.2139/ssrn.3663305 17. Khandelwal P., Khandelwal A., Agarwal S.. Using Computer Vision to enhance Safety of Workforce in Manufacturing in a Post COVID World. arXiv (2020.0) 18. Jones D., Christoforou C.. Mask Recognition with Computer Vision in the Age of a Pandemic. Int. FLAIRS Conf. Proc. (2021.0) 34 1-6. DOI: 10.32473/flairs.v34i1.128486 19. Deshmukh M., Deshmukh G., Pawar P., Deore P.. Covid-19 Mask Protocol Violation Detection Using Deep Learning and Computer Vision. Int. Res. J. Eng. Technol. (2021.0) 8 3292-3295 20. Vibhuti N., Jindal H., Singh P.. Face mask detection in COVID-19: A strategic review. Multimed. Tools Appl. (2022.0) 81 40013-40042. DOI: 10.1007/s11042-022-12999-6 21. Rudraraju S., Suryadevara N., Negi A.. Face Mask Detection at the Fog Computing Gateway. Proceedings of the 2020 Federated Conference on Computer Science and Information Systems 22. Singh S., Ahuja U., Kumar M., Kumar K., Sachdeva M.. Face mask detection using YOLOv3 and faster R-CNN models: COVID-19 envirorment. Multimed. Tools Appl. (2021.0) 80 19753-19768. DOI: 10.1007/s11042-021-10711-8 23. Yu J., Zhang W.. Face Mask Wearing Detection Algorithm Based on Improved YOLO-v4. Sensors (2021.0) 21. DOI: 10.3390/s21093263 24. Jiang X., Gao T., Zhu Z., Zhao Y.. Real-Time Face Mask Detection Method Based on YOLOv3. Electronics (2021.0) 10. DOI: 10.3390/electronics10070837 25. Wang B., Zhao Y., Chen P.. Hybrid Transfer Learning and Broad Learning System for Wearing Mask Detection in the COVID-19 Era. IEEE Trans. Instrum. Meas. (2021.0) 70 1-12. DOI: 10.1109/TIM.2021.3069844 26. Bhattarai B., Raj Pandeya Y., Lee J.. Deep Learning-based Face Mask Detection Using Automated GUI for COVID-19. Proceedings of the 6th International Conference on Machine Learning Technologies 27. Pham-Hoang-Nam A., Le-Thi-Tuong V., Phung-Khanh L., Ly-Tu N.. Densely Populated Regions Face Masks Localization and Classification Using Deep Learning Models. Proceedings of the Sixth International Conference on Research in Intelligent and Computing 28. Soto-Paredes C., Sulla-Torres J.. Hybrid Model of Quantum Transfer Learning to Classify Face Images with a COVID-19 Mask. Int. J. Adv. Comput. Sci. Appl. (2021.0) 12 826-836. DOI: 10.14569/IJACSA.2021.0121092 29. Das A., Wasif Ansari M., Basak R.. Covid-19 Face Mask Detection Using TensorFlow, Keras and OpenCV. Proceedings of the 2020 IEEE 17th India Council International Conference (INDICON) 30. Kaur G., Sinha R., Tiwari P., Yadav S., Pandey P., Raj R., Vashisth A., Rakhra M.. Face mask recognition system using CNN model. Neurosci. Inform. (2022.0) 2 100035. DOI: 10.1016/j.neuri.2021.100035 31. Sethi S., Kathuria M., Mamta T.. A Real-Time Integrated Face Mask Detector to Curtail Spread of Coronavirus. Comput. Model. Eng. Sci. (2021.0) 127 389-409. DOI: 10.32604/cmes.2021.014478 32. Larxel. Face Mask Detection 33. Jangra A.. Face Mask Detection 12K Images Dataset 34. Aydemir E., Yalcinkaya M., Barua P., Baygin M., Faust O., Dogan S., Chakraborty S., Tuncer T., Acharya R.. Hybrid Deep Feature Generation for Appropriate Face Mask Use Detection. Int. J. Environ. Res. Public Health (2022.0) 19. DOI: 10.3390/ijerph19041939 35. Wu E.Q., Tang Z.R., Xiong P., Wei C.F., Song A., Zhu L.M.. ROpenPose: A Rapider OpenPose Model for Astronaut Operation Attitude Detection. IEEE Trans. Ind. Electron. (2022.0) 69 1043-1052. DOI: 10.1109/TIE.2020.3048285 36. Hou Y., Wu E.Q., Cao Z., Xu X., Zhu L.M., Yu M.. Spherical-Orthogonal-Symmetric Haar Wavelet for Driver’ s Visual Detection. IEEE Transactions on Intelligent Vehicles (2022.0) 1-11 37. Tong W., Sun Z.H., Wu E., Wu C., Jiang Z.. Adaptive Cost Volume Representation for Unsupervised High-resolution Stereo Matching. IEEE Transactions on Intelligent Vehicles (2022.0) 1 38. Tong W., Guan X., Kang J., Sun P.Z.H., Law R., Ghamisi P., Wu E.Q.. Normal Assisted Pixel-Visibility Learning With Cost Aggregation for Multiview Stereo. IEEE Trans. Intell. Transp. Syst. (2022.0) 23 24686-24697. DOI: 10.1109/TITS.2022.3193421 39. Jia Y., Shelhamer E., Donahue J., Karayev S., Long J., Girshick R., Guadarrama S., Darrell T.. Caffe: Convolutional Architecture for Fast Feature Embedding. Proceedings of the MM 2014—ACM Conference on Multimedia 40. Chen Y.-Q., Sun Z.-L., Lam K.-M.. An Effective Subsuperpixel-Based Approach for Background Subtraction. IEEE Trans. Ind. Electron. (2020.0) 67 601-609. DOI: 10.1109/TIE.2019.2893824
--- title: 'Investigation on the Essential Oils of the Achillea Species: From Chemical Analysis to the In Silico Uptake against SARS-CoV-2 Main Protease' authors: - Hossein Rabbi Angourani - Armin Zarei - Maryam Manafi Moghadam - Ali Ramazani - Andrea Mastinu journal: Life year: 2023 pmcid: PMC9967057 doi: 10.3390/life13020378 license: CC BY 4.0 --- # Investigation on the Essential Oils of the Achillea Species: From Chemical Analysis to the In Silico Uptake against SARS-CoV-2 Main Protease ## Abstract In this study, phytochemicals extracted from three different Achillea genera were identified and analyzed to be screened for their interactions with the SARS-CoV-2 main protease. In particular, the antiviral potential of these natural products against the SARS-CoV-2 main protease was investigated, as was their effectiveness against the SARS-CoV-1 main protease as a standard (due to its high similarity with SARS-CoV-2). These enzymes play key roles in the proliferation of viral strains in the human cytological domain. GC-MS analysis was used to identify the essential oils of the Achillea species. Chemi-informatics tools, such as AutoDock 4.2.6, SwissADME, ProTox-II, and LigPlot, were used to investigate the action of the pharmacoactive compounds against the main proteases of SARS-CoV-1 and SARS-CoV-2. Based on the binding energies of kessanyl acetate, chavibetol (m-eugenol), farnesol, and 7-epi-β-eudesmol were localized at the active site of the coronaviruses. Furthermore, these molecules, through hydrogen bonding with the amino acid residues of the active sites of viral proteins, were found to block the progression of SARS-CoV-2. Screening and computer analysis provided us with the opportunity to consider these molecules for further preclinical studies. Furthermore, considering their low toxicity, the data may pave the way for new in vitro and in vivo research on these natural inhibitors of the main SARS-CoV-2 protease. ## 1. Introduction Iran is home to a significant share of plant species and endless natural habitats characterized by numerous special plants and centers of local endemic species [1]. A. millefolium, with the common name yarrow, is a member of the Asteraceae family and is a flowering plant [2] comprising 130 species, 19 of which are grow wild in Iran [1]. As far as the Achillea genus has existed across the world (from Asia and Europe to North America), it has been used in ancient life in traditional or folk medicine. Bumadaran is a common name for some species of Achillea in Persian culture [3]. The genus Achillea has a long history of use in traditional medicine as an anti-inflammatory, diaphoretic, anti-spasmodic, tonic, diuretic, and emmenagogic agent [3], and it has been used as a natural remedy (in Iranian traditional medicine) for the treatment of bleeding, headaches, respiratory infections, inflammation, spasmodic diseases, flatulence, dyspepsia [4,5], pneumonia, hemorrhaging, rheumatic pain, and wounds, and it is useful for liver disease and acts as a mild sedative [6]. Currently, the different medicinal functions of yarrow, such as its use as a spasmolytic, a choleretic, a wound-healing treatment, and an anti-inflammatory treatment, as well as its antifungal activities, have made it an important medicinal plant [3,7,8,9,10,11]. A. willhelmsii C. Koch (Asteraceae) possesses antioxidant, anti-inflammatory, antimicrobial, and antiulcerogenic properties, and it has been used to treat multiple diseases, as well as infections, hemorrhaging, pneumonia, rheumatic pain, and wounds [12]. The phytochemical characterization of the essential oil (EO) of A. wilhelmsii has been the subject of very few studies, most of which have focused on a small number of its molecules, including camphor, linalool, borneol, carvacrol, p-cymene, 1,8-cineole, and thymol, all of which show high potential for antioxidant activities [1]. A. tenuifolia Lam. ( AT) is another member of the Asteraceae family that has been used in many cultures for more than 3000 years [13]. The attractive properties of AT extract include its anti-inflammatory, antitumor, antioxidant, and antimicrobial activities [14]. The antioxidant activity of AT is attributed to the presence of flavonoids and phenolic components [13,15]. Several studies have indicated that the biological function of AT is associated with its chemical composition. In addition, a correlation was found between the total phenol content and the antioxidant capacity of AT extract [16]. The conservation of plant species (Figure 1) and innumerable natural habitats characterized by many unique plants and endemic centers is essential for pharmacological research [17,18,19,20,21,22,23,24,25,26]. Aromatic and medicinal plants produce a wide variety of volatile terpene hydrocarbons (aliphatic and cyclic) in addition to their corresponding oxygenated isoprenoid derivatives and analogs. Mixtures of these compounds, known as essential oils (EOs), can be isolated from diverse parts of plants by steam distillation and have excellent bioactivities, such as antimicrobial properties [2,27,28,29,30,31]. Moreover, EOs account for only a small portion of a plant’s composition; however, they determine the vital characteristics of aromatic plants [32]. As secondary metabolites, EOs involve complex mixtures of natural compounds with versatile organic structures that represent useful medicinal properties [33,34,35] which can be extracted from different parts of plant materials using classical and advanced techniques [36]. The composition of EOs is greatly influenced by diverse parameters, including the time and season of harvesting of the plants, the type of plant organs and the plant’s corresponding family, the plant’s geographical and climatic conditions, the plant’s physiological age, and the plant’s growth stage [10]. In addition, EOs and aromatic extracts are typically used as perfuming agents, pharmaceuticals, and food flavors, as well as in aromatherapy. Although the sophistication and complexity of EOs are undesirable properties for drug discovery, they are valuable in medical therapy and may be used as promising sources of novel drugs. Extracted compounds from EOs may be useful options against COVID-19, according to a recent study in which some signs of the inhibitory potential of EO constituents against the SARS-CoV-2 main protease have been highlighted [37]. Over the last few decades, viral diseases such as H1N1 flu, SARS-CoV-1, SARS-CoV-2, Ebola, and Mers have posed a serious threat to human society [38]. In 2019, a novel coronavirus disease (COVID-19) outbreak occurred in Wuhan, China. Since then, this viral disease has spread quickly worldwide, with millions of casualties [24,25,39]. This viral disease is caused by SARS-CoV-2 (syndrome acute respiratory CoronaVirus-2), which belongs to the beta-coronavirus family and shares a similarity of up to $79\%$ to SARS-CoV-1, which underlies the next-generation sequencing technology. The SARS-CoV-2 3C-like protease (3CLpro), or the main protease (Mpro), plays a vital role in the proteolytic processing of coronaviruses and creates other significant proteins for viral replication [40]. Hence, this protease may be the best target for therapeutic repositioning to identify promising antiviral drugs against SARS-CoV-2. Although pharmaceutical companies and global research groups have made efforts to find promising antiviral drugs to prevent the spread of COVID-19, some monoclonal antibodies, such as bamlanivimab–etesevimab [41] and a cocktail of casirivimab–imdevimab (REGEN-COV™) [42], and single monoclonal sotrovimab [43] have been authorized by the United States Food and Drug Administration (FDA) under emergency use authorization (EUA) for the treatment of patients (≥12 years of age) suffering from mild to moderate COVID-19 who are at high risk of progression to severe COVID-19 and/or hospitalization. However, this authorization was revoked on April 16, 2021 owing to the resistance of variants to bamlanivimab [44]. Remdesivir has been authorized for use in pediatric and adult patients (≥12 years of age) that require hospitalization, though it requires administration by injection or infusion in a therapeutic setting, with frequent monitoring [44,45]. In December 2021, two new oral antiviral agents, molnupiravir and nirmatrelvir/ritonavir (Paxlovid), gained emergency use authorization from the FDA for adult patients with mild to moderate COVID-19 who were at high risk of progression to severe COVID-19 under certain limitations [44,46,47]. Tocilizumab [48,49] and baricitinib [50,51] (both anti-inflammatory agents) have also gained EAU from the FDA for people of ≥ 18 years of age with COVID-19 who were admitted to a hospital [52]. The most promising drug target is the main protease of SARS-CoV-2 (Mpro or 3CLpro) because it plays a vital role in viral replication and transcription. Thus, the main axis of recent research has focused on the rapid development of SARS-CoV-2 Mpro inhibitors as drug candidates. Several techniques have been used for the discovery of SARS-CoV-2 Mpro inhibitors, including the high-throughput screening of structurally diverse compound libraries [53,54], drug repurposing [55,56], structure-based drug design [57,58], and in silico studies [59,60]. Despite considerable efforts, only one compound, nirmatrelvir/ritonavir (Paxlovid), has gained EUA by the FDA for adult patients with mild to moderate COVID-19 who are at a high risk of hospitalization [46], and a new compound (S-217622) is undergoing clinical trials (NCT05305547) [59]. Owing to their long history of folk use in Iranian traditional medicine and because of their low toxicity and excellent pharmacokinetics, the Achillea genera, alongside the reported biological activities of three Achillea species (A. millefolium, A. wilhelmsii C. Koch, and A. tenuifolia Lam), were the subject of this research. We wanted to address the question of whether the extracted compounds from three different EOs of the Achillea species can be the inhibitors, or have the potency to be a possible source of strong and/or effective inhibitors, of the SARS-CoV-2 3CL protease. The antiviral potential of the extracted compounds (from their EOs) against SARS-CoV-2 3CLpro (and SARS-CoV-1 3CLpro, as a standard) was investigated for the first time. Initially, the EOs of A. millefolium, A. wilhelmsii C. Koch, and A. tenuifolia Lam were obtained and their compounds were identified using GC-MS. The identified compounds were selected to investigate their antiviral potential against viral targets using AutoDcock 4.2.6. These compounds also showed low toxicity and very good pharmacokinetic properties in the SwissADME and ProTox-II evaluations. The compounds with the lowest binding energies (with the viral targets) were the best antiviral candidates against the viral protease and are worthy of further in vitro and in vivo analyses to clarify their antiviral potential. ## 2.1. Samples (Plant Materials) The plants (three species of the Achillea genus: A. millefolium, A. wilhelmsii C. Koch, and A. Tenuifolia Lam) were collected during June and July 2021 from their natural habitat in the village of Homayoun in Zanjan province, Iran (elevation 1868 m, latitude east 48°28′35.36″, longitude 36°45′5.28″). A voucher specimen was maintained at the Research Institute of Modern Biological Techniques (RIMBT), University of Zanjan. Mr. Angourani deposited the botany herbarium voucher specimens under the following numbers: A. millefolium, number 1518; A. tenuifolia Lam, number 1522; and A. wilhelmsii C. Koch, number 1528. All samples were prepared from the whole flowers and leaves of the plants that were harvested at the full bloom stage of growth and then air-dried at room temperature in the shade. ## 2.2. Essential Oil Extraction The essential oils from the flowers and leaves of the plants (A. millefolium, A. wilhelmsii C. Koch, and A. tenuifolia Lam) were extracted by hydrodistillation for approximately 4 h using a Clevenger apparatus in accordance with the British Pharmacopeia [1988] [61]. The finely cut plant material was fully submerged in distilled water (600 mL) in a round-bottom flask, which was then placed in a heating mantle. The condensed vapors were collected in a separating funnel and the resulting EO was separated, dehydrated with an appropriate amount of Na2SO4, filtered, and stored in brown vials at 4 °C. ## 2.3. GC-MS Analysis of the Essential Oil Components The EOs were dehydrated with Na2SO4, and the samples were injected into the device under the thermal planning of the column. The compounds of the EOs were identified using the retention time (R.T) and the Kovats index (R.I) and by studying the mass spectra and checking these parameters with the standard compounds in the GC/MS library. The EO extraction yields were estimated for 100 g of fresh plant material. The EOs were analyzed using gas chromatography-mass spectrometry (GC–MS). This device includes a gas chromatograph (model 7890B) and mass spectrometer (model 5977A) made by Agilent Company of America that was equipped with a split/splitless injection system and an ion bombardment ionization model (possessing the NIST and WILEY mass libraries). To analyze the fatty acids, an HP5-MS column with a length of 60 m, an inner diameter of 0.25 mm, and a thickness of 0.25 μm was used. The injection site, interface, and ionization site temperatures were set to 280, 290, and 250 °C, respectively. The temperature program of the column was triggered at an initial temperature of 60 °C, which remained constant for 5 min. The temperature reached 180 °C (with a slope of 15 °C/min) for 2 min before increasing to 280 °C (with a slope of 20 °C/ min), and it remained at this temperature for 10 min. The split ratio was set to 1 to 20 and the injection volume was half of a microliter. ## 2.4. Molecular Docking The molecular and 3D structures of the targets (SARS-CoV-2 3CLpro: PDB code 6LU7 and SARS-CoV-1 3CLpro: PDB code 2H2Z) were extracted from the Protein Data Bank (www.rcsb.org (accessed on 1 January 2022)). To prepare the molecular structures of the natural compounds, their 2D chemical conformations were first drawn using the ChemSketch tool of the ACD/LAB package (www.acdlabs.com (accessed on 1 December 2022)) before transferring them to the Avogadro package to minimize energy and optimize their conformation via a steep algorithm. ## Docking Studies Molecular docking is a highly important sub-technique of molecular modelling that plays a significant role in computer-based drug design. The docking protocol was divided into two general stages using the AutoDock Autogrid package. Initially, blind docking was performed for all viral targets. Structures that bound to the active site of proteases with binding energies of >5.5 kcal/mol were chosen for the second stage (targeted docking). First, gastiger charges and polar hydrogen atoms were added to all ligands using the MGLtools package [62]. All bonds were set as active for ligands, and energetic maps were estimated for each atom type using Autogrid 4. For blind docking, the search space was set as large as all proteins were accessible for ligand binding. The active sites of the viral targets were set as search spaces for targeted docking. Finally, 250 runs of molecular docking according to the Lamarckian genetic algorithm were performed for ligands that met the conditions of the first step of the docking protocol [63]. Discovery studio 4.5 was used for the 3D visualizations of the results, and LigPlot+ V.2.2 was applied for the 2D presentations. ## 3.1. Achillea millefolium (Yarrow) The essential oils of the flowering and leafy branches of the dried vegetative body of A. millefolium were obtained via water distillation and a Clevenger apparatus, and the oil yield was $0.74\%$ (content (% v/w)). The GC-MS results showed that the EO of the plant was composed of 105 compounds, of which 18 compounds accounted for $60.5\%$ of the essential oil (Table 1 and Figure 2). ## 3.2. A. wilhelmsii C. Koch The essential oil of A. wilhelmsii C. Koch was yellow in color, with a yield of $0.89\%$ (% v/w). The GC-MS results revealed that the essential oil of this plant in the target area comprised 106 substances, of which 21 compounds represented $66.93\%$ of the total essential oil (Table 2 and Figure 3). ## 3.3. A. tenuifolia Lam According to the GC-MS data, the essential oil of this plant contained 88 compounds. The essential oil of A. tenuifolia Lam was orange-yellow in color, with a yield of $0.82\%$ (% v/w). The main chemical composition of the desert yarrow essential oil is presented in Table 3 and Figure 4. We identified 21 compounds that constituted $73.48\%$ of the essential oil. ## 3.4. Docking Studies of SARS-CoV-2 3CLpro Molecular docking is a potent approach used to elucidate protein and small-molecule interactions [24,25,64]. Among the identified compounds from the EOs of the three different types of yarrow tested against SARS-CoV-2 3CLpro, four natural inhibitors (Kessanyl acetate, Chavibetol (m-Eugenol), Farnesol, and 7-epi-β-Eudesmol) extracted from the EO of A. millefolium exhibited strong inhibitory effects against the viral target (Table 4). The binding energies of the proteases are listed in Table 4. It was clear that 3CLpro (or the main protease) catalyzed the most important maturation cleavage phenomenon and also played a significant role in the viral replication of SARS-CoV-2, signifying its importance as a drug target [65]. As shown in Figure 5A (for the 2D version, see Figure S4), kessanyl acetate formed three H-bonds with the viral target: two with Ser 144 and one with the important residue Cys 145 (located in domain II and acting as a nucleophile in the initial step of the hydrolysis reaction in the catalytic region of the enzyme with the assistance of His 41 as a base catalyst) [65]; therefore, the interaction with this residue may have remarkably caused a malfunction in its enzymatic activities. Moreover, the hydrophobic interactions between kessanyl acetate and the important residues located in domains I and II [65], including His 41, Met 49, Leu 141, Asn 142, Gly 143, and Gln 189, can ratify the potent inhibitory effect of this natural product on the active site of the enzyme. Cys 145 is located in the active site positioned in the oxyanion loop (residues 138–146), which, along with Gly 143, defines an oxyanion hole (via interactions between their amide groups and the carbonyl group of peptides), and, together with the β-strand segment (His163-Pro168), they are highly important for the preparation of active sites and substrate binding. Furthermore, residues near the loop segment play a significant role in expanding the active site [65]. Based on this information, it is worth pointing out that chavibetol (m-eugenol) binds to the active site of 3CLpro through five strong hydrogen bonds with the residues Leu 141, Gly 143, Ser 144, and His 163, showing that this natural inhibitor may have high impairment and disruptive effects on the active site, and it may potentially disturb active site formation and accurate substrate binding (Figure 5B). These disruptive effects on the active site formation and substrate binding of 3CLpro can be seen for the other two remaining natural inhibitors (Farnesol and 7-epi-β-Eudesmol) because each of them form H-bonds with Leu 141 and Ser 144, although the former shows an additional H-bond with His 163. Interestingly, both had hydrophobic interactions with Cys 145 and His 41, showing their strong antiviral effects on the catalytic function of this enzyme (Figure 5C,D). Taken together, these results show that the binding of the four natural inhibitors (extracted from A. millefolium) to the active site of 3CLpro may disrupt substrate binding and active site formation, thereby impeding catalytic functions. ## 3.5. Docking Studies of SARS-CoV-1 3CLpro To investigate the potency of the extracted natural products against other major proteases, we selected SARS-CoV-1 3CLpro by using the same docking procedure. Based on the docking data, it could be asserted that four of the tested ligands showed higher binding energies (as they did against SARS-CoV-2 3CLpro). The 3CL proteases of SARS-CoV-1 and SARS-CoV-2 share approximately $96\%$ sequence identity, indicating that they are similar to each other [66], and SARS-CoV-1 3CLpro also possesses the three main domains, I, II, and III, and its catalytic dyad is composed of His 41 and Cys 145, in which its substrate-binding pocket is located in the cleft between I and II [67]. Interestingly, kessanyl acetate binds to this viral receptor through three H-bonds with Gly 143, Ser144, and Cys 145, which, in turn, may hinder the catalytic site and cause strong disorganization in the enzymatic function of the protease. In addition, hydrophobic interactions with other residues located in the active site (Leu 141 and Asn 142) can prove its high preventive effects on the catalytic site (Figure 5E). As shown in Figure 5F, all interacting residues were located at the active site. Moreover, Chavibetol (m-eugenol) formed five strong hydrogen bonds with amino acid residues located in the catalytic region of domain II (Gly 143, Gly 143, Ser144, and His 163), and the active site entry may have been impaired [67], and the catalytic activities of such enzymes can be remarkably diminished. With regard to the interaction of SARS-CoV-2 3CLpro with Farnesol and 7-epi-β-Eudesmol, it can be mentioned that both have formed H-bonds with Leu 141 and Ser 144, along with hydrophobic interactions with the significant catalytic dyad residues His 41 and Cys 145, indicating that these two natural inhibitors may possess highly potent activity for hindering the catalytic region of the protease, and they can potentially disrupt its enzymatic function (Figure 5G,H). Thus, these four natural inhibitors may have strong inhibitory effects on the active site of SARS-CoV-1 3CLpro and could have a preventive and momentous disruptive impact on its catalytic function. The four selected natural products were pharmacokinetically investigated using SwissADME. Interestingly, according to Table 5 and Table 6, all four natural inhibitors satisfied the drug-likeliness properties, indicating a high bioavailability score. Remarkably, none of them served as P-glycoprotein (P-gp) substrates, showing excellent oral availability of these ligands and indicating that they have no inhibitory or inductive effect on metabolizing enzymes [68,69]. More importantly, all showed high gastrointestinal absorption and could cross the blood–brain barrier, indicating that passing the cell membrane is vital for the components to enter the cells infected with SARS-CoV-2 and obtain successful access to the viral 3CLpro of SARS-CoV-2. Consistent with our results, it should be noted that kessanyl acetate and farnesol possessed better pharmacokinetic characteristics, such as pharmacological and drug-likeliness properties. Although Kessanyl acetate did not interact with other medications, Farnesol and Chavibetol (m-eugenol) showed inhibitory effects on CYP1A2, which can be anticipated to interact with the metabolism of some typical anxiety or dispersive disorder drugs. Considering all of these properties, it can be postulated that these four naturally occurring compounds may be effectively absorbed, distributed, and diffused within the body. Finally, the toxicities of the four compounds were predicted using the ProTox-II online tool. As shown in Table 6, all compounds exhibited modest or light toxicity, and Farnesol and Kessanyl acetate showed higher LD50 values. Based on the results obtained in this study, it can be asserted that all four natural compounds (particularly kessanyl acetate and farnesol) may be the most desirable candidates for further in vitro and in vivo investigations of SARS-CoV-2. ## 4. Conclusions The EO compounds from three different types of Achillea (A. wilhelmsii C. Koch, A. tenuifolia Lam, and A. millefolium) were explored using a gas chromatography device connected to a mass spectrometer. The antiviral potential of the identified compounds from the EOs of three different types of Achillea against SARS-CoV-2 3CLpro (and SARS-CoV-1 3CLpro, as a standard) was computationally evaluated using Autodock 4.2.6. Our findings revealed that among the identified compounds (from the EOs of Achillea genera), four compounds extracted from the EO of A. millefolium (Kessanyl acetate, Chavibetol (m-Eugenol), farnesol, and 7-epi-β-Eudesmol) showed the highest binding energies and the maximum number in the clusters, revealing a high potential to demolish the active site of the proteases and ruin their catalytic functions by forming H-bonds with two important amino acid residues of CYS 145 and HIS 41. Furthermore, the ADME properties of the final compounds were determined to evaluate their pharmacokinetic characteristics. Our results revealed that kessanyl acetate and farnesol are the most suitable natural compounds with regard to drug pharmacokinetics. Based on the findings of the present study, it should be mentioned that kessanyl acetate and farnesol are more potent inhibitors of 3CL-proteases than the other studied natural compounds, and they may be applied in future experimental studies to identify effective anti-SARS-CoV-2 compounds. ## References 1. Mirdavoudi H., Ghorbanian D., Zarekia S., Soleiman J.M., Ghonchepur M., Sweeney E.M., Mastinu A.. Ecological Niche Modelling and Potential Distribution of. Land (2022) 11. DOI: 10.3390/land11122315 2. Magiatis P., Skaltsounis A.L., Chinou I., Haroutounian S.A.. Chemical composition and in-vitro antimicrobial activity of the essential oils of three Greek Achillea species. Z. Naturforsch. C J. Biosci. (2002) 57 287-290. DOI: 10.1515/znc-2002-3-415 3. Saeidnia S., Gohari A., Mokhber-Dezfuli N., Kiuchi F.A.. A review on phytochemistry and medicinal properties of the genus Achillea. DARU J. Fac. Pharm. Tehran Univ. Med. Sci. (2011) 19 173 4. Final report on the safety assessment of Yarrow (. Int. J. Toxicol. (2001) 20 79-84. DOI: 10.1080/10915810160233785 5. Chandler R., Hooper S., Harvey M.J.. Ethnobotany and phytochemistry of yarrow,. Econ. Bot. (1982) 36 203-223. DOI: 10.1007/BF02858720 6. Applequist W.L., Moerman D.E.. Yarrow (. Econ. Bot. (2011) 65 209-225. DOI: 10.1007/s12231-011-9154-3 7. Ali S.I., Gopalakrishnan B., Venkatesalu V.. Pharmacognosy, phytochemistry and pharmacological properties of. Phytother. Res. (2017) 31 1140-1161. DOI: 10.1002/ptr.5840 8. Benedek B., Kopp B., Melzig M.F.. Achillea millefolium L. sl–Is the anti-inflammatory activity mediated by protease inhibition?. J. Ethnopharmacol. (2007) 113 312-317. DOI: 10.1016/j.jep.2007.06.014 9. Mehraliyevaa S., Valiyeva M., Abbasli N., Suleymanova T., Musayeva S., Davaran S., Khalilov R., Eftekhari A.. Development of novel antibacterial gel using clove and calendula extracts with colloidal silver nanoparticles. ECC (2021) 3 170-179 10. Mohammadhosseini M., Sarker S.D., Akbarzadeh A.. Chemical composition of the essential oils and extracts of Achillea species and their biological activities: A review. J. Ethnopharmacol. (2017) 199 257-315. DOI: 10.1016/j.jep.2017.02.010 11. Rehan T., Al-Lami N., Alanee R.. Anti-cancer and antioxidant activities of some new synthesized 3-secondary amine derivatives bearing imidazo [1, 2-A] pyrimidine. Eurasian Chem. Commun. (2021) 3 339-351 12. Mirzaei M., Hajali N., Pourhadi M.. In Silico Interactions of Morphine Opioids with TLR4 and TRPV1 Targets. Biointerface Res. Appl. Chem. (2022) 13 208 13. Manayi A., Mirnezami T., Saeidnia S., Ajani Y.. Pharmacognostical evaluation, phytochemical analysis and antioxidant activity of the roots of. Pharmacogn. J. (2012) 4 14-19. DOI: 10.5530/pj.2012.30.3 14. Moradkhani S., Kobarfard F., Ayatollahi S.A.M.. Phytochemical investigations on chemical constituents of. IJPR (2014) 13 1049. PMID: 25276207 15. Gharibi S., Tabatabaei B.E.S., Saeidi G.. Comparison of essential oil composition, flavonoid content and antioxidant activity in eight. J. Essent. Oil Bear. Plants (2015) 18 1382-1394. DOI: 10.1080/0972060X.2014.981600 16. Kazemi M.. Gas chromatography-mass spectrometry analyses for detection and identification of antioxidant constituents of. Int. J. Food Prop. (2015) 18 1936-1941. DOI: 10.1080/10942912.2014.951893 17. Amirkhani R., Zarei A., Gholampour M., Tavakoli H., Ramazani A.. Honey and royal jelly natural products as possible antiviral nominations to combat SARS-CoV-2 main protease. ECC (2022) 4 567-579 18. Dodangeh M., Ramazani A., Maghsoodlou M.T., Zarei A., Rezayati S.. Application of readily available metals for CH activation. Curr. Org. Chem. (2020) 24 1582-1609. DOI: 10.2174/1385272824999200616114037 19. Fadel Z., Al-Azzawi A.M.. Design, Synthesis and Antimicrobial Activity Evaluation of New Bisimidyl Sulfonamido Ketone Comprising Drug component. Chem. Methodol. (2021) 5 464-470 20. Mohammadi S.S., Ghasemi N., Ramezani M.. Bio-fabrication of silver nanoparticles using naturally available wild herbaceous plant and its antibacterial activity. Eurasian Chem. Commun. (2020) 2 87-102 21. Phillipson J.D.. Phytochemistry and pharmacognosy. Phytochemistry (2007) 68 2960-2972. DOI: 10.1016/j.phytochem.2007.06.028 22. Ramazani A., Shahkarami F., Zarei A., Rezayati S., Rezaei A., Bodaghi A., Youseftabar-Miri L.. Shikimic acid from staranise (Illicium verum Hook): Extraction, purification and determination. ECC (2021) 452-460 23. Taghvaei M., Fazilati M., Nazem H., Habibollahi S., Bastani F.. Comparison of Different Extracts of Nettle in Quercetin and Evaluation of its Antimicrobial Activity. Chem. Methodol. (2020) 4 572-583 24. Zarei A., Ramazani A., Pourmand S., Sattari A., Rezaei A., Moradi S.. In silico evaluation of COVID-19 main protease interactions with honeybee natural products for discovery of high potential antiviral compounds. Nat. Prod. Res. (2021) 36 4254-4260. DOI: 10.1080/14786419.2021.1974435 25. Zarei A., Ramazani A., Rezaei A., Moradi S.. Screening of honey bee pollen constituents against COVID-19: An emerging hot spot in targeting SARS-CoV-2-ACE-2 interaction. Nat. Prod. Res. (2022) 1-7. DOI: 10.1080/14786419.2022.2092865 26. Zarei A., Fardood S.T., Moradnia F., Ramazani A.. A Review on coronavirus family persistency and considerations of novel type, COVID-19 features. Eurasian Chem. Commun. (2020) 2 798-811. DOI: 10.33945/SAMI/ECC.2020.7.7 27. Angourani H.R., Heydari M., Yousefi A.R., Pashaei B., Mastinu A.. Nanoparticles Based-Plant Protein Containing Rosmarinus officinalis Essential Oil; Fabrication, Characterization, and Evaluation. Appl. Sci. (2022) 12. DOI: 10.3390/app12199968 28. Ghanaat J., Khalilzadeh M., Zareyee D.. KF/CP NPs as an efficient nanocatalyst for the synthesis of 1, 2, 4-triazoles: Study of antioxidant and antimicrobial activity. ECC (2020) 2 2020-2212 29. Gharibvandi A., Karimmojeni H., Ehsanzadeh P., Rahimmalek M., Mastinu A.. Weed management by allelopathic activity of Foeniculum vulgare essential oil. Plant Biosyst. Int. J. Deal. All Asp. Plant Biol. (2022) 156 1298-1306. DOI: 10.1080/11263504.2022.2036848 30. Ammarellou A., Yousefi A.R., Heydari M., Uberti D., Mastinu A.. Biochemical and Botanical Aspects of. BioTech (2022) 11. DOI: 10.3390/biotech11020016 31. Jam B.J., Shekari F., Andalibi B., Fotovat R., Jafarian V., Najafi J., Uberti D., Mastinu A.. Impact of Silicon Foliar Application on the Growth and Physiological Traits of. Silicon (2022). DOI: 10.1007/s12633-022-02090-y 32. Moradi P., Aghajanloo F., Moosavi A., Monfared H.H., Khalafi J., Taghiloo M., Khoshzaman T., Shojaee M., Mastinu A.. Anthropic Effects on the Biodiversity of the Habitats of. Sustainability (2021) 13. DOI: 10.3390/su13147874 33. Mohammadhosseini M., Nekoei M.. Chemical compositions of the essential oils and volatile compounds from the aerial parts of Ferula ovina using hydrodistillation, MAHD, SFME and HS-SPME methods. J. Essent. Oil Bear. Plants (2014) 17 747-757. DOI: 10.1080/0972060X.2014.884951 34. Paidi R.K., Jana M., Raha S., McKay M., Sheinin M., Mishra R.K., Pahan K.. Eugenol, a Component of Holy Basil (Tulsi) and Common Spice Clove, Inhibits the Interaction Between SARS-CoV-2 Spike S1 and ACE2 to Induce Therapeutic Responses. J. Neuroimmune Pharmacol. (2021) 16 743-755. DOI: 10.1007/s11481-021-10028-1 35. Panikar S., Shoba G., Arun M., Sahayarayan J.J., Usha Raja Nanthini A., Chinnathambi A., Alharbi S.A., Nasif O., Kim H.J.. Essential oils as an effective alternative for the treatment of COVID-19: Molecular interaction analysis of protease (M(pro)) with pharmacokinetics and toxicological properties. J. Infect. Public Health (2021) 14 601-610. DOI: 10.1016/j.jiph.2020.12.037 36. Mohammadhosseini M.. Essential oils extracted using microwave-assisted hydrodistillation from aerial parts of eleven Artemisia species: Chemical compositions and diversities in different geographical regions of Iran. Rec. Nat. Prod. (2017) 11 114 37. Strub D.J., Talma M., Strub M., Rut W., Zmudzinski M., Brud W., Neyts J., Vangeel L., Zhang L., Sun X.. Evaluation of the anti-SARS-CoV-2 properties of essential oils and aromatic extracts. Sci. Rep. (2022) 12 14230. DOI: 10.1038/s41598-022-18676-w 38. Gao Y., Yan L., Huang Y., Liu F., Zhao Y., Cao L., Wang T., Sun Q., Ming Z., Zhang L.. Structure of the RNA-dependent RNA polymerase from COVID-19 virus. Science (2020) 368 779-782. DOI: 10.1126/science.abb7498 39. Pandey P., Rane J.S., Chatterjee A., Kumar A., Khan R., Prakash A., Ray S.. Targeting SARS-CoV-2 spike protein of COVID-19 with naturally occurring phytochemicals: An in silico study for drug development. J. Biomol. Struct. Dyn. (2020) 39 6306-6316. DOI: 10.1080/07391102.2020.1796811 40. Vuong W., Khan M.B., Fischer C., Arutyunova E., Lamer T., Shields J., Saffran H.A., McKay R.T., van Belkum M.J., Joyce M.A.. Feline coronavirus drug inhibits the main protease of SARS-CoV-2 and blocks virus replication. Nat. Commun. (2020) 11 4282. DOI: 10.1038/s41467-020-18096-2 41. Gottlieb R.L., Nirula A., Chen P., Boscia J., Heller B., Morris J., Huhn G., Cardona J., Mocherla B., Stosor V.. Effect of bamlanivimab as monotherapy or in combination with etesevimab on viral load in patients with mild to moderate COVID-19: A randomized clinical trial. JAMA (2021) 325 632-644. DOI: 10.1001/jama.2021.0202 42. Casirivimab and imdevimab in patients admitted to hospital with COVID-19 (RECOVERY): A randomised, controlled, open-label, platform trial. Lancet (2022) 399 665-676. PMID: 35151397 43. Gupta A., Gonzalez-Rojas Y., Juarez E., Crespo Casal M., Moya J., Falci D.R., Sarkis E., Solis J., Zheng H., Scott N.. Early treatment for COVID-19 with SARS-CoV-2 neutralizing antibody sotrovimab. N. Engl. J. Med. (2021) 385 1941-1950. DOI: 10.1056/NEJMoa2107934 44. Singh A.K., Singh A., Singh R., Misra A.. An updated practical guideline on use of molnupiravir and comparison with agents having emergency use authorization for treatment of COVID-19. Diabetes Metab. Syndr. (2022) 16 102396. DOI: 10.1016/j.dsx.2022.102396 45. Beigel J.H., Tomashek K.M., Dodd L.E., Mehta A.K., Zingman B.S., Kalil A.C., Hohmann E., Chu H.Y., Luetkemeyer A., Kline S.. Remdesivir for the treatment of COVID-19. N. Engl. J. Med. (2020) 383 1813-1826. DOI: 10.1056/NEJMoa2007764 46. Owen D.R., Allerton C.M.N., Anderson A.S., Aschenbrenner L., Avery M., Berritt S., Boras B., Cardin R.D., Carlo A., Coffman K.J.. An oral SARS-CoV-2 Mpro inhibitor clinical candidate for the treatment of COVID-19. Science (2021) 374 1586-1593. DOI: 10.1126/science.abl4784 47. Stephenson J.. FDA Authorizes Pharmacists to Prescribe Oral Antiviral Medication for COVID-19. JAMA Health Forum (2022) 48. Effect of hydroxychloroquine in hospitalized patients with COVID-19. N. Engl. J. Med. (2020) 383 2030-2040. DOI: 10.1056/NEJMoa2022926 49. Tocilizumab in patients admitted to hospital with COVID-19 (RECOVERY): A randomised, controlled, open-label, platform trial. Lancet (2021) 397 1637-1645. PMID: 33933206 50. Kalil A.C., Patterson T.F., Mehta A.K., Tomashek K.M., Wolfe C.R., Ghazaryan V., Marconi V.C., Ruiz-Palacios G.M., Hsieh L., Kline S.. Baricitinib plus remdesivir for hospitalized adults with COVID-19. N. Engl. J. Med. (2021) 384 795-807. DOI: 10.1056/NEJMoa2031994 51. Lilly E.. A Study of Baricitinib (LY3009104) in Participants with COVID-19. (2020) 52. Yoo S.H., Kim L., Lu M., Nagoshi K., Namchuk M.N.. A review of clinical efficacy data supporting emergency use authorization for COVID-19 therapeutics and lessons for future pandemics. Clin. Transl. Sci. (2022) 15 2279-2292. DOI: 10.1111/cts.13384 53. Ma C., Sacco M.D., Xia Z., Lambrinidis G., Townsend J.A., Hu Y., Meng X., Szeto T., Ba M., Zhang X.. Discovery of SARS-CoV-2 papain-like protease inhibitors through a combination of high-throughput screening and a FlipGFP-based reporter assay. ACS Cent. Sci. (2021) 7 1245-1260. DOI: 10.1021/acscentsci.1c00519 54. Zhu W., Xu M., Chen C.Z., Guo H., Shen M., Hu X., Shinn P., Klumpp-Thomas C., Michael S.G., Zheng W.. Identification of SARS-CoV-2 3CL protease inhibitors by a quantitative high-throughput screening. ACS Pharmacol. Transl. Sci. (2020) 3 1008-1016. DOI: 10.1021/acsptsci.0c00108 55. Jang W.D., Jeon S., Kim S., Lee S.Y.. Drugs repurposed for COVID-19 by virtual screening of 6,218 drugs and cell-based assay. Proc. Natl. Acad. Sci. USA (2021) 118 e2024302118. DOI: 10.1073/pnas.2024302118 56. Narayanan A., Narwal M., Majowicz S.A., Varricchio C., Toner S.A., Ballatore C., Brancale A., Murakami K.S., Jose J.. Identification of SARS-CoV-2 inhibitors targeting Mpro and PLpro using in-cell-protease assay. Commun. Biol. (2022) 5 169. DOI: 10.1038/s42003-022-03090-9 57. Jin Z., Du X., Xu Y., Deng Y., Liu M., Zhao Y., Zhang B., Li X., Zhang L., Peng C.. Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors. Nature (2020) 582 289-293. DOI: 10.1038/s41586-020-2223-y 58. Ma C., Xia Z., Sacco M.D., Hu Y., Townsend J.A., Meng X., Choza J., Tan H., Jang J., Gongora M.V.. Discovery of di-and trihaloacetamides as covalent SARS-CoV-2 main protease inhibitors with high target specificity. J. Am. Chem. Soc. (2021) 143 20697-20709. DOI: 10.1021/jacs.1c08060 59. Unoh Y., Uehara S., Nakahara K., Nobori H., Yamatsu Y., Yamamoto S., Maruyama Y., Taoda Y., Kasamatsu K., Suto T.. Discovery of S-217622, a Noncovalent Oral SARS-CoV-2 3CL Protease Inhibitor Clinical Candidate for Treating COVID-19. J. Comput. Chem. (2022) 65 6499-6512. DOI: 10.1021/acs.jmedchem.2c00117 60. Zaidman D., Gehrtz P., Filep M., Fearon D., Gabizon R., Douangamath A., Prilusky J., Duberstein S., Cohen G., Owen C.D.. An automatic pipeline for the design of irreversible derivatives identifies a potent SARS-CoV-2 Mpro inhibitor. Cell Chem. Biol. (2021) 28 1795-1806.e5. DOI: 10.1016/j.chembiol.2021.05.018 61. Pharmacopoeia B.. Bernan Press (1988) Volume 2 A137-A138 62. Weiner S.J., Kollman P.A., Case D.A., Chandra Singh U., Ghio C., Alagona G., Profeta S., Weiner P.. A new force field for molecular mechanical simulation of nucleic acids and proteins. J. Am. Chem. Soc. (1984) 106 765-784. DOI: 10.1021/ja00315a051 63. Ribaudo G., Yun X., Ongaro A., Oselladore E., Ng J.P.L., Haynes R.K., Law B.Y.K., Memo M., Wong V.K.W., Coghi P.. Combining computational and experimental evidence on the activity of antimalarial drugs on papain-like protease of SARS-CoV-2: A repurposing study. Chem Biol Drug Des. (2022). DOI: 10.1111/cbdd.14187 64. Mufti A., Tir M., Zarei A., del Mar Contreras M., Gómez-Cruz I., Feriani A., Ghazouani L., Saadaoui E., Salah Allagui M., Harrath A.H.. Phytochemical Profiling of Ephedra alata subsp. alenda Seeds by High-Performance Liquid Chromatography—Electrospray Ionization—Quadrupole-Time-of-Flight-Mass Spectrometry (HPLC-ESI-QTOF-MS), Molecular Docking, and Antioxidant, Anti-diabetic, and Acetylcholinesterase Inhibition. Anal. Lett. (2022) 55 2450-2466 65. Suárez D., Díaz N.. SARS-CoV-2 main protease: A molecular dynamics study. J. Chem. Inf. Model. (2020) 60 5815-5831. DOI: 10.1021/acs.jcim.0c00575 66. Zhang L., Lin D., Sun X., Curth U., Drosten C., Sauerhering L., Becker S., Rox K., Hilgenfeld R.. Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved α-ketoamide inhibitors. Science (2020) 368 409-412. DOI: 10.1126/science.abb3405 67. Xue X., Yang H., Shen W., Zhao Q., Li J., Yang K., Chen C., Jin Y., Bartlam M., Rao Z.. Production of authentic SARS-CoV Mpro with enhanced activity: Application as a novel tag-cleavage endopeptidase for protein overproduction. J. Mol. Biol. (2007) 366 965-975. DOI: 10.1016/j.jmb.2006.11.073 68. Sun J., He Z.G., Cheng G., Wang S.J., Hao X.H., Zou M.J.. Multidrug resistance P-glycoprotein: Crucial significance in drug disposition and interaction. Med. Sci. Monit. (2004) 10 RA5-RA14. PMID: 14704647 69. Ribaudo G., Coghi P., Yang L.J., Ng J.P.L., Mastinu A., Memo M., Wong V.K.W., Gianoncelli A.. Computational and experimental insights on the interaction of artemisinin, dihydroartemisinin and chloroquine with SARS-CoV-2 spike protein receptor-binding domain (RBD). Nat. Prod. Res. (2022) 36 5358-5363. DOI: 10.1080/14786419.2021.1925894
--- title: 'Sustainable Dietary Score: Methodology for Its Assessment in Mexico Based on EAT-Lancet Recommendations' authors: - Fabricio Campirano - Nancy López-Olmedo - Paula Ramírez-Palacios - Jorge Salmerón journal: Nutrients year: 2023 pmcid: PMC9967068 doi: 10.3390/nu15041017 license: CC BY 4.0 --- # Sustainable Dietary Score: Methodology for Its Assessment in Mexico Based on EAT-Lancet Recommendations ## Abstract We developed a Sustainable Dietary Score (SDS) based on the EAT-Lancet commission’s recommendations and evaluated its adherence in a sample of Mexican adults. We used data on 1908 men and women aged 19 to 59 participating in the Health Workers Cohort Study in 2004. Fourteen of the healthy reference diet components were used to develop the SDS. We computed an individual SDS for each food component with scales from 0 (non-adherence) to 10 (perfect adherence), as well as a total SDS including all components, ranging from 0 to 140, based on a food frequency questionnaire. Our score incorporates characteristics of the context in which the score is applied, such as the high consumption of tortillas and eggs, and cut-off points that consider the nutrient deficiencies that prevail in the Mexican population. We propose a practical methodology to estimate a SDS incorporating a gradual score for a better distinction between the degrees of adherence to the reference diet proposed by the EAT-Lancet Commission. ## 1. Introduction One of the most used strategies to prevent chronic diseases is the promotion of healthy diets, which have focused on reducing the problems of malnutrition and obesity, as well as reducing the effects of nutrient deficiencies [1]. However, until recent years, only a few countries, including Brazil and Sweden, had considered the potential impact on the environment in their dietary guidelines, despite the contribution of food systems to the anthropogenic greenhouse gas emissions, which nowadays is $34\%$ globally [2]. One of the United Nations Sustainable Development Goals is the improvement of sustainable food systems [3]. To support this initiative, the EAT-Lancet Commission proposed a healthy reference diet (HRD) to promote human and environmental health [4]. The HRD proposes a moderate consumption of whole grains, starchy vegetables, vegetables, fruits, and legumes; moderate or low amounts of seafood and poultry; and low or no amounts of red meat, refined grains, and added sugars. Dietary indices based on EAT-Lancet recommendations have been proposed to evaluate the dietary quality of populations, such as developed by Knuppel in the UK [5] and later taken up by Ibsen in Denmark [6]. These indices evaluate the adherence to the recommendations establishing minimum intake values for multiple nutrient-dense food groups at 0 g/d. A limitation of this approach is that it does not take into account that nutrient deficiencies are still prevalent in some populations, and therefore, minimum intakes of rich-nutrient foods need to be established in specific contexts [7]. In addition to the necessity to develop sustainable dietary indices to evaluate the diet quality considering the context, it is important to take into account the complexity of the EAT-Lancet recommendations, which consider a range of optimal intakes. The latter means that both diets below and above the range should be penalized in a dietary index. However, most of the indices developed so far to determine adherence to EAT-Lancet recommendations assigned participants simple cut-off points for meeting the minimum or the median intake values recommended for each component [5,6,7,8,9]. In México, nutrient deficiencies are still prevalent, and despite the globalization of food systems, this country maintains some foods, such as corn tortillas and corn-based dishes, foods, and eggs as staples [10]. Therefore, it is necessary to develop a specific dietary index to evaluate the Mexican population’s diet quality considering the complexity of the sustainable recommendations. The development of a sustainable dietary score (SDS) is also relevant to inform policies that allow for achieving sustainable development goals. This study mainly documents the methodology used to develop a SDS based on the EAT-Lancet recommendations and how it works in a sample of Mexican adults. ## 2.1. Study Design and Participants The Health Workers Cohort Study (HWCS) is an ongoing cohort study conducted in Central Mexico. Participants are mostly health care workers from the Mexican Social Security Institute and their relatives in Cuernavaca, Morelos. At the baseline assessment, participants were asked to complete an extensive self-administered food frequency questionnaire at home and visited a research center for a physical examination. The goal of the HWCS is to examine the effect of genetic and lifestyle factors on the occurrence of different health outcomes of interest in the Mexican population. The study protocol, questionnaires, and informed consent forms were approved by the Institutional Review Board of the Mexican Social Security Institute (12CEI 09 006 14). Written informed consent was obtained from all study participants. Further details regarding the design and the methods are described in detail elsewhere [11]. For the present analysis, we used the baseline data of 2161 participants aged 19 to 70 collected during 2004–2006. The response rate to the study was just over $75\%$. We excluded participants that did not complete all sections of the semiquantitative food frequency questionnaire (FFQ) ($$n = 234$$) and participants with implausible energy intakes (<500 kcal/day or >6500 kcal/day, $$n = 19$$) [12]. Therefore, the analytic sample was composed of 1908 individuals. Among the clinical and demographic data collected by the HWCS, we considered age, sex, educational level, and body mass index (BMI) information. Educational level was categorized as elementary school, middle school, or high school or more. Due to the frequency of missing data for education level ($1.7\%$), we used a missing indicator category for this variable to minimize sample size reduction. ## 2.2. Development of the Sustainable Dietary Score (SDS) We developed a SDS based on the EAT-Lancet proposed sustainable HRD [4]. We considered the 14 components of the HRD: [1] whole grains; [2] tubers and starchy vegetables (tubers); [3] vegetables; [4] fruits: [5] dairy foods (dairy); [6] beef, lamb, pork and processed meat (red meat); [7] chicken and other poultry (poultry); [8] eggs; [9] fish; [10] legumes; [11] tree nuts (nuts); [12] unsaturated fats (unsaturated fats); [13] saturated fats; and [14] added sugars [13]. We replaced whole grains with high-fiber cereal (HF cereals) components due to a very low intake of whole grains and lack of variation in whole grains intake in the Mexican diet [14]. We defined individual scores for each component between 0 and 10 points. We assigned 10 points when the intakes were within the recommended range for each 2500 kcal/d of total energy intake, except for saturated fats and added sugars (Table 1). The latter means that, if someone consumes more or less than 2500 kcal/d, the intake of each food component is rescaled. That is, the score of each individual is based on their total energy intake. For vegetables, fruits, and unsaturated fats, we used the recommended intake range as indicated in the EAT-Lancet recommendations. For tubers, dairy, red meat, poultry, fish, legumes, and nuts, we used the median value of the recommended intake range as the lower limit (instead of nonconsumption, as established by EAT-Lancet) and the upper range as the upper limit (Table 1). We considered these cut-off points because nutrient deficiencies are still a problem in Mexico [15]. We modified the recommended intake for HF cereals, eggs, saturated fat, and added sugars with more appropriate lower or upper limits, as described below. We classified HF cereals as those cereals with more than 2.5 g of fiber per serving, including high-fiber bread, oatmeal, and tortillas, among others, in agreement with the Official Mexican Standard NOM-086-SSA1-1994 [16]. The recommended intake of HF cereal was established in a range between 125 g/d and 232 g/d. We based the lower limit on the Nutrition and Chronic Diseases Expert Group (NutriCoDE) recommendations [17]. The upper limit for HF cereals was based on the EAT-Lancet recommendation. For eggs, we established <13 g/d (~2 small-sized eggs per week) as the lower limit and >40 g/d (~7 small-sized eggs per week) as the upper limit. We used the median value recommended by EAT-Lancet as the lower limit for egg intake and the upper limit considering The American Heart Association (AHA) recommendation of the intake of one egg (or two egg whites) per day as part of a healthy diet [13,18]. EAT-Lancet indicates that a higher intake of some components can be safe and beneficial for low- and middle-income populations with poor dietary qualities [4,5]. Similar to poultry, eggs are among the most consumed proteins of high biological value in Mexico, where egg consumption per person per year is 345 units, making Mexico the country with the highest consumption in Latin America and one of the main ones worldwide [19]. For saturated fat, the lower limit was 11.8 g/d, using the EAT-Lancet median value. The upper limit was 27.7 g/d based on the World Health Organization (WHO) recommendation; saturated fat intake should not exceed $10\%$ of the total caloric intake [18]. The score for saturated fat was assigned such that the closer the saturated fat intake to the upper limit, the lower the score. Finally, for added sugars, we only considered the consumption of 31 g/d as the cut-off point; those with consumptions above the cut-off point had zero points, while those with consumptions below the cut-off point were assigned a score closer to 10 points, as the consumption was closer to zero. The scores below the lower limits were assigned according to two groups: essential foods (whose daily consumption is recommended according to the EAT-Lancet Commission [4] (HF cereals, tubers, fruits, vegetables, legumes, nuts, and unsaturated fats) and non-essential foods (whose consumption may not be daily or may be substituted by other sources such as dairy, chicken, eggs, red meat, and fish). We considered no consumption of these components as inappropriate for the Mexican population, since a high prevalence of chronic malnutrition associated with extreme poverty conditions persists in many places, and the prevalence of moderate and severe levels of food insecurity is still high in Mexican households ($43\%$) [20,21]. For non-essential foods, scores below the lower limit also diminished linearly as the intake did, but the lower score was 5 to participants with no consumption. The intakes above the upper limit for all components except saturated fats and added sugars were classified into ten categories using deciles. The individuals classified in the decile closest to the upper limit received a score of 10, while those in the more distant decile had 0 points. We generated the total SDS score by summing all individual component scores, ranging from 0 (nonadherence) to 140 (perfect adherence). ## 2.3. Diet Assessment Using a semiquantitative food frequency questionnaire (FFQ), we estimated the dietary intakes and determined the adherence of each participant to the SDS. The validity and reproducibility of the FFQ in the Mexican population have been previously published [11]. Briefly, the FFQ was administered twice, at a 1-y interval, to 134 women residing in Mexico City, and the results were then compared with those from the set of 4 recall tests given at 3-mo intervals. For the first FFQ, the deattenuated coefficients varied from 0.65 for saturated fatty acids to 0.12 for polyunsaturated fatty acids, whereas, for the second FFQ, the coefficients ranged from 0.63 for total fat to 0.21 for polyunsaturated fatty acids [22]. This questionnaire included data regarding the consumption of 116 food items. For each food, a commonly used portion size (e.g., 1 slice of bread or 1 cup of coffee) was specified on the FFQ, and participants reported their frequency of consumption of each specific food over the previous year. Participants chose from 10 possible responses, ranging from “never” to “6 or more times per day”. Grams consumed of each food item per day were calculated by multiplying the frequencies of consumption reported by the portion size of each food. The nutritional composition of each food included in the questionnaire was derived from the US Department of Agriculture (USDA) food composition tables and, when necessary, complemented by the nutrient database developed by the National Institute of Nutrition [23]. ## 2.4. Statistical Analysis We first described the study population by sociodemographic variables (age, sex, and educational level); total energy intake; total SDS; and BMI as a continuous and categorical variable (normal, overweight, and obesity). Means and standard deviations (SD) are presented for continuous variables and frequencies and percentages for categorical variables. Then, we calculated the medians and interquartile range of each individual score component of the SDS by sex. Finally, to understand the differences in the adherence to each SDS by sex, we estimated the percentage of subjects classified in the different categories of the score (below, within, and above the optimal intake range). We used the Mann–Whitney test to evaluate the differences in SDS between men and women. Poisson models were run to test the differences between percentages of consumption by sex and category. We conducted all the analyses in Stata 15.0 (Stata Corp, Stata Statistical Software, Release 15, 2017). ## 3. Results The mean age of the study sample was 45.5 ± 12.8 years, with an average BMI of 26.5 ± 4.3. Over half of the population was classified as overweight or obese ($61\%$), and a large proportion of the subjects studied high school or more ($41\%$). The median of the SDS was 80.5 (p25, p75 = 72.7, 88.0) out of a total of 140. It is also important to note that the median of SDS was 2 pp higher in women than in men (Table 2). The median SDS was greater than 8 for almost half of the individual food components. The food components with a higher score were unsaturated fats, poultry, eggs, fish, dairy, vegetables, and cereals high in fiber. The lowest scores were for saturated fat, legumes, nuts, and added sugars. Significant differences were found when comparing by sex. For the following food groups, the median score was higher in women than men: vegetables (9.0 vs. 6.8, $p \leq 0.001$), tubers (5.5 vs. 5.2, $p \leq 0.05$), and red meat (6.0 vs. 4.0, $p \leq 0.001$). Only the fruit score was higher in men than in women (7.0 vs. 5.0, $p \leq 0.001$ (Table 3). More men than women were classified with a HF cereals intake within the range ($36.4\%$ vs. $35.3\%$, respectively; p-value < 0.001). On the contrary, more women than men had tubers and fruit intakes above the recommended ($8.4\%$ vs. $3.5\%$ for tubers and 85.9 vs. $71.6\%$ for fruits, respectively; p-value < 0.001). Additionally, more women than men had a consumption within the recommended range for vegetables (43.9 vs. $24.6\%$; p-value < 0.001). Red meat intake was above the recommended for more than $80\%$ of men and women, while the fish intake was below the recommended for almost $60\%$ of adults. Likewise, more than $90\%$ of men and women had legumes and nut intakes below the recommended, and above $35\%$ consumed more than the recommended saturated fat (Table 4). Finally, more than $80\%$ of study participants consumed above the recommended for added sugars. ## 4. Discussion We developed a Sustainable Dietary Score based on the EAT-Lancet commission’s recommendations in the present study. This dietary score considers the complexity of the EAT-Lancet Commission recommendations and the Mexican context in which the score could be applied. Our results showed an overall median of 80.5 out of 140 points for the SDS in a sample of Mexican adults, representing an adherence (calculated as the mean or median score obtained in the sample divided by the total possible points in the score per 100) of $57.5\%$. The adherence to EAT-Lancet recommendations observed in our study was higher than that of Shamah et al. [ 2020]. The differences observed between studies reinforce the need to develop specific indices for each country, as well as the importance of considering minimum and maximum limits. We specifically observed high adherence to various food components. The highest score was for unsaturated fats; virtually all the study samples had an intake for this macronutrient within the recommended range. A possible explanation is that the EAT-Lancet recommendation for unsaturated fats is low (11.8–27.7 g per 2500 kcal). Historically, the Mexican population consumes little unsaturated fats, which can be good for the planet but not necessarily healthy if the saturated fat intake remains high. We found a median consumption of saturated fats of 18.2 g/d and a median score of 2.0; it was one of the lowest-rated components, along with the added sugars. Previous studies have also indicated a high consumption of saturated fat among Mexicans. The National Health and Nutrition Survey 2012 showed that the mean of the usual saturated fat intake was 27 g/d and 22 g/d for men and women, respectively. Moreover, this study found that more than $50\%$ of adults consumed more than the recommended saturated fats [24]. The Global Burden of Diseases Nutrition and Chronic Diseases Expert Group report also showed that the contribution of saturated fat intake to the total energy intake among Mexican adults in 2010 was between 7.0 and $8.4\%$, relatively close to the maximum recommended ($10\%$) [25]. The median score for poultry, eggs, dairy, and fish was higher than 8 points, which might reflect a moderate consumption of those foods due to costs or cultural reasons. Although the consumption of meat in Mexico does not reach the levels of other countries such as the USA, Argentina, Brazil, or Uruguay [26], it was enough to exceed the EAT-Lancet recommendations for more than $80\%$ of the sample, resulting in a score relatively low (median of 4 points). A high intake of red meat is not recommended, because it represents an important risk factor for chronic diseases such as cancer, as shown in multiple observational studies. It is also not recommended for planetary health, since meat is one of the foods that produces the most greenhouse gases per kilogram produced [27,28]. We found differences by sex in the scores for the fruits, vegetables, tubers, and red meat groups. The fruit score was higher in men than in women, which may be explained, in part, because a higher percentage of women than men consume more than recommended, which can be good for health but not the planet. On the other hand, the median scores for vegetables, tubers, and red meat were higher in women than in men. A potential explanation of these findings might be gender-related sociocultural factors. Some cultures highlight the relevance of physical appearance among women, likely making them more concerned with maintaining healthy eating behaviors (such as a higher intake of fruits, vegetables, and tubers) to stay in good physical shape [29]. Likewise, meat consumption has been associated with masculinity, which could explain, at least in part, why men had a lower score (higher consumption) for meat than women. Given the potential gender roles in diet, we cannot rule out the possibility that the differences observed by sex are also explained by social desirability bias. This term refers to the tendency to underreport socially undesirable attitudes and behaviors and to overreport more desirable attributes [27]. Finally, we found a high consumption of added sugars in men and women. This result is in line with Sánchez-Pimienta et al. They found that the contribution of added sugars to the total energy intake was $12.6\%$ when the WHO recommends that added sugars represent <$10\%$ of the total energy intake [28]. The main limitation of this study was the use of an FFQ. Unlike other dietary methods, the FFQ may raise the subject’s burden and increase response error. Moreover, the FFQ can be affected by bias of over- or underreporting, as previously described. Given the high prevalence of overweight and obesity in the sample analyzed, it is more likely the underreporting of unhealthy foods and the overreporting of healthy foods in this sample. Therefore, we do not rule out the possibility that the scores of some components, such as those less healthy, are overestimated. For more healthy components, the direction of bias is less clear, since high intakes are penalized through the SDS. Despite the limitations, the FFQ is a useful tool to assess adherence to dietary recommendations, such as those determined by the EAT-Lancet Commission for sustainable diets. Another possible limitation refers to the representativeness of the sample; we tested the SDS in a sample of Mexican adults, which is far from representing the entire population of Mexican adults. The Health Workers Cohort Study includes mainly women nurses and administrative service workers. Future studies to test the SDS using national samples is desirable. ## 5. Conclusions Although a global sustainable diet quality index can allow for comparing the healthiness and environmental sustainability of the diet across various countries, our results suggest that regional adaptations to the EAT-Lancet recommendations are necessary. Our index proposes a practical methodology, a gradual score that allows for a better distinction between the degrees of adherence to a sustainable diet, considering the context in which the index is intended to be applied. This tool can help inform food policies to improve health and the environment. We also expect that applying this index can help monitor the progress of interventions to be implemented. ## References 1. Watts N., Amann M., Arnell N., Ayeb-Karlsson S., Beagley J., Belesova K., Boykoff M., Byass P., Cai W., Campbell-Lendrum D.. The 2020 Report of The Lancet Countdown on Health and Climate Change: Responding to Converging Crises. Lancet (2021.0) 397 129-170. DOI: 10.1016/S0140-6736(20)32290-X 2. Crippa M., Solazzo E., Guizzardi D., Monforti-Ferrario F., Tubiello F.N., Leip A.. Food Systems Are Responsible for a Third of Global Anthropogenic GHG Emissions. Nat. Food (2021.0) 2 198-209. DOI: 10.1038/s43016-021-00225-9 3. Lee B.X., Kjaerulf F., Turner S., Cohen L., Donnelly P.D., Muggah R., Davis R., Realini A., Kieselbach B., MacGregor L.S.. Transforming Our World: Implementing the 2030 Agenda Through Sustainable Development Goal Indicators. J. Public Health Policy (2016.0) 37 13-31. DOI: 10.1057/s41271-016-0002-7 4. Willett W., Rockström J., Loken B., Springmann M., Lang T., Vermeulen S., Garnett T., Tilman D., DeClerck F., Wood A.. Food in the Anthropocene: The EAT–Lancet Commission on Healthy Diets from Sustainable Food Systems. Lancet (2019.0) 393 447-492. DOI: 10.1016/S0140-6736(18)31788-4 5. Knuppel A., Papier K., Key T.J., Travis R.C.. EAT-Lancet Score and Major Health Outcomes: The EPIC-Oxford Study. Lancet (2019.0) 394 213-214. DOI: 10.1016/S0140-6736(19)31236-X 6. Ibsen D.B., Christiansen A.H., Olsen A., Tjønneland A., Overvad K., Wolk A., Mortensen J.K., Dahm C.C.. Adherence to the EAT-Lancet Diet and Risk of Stroke and Stroke Subtypes: A Cohort Study. Stroke (2022.0) 53 154-163. DOI: 10.1161/STROKEAHA.121.036738 7. Hanley-Cook G.T., Argaw A.A., de Kok B.P., Vanslambrouck K.W., Toe L.C., Kolsteren P.W., Jones A.D., Lachat C.K.. EAT- Lancet Diet Score Requires Minimum Intake Values to Predict Higher Micronutrient Adequacy of Diets in Rural Women of Reproductive Age from Five Low- and Middle-Income Countries. Br. J. Nutr. (2020.0) 126 92-100. DOI: 10.1017/S0007114520003864 8. Marchioni D.M., Cacau L.T., De Carli E., de Carvalho A.M., Rulli M.C.. Low Adherence to the EAT-Lancet Sustainable Reference Diet in the Brazilian Population: Findings from the National Dietary Survey 2017–2018. Nutrients (2022.0) 14. DOI: 10.3390/nu14061187 9. Vallejo R.M., Schulz C.-A., van de Locht K., Oluwagbemigun K., Alexy U., Nöthlings U.. Associations of Adherence to a Dietary Index Based on the EAT-Lancet Reference Diet with Nutritional, Anthropometric, and Ecological Sustainability Parameters: Results from the German DONALD Cohort Study. J. Nutr. (2022.0) 152 1763-1772. DOI: 10.1093/jn/nxac094 10. Gálvez A.. Eating NAFTA: Trade, Food Policies, and the Destruction of Mexico (2018.0) 11. Denova-Gutiérrez E., Flores Y.N., Gallegos-Carrillo K., Ramírez-Palacios P., Rivera-Paredez B., Muñoz-Aguirre P., Velázquez-Cruz R., Torres-Ibarra L., Meneses-León J., Méndez-Hernández P.. Health Workers Cohort Study: Methods and Study Design. Salud Pública Méx. (2016.0) 58 708-716. DOI: 10.21149/spm.v58i6.8299 12. Willett W., Stampfer M.J.. Total Energy Intake: Implications For Epidemiologic Analysis. Am. J. Epidemiol. (1986.0) 124 17-27. DOI: 10.1093/oxfordjournals.aje.a114366 13. EAT-Lancet_Commission_Summary_Report. Food Planet Health. Healthy Diets from Sustainable Food Systems Summary Report of the EAT-Lancet Commission. Lancet (2019.0) 14. Marrón-Ponce J.A., Tolentino-Mayo L., Hernández-F M., Batis C.. Trends in Ultra-Processed Food Purchases from 1984 to 2016 in Mexican Households. Nutrients (2018.0) 11. DOI: 10.3390/nu11010045 15. Ramírez-Silva I., Rodríguez-Ramírez S., Barragán-Vázquez S., Castellanos-Gutiérrez A., Reyes-García A., Martínez-Piña A., Pedroza-Tobías A.. Prevalence of Inadequate Intake of Vitamins and Minerals in the Mexican Population Correcting by Nutrient Retention Factors, Ensanut 2016. Salud Pública Méx. (2020.0) 62 521-531. DOI: 10.21149/11096 16. DOF-Diario Oficial de La Federación 17. Micha R., Shulkin M.L., Peñalvo J.L., Khatibzadeh S., Singh G.M., Rao M., Fahimi S., Powles J., Mozaffarian D.. Etiologic Effects and Optimal Intakes of Foods and Nutrients for Risk of Cardiovascular Diseases and Diabetes: Systematic Reviews and Meta-Analyses from the Nutrition and Chronic Diseases Expert Group (NutriCoDE). PLoS ONE (2017.0) 12. DOI: 10.1371/journal.pone.0175149 18. Healthy Diet 19. Situation of the Mexican Poultry Industry 2021 20. Mundo-Rosas V., Unar-Munguía M., Hernández-F M., Pérez-Escamilla R., Shamah-Levy T.. La seguridad alimentaria en los hogares en pobreza de México: Una mirada desde el acceso, la disponibilidad y el consumo. Salud Pública Méx. (2019.0) 61 866-875. DOI: 10.21149/10579 21. Cuevas-Nasu L., Gaona-Pineda E.B., Rodríguez-Ramírez S., Morales-Ruán M.D.C., González-Castell L.D., García-Feregrino R., Gómez-Acosta L.M., Ávila-Arcos M.A., Shamah-Levy T., Rivera-Dommarco J.. Stunting in children population in localities under 100 000 inhabitants in Mexico. Salud Pública Méx. (2019.0) 61 833-840. DOI: 10.21149/10642 22. Hernández-Avila M., Romieu I., Parra S., Hernández-Avila J., Madrigal H., Willett W.. Validity and Reproducibility of a Food Frequency Questionnaire to Assess Dietary Intake of Women Living in Mexico City. Salud Pública Méx. (1998.0) 40 133-140. DOI: 10.1590/S0036-36341998000200005 23. Composition of Foods–Raw, Processed, and Prepared. Agricultural Hand Book (1985.0) 24. Lopez-Olmedo N., Carriquiry A.L., Rodríguez-Ramírez S., Ramírez-Silva I., Espinosa-Montero J., Hernández-Barrera L., Campirano F., Martínez-Tapia B., Rivera J.A.. Usual Intake of Added Sugars and Saturated Fats Is High While Dietary Fiber Is Low in the Mexican Population. J. Nutr. (2016.0) 146 1856S-1865S. DOI: 10.3945/jn.115.218214 25. Micha R., Khatibzadeh S., Shi P., Fahimi S., Lim S., Andrews K.G., Engell R.E., Powles J., Ezzati M., Mozaffarian D.. Global, Regional, and National Consumption Levels of Dietary Fats and Oils in 1990 and 2010: A Systematic Analysis Including 266 Country-Specific Nutrition Surveys. BMJ (2014.0) 348 g2272. DOI: 10.1136/bmj.g2272 26. Infographic: The Countries That Eat The Most Meat 27. Latkin C.A., Edwards C., Davey-Rothwell M.A., Tobin K.E.. The Relationship between Social Desirability Bias and Self-Reports of Health, Substance Use, and Social Network Factors among Urban Substance Users in Baltimore, Maryland. Addict. Behav. (2017.0) 73 133-136. DOI: 10.1016/j.addbeh.2017.05.005 28. Sánchez-Pimienta T.G., Batis C., Lutter C.K., Rivera J.A.. Sugar-Sweetened Beverages Are the Main Sources of Added Sugar Intake in the Mexican Population. J. Nutr. (2016.0) 146 1888S-1896S. DOI: 10.3945/jn.115.220301 29. Love H.J., Sulikowski D.. Of Meat and Men: Sex Differences in Implicit and Explicit Attitudes Toward Meat. Front. Psychol. (2018.0) 9 559. DOI: 10.3389/fpsyg.2018.00559
--- title: A Low Concentration of Citreoviridin Prevents Both Intracellular Calcium Deposition in Vascular Smooth Muscle Cell and Osteoclast Activation In Vitro authors: - Seongtae Jeong - Bok-Sim Lee - Seung Eun Jung - Yoojin Yoon - Byeong-Wook Song - Il-Kwon Kim - Jung-Won Choi - Sang Woo Kim - Seahyoung Lee - Soyeon Lim journal: Molecules year: 2023 pmcid: PMC9967071 doi: 10.3390/molecules28041693 license: CC BY 4.0 --- # A Low Concentration of Citreoviridin Prevents Both Intracellular Calcium Deposition in Vascular Smooth Muscle Cell and Osteoclast Activation In Vitro ## Abstract Vascular calcification (VC) and osteoporosis are age-related diseases and significant risk factors for the mortality of elderly. VC and osteoporosis may share common risk factors such as renin-angiotensin system (RAS)-related hypertension. In fact, inhibitors of RAS pathway, such as angiotensin type 1 receptor blockers (ARBs), improved both vascular calcification and hip fracture in elderly. However, a sex-dependent discrepancy in the responsiveness to ARB treatment in hip fracture was observed, possibly due to the estrogen deficiency in older women, suggesting that blocking the angiotensin signaling pathway may not be effective to suppress bone resorption, especially if an individual has underlying osteoclast activating conditions such as estrogen deficiency. Therefore, it has its own significance to find alternative modality for inhibiting both vascular calcification and osteoporosis by directly targeting osteoclast activation to circumvent the shortcoming of ARBs in preventing bone resorption in estrogen deficient individuals. In the present study, a natural compound library was screened to find chemical agents that are effective in preventing both calcium deposition in vascular smooth muscle cells (vSMCs) and activation of osteoclast using experimental methods such as Alizarin red staining and Tartrate-resistant acid phosphatase staining. According to our data, citreoviridin (CIT) has both an anti-VC effect and anti-osteoclastic effect in vSMCs and in Raw 264.7 cells, respectively, suggesting its potential as an effective therapeutic agent for both VC and osteoporosis. ## 1. Introduction Defined as extracellular deposition of calcium-phosphate complexes in the arterial wall, vascular calcification (VC) is known to be an independent predictor of cardiovascular disease-related mortality [1,2,3]. VC once was considered as a passive process, but now it is regarded as an active and regulated pathological process, where vascular smooth muscle cells (vSMCs) play an important role [4]. In stimulated vascular walls, vSMCs can trans-differentiate into an osteoblast-like phenotype by the changes of various VC stimulating factors and/or inhibitors involved in bone metabolism [5]. Osteoporosis is characterized by demineralization of bone tissue caused by the imbalance between the osteoblast-mediated bone formation and the osteoclast-mediated resorption, and the resultant increase of bone fragility increases a risk of fracture in return [6]. Pro-inflammatory cytokines, such as tumor necrosis factor alpha (TNF-α), interleukin-1 (IL-1), IL-6, and receptor activator of nuclear factor-kappa B ligand (RANKL), are known to increase bone resorption by stimulating osteoclast differentiation and activation, and consequently, inhibit bone formation [7]. Interestingly, these bone metabolic mediators have been associated with the development of VC as well [8,9]. Since the late 1990s, clinical and laboratory findings have demonstrated the association between osteoporosis and vascular calcification, and it also has been reported that they share a number of common risk factors such as aging, estrogen deficiency, chronic inflammation, oxidative stress, and lipid metabolism [10,11,12,13,14]. Furthermore, it has been suggested that, especially under hypertensive conditions, vascular calcification and osteoporosis share pathophysiological mechanisms involving the renin-angiotensin system (RAS) [15]. Angiotensin II (Ang II)-induced excessive reactive oxygen species (ROS) can activate Src-family kinases, protein kinase C (PKC), and mitogen-activated protein kinases (MAPKs) in vSMCs [16]. This signaling cascade, in turn, activates a series of transcription factors responsible for osteogenic gene expressions, such as nuclear factor kappa-B (NF-κB), Runx2, and activator protein 1 (AP-1) contributing to VC [17,18,19]. In addition to directly inducing osteoblast-like differentiation of vSMCs leading to VC, Ang II can also indirectly contribute to bone resorption by activating osteoclast via secretion of RANKL. Upon binding to RANK, RANKL recruits adaptor molecules such as tumor necrosis factor receptor-associated factor 6 (TRAF6), and this sequentially activates MAPKs, NF-κB, and AP-1. In turn, the activated NF-κB induces the expression of nuclear factor of activated T-cells cytoplasmic 1 (NFATc1), a major osteoclastogenesis regulator [20,21]. Furthermore, NFATc1 also increases a number of genes including tartrate-resistant acid phosphatase (TRAP), Cathepsin K (CTSK), calcitonin receptor (CALCR), and dendritic-cell-specific transmembrane protein (DC-STAMP) to regulate osteoclast differentiation [22,23], leading to the formation and activation of osteoclasts. It also has been demonstrated that the expression of RANKL can be increased in calcifying vSMCs in response to oxidative stress [24], and Ang II induced expression of RANKL in human vSMCs and in ApoE knockout mice has been demonstrated as well [25]. Additionally, Ang II can also induce the expression of RANKL and subsequent extracellular secretion in non-vSMCs such as osteoblasts or synovial cells indirectly contributing to osteoclast activation and osteoporosis. [ 26,27,28]. As circumstantial evidence of the involvement of RAS in both VC and osteoporosis, preclinical studies have demonstrated that Angiotensin type 1 receptor blockers (ARBs) have a protective effect on vascular calcification [25,29]. In addition, a beneficial effect of ARBs on hip fracture in elderly has been demonstrated in a cohort study [30]. However, upon a closer look, this particular cohort study also demonstrated that there may a sex-dependent discrepancy in the responsiveness to ARB treatment. To be more specific, the risk of fracture was more decreased in men than women following ARB treatment, and especially older women in postmenopausal period showed much higher risk of fracture [30]. Considering estrogen deficiency can lead to a prolonged bone loss by promoting osteoclast formation and life span [31,32], the reported sex-dependent discrepancy may stem from the estrogen deficiency in older women. This suggested that blocking Ang II signaling pathway may not be effective to suppress bone resorption especially if an individual has underlying osteoclast activating conditions such as estrogen deficiency. Therefore, it has its own significance to find alternative chemical agents to inhibit the development of both VC and osteoporosis, possibly by directly targeting osteoclast activation to circumvent the shortcoming of ARBs in preventing bone resorption in estrogen deficient individuals. In the present study, a natural compound library was screened to find chemical agents that are effective in preventing both calcium deposition in vSMCs and activation of osteoclast using experimental methods such as Alizarin red staining (calcium deposition) and TRAP staining (osteoclast activation). According to our data, citreviridin (CIT) has both an anti-VC effect and anti-osteoclastic effect in vSMCs and in Raw 264.7 cells, respectively, suggesting its potential as an effective therapeutic agent for both VC and osteoporosis. ## 2.1. CIT Attenuates Both Ang II-Induced VC and RANKL-Induced Osteogenic Differentiation Using a natural product library composed of 502 natural compounds, primary screening for agents (2 μg/mL, each) that could suppress calcium deposition in vSMCs was conducted. A selection threshold was ≥$70\%$ of calcium deposition inhibition, and 21 candidate compounds were selected (Figure 1). As a secondary screening, the selected compounds were further screened for their anti-osteoclastic effect using RANKL-treated Raw264.7 cells. The results indicated that six compounds (Figure 2) showed the protective effect. However, CIT was the only compound that has never been reported to be effective on VC and/or osteoporosis and, thus, the effect of CIT on both VC and osteoporosis was further examined for this study. ## 2.2. Effect of CIT on vSMC Viability Upon morphological examination of vSMCs treated with CIT (1 and 2 μg/mL) with or without Ang II (500 nM), 2 μg/mL of CIT showed a mild cytotoxicity (Figure 3A and 3B), which was not observed during the first screening. Since the major difference in the protocols used for the first screening and this cytotoxicity testing was the size culture dishes (48 well vs. 60 mm, respectively), this unexpected cytotoxicity was possibly due to the absolute amount of CIT used for this experiment (0.8 μg for 48 well plates vs. 4 μg for 60 mm plates). Nevertheless, 1 μg/mL of CIT did not show any significant decrease of viability (Figure 3C) or cytotoxicity (Figure 3D). ## 2.3. CIT Inhibits Ang II-Induced Calcium Deposition in vSMCs First, Alizarin red S staining indicated that Ang II treatment apparently increased calcium deposition in vSMCs, and this was suppressed by CIT treatment (Figure 4A). Quantification of Alizarin red S showed that 1 μg/mL of CIT significantly decreased the amount of calcium deposited in vSMCs compared to the Ang II treated control (4.91 ± 1.58 fold increase vs. 1.46 ± 0.19 fold increase compared to the untreated control). Lower concentrations of CIT (0.1 and 0.5 μg/mL) also inhibited Ang II-induced calcium deposition, and it showed a concentration-dependent tendency of decrease (Figure 4B). However, additional calcium assay results indicated that only 1 μg/mL of CIT was significantly effective (Figure 4C) and, therefore, 1 μg/mL of CIT was used for the rest of the experiments. ## 2.4. CIT Attenuates Ang II-Induced Osteogenic Marker Expressions in vSMCs According to our Western blot data, Ang II significantly decreased the expressions of SMC differentiation markers, namely Calponin and smooth muscle protein 22-alpha (SM22α), but this was attenuated by CIT (Figure 5A). On the other hand, Ang II significantly increased the expression of Osterix, an osteoblast marker, at both protein and mRNA level, and this was also attenuated by CIT (Figure 5B). Additionally, Ang II significantly increase the mRNA expression of ALP, an osteoblast activity marker, and this was again significantly suppressed by CIT (Figure 5C). The amount of ALP secreted from the cells was estimated by the ALP activity of culture media, and the results showed that CIT was able to significantly suppress the Ang II-induced ALP secretion from vSMCs (Figure 5D). ## 2.5. CIT Inhibits Ang II-Induced ROS Production in vSMCs Since excessive ROS can lead to VC and, thus, a ROS-lowering agent can prevent VC [33,34], the effect of CIT on Ang II-induced ROS production in vSMCs was examined. Flow cytometry data indicated that Ang II significantly increased ROS production in vSMCs, but it was significantly attenuated by CIT (Figure 6A). As modulators of ROS production, the expressions of NADPH oxidase (NOX) and phosphorylation status of PKC and Src were examined. As shown in the Figure 6B, the expression of NOX was significantly increased by Ang II, but this was significantly attenuated by CIT. Ang II also significantly increased the phosphorylation of PKCδ and Src (Tyr416), which was abrogated by CIT (Figure 6C). ## 2.6. CIT Inhibits Ang II-Induced MAPK Signaling and Subsequent NF-κB and AP-1 Activation According to our data, Ang II significantly increased the phosphorylation of MAPKs (extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinase (JNK) and p38). Although CIT showed no significant inhibitory effect on the Ang II-induced ERK phosphorylation, it did significantly inhibit the Ang II-induced phosphorylation of JNK and P38 (Figure 7A). Furthermore, CIT significantly suppressed the Ang II-induced phosphorylation of downstream transcription factors, namely NF-κB (Figure 7B) and AP-1 (Figure 7C). ## 2.7. CIT Inhibits RANKL-Induced Osteoclast Differentiation of Raw264.7 Cells First, any possible cytotoxicity of CIT on Raw264.7 cells was evaluated by using LDH assay and Annexin V/PI staining. CIT up to 1 μg/mL did not show any significant cytotoxic effect on Raw264.7 cells (Figure 8A), and Annexin V/PI staining also indicated that 1 μg/mL of CIT did not cause apoptosis in Raw264.7 cells (Figure 8B). The results of TRAP staining to visualize the osteoclast marker enzyme demonstrated that RANKL apparently increased the number of TRAP positive cells, while CIT abrogated the pro-osteoclastic effect of RANKL (Figure 9A). Quantification of TRAP staining indicated that CIT significantly suppressed RANKL-induced osteoclast differentiation of Raw264.7 cells as evidenced by the decrease of both TRAP-positive cells (Figure 9B) and TRAP-positive multinuclear cells (Figure 9C,D) in a concentration-dependent manner. Furthermore, RANK increased the mRNA expressions of osteogenic markers such as TRAP, CTSK, CALCR, DC-STAMP, and NFATc1, but these were significantly abrogated by CIT (Figure 9E). ## 2.8. CIT Inhibits RANKL-Induced Nuclear Translocation of NFκB and NFATc1 According to our data, RANKL significantly increased both the phosphorylation of NF-κB (Figure 10A) and expression of NFATc1 (Figure 10C). However, CIT significantly abrogated the RANKL-induced increase of NF-κB phosphorylation and NFATc1 expression. Furthermore, immunocytochemistry using NF-κB (Figure 10B) and NFATc1 (Figure 10D) specific antibodies demonstrated that CIT also significantly suppressed the RANKL-induced nuclear translocation of these transcription factors in Raw264.7 cells. ## 2.9. CIT Inhibits RANKL-Induced Osteoclast Activity The results of Pit formation assay indicated that RANKL significantly increased the bone resorptive activity of Raw264.7 cells (Figure 11A), but this was significantly suppressed by CIT reducing both Pit diameter (Figure 11B) and area (Figure 11C). ## 2.10. Ang II-Induced RANKL Secretion from vSMCs May Connect VC and Osteoporosis To examine the possibility that RANKL being a mediator linking VC and osteoporosis, the expression of RANKL in vSMCs following Ang II stimulation was examined. As shown in the Figure 12A, Ang II increased the mRNA expression of RANKL in vSMCs, and this was attenuated by CIT. Furthermore, the amount of RANKL protein presented in the culture medium of Ang II-stimulated vSMCs was increased compared to that in the untreated control (Figure 12B), suggesting that the RANKL released from the vSMCs upon Ang II stimulation may, in turn, promote the activation of osteoclasts leading to the development of osteoporosis. ## 3. Discussion In the present study, we report that a low concentration of natural compound CIT effectively suppressed the Ang II-induced calcium deposition in HAoSMCs and RANKL-induced osteoclastogenesis of Raw264.7 cells. However, its reported toxicity can be an issue if it were to be further evaluated for its potential as a therapeutic agent for both VC and osteoporosis. CIT, a toxic secondary metabolite derived from Penicillium strains, has been associated with the development of cardiovascular diseases such as atherosclerosis and cardiac beriberi [35,36], and its biological effects including cytotoxicity have been evaluated in different in vitro and in vivo systems. For example, disruption of nerve and muscle metabolism by CIT was reported to be responsible for the development of beriberi [36], and it is also known to cause Keshan disease through oxidative stress [37]. Furthermore, CIT induced autophagic cell death through the lysosomal–mitochondrial axis in hepatocytes [38]. Especially pertaining to cardiovascular system, it has been reported that CIT enhanced tumor necrosis factor-α (TNFα)-induced endothelial adhesion by increasing the expression of adhesion molecules such as ICAM-1, VCAM-1, and E-selectin in human umbilical vein endothelial cells [39], and it caused myocardial apoptosis via activating autophagic pathway [40]. However, currently available empirical data on the cytotoxicity of CIT is just not sufficient to draw an exclusive conclusion on CIT’s cytotoxicity, and it needs to be carefully examined, especially in terms of the concentration used for the experiments. First of all, there are not that many in vitro studies examined the cytotoxicity of CIT over a range of varying concentrations to which our data can be compared, not to mention in vitro studies that used vSMCs or Raw264.7 cells. For example, the above-mentioned study examined the CIT-induced autophagy-dependent apoptosis of hepatocytes used 5 μM of CIT throughout the study [38], and yet another study reported CIT-induced myocardial apoptosis through autophagic pathway used 0.1–0.3 mg/kg of CIT to demonstrate myocardial apoptosis in vivo, but it did not directly show the cytotoxicity of CIT on in vitro cultured cardiomyocytes. Although they used 2–8 μM of CIT for some in vitro experiments, those experiments were mainly to examine the effect of CIT on autophagy in cardiomyocytes rather than on cell viability [40]. On the other hand, although the cell used was porcine kidney cells, a concentration–response curve for CIT was generated in one study, and the results indicated that even up to 101.5 μM of CIT (approximately 32 μM) did not show any significant cytotoxicity in MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide, a tetrazole) assay [41]. In the present study, 1 μg/mL of CIT was used for most of the experiments, and it is approximately equivalent to 2.5 μM of CIT (molecular weight of CIT is 402.5). More importantly, this particular concentration of CIT did not show any significant cytotoxic effect on both vSMCs and Raw264.7 cells, although 2 μg/mL of CIT did show mild cytotoxicity (Figure 3). Certainly, this does not necessarily mean that a low concentration of CIT cleared the issue of possible cytotoxicity. However, at this point, our data simply indicated that, similar to other potentially toxic substances used for therapeutic purposes such as Botulinum toxin [42], CIT may do more good than harm at certain given concentrations. Therefore, it will be worthwhile to systemically examine the effects of varying concentrations of CIT, especially on VC and osteoporosis, to verify the narrow therapeutic window of CIT on VC and osteoporosis that may or may not exist in further in vivo studies. Prior to the development of VC, a phenotypic change of vSMCs from contractile to osteogenic occurs accompanying the down-regulation of vSMC contractile markers such as smooth muscle α-actin and up-regulation of osteogenic markers such as ALP [43]. Our data demonstrated that CIT significantly suppressed Ang II-induced calcium deposition in vSMCs (Figure 4) and osteogenic trans-differentiation of vSMCs (Figure 5), indicating CIT can attenuate calcium deposition in vSMCs by preventing phenotype switching of vSMCs. In a mechanistic point of view, elevated ROS production has been associated with VC leading to activation of osteogenic signaling pathways [33]. Ang II is well known to induce excessive production of ROS [44], and the Ang II-induced VC in the present study might as well be the result of the excessive ROS produced by Ang II. Therefore, it is reasonable to assume that inhibition of ROS production can be an effective approach to attenuate phenotype switching of vSMCs and resultant VC. Agreeing with such assumption, CIT significantly suppressed Ang II-induced production of ROS in the present study (Figure 6). In addition, at transcriptional level, osteogenic transcriptional factors activated via MAPK signaling pathway are known to drive the transcription of osteogenic genes such as Runx2 and Osterix [45]. Our data also indicate that CIT significantly suppressed the Ang II-induced phosphorylation of JNK and p38, and subsequent activation of the osteogenic transcription factor NFκB and AP-1 (Figure 7). These data suggested that CIT can be an agent that effectively suppress the Ang II-induced VC in vSMCs. One of the main cause of osteoporosis, the other pathologic condition on which the effect of CIT was examined in the present study, is pathologically enhanced activity of osteoclast and resultant increase of bone resorption [46]. It is a well-established fact that monocyte/macrophage lineage precursor cells can differentiate into osteoclasts by RANKL [47], and RANKL-induced osteoclast differentiation of Raw264.7 cell was also demonstrated in the present study. However, such RANKL-induced osteoclast differentiation was significantly abrogated by CIT, along with down-regulated expression of osteoclast-specific genes such as TRAP, CTSK, CALCR, and DC-STAMP (Figure 9) and attenuated translocation of NF-κB/NFATc1 signaling (Figure 10). DC-STAMP is known to play an important role in cell–cell fusion to create multi-nucleated cells [48], and TRAP, CALCR, and CTSK are the acidic substances secreted by mature osteoclasts contributing to the bone resorptive activity by degrading bone surface [49,50]. Therefore, it was speculated that the activity of osteoclast could be suppressed by CIT, and our data on bone resorption assay strongly supported such speculation (Figure 11). Although our results demonstrated that CIT could effectively inhibit both Ang II-induced osteogenic differentiation of vSMCs and RANKL-induced osteoclast differentiation of Raw264.7 cells so far, those data were obtained from two independent in vitro systems and, thus, not sufficient enough to claim that CIT will simultaneously act on both VC and osteoporosis in a single entity. Therefore, as a possible missing link between these two systems, the expression of RANKL production from Ang II-stimulated vSMCs was examined in the present study. In fact, Ang II has been reported to increase the expression of RANKL in vSMCs [25], and it can promote macrophage migration and differentiation into osteoclast-like cells [24]. Our data also indicated that Ang II significantly increased mRNA expression of RANKL, which was attenuated by CIT, and secretion of RANKL as well (Figure 12). Such a role of RANKL as a linker that connects VC and osteoporosis has been demonstrated in a previous study where PKA agonist-treated aortic SMCs were co-cultured with Raw264.7 cells that eventually differentiated into osteoclast by the RANKL secreted by the aortic SMCs [51]. Additionally, Ang II stimulation is known to induce a variety of pro-inflammatory factors, such as interleukin 1β (IL-1β), IL-6, tumor necrosis factor-α (TNF-α), and high mobility group box 1 protein (HMGB1) from various cells [52,53,54]. In turn, these factors can affect the expression of RANKL. For example, extracellular HMGB1 can enhance the expression of RANKL that can act as a ligand for toll-like receptors (TLRs) and the receptor for advanced glycation end products (RAGE) in osteoblastogenic bone marrow stromal cell cultures, osteocytes, and osteoblasts [55,56]. Therefore, a crosstalk between VC and osteoporosis can be facilitated by many different pathways and cell types involving production of RANKL in vivo. Nevertheless, since CIT worked on RANKL-induced osteoclast differentiation of Raw264.7 cells even without the presence of Ang II, CIT is expected to work on both pathologic conditions in a single entity, as long as they are mainly linked by RANKL. ## 4.1.1. Culture and Calcium Deposition of Human Aortic Smooth Muscle Cells (HAoSMCs) HAoSMCs were purchased from Lifeline cell technology (Frederick, MD, USA; FC-0015) and cultured in VascuLife® Basal Medium (LM-0002; Lifeline) with supplements (VascuLife® SMC LifeFactors kit; LS-1040; Lifeline). For antibiotics, 100 U/mL penicillin (15140-122; Thermo Fisher Scientific, Waltham, MA, USA) and 100 μg/mL streptomycin (15140-122; Thermo Fisher Scientific) were used. To induce vascular calcification or calcium deposition, cells of passage 6–8 were grown in a 48-well culture plate at density of 6 × 104 cells per wells in a HAoSMC culture medium. After cells were attached, the medium was changed to Ang II-based calcification induction media; mixture of Ang II (500 nM), β-Glycerophosphate (β-GP; 4 mM), CaCl2 (3.8 mM) and $15\%$ fetal bovine serum (FBS) in the Basal medium. The cells were cultured for 6–8 days for calcium deposition. ## 4.1.2. Culture and Osteoclast Differentiation of Raw264.7 Cells Raw264.7 cells (TIB-71; ATCC, Manassas, VA, USA), an established macrophage cell line for osteoclastic differentiation, were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (30-2002; ATCC) with $10\%$ FBS (16000-044, Thermo Fisher Scientific, Waltham, MA, USA), 100 U/mL penicillin (15140-122, Thermo Fisher Scientific), and 100 μg/mL streptomycin (15140-122; Thermo Fisher Scientific) at 37 °C in a humidified incubator under $5\%$ CO2 atmosphere. For cell differentiation and drug treatment, cells were used between passage 5 and 15. To differentiate osteoclast cells from Raw264.7 cells, the cells were seeded on 48-well plates (3 × 103 cells/well) in complete DMEM media. After cell attachment, complete DMEM medium was changed to complete alpha modified minimal essential medium (α-MEM; 11900-024, Thermo Fisher Scientific) as differentiation medium. Raw264.7 cells were treated 40 ng/mL recombinant RANKL (ALX-522-131; ENZO Life Science, Farmingdale, NY, USA) every 2 days for 4 days. ## 4.2. Screening of Natural Compounds for Suppressing Calcium Deposition of Human Aortic Smooth Muscle Cells (HAoSMCs) Using Alizarin Red Staining To screen natural compounds for suppressing Ang II-induced calcium deposition in HAoSMCs, ENZo library was used (Screen-well® Natural product library, BML-2865). Natural compounds were added upon induction with Ang II. Intracellular calcium deposition was assessed by alizarin red staining. Briefly, the cells were fixed in cold $70\%$ ethanol for 30 min at 4 °C. The fixed cells were stained with $2\%$ alizarin red S (A5533; Sigma-Aldrich, Seoul, Republic of Korea) for 10 min at room temperature. The cells were observed under the microscope (CKX41; Olympus, Tokyo, Japan) and images were obtained by a digital camera (eXcope T300; Olympus) at 40× magnification. For quantification, the stained cells were destained in $10\%$ cetylpyridinum chloride (C0732; Sigma-Aldrich). The absorbance was measured at 595 nm using a microplate reader (Multiskan FC; Thermo Fisher Scientific). ## 4.3. Evaluation of Cell Viability (WST-1) and Toxicity (LDH) Cell viability was assessed by water-soluble tetrazolium 1 (WST-1) assay. The Cell viability was checked using EZ-Cytox water-soluble tetrazolium salt (WST) assay kit (EZ-3000; Dogenbio, Seoul, Republic of Korea). The absorbance was measured at 450 nm by microplate reader (Multiskan FC; Thermo Fisher Scientific). Cytotoxicity was evaluated by using Lactate dehydrogenase (LDH) Cytotoxicity Detection Kit (MK401; TAKARA, Otsu, Japan) according to the manufacturer’s protocols. Briefly, a total of 100 μL of LDH reaction reagent was added into 100 μL of supernatants and then incubated at 37 °C for 30 min. Absorbance was determined at 450 nm using microplate reader (Multiskan FC; Thermo Fisher Scientific). ## 4.4. Annexin V/PI Staining To assess cell death, Annexin V/propodium iodide (PI) staining was performed using FITC Annexin V Apoptosis Detection Kit I (556547; BD biosciences, Becton, NJ, USA). Briefly, the cells were stained with 5 μL of Annexin V and 5 μL of PI at room temperature for 15 min. The cells were diluted with the binding buffer and analyzed with BD Accuri™ C6 flow cytometer (BD biosciences). ## 4.5. Calcium Assay To quantify the amount of calcium deposited, HAoSMCs were incubated overnight with 0.6 N HCl at 4 °C to dissolve the deposited calcium. The amount of calcium present in the supernatants was determined by using QuantiChrom™ Calcium Assay Kit (DICA-500; Bioassay systems, Hayward, CA, USA). Briefly, a total of 200 μL of working reagent mixture was added into 3 μL of supernatants and then incubated at room temperature for 3 min. Absorbance was determined at 612 nm using microplate reader (Multiskan FC; Thermo Fisher Scientific). ## 4.6. Alkaline Phosphatase (ALP) Assay Osteoblast-like differentiation of HAoSMCs was evaluated by using an alkaline phosphatase assay kit (ab83369; Abcam, Cambridge, UK) following the manufacturer’s protocol. The absorbance of supernatants was determined at 405 nm using a microplate reader (Multiskan FC; Thermo Fisher Scientific). ## 4.7. Intracellular ROS Detection For reactive oxygen species (ROS) detection, the cells were harvested with accutase (A6964; Sigma-Aldrich). A total of 5 × 105 cells in cell suspension were incubated with 5 µM CM-H2DCFDA in the dark for 10 min at 37 °C. The cell suspension was collected by centrifugation at 1600× g rpm and supernatants were removed. Collected HAoSMCs were resuspended with 500 µL pre-warmed medium. Intracellular ROS levels in HAoSMCs were then immediately determined via BD Accuri™ C6 flow cytometer (BD biosciences). The result was calculated from four independent experiments and analyzed as fold change compared to control. ## 4.8. TRAP Staining Tartrate-Resistant Acid Phosphatase (TRAP) staining was performed using the TRACP & ALP double-stain Kit (MK300; TAKARA) according to the manufacturer’s instructions. The cells were gently washed with PBS. A total of 120 μL of fixation solution was added to each well for 5 min at room temperature. After that, the cells were washed using sterile distilled water (D.W.). A total of 120 μL of substrate solution was added to each well and covered with parafilm to prevent drying. The plate was incubated at 37 °C for 45 min and washed three times with D.W. to stop the reaction. TRAP-positive cells were visualized and captured via microscope (CKX41; Olympus, Tokyo, Japan) and digital camera (eXcope T300; Olympus). Osteoclast cells that stained as a purple color were counted as TRAP-positive cells from three or more independent experiments. The number of the TRAP-positive cells that have three more nucleus was measured by NIH ImageJ 1.52a software (Silk Scientific Corp., Orem, Utah). ## 4.9. Pit Assay The bone resorption assay kit (CSR-BRA-48KIT; COSMOBIO, Tokyo, Japan) was used to check resorption activity of osteoclast cells according to the manufacturer’s protocol. Briefly, Raw264.7 cells (5 × 103 cells/well) were seeded into calcium phosphate (CaP)-coated 48-well culture plate. Additional collagen type I coating was performed using 50 μg/mL collagen type I (3447-020-01; R&D System, Minneapolis, MN, USA) to mimic a bone biomimetic surface. Raw264.7 cells were cultured at 37 °C and $5\%$ CO2 in DMEM containing $10\%$ FBS. After 24 h, the medium was changed to complete α-MEM medium without phenol-red and then pretreated with a vehicle (DMSO) and citreoviridin (0.5 and 1μg/mL) for 1 h. Subsequently, RANKL (100 ng/mL) was added in the medium to induce an osteoclast differentiation. RANKL and citreoviridin were re-treated every 2 days for 6 days. On day 6, the cells were washed with PBS and treated with $5\%$ sodium hypochlorite for 5 min to remove the cells. Then, the plate was washed with D.W. and dried. Osteoclast resorbing areas were captured using a digital camera (Olympus) attached to microscope (Olympus) and the pit areas obtained from 20 different regions by 4 independent experiments were measured by NIH ImageJ 1.52a software. ## 4.10. Reverse Transcription PCR (RT-PCR) For RT-PCR, total RNA was isolated from the cells using Hybrid-R (305-101; GeneAll Biotechnology, Korea) according to manufacturer’s protocol. The amount of RNA was measured by Nanodrop one (Thermo Fisher Scientific), and 1 μg of RNA was used to synthesize cDNA using the Maxime RT PreMix Kit (25081; iNtRON Biotechnology, Seongnam, Korea). For PCR reaction, AccuPower® PCR PreMix (K-2016; Bioneer, Daejeon, Korea) was used. PCR was performed under the following conditions using a PCR machine (C1000 touch Thermal cycler; Bio-Rad, Hercules, CA, USA): denaturation at 95 °C for 20 s, annealing at 56 °C for 30 s, and extension at 72 °C for 30 s for 35 cycles and, final extension at 72 °C for 5 min. The human (h) and mouse (m) specific primers used are the followings: (m)Calcitonin receptor (forward: 5′-AGC TTG TTG GCA CTT TGT AT-3′; reverse: 5′-TTG CCT ATG CCA GGA CCA AT-3′), (m)Cathepsin K (forward: 5′-GCA GAT GTT TGT GTT GGT CTC T-3′; reverse: 5′-TGG TGG AAA GGT GTG ACA GG-3′), (m)DC-STAMP (forward: 5′-TTG AAC CGA GCT GCA TTC CT-3′; reverse: 5′-GCA CTA CCT TGG CCT TAC CT-3′), (m)NFTAc1 (forward: 5′-GGA GAG TCC GAG AAT CGA GAT-3′; reverse: 5′-TTG CAG CTA GGA AGT ACG TCT -3′), (m)TRAP (forward: 5′-CTC CTG CCT GTT CTC TTC CCA-3′; reverse: 5′- AAG AGA GAA AGT CAA GGG AGT GGC-3′), (m)GAPDH (forward: 5′-CAA GGT CAT CCA TGG ACA ACT TTG-3′; reverse: 5′-GTC CAC CAC CCT GTT GCT GTA G-3′). (h)RANKL (118bp; forward: 5′-CCC AAG TTC TCA TAC CCT GAT G-3′; reverse: 5′-TTC CTC TCC AGA CCG TAA CT-3′), (h)ALP (108bp; forward: 5′-ATG GGA TGG GTC TCC ACA-3′; reverse: 5′-CCA CGA AGG GGA ACT TGT C-3′), (h)Osterix (711bp; forward: 5′- GCT TGA GGA GGA AGT TCA CTA T-3′; reverse: 5′-CCT TCT AGC TGC CCA CTA TTT-3′), (h)GAPDH (133bp; forward: 5′-CAT GGG TGT GAA CCA TGA GA-3′; reverse: 5′-GGT CAT GAG TCC TTC CAC GA-3′). The PCR products from 3 or more independent experiments were electrophoresed on a $1.5\%$ agarose gel. The gel images were captured by a BioRad ChemiDoc XRS imaging system (Bio-Rad) and analyzed with NIH ImageJ 1.52a software. The relative expression in each gene was then calculated using Glyceraldehyde 3-phosphate dehydrogenase (GAPDH). ## 4.11. Western Blot Proteins were extracted from the cells using RIPA buffer (25 mM Tris pH 7.6, 150 mM NaCl, $1\%$ NP-40, $1\%$ sodium deoxycholate, $0.1\%$ SDS) containing protease inhibitor (sc-11697498001; Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and phosphatase inhibitors (4906845001; Thermo Fisher Scientific). The protein concentration was determined using the Bicinchoninic Acid Assay (BCA) method. The protein samples were separated on $12\%$ sodium dodecyl sulfate (SDS)-polyacrylamide gel and then transferred to immobilon-P PVDF membranes (IPVH00010; Merk Millipore, Burlington, MA, USA). The membranes were blocked with $5\%$ skim milk for 30 min at room temperature and incubated overnight at 4 °C with the following antibodies; ERK (1:1000; Cell signaling Technology, Danver, MA, USA, 9102), phospho-ERK (1:1000; Santa Cruz, sc-7383), PKCδ (1:1000; Cell signaling, 2085), phospho-PKCδ (1:500; Cell signaling, 9374), Src (1:1000; Cell signaling, 2110), phospho-Src (1:1000, Cell signaling, 2101), p38 (1:1000; Cell signaling, 9212), phospho-p38 (1:1000; Cell signaling, 9211), JNK (1:1000; Cell signaling, 9252), phospho-JNK (1:1000; Cell signaling, 9251), NF-κB p65 (1:1000; Cell signaling, 6956), phospho-NF-κB p65 (1:1000; Cell signaling, 3033), c-jun (1:1000, Cell signaling, 9165), phospho-c-jun (1:500; Cell signaling, 2361), SM22α (1:5000; Abcam, ab14106), Calponin (1:1000; Abcam, ab46794), α-SMA (1:1000, Abcam, ab5694), osterix (1:500, Santa Cruz, sc393325), RANKL (1:1000; Santa Cruz, sc377079), GAPDH (1:5000; Santa Cruz, sc32233), beta actin (1:5000; Santa Cruz, sc47778). For secondary antibodies, the following secondary antibodies were used; anti-mouse (ADI-SAB-100J; ENZO) or anti-rabbit (ADI-SAB-300-J; ENZO). Secondary antibodies were used at 1:2000 dilution in $5\%$ skim milk in TBST for 1 h at room temperature. Developed using the enhanced chemiluminescence method (AbClon, Seoul, Republic of Korea) and then detected protein expression using a BioRad ChemiDoc XRS imaging system (Bio-Rad). The protein bands were quantified by ImageJ 1.52a software. To detect a soluble RANKL level from HAoSMCs, culture media was concentrated by 100x using Vivaspin® Turbo 4 (VS04T91; Satorius, Goettingen, Germany). ## 4.12. Immunocytochemistry To determine of NF-κB p65 and NFATc1 translocation, immunocytochemistry was performed. After treatment, the cells were rinsed with PBS and fixed with $4\%$ paraformaldehyde in PBS for 10 min at room temperature. The cells were then permeabilized with $0.2\%$ Triton X-100 in PBS for 10 min at room temperature. Nonspecific staining was blocked by $2.5\%$ normal horse serum blocking solution (S-2012-50; Vector Laboratories, Burlingame, CA, USA) for 10 min. After blocking, the cells were incubated with primary antibodies against total NF-κB (3039; Cell Signaling) or NFATc1 (sc-7294; Santa Cruz). All antibodies were diluted 1:200 in $2.5\%$ normal horse serum for overnight at 4 °C. Rhodamine-conjugated anti mouse (1:500; AP124R, Merk Millipore) secondary antibodies were used. The cell nuclei were stained with diamidino-2-pehnylindole (DAPI) (D21490; Thermo Fisher Scientific). The images of NF-κB and NFATc1 translocation were taken using a LSM 700 laser scanning confocal microscope (Carl zeiss, Overkochen, Germany). ## 4.13. Statistical Analysis The data were expressed as the mean ± standard error of the mean (SEM). Statistical analyses were performed using GraphPad Prism 7 software (GraphPad Software, San Diego, CA, USA). One-way analysis of variance (ANOVA) was used to compare three or more groups followed by a Bonferroni post hoc test. Student’s t test was used to compare two groups. Differences with p values of less than 0.05 were considered statistically significant. ## 5. Conclusions In the present study, a novel biological effect of CIT on VC and osteoporosis that has not been reported anywhere has been demonstrated. Our data suggest that CIT can be an effective agent to simultaneously control both VC and osteoporosis and call for further studies to validate its in vivo efficacy and to find an optimal therapeutic window. ## References 1. Blacher J., Guerin A.P., Pannier B., Marchais S.J., London G.M.. Arterial calcifications, arterial stiffness, and cardiovascular risk in end-stage renal disease. Hypertension (2001) 38 938-942. DOI: 10.1161/hy1001.096358 2. Radford N.B., DeFina L.F., Barlow C.E., Lakoski S.G., Leonard D., Paixao A.R., Khera A., Levine B.D.. Progression of CAC Score and Risk of Incident CVD. JACC Cardiovasc. Imaging (2016) 9 1420-1429. DOI: 10.1016/j.jcmg.2016.03.010 3. Yamamoto H., Shavelle D., Takasu J., Lu B., Mao S.S., Fischer H., Budoff M.J.. Valvular and thoracic aortic calcium as a marker of the extent and severity of angiographic coronary artery disease. Am. Heart J. (2003) 146 153-159. DOI: 10.1016/S0002-8703(03)00105-4 4. Leopold J.A.. Vascular calcification: Mechanisms of vascular smooth muscle cell calcification. Trends Cardiovasc. Med. (2015) 25 267-274. DOI: 10.1016/j.tcm.2014.10.021 5. Cannata-Andia J.B., Carrillo-Lopez N., Messina O.D., Hamdy N.A.T., Panizo S., Ferrari S.L.. Pathophysiology of Vascular Calcification and Bone Loss: Linked Disorders of Ageing?. Nutrients (2021) 13. DOI: 10.3390/nu13113835 6. Drake M.T., Clarke B.L., Lewiecki E.M.. The Pathophysiology and Treatment of Osteoporosis. Clin. Ther. (2015) 37 1837-1850. DOI: 10.1016/j.clinthera.2015.06.006 7. O’Brien W., Fissel B.M., Maeda Y., Yan J., Ge X., Gravallese E.M., Aliprantis A.O., Charles J.F.. RANK-Independent Oste oclast Formation and Bone Erosion in Inflammatory Arthritis. Arthritis Rheumatol. (2016) 68 2889-2900. DOI: 10.1002/art.39837 8. Henaut L., Massy Z.A.. New insights into the key role of interleukin 6 in vascular calcification of chronic kidney disease. Nephrol. Dial. Transpl. (2018) 33 543-548. DOI: 10.1093/ndt/gfx379 9. Alexander M.Y.. RANKL links arterial calcification with osteolysis. Circ. Res. (2009) 104 1032-1034. DOI: 10.1161/CIRCRESAHA.109.198010 10. Banks L.M., Lees B., MacSweeney J.E., Stevenson J.C.. Effect of degenerative spinal and aortic calcification on bone density measurements in post-menopausal women: Links between osteoporosis and cardiovascular disease?. Eur. J. Clin. Investig. (1994) 24 813-817. DOI: 10.1111/j.1365-2362.1994.tb02024.x 11. Frye M.A., Melton L.J., Bryant S.C., Fitzpatrick L.A., Wahner H.W., Schwartz R.S., Riggs B.L.. Osteoporosis and calci fication of the aorta. Bone Miner. (1992) 19 185-194. DOI: 10.1016/0169-6009(92)90925-4 12. Hofbauer L.C., Brueck C.C., Shanahan C.M., Schoppet M., Dobnig H.. Vascular calcification and osteoporosis--from clinical observation towards molecular understanding. Osteoporos. Int. (2007) 18 251-259. DOI: 10.1007/s00198-006-0282-z 13. Garcia-Gomez M.C., Vilahur G.. Osteoporosis and vascular calcification: A shared scenario. Clin. Investig. Arterioscler. (2020) 32 33-42. DOI: 10.1016/j.artere.2019.03.008 14. Veronese N., Stubbs B., Crepaldi G., Solmi M., Cooper C., Harvey N.C., Reginster J.Y., Rizzoli R., Civitelli R., Schofield P.. Relationship Between Low Bone Mineral Density and Fractures with Incident Cardiovascular Disease: A Systematic Review and Meta-Analysis. J. Bone Miner. Res. (2017) 32 1126-1135. DOI: 10.1002/jbmr.3089 15. Shimizu H., Nakagami H., Morishita R.. [Bone metabolism and cardiovascular function update. Cross link of hypertension, bone loss and vascular calcification—Common back grounds in renin angiotensin system with anti-aging aspect -]. Clin. Calcium (2014) 24 53-62 16. Sato H., Sato M., Kanai H., Uchiyama T., Iso T., Ohyama Y., Sakamoto H., Tamura J., Nagai R., Kurabayashi M.. Mito chondrial reactive oxygen species and c-Src play a critical role in hypoxic response in vascular smooth muscle cells. Cardiovasc. Res. (2005) 67 714-722. DOI: 10.1016/j.cardiores.2005.04.017 17. Suryavanshi S.V., Kulkarni Y.A.. NF-kappabeta: A Potential Target in the Management of Vascular Complications of Diabetes. Front. Pharmacol. (2017) 8 798. DOI: 10.3389/fphar.2017.00798 18. Jia G., Stormont R.M., Gangahar D.M., Agrawal D.K.. Role of matrix Gla protein in angiotensin II-induced exacerbation of vascular calcification. Am. J. Physiol. Heart Circ. Physiol. (2012) 303 H523-H532. DOI: 10.1152/ajpheart.00826.2011 19. Das S., Senapati P., Chen Z., Reddy M.A., Ganguly R., Lanting L., Mandi V., Bansal A., Leung A., Zhang S.. Regu lation of angiotensin II actions by enhancers and super-enhancers in vascular smooth muscle cells. Nat. Commun. (2017) 8 1467. DOI: 10.1038/s41467-017-01629-7 20. Park J.H., Lee N.K., Lee S.Y.. Current Understanding of RANK Signaling in Osteoclast Differentiation and Maturation. Mol. Cells (2017) 40 706-713. DOI: 10.14348/molcells.2017.0225 21. Kobayashi N., Kadono Y., Naito A., Matsumoto K., Yamamoto T., Tanaka S., Inoue J.. Segregation of TRAF6-mediated signaling pathways clarifies its role in osteoclastogenesis. EMBO J. (2001) 20 1271-1280. DOI: 10.1093/emboj/20.6.1271 22. Yagi M., Ninomiya K., Fujita N., Suzuki T., Iwasaki R., Morita K., Hosogane N., Matsuo K., Toyama Y., Suda T.. Induction of DC-STAMP by alternative activation and downstream signaling mechanisms. J. Bone Miner. Res. (2007) 22 992-1001. DOI: 10.1359/jbmr.070401 23. Matsumoto M., Kogawa M., Wada S., Takayanagi H., Tsujimoto M., Katayama S., Hisatake K., Nogi Y.. Essential role of p38 mitogen-activated protein kinase in cathepsin K gene expression during osteoclastogenesis through association of NFATc1 and PU.1. J. Biol. Chem. (2004) 279 45969-45979. DOI: 10.1074/jbc.M408795200 24. Byon C.H., Sun Y., Chen J., Yuan K., Mao X., Heath J.M., Anderson P.G., Tintut Y., Demer L.L., Wang D.. Runx2-upregulated receptor activator of nuclear factor kappaB ligand in calcifying smooth muscle cells promotes migration and osteoclastic differentiation of macrophages. Arterioscler. Thromb. Vasc. Biol. (2011) 31 1387-1396. DOI: 10.1161/ATVBAHA.110.222547 25. Osako M.K., Nakagami H., Shimamura M., Koriyama H., Nakagami F., Shimizu H., Miyake T., Yoshizumi M., Rakugi H., Morishita R.. Cross-talk of receptor activator of nuclear factor-kappaB ligand signaling with renin-angiotensin system in vascular calcification. Arterioscler. Thromb. Vasc. Biol. (2013) 33 1287-1296. DOI: 10.1161/ATVBAHA.112.301099 26. Asaba Y., Ito M., Fumoto T., Watanabe K., Fukuhara R., Takeshita S., Nimura Y., Ishida J., Fukamizu A., Ikeda K.. Activation of renin-angiotensin system induces osteoporosis independently of hypertension. J. Bone Miner. Res. (2009) 24 241-250. DOI: 10.1359/jbmr.081006 27. Sato K., Suematsu A., Okamoto K., Yamaguchi A., Morishita Y., Kadono Y., Tanaka S., Kodama T., Akira S., Iwakura Y.. Th17 functions as an osteoclastogenic helper T cell subset that links T cell activation and bone destruction. J. Exp. Med. (2006) 203 2673-2682. DOI: 10.1084/jem.20061775 28. Ono T., Hayashi M., Sasaki F., Nakashima T.. RANKL biology: Bone metabolism, the immune system, and beyond. Inflamm. Regen. (2020) 40 2. DOI: 10.1186/s41232-019-0111-3 29. Armstrong Z.B., Boughner D.R., Drangova M., Rogers K.A.. Angiotensin II type 1 receptor blocker inhibits arterial calcification in a pre-clinical model. Cardiovasc. Res. (2011) 90 165-170. DOI: 10.1093/cvr/cvq391 30. Ruths S., Bakken M.S., Ranhoff A.H., Hunskaar S., Engesaeter L.B., Engeland A.. Risk of hip fracture among older people using antihypertensive drugs: A nationwide cohort study. BMC Geriatr. (2015) 15. DOI: 10.1186/s12877-015-0154-5 31. Jilka R.L., Hangoc G., Girasole G., Passeri G., Williams D.C., Abrams J.S., Boyce B., Broxmeyer H., Manolagas S.C.. Increased osteoclast development after estrogen loss: Mediation by interleukin-6. Science (1992) 257 88-91. DOI: 10.1126/science.1621100 32. Zhao R.. Immune regulation of osteoclast function in postmenopausal osteoporosis: A critical interdisciplinary perspective. Int. J. Med. Sci. (2012) 9 825-832. DOI: 10.7150/ijms.5180 33. Toth A., Balogh E., Jeney V.. Regulation of Vascular Calcification by Reactive Oxygen Species. Antioxidants (2020) 9. DOI: 10.3390/antiox9100963 34. Chao C.T., Yeh H.Y., Tsai Y.T., Chuang P.H., Yuan T.H., Huang J.W., Chen H.W.. Natural and non-natural antioxidative compounds: Potential candidates for treatment of vascular calcification. Cell Death Discov. (2019) 5 145. DOI: 10.1038/s41420-019-0225-z 35. Hou H.F., Yuan N., Guo Q., Sun T., Li C., Liu J.B., Li Q.W., Jiang B.F.. Citreoviridin Enhances Atherogenesis in Hypercho lesterolemic ApoE-Deficient Mice via Upregulating Inflammation and Endothelial Dysfunction. PLoS ONE (2015) 10. DOI: 10.1371/journal.pone.0125956 36. Almeida M.I., Almeida N.G., Carvalho K.L., Goncalves G.A., Silva C.N., Santos E.A., Garcia J.C., Vargas E.A.. Co-occur rence of aflatoxins B(1), B(2), G(1) and G(2), ochratoxin A, zearalenone, deoxynivalenol, and citreoviridin in rice in Brazil. Food Addit. Contam. Part A Chem. Anal. Control Expo. Risk Assess. (2012) 29 694-703. DOI: 10.1080/19440049.2011.651750 37. Sun S.. Chronic exposure to cereal mycotoxin likely citreoviridin may be a trigger for Keshan disease mainly through oxidative stress mechanism. Med. Hypotheses (2010) 74 841-842. DOI: 10.1016/j.mehy.2009.11.043 38. Wang Y., Liu Y., Liu X., Jiang L., Yang G., Sun X., Geng C., Li Q., Yao X., Chen M.. Citreoviridin Induces Autophagy-Dependent Apoptosis through Lysosomal-Mitochondrial Axis in Human Liver HepG2 Cells. Toxins (2015) 7 3030-3044. DOI: 10.3390/toxins7083030 39. Hou H., Zhou R., Jia Q., Li Q., Kang L., Jiao P., Li D., Jiang B.. Citreoviridin enhances tumor necrosis factor-alpha-induced adhesion of human umbilical vein endothelial cells. Toxicol. Ind. Health (2015) 31 193-201. DOI: 10.1177/0748233713483194 40. Feng C., Li D., Chen M., Jiang L., Liu X., Li Q., Geng C., Sun X., Yang G., Zhang L.. Citreoviridin induces myocardial apoptosis through PPAR-gamma-mTORC2-mediated autophagic pathway and the protective effect of thiamine and selenium. Chem. Biol. Interact. (2019) 311 108795. DOI: 10.1016/j.cbi.2019.108795 41. Heussner A.H., Dietrich D.R., O’Brien E.. In vitro investigation of individual and combined cytotoxic effects of ochratoxin A and other selected mycotoxins on renal cells. Toxicol. In Vitro (2006) 20 332-341. DOI: 10.1016/j.tiv.2005.08.003 42. Dressler D., Johnson E.A.. Botulinum toxin therapy: Past, present and future developments. J. Neural. Transm. (2022) 129 829-833. DOI: 10.1007/s00702-022-02494-5 43. Poetsch F., Henze L.A., Estepa M., Moser B., Pieske B., Lang F., Eckardt K.U., Alesutan I., Voelkl J.. Role of SGK1 in the Osteogenic Transdifferentiation and Calcification of Vascular Smooth Muscle Cells Promoted by Hyperglycemic Conditions. Int. J. Mol. Sci. (2020) 21. DOI: 10.3390/ijms21197207 44. Dikalov S.I., Nazarewicz R.R.. Angiotensin II-induced production of mitochondrial reactive oxygen species: Potential mecha nisms and relevance for cardiovascular disease. Antioxid. Redox Signal. (2013) 19 1085-1094. DOI: 10.1089/ars.2012.4604 45. Jeong J., Cho S., Seo M., Lee B.S., Jang Y., Lim S., Park S.. Soluble RAGE attenuates Ang II-induced arterial calcification via inhibiting AT1R-HMGB1-RAGE axis. Atherosclerosis (2022) 346 53-62. DOI: 10.1016/j.atherosclerosis.2022.02.022 46. Liang B., Burley G., Lin S., Shi Y.C.. Osteoporosis pathogenesis and treatment: Existing and emerging avenues. Cell. Mol. Biol. Lett. (2022) 27 72. DOI: 10.1186/s11658-022-00371-3 47. Udagawa N., Koide M., Nakamura M., Nakamichi Y., Yamashita T., Uehara S., Kobayashi Y., Furuya Y., Yasuda H., Fukuda C.. Osteoclast differentiation by RANKL and OPG signaling pathways. J. Bone Miner. Metab. (2021) 39 19-26. DOI: 10.1007/s00774-020-01162-6 48. Yagi M., Miyamoto T., Sawatani Y., Iwamoto K., Hosogane N., Fujita N., Morita K., Ninomiya K., Suzuki T., Miyamoto K.. DC-STAMP is essential for cell-cell fusion in osteoclasts and foreign body giant cells. J. Exp. Med. (2005) 202 345-351. DOI: 10.1084/jem.20050645 49. Kim J.H., Kim N.. Regulation of NFATc1 in Osteoclast Differentiation. J. Bone Metab. (2014) 21 233-241. DOI: 10.11005/jbm.2014.21.4.233 50. Hsiao C.Y., Chen T.H., Chu T.H., Ting Y.N., Tsai P.J., Shyu J.F.. Calcitonin Induces Bone Formation by Increasing Expression of Wnt10b in Osteoclasts in Ovariectomy-Induced Osteoporotic Rats. Front. Endocrinol. (2020) 11 613. DOI: 10.3389/fendo.2020.00613 51. Tseng W., Graham L.S., Geng Y., Reddy A., Lu J., Effros R.B., Demer L., Tintut Y.. PKA-induced receptor activator of NF-kappaB ligand (RANKL) expression in vascular cells mediates osteoclastogenesis but not matrix calcification. J. Biol. Chem. (2010) 285 29925-29931. DOI: 10.1074/jbc.M110.117366 52. Maranduca M.A., Tanase D.M., Branisteanu D.C., Serban D.N., Branisteanu D.E., Serban I.L.. Involvement of proinflamma tory cytokines in angiotensin II-induced hypertension in rat. Exp. Ther. Med. (2020) 20 3541-3545. DOI: 10.3892/etm.2020.9100 53. Zhou S., Lu H., Chen R., Tian Y., Jiang Y., Zhang S., Ni D., Su Z., Shao X.. Angiotensin II enhances the acetylation and release of HMGB1 in RAW264.7 macrophage. Cell. Biol. Int. (2018) 42 1160-1169. DOI: 10.1002/cbin.10984 54. Lim S., Lee M.E., Jeong J., Lee J., Cho S., Seo M., Park S.. sRAGE attenuates angiotensin II-induced cardiomyocyte hyper trophy by inhibiting RAGE-NFkappaB-NLRP3 activation. Inflamm. Res. (2018) 67 691-701. DOI: 10.1007/s00011-018-1160-9 55. Bidwell J.P., Yang J., Robling A.G.. Is HMGB1 an osteocyte alarmin?. J. Cell. Biochem. (2008) 103 1671-1680. DOI: 10.1002/jcb.21572 56. Sakamoto Y., Okui T., Yoneda T., Ryumon S., Nakamura T., Kawai H., Kunisada Y., Ibaragi S., Masui M., Ono K.. High-mobility group box 1 induces bone destruction associated with advanced oral squamous cancer via RAGE and TLR4. Biochem. Biophys. Res. Commun. (2020) 531 422-430. DOI: 10.1016/j.bbrc.2020.07.120
--- title: Systematic Breakfast Consumption of Medium-Quantity and High-Quality Food Choices Is Associated with Better Vascular Health in Individuals with Cardiovascular Disease Risk Factors authors: - Eirini D. Basdeki - Antonios A. Argyris - Olga Efthymiou - Elpida Athanasopoulou - Petros P. Sfikakis - Athanase D. Protogerou - Kalliopi Karatzi journal: Nutrients year: 2023 pmcid: PMC9967081 doi: 10.3390/nu15041025 license: CC BY 4.0 --- # Systematic Breakfast Consumption of Medium-Quantity and High-Quality Food Choices Is Associated with Better Vascular Health in Individuals with Cardiovascular Disease Risk Factors ## Abstract Background: Breakfast consumption has been associated with the improvement of many cardiovascular disease (CVD) risk factors, yet data regarding its association with subclinical vascular damage, which precedes the onset of CVD, are scarce. The aim of this study is to investigate this association in a large sample of adults with CVD risk factors. Methods: Anthropometric measurements, vascular biomarkers and dietary intake with two 24-h dietary recalls, focusing on breakfast frequency and its quantity and content, were assessed in 902 adults ($45.2\%$ males). Breakfast quality was assessed by identifying a posteriori breakfast dietary pattern (DP) by using principal component analysis (PCA). Results: Systematic breakfast consumption (SBC) was inversely associated with central systolic blood pressure (b: −3.28, $95\%$ C.I.: −5.7 to −0.86), diastolic blood pressure (b: −1.85, $95\%$ C.I.: −3.34 to −0.36), augmentation index (b: −3.17, $95\%$ C.I.:−4.98 to 1.35) and left carotid intima media thickness (b: −0.03, $95\%$ C.I.:−0.06 to −0.01) compared to breakfast skipping independently of age, sex, hypertension, diabetes, dyslipidemia, smoking, and BMI. SBC of 10–$20\%$ of daily total energy intake (dTEI) was inversely associated with Aix (b: −2.31, $95\%$ C.I.:−4.05 to −0.57) compared to <$10\%$ dTEI after adjustment for the aforementioned confounders. DP1 (high coffee and sugar consumption, low consumption of low- and full-fat dairy products, fruits, and fresh juices) was positively associated with Aix (b: 1.19, $95\%$ C.I.: 0.48 to 1.90). Conclusion: SBC comprised of medium-energy density and high-nutrient content food items may be a simple daily habit associated with better vascular health. ## 1. Introduction Breakfast is defined either as the first meal of the day which breaks the night fast, or the meal that is consumed within 2–3 h after waking, including at least one food item or drink and can be consumed in any location [1]. Moreover, its caloric content usually does not exceed 20–$35\%$ of daily total energy intake (dTEI) [2]. Breakfast consumption has been associated with a variety of beneficial health effects, such as lower body mass index (BMI) and better insulin sensitivity, which are significant risk factors for cardiovascular disease (CVD) and Type 2 diabetes mellitus [2]. Moreover, breakfast skipping has been associated with increased risk for CVD or development of risk factors for CVD [3,4,5]. Another important aspect of breakfast consumption is its content in high-quality food items, which is positively associated with mental health, cardiometabolic indices and quality of life, and inversely associated with anxiety and depression [6,7,8]. Early detection of preclinical vascular abnormalities is an important step for efficient prevention of CVD, a leading cause of global mortality [9,10]. Several hemodynamic and vascular biomarkers exist that depict arterial functional and structural damage. Of these, the most commonly used are peripheral and central blood pressure (BP) (systolic and diastolic (SBP, DBP)), augmentation index (Aix) (pressure wave reflections), and pulse wave velocity (PWV) (arterial stiffness), as indices of vascular function, and carotid intima media thickness (IMT) as an index of arterial remodeling and structural damage [10,11,12]. Until now, several markers of subclinical vascular damage (SVD) have been associated with unhealthy dietary behavior; however the possible association between breakfast consumption and subclinical vascular biomarkers has been understudied [13,14]. The only two available studies included a mixed adult population and assessed only breakfast consumption, but not the quality or quantity of the breakfast, and limited vascular biomarkers were employed. Both of them concluded that skipping breakfast was inversely associated with arterial stiffness and atheromatosis. The aim of the present study was to investigate the potential association between breakfast consumption and vascular health in a large sample of adults with at least one CVD risk factor. Our hypothesis was that breakfast consumption is associated with better vascular structure and function. To test this hypothesis, the frequency, the quantity, and the quality of breakfast were taken into consideration, and multiple vascular indices were employed regarding arterial function and structure. ## 2.1. The Study The present cross-sectional study was conducted at the Cardiovascular Prevention and Research Unit, Athens, Greece, in accordance with the Declaration of Helsinki and its later amendments and was approved by the Bioethics Committee, and all participants provided written informed consent before entering the study. ## 2.2. Study Population In the present study, 902 adults participated ($45.2\%$ males), with one or more CVD risk factor, but free of established CVD. Exclusion criteria were cancer, liver disease, or eating disorders. All patients abstained from food, drink, and any medication 12 h prior to their visit for anthropometric and vascular measurements. ## 2.3. Definition of CVD Risk Factors Hypertension was defined by the use of antihypertensive drugs and/or office BP measurement higher than $\frac{139}{89}$ mm Hg. Dyslipidemia was defined on the basis of treatment with lipid-modifying drugs or low-density lipoprotein cholesterol level >160 mg/dL [15]. Current smoking was defined by the use of at least one cigarette per day each day of the week; ex-smoking was defined as interruption for more than six months. Family history of premature CVD was defined as the presence of coronary heart disease in a first-degree relative under the age of 55 years for males and 65 years for females [16]. ## 2.4. Anthropometric Measurements A trained dietician performed a physical examination on every participant of the study, using the same protocol and equipment. Participants’ body weight was measured without shoes or heavy clothes to the nearest 10 g by using the Tanita Body Composition Analyzer, BC-418 weight scale. Participants’ height was measured without shoes, with the participants standing with their shoulders relaxed, their arms hanging freely, and their head in Frankfurt horizontal plane, to the nearest 0.1 cm by using stadiometer SECA 213. Body mass index (BMI) was calculated by using both measurements as weight/(height2) (Kg/m2). ## 2.5. Vascular Assessment All participants underwent vascular assessment, performed by the same examiner. BP was assessed, after a 10-min rest in supine position and the average of three sequential readings with a one-min interval was used in the analysis (MicrolifeWatchBP Office, Microlife AG, Widnau, Switzerland). Central BP, Aix (also normalized for the heart rate of 75 bpm -Aix) and carotid to femoral PWV were assessed by a common tonometric method and the use of a generalized transfer function (Sphygmocor; Atcor Medical, Sydney, Australia), as previously described [17]. High-resolution ultrasonography was performed at the carotid and femoral arteries for the assessment of IMT (Vivid 7 Pro, General Electric Healthcare, Little Chalfont, United Kingdom). IMT was calculated in the right and left common carotid artery and the carotid bulb and internal carotid artery, as well as at the femoral artery. Two sequential images were obtained, and the average measurements were used [18]. ## 2.6. Dietary Intake Assessment and Definition of Breakfast Consumption Two 24-h dietary recalls were conducted (one weekday and one weekend day) with an interval of at least one week in between, by trained dieticians. Data were analyzed in food groups, where separation was based on food equivalents with further separation in subgroups where needed. Moreover, analysis of energy content, macronutrients, and micronutrients was conducted by using Nutritionist Pro software (Axxya Systems Nutritionist Pro TM 2011). Breakfast consumption was defined as the first meal of the day consumed in the morning after a night’s sleep. Systematic breakfast skipping was defined as consumption of breakfast ≤1 times out of 2 total 24 h recalls, and breakfast consumption in both recalls was considered as systematic breakfast consumption. Additionally, consumption of solely coffee was not considered as breakfast consumption. Breakfast quality was indirectly assessed by identifying a posteriori dietary patterns. ## 2.7. Statistical Analysis Statistical analysis was conducted by using statistical package 21.0 (SPSS: Statistical Package for Social Sciences, SPSS Inc., Chicago, IL, USA). The level of statistical significance was set at p ≤ 0.05 for all analyses. Normality of distribution of continuous variables was checked by using the Kolmogorov–Smirnov test. Categorical variables are presented as relative frequencies (%) and continuous as mean ± standard deviation (SD). Breakfast caloric intake was divided into three categories depending on the percentage of dTEI consumed in that meal: systematic breakfast consumption of >0–$10\%$ of dTEI (first category); systematic breakfast consumption of 10–$20\%$ of dTEI (second category); systematic breakfast consumption of >$20\%$ of dTEI (third category) [19]. Differences between systematic breakfast consumers and systematic breakfast skippers were tested with independent sample t-tests for continuous variables and with chi-square for categorical variables. Analysis of variance (one-way ANOVA) was conducted to address differences between the three breakfast categories (related to the % of dTEI consumed on breakfast) among systematic breakfast consumers. Multiple linear regression analysis was performed to determine the independent associations between vascular biomarkers and systematic breakfast consumption. Results are presented as standardized beta coefficients and $95\%$ confidence intervals (CI) (for peripheral and central SBP, DBP, PWV, Aix, right and left IMT). Moreover, univariate analysis of variance was performed to check the associations between vascular biomarkers and categories of breakfast according to percentages of dTEI consumed at breakfast. Three different models were applied to assess all of the above associations:model 1: adjusted for age and sex;model 2: adjusted for age, sex, hypertension, diabetes mellitus, dyslipidemia, and smoking; andmodel 3: adjusted for age, sex, hypertension, diabetes mellitus, dyslipidemia, smoking, and BMI. Principal component analysis (PCA) was applied to identify dietary patterns (DP) among study participants with systematic breakfast consumption. The Kaiser–Meyer–Olkin (KMO) criterion was applied, and it was equal to 0.432. The orthogonal rotation (varimax option) was used to derive optimal noncorrelated components (DPs) and the Bartlett’s method was used to estimate factor scores. The selection of the optimal number of components was based on an eigenvalue >1 (Kaiser criterion), and it was further corroborated by visual assessment of the scree plot, retaining only components on the steep slope. If one of the initial variables correlated with more than one component, it participated in the interpretation of the DP in which it displayed the highest coefficient value. Three DPs emerged after PCA. Multiple linear regression analysis was applied to determine possible associations between DPs and subclinical vascular biomarkers. Results are presented as standardized beta coefficients and $95\%$ confidence intervals (CI) (for peripheral and central SBP, DBP, PWV, Aix, and right and left IMT). The aforementioned models that were applied in breakfast quantity analyses were also applied, in order to assess the above associations regarding breakfast quality. ## 3. Results The present study included 902 subjects ($45.2\%$ males) with at least one risk factor for CVD, mean age 52.4 ± 13.8 years, and mean BMI 27.97 ± 5.52 Kg/m2. The total sample consisted of 765 systematic and 136 nonsystematic breakfast consumers. Nonsystematic breakfast consumers had significantly higher total cholesterol ($$p \leq 0.003$$), LDL ($$p \leq 0.002$$), triglycerides ($$p \leq 0.01$$), and DBP levels ($$p \leq 0.026$$) compared to systematic breakfast consumers. On the other hand, systematic breakfast consumers had significantly higher daily energy intake levels ($$p \leq 0.000$$). Table 1 summarizes characteristics of the total study sample and systematic and nonsystematic breakfast consumers. Table 2 summarizes vascular biomarkers of systematic breakfast consumers ($$n = 765$$) according to percentage of energy intake consumed at breakfast. There were no differences in vascular indices between the three groups, except for PWV. Specifically, PWV tended to be higher ($$p \leq 0.051$$) in the group that consumed >$20\%$ compared to the group that consumed >0–$10\%$ of dTEI at breakfast. Data regarding dietary intake of systematic breakfast consumers according to dTEI at breakfast is shown in Table 3. Participants at the lowest group of dTEI at breakfast (>0–$10\%$ of dTEI) consumed a significantly higher percentage of energy from fat ($$p \leq 0.029$$) and had higher dTEI ($p \leq 0.001$). Few differences between the three groups were noted regarding food groups. Specifically, breakfast consumers with >$20\%$ of dTEI at breakfast consumed significantly more low-fat ($$p \leq 0.004$$) and full-fat ($$p \leq 0.006$$) dairy products and more fruits ($$p \leq 0.019$$) compared to the group that consumed >0–$10\%$ of dTEI at breakfast. Associations of subclinical vascular biomarkers and systematic breakfast consumption are presented in Table 4. Systematic breakfast consumption was inversely associated with central SBP (B: −3.28, $95\%$ C.I.: −5.7–(−0.86)), DBP (B: −1.85, $95\%$ C.I.: −3.34–(−0.36)), Aix (B: −3.17, $95\%$ C.I.: −4.98–1.35), and left IMT (B: −0.03, $95\%$ C.I.: −0.06–(−0.01)) independently of age, sex, hypertension, diabetes, dyslipidemia, smoking, and BMI. Table 5 shows the associations between vascular biomarkers and systematic breakfast consumption categories according to percentage of dTEI consumed at breakfast. Consumption of 10–$20\%$ of dTEI at breakfast was inversely associated with Aix levels (B: −2.31, $95\%$ C.I.: −4.05–(−0.57)) after adjustment for all possible confounders, as presented above. Table 6 presents the three DPs that emerged from PCA analysis. DP1 consists of high coffee and sugar consumption, low consumption of low-fat and full-fat dairy products and low consumption of fruits and fresh juices; DP2 consists of high consumption of refined grains, cold cuts and full-fat cheese and DP3 consists of high consumption of sugar/honey/jam substitutes, whole wheat grains and tahini/peanut butter and margarine. Associations between the three identified DPs and subclinical vascular biomarkers are presented in Table 7. DP1 was positively associated with Aix (B: 1.19, $95\%$ C.I.: 0.48–1.90) independently of age, sex, hypertension, diabetes, dyslipidemia, smoking, and BMI. It was also positively associated with DBP levels (0.66 (0.10, 1.23)) after adjusting for age, sex, hypertension, diabetes, dyslipidemia, and smoking, but after adding BMI to the model significance was lost. DP2 and DP3 showed no statistically significant associations with subclinical vascular biomarkers. ## 4. Discussion The present study aimed to examine the possible association between breakfast consumption and SVD in adults, free of overt CVD but with at least one CVD risk factor. Both breakfast quality and quantity were examined: quantity as percentage of dTEI consumed at breakfast and quality through a posteriori breakfast dietary patterns. The main findings of this study are: (i) the positive association of systematic breakfast consumption with better hemodynamic indices compared to breakfast skipping; (ii) the inverse association of moderate compared to lower energy consumption at breakfast with Aix; and (iii) the positive association of DP1 (high coffee and sugar consumption, low consumption of low-fat and full-fat dairy products, low consumption of fruits and fresh juices) with Aix. Systematic breakfast consumption was found to be inversely associated with central SBP, peripheral DBP, Aix, and IMT compared to breakfast skipping. Previous studies, examining associations between breakfast eating/skipping and BP levels or risk of hypertension, are in agreement with the present findings [20,21]. Interestingly, there are no available studies examining the association of breakfast consumption with arterial stiffness using Aix, but there is a relevant cross-sectional study conducted in adults with established CVD or with CVD risk factors, showing an inverse association between breakfast consumption and PWV, which is in alignment with the findings of the present study [13]. The same study also showed that breakfast skippers had significantly higher carotid IMT compared to systematic breakfast consumers, which is also in line with the present findings. Considering all the mentioned findings, it is possible that systematic breakfast consumption might lead to a reduction not only in peripheral resistance, but also, and mainly, in pressure wave reflection coefficients, resulting in lower pressure wave reflections as quantified by Aix, which in turn decreases aortic pressures. This observation explains why in the present study aortic, but not peripheral SBP, was associated with systematic breakfast consumption. As recently shown in a systematic metanalysis, lower Aix are associated with lower ventricular mass [22], and this may explain the well-described association with reduced mortality [11]. The exact mechanism behind the benefits of systematic breakfast eating on central hemodynamics must be elucidated by future relevant studies. Among systematic breakfast consumers, it was found that moderate (10–$20\%$) energy intake at breakfast was associated with lower pressure wave reflections (Aix) compared to lower energy intake consumption. There are no available studies examining the impact of breakfast energy content on central haemodynamics, pressure wave reflections, and arterial stiffness. However, there is one cross-sectional study conducted in either healthy middle-aged adults or middle-aged adults having CVD risk factors, which investigated the association of energy consumed at breakfast with atheromatosis [14]. In contrast with the present study, it was found that moderate energy consumption (5–$20\%$) at breakfast had higher risk of subclinical atheromatosis, compared to the group that consumed a higher amount of energy at breakfast. This discrepancy might be attributed to differences in the studied populations (in the aforementioned study, healthy adults were also included), and also to the quite distinct pathologies that induce atheromatosis and modify pressure wave reflections. Regarding breakfast quality, out of the three DPs that were identified, DP1 (high coffee and sugar consumption, low consumption of low-fat and full-fat dairy products, low consumption of fruits and fresh juices) was positively associated with Aix. There are no available studies investigating the association between DPs at breakfast and vascular health, but some of the DPs’ components have been previously associated with vascular biomarkers. Data regarding coffee consumption and Aix are insufficient, because there are studies in agreement with our findings [23] and studies that found an inverse association between coffee consumption and Aix [24,25]. The positive association of coffee consumption with Aix could be explained by the increased wave reflections caused by caffeine intake, which in turn increase Aix [23]. Dairy product consumption has also been inversely associated with arterial stiffness, which is attributed to bioactive peptides released during milk-protein digestion [26], but there are also opposite findings reported [27]; hence, this issue needs further research. The positive association of low consumption of fruits and fresh juices in the DP with Aix could be supported by the fact that increased concentration of flavonoids in fruits, have been described as having a protective role against unfavorable vascular alterations [28,29]. The present study is the first to assess the association between breakfast quality and SVD. It is also the first study that assessed the association of breakfast consumption by using a large number of subclinical vascular biomarkers. Vascular assessment was conducted by the same examiner with the same equipment, which reduces observer’s variability. Moreover, 24-h recalls were used for the nutritional assessment, which provide useful information regarding participants’ meal patterns, quality, and quantity of food. However, this study has some certain limitations. Its cross-sectional design does not allow us to establish a causative relationship. Furthermore, 24-h recalls are based on participants’ memory and are prone to recall bias. Social desirability bias should also be added to the study’s limitations. Finally, dietary intake analysis was conducted by the use of Nutritionist Pro, a program based on US food databases, and there may be differences in the content of some nutrients compared with the food items consumed in Greece. ## 5. Conclusions As a conclusion, systematic consumption of a high-quality, moderate-in-energy breakfast is beneficially associated with BP levels and particularly with aortic SBP due to lower pressure wave reflections. Breakfast consumption might be a simple, everyday dietary habit that could promote vascular health; however, these data need to be further investigated and confirmed. ## References 1. O’Neil C.E., Byrd-Bredbenner C., Hayes D., Jana L., Klinger S.E., Stephenson-Martin S.. The Role of Breakfast in Health: Definition and Criteria for a Quality Breakfast. J. Acad. Nutr. Diet. (2014) 114 S8-S26. DOI: 10.1016/j.jand.2014.08.022 2. Timlin M.T., Pereira M.A.. Breakfast frequency and quality in the etiology of adult obesity and chronic diseases. Nutr. Rev. (2007) 65 268-281. DOI: 10.1111/j.1753-4887.2007.tb00304.x 3. Ofori-Asenso R., Owen A.J., Liew D.. Skipping Breakfast and the Risk of Cardiovascular Disease and Death: A Systematic Review of Prospective Cohort Studies in Primary Prevention Settings. J. Cardiovasc. Dev. Dis. (2019) 6. DOI: 10.3390/jcdd6030030 4. Cahill L.E., Chiuve S., Mekary R.A., Jensen M.K., Flint A.J., Hu F.B., Rimm E.B.. Prospective Study of Breakfast Eating and Incident Coronary Heart Disease in a Cohort of Male US Health Professionals. Circulation (2013) 128 337-343. DOI: 10.1161/CIRCULATIONAHA.113.001474 5. Yokoyama Y., Onishi K., Hosoda T., Amano H., Otani S., Kurozawa Y., Tamakoshi A.. Skipping Breakfast and Risk of Mortality from Cancer, Circulatory Diseases and All Causes: Findings from the Japan Collaborative Cohort Study. Yonago Acta Med. (2016) 59 55-60. PMID: 27046951 6. Ferrer-Cascales R., Sánchez-SanSegundo M., Ruiz-Robledillo N., Albaladejo-Blázquez N., Laguna-Pérez A., Zaragoza-Martí A.. Eat or Skip Breakfast? The Important Role of Breakfast Quality for Health-Related Quality of Life, Stress and Depression in Spanish Adolescents. Int. J. Environ. Res. Public Health (2018) 15. DOI: 10.3390/ijerph15081781 7. A O’Sullivan T., Robinson M., E Kendall G., Miller M., Jacoby P., Silburn S.R., Oddy W.H.. A good-quality breakfast is associated with better mental health in adolescence. Public Health Nutr. (2009) 12 249-258. DOI: 10.1017/S1368980008003935 8. Iqbal K., Schwingshackl L., Gottschald M., Knüppel S., Stelmach-Mardas M., Aleksandrova K., Boeing H.. Breakfast quality and cardiometabolic risk profiles in an upper middle-aged German population. Eur. J. Clin. Nutr. (2017) 71 1312-1320. DOI: 10.1038/ejcn.2017.116 9. 9. WHO Cardiovascular DiseasesWHOGeneva, Switzerland2021. Cardiovascular Diseases (2021) 10. Vlachopoulos C., Xaplanteris P., Aboyans V., Brodmann M., Cífková R., Cosentino F., De Carlo M., Gallino A., Landmesser U., Laurent S.. The role of vascular biomarkers for primary and secondary prevention. A position paper from the European Society of Cardiology Working Group on peripheral circulation: Endorsed by the Association for Research into Arterial Structure and Physiology (ARTERY) Society. Atherosclerosis (2015) 241 507-532. PMID: 26117398 11. Vlachopoulos C., Aznaouridis K., O’Rourke M.F., Safar M.E., Baou K., Stefanadis C.. Prediction of cardiovascular events and all-cause mortality with central haemodynamics: A systematic review and meta-analysis. Eur. Heart J. (2010) 31 1865-1871. DOI: 10.1093/eurheartj/ehq024 12. Nambi V., Chambless L., He M., Folsom A.R., Mosley T., Boerwinkle E., Ballantyne C.M.. Common carotid artery intima-media thickness is as good as carotid intima-media thickness of all carotid artery segments in improving prediction of coronary heart disease risk in the Atherosclerosis Risk in Communities (ARIC) study. Eur. Heart J. (2011) 33 183-190. DOI: 10.1093/eurheartj/ehr192 13. Oikonomou E., Lazaros G., Christoforatou E., Chasikidis C., Vavouranaki G., Vogiatzi G., Papamikroulis G.A., Tsalamandris S., Gergiopoulos G., Mazaris S.. Breakfast association with arterial stiffness and carotid atherosclerotic burden. Insights from the ‘Corinthia’ study. Nutr. Metab. Cardiovasc. Dis. (2019) 29 744-750. DOI: 10.1016/j.numecd.2019.04.008 14. Uzhova I., Fuster V., Fernández-Ortiz A., Ordovás J.M., Sanz J., Fernández-Friera L., López-Melgar B., Mendiguren J.M., Ibáñez B., Bueno H.. The Importance of Breakfast in Atherosclerosis Disease: Insights From the PESA Study. J. Am. Coll. Cardiol. (2017) 70 1833-1842. DOI: 10.1016/j.jacc.2017.08.027 15. Catapano A.L., Graham I., De Backer G., Wiklund O., Chapman M.J., Drexel H., Hoes A.W., Jennings C.S., Landmesser U., Pedersen T.R.. 2016 ESC/EAS Guidelines for the Management of Dyslipidaemias. Eur. Heart J. (2016) 37 2999-3058. DOI: 10.1093/eurheartj/ehw272 16. Nasir K., Budoff M.J., Wong N.D., Scheuner M., Herrington D., Arnett D.K., Szklo M., Greenland P., Blumenthal R.S.. Family history of premature coronary heart disease and coronary artery calcification: Multi-Ethnic Study of Athero-sclerosis (MESA). Circulation (2007) 116 619-626. DOI: 10.1161/CIRCULATIONAHA.107.688739 17. Stamatelopoulos K., Karatzi K., Sidossis L.S.. Noninvasive methods for assessing early markers of atherosclerosis: The role of body composition and nutrition. Curr. Opin. Clin. Nutr. Metab. Care (2009) 12 467-473. DOI: 10.1097/MCO.0b013e32832f0d99 18. Roman M.J., Moeller E., Davis A., Paget S.A., Crow M.K., Lockshin M.D., Sammaritano L., Devereux R.B., Schwartz J., Levine D.M.. Preclinical Carotid Atherosclerosis in Patients with Rheumatoid Arthritis. Ann. Intern. Med. (2006) 144 249-256. DOI: 10.7326/0003-4819-144-4-200602210-00006 19. Preziosi P., Galan P., Deheeger M., Yacoub N., Drewnowski A., Hercberg S.. Breakfast Type, Daily Nutrient Intakes and Vitamin and Mineral Status of French Children, Adolescents and Adults. J. Am. Coll. Nutr. (1999) 18 171-178. DOI: 10.1080/07315724.1999.10718846 20. Li Z., Li H., Xu Q., Long Y.. Skipping Breakfast Is Associated with Hypertension in Adults: A Meta-Analysis. Int. J. Hypertens. (2022) 2022 7245223. DOI: 10.1155/2022/7245223 21. Rong S., Snetselaar L.G., Xu G., Sun Y., Liu B., Wallace R.B., Bao W.. Association of Skipping Breakfast With Cardiovascular and All-Cause Mortality. J. Am. Coll. Cardiol. (2019) 73 2025-2032. DOI: 10.1016/j.jacc.2019.01.065 22. Yofoglu L.K., Karachalias F., Georgakis M.K., Tountas C., Argyris A.A., Zhang Y., Papaioannou T.G., Blacher J., Weber T., Vlachopoulos C.. Association of Pressure Wave Reflections With Left Ventricular Mass: A Systematic Review and Meta-Analysis. Hypertension (2022) 80 e29-e42. DOI: 10.1161/HYPERTENSIONAHA.122.19980 23. Vlachopoulos C.V., Vyssoulis G.G., A Alexopoulos N., I Zervoudaki A., Pietri P.G., A Aznaouridis K., Stefanadis C.. Effect of chronic coffee consumption on aortic stiffness and wave reflections in hypertensive patients. Eur. J. Clin. Nutr. (2006) 61 796-802. DOI: 10.1038/sj.ejcn.1602577 24. Del Giorno R., Scanzio S., De Napoli E., Stefanelli K., Gabutti S., Troiani C., Gabutti L.. Habitual coffee and caffeinated beverages consumption is inversely associated with arterial stiffness and central and peripheral blood pressure. Int. J. Food Sci. Nutr. (2021) 73 106-115. DOI: 10.1080/09637486.2021.1926935 25. Tsirimiagkou C., Basdeki E.D., Korovesi A.A.K., Chairistanidou C., Ouamer D.S., Argyris A., Sfikakis P.P., Karatzi K., Protogerou A.D.. Habitual consumption of instant coffee is favorably associated with arterial stiffness but not with atheromatosis. Clin. Nutr. ESPEN (2021) 45 363-368. DOI: 10.1016/j.clnesp.2021.07.018 26. Ribeiro A.G., Mill J.G., Cade N.V., Velasquez-Melendez G., Matos S.M.A., Molina M.D.C.B.. Associations of Dairy Intake with Arterial Stiffness in Brazilian Adults: The Brazilian Longitudinal Study of Adult Health (ELSA-Brasil). Nutrients (2018) 10. DOI: 10.3390/nu10060701 27. Cortés-Valencia A., Monge A., Tamayo-Ortiz M., Lopez-Ridaura R., Rodriguez B.L., Cantú-Brito C., Catzin-Kuhlmann A., Lajous M.. Dairy consumption and subclinical atherosclerosis: A cross-sectional study among middle-aged Mexican women. Nutr. Metab. Cardiovasc. Dis. (2021) 31 1747-1755. DOI: 10.1016/j.numecd.2021.02.024 28. Li R.-L., Wang L.-Y., Liu S., Duan H.-X., Zhang Q., Peng W., Huang Y., Wu C.. Natural Flavonoids Derived From Fruits Are Potential Agents Against Atherosclerosis. Front. Nutr. (2022) 9 862277. DOI: 10.3389/fnut.2022.862277 29. Tzelefa V., Tsirimiagkou C., Argyris A., Moschonis G., Perogiannakis G., Yannakoulia M., Sfikakis P., Protogerou A.D., Karatzi K.. Associations of dietary patterns with blood pressure and markers of subclinical arterial damage in adults with risk factors for CVD. Public Health Nutr. (2021) 24 6075-6084. DOI: 10.1017/S1368980021003499
--- title: 'Decreased Vitamin D Levels and Altered Placental Vitamin D Gene Expression at High Altitude: Role of Genetic Ancestry' authors: - Eugenia Mata-Greenwood - Hans C. A. Westenburg - Stacy Zamudio - Nicholas P. Illsley - Lubo Zhang journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9967090 doi: 10.3390/ijms24043389 license: CC BY 4.0 --- # Decreased Vitamin D Levels and Altered Placental Vitamin D Gene Expression at High Altitude: Role of Genetic Ancestry ## Abstract High-altitude hypoxia challenges reproduction; particularly in non-native populations. Although high-altitude residence is associated with vitamin D deficiency, the homeostasis and metabolism of vitamin D in natives and migrants remain unknown. We report that high altitude (3600 m residence) negatively impacted vitamin D levels, with the high-altitude Andeans having the lowest 25-OH-D levels and the high-altitude Europeans having the lowest 1α,25-(OH)2-D levels. There was a significant interaction of genetic ancestry with altitude in the ratio of 1α,25-(OH)2-D to 25-OH-D; with the ratio being significantly lower in Europeans compared to Andeans living at high altitude. Placental gene expression accounted for as much as $50\%$ of circulating vitamin D levels, with CYP2R1 (25-hydroxylase), CYP27B1 (1α-hydroxylase), CYP24A1 (24-hydroxylase), and LRP2 (megalin) as the major determinants of vitamin D levels. High-altitude residents had a greater correlation between circulating vitamin D levels and placental gene expression than low-altitude residents. Placental 7-dehydrocholesterol reductase and vitamin D receptor were upregulated at high altitude in both genetic-ancestry groups, while megalin and 24-hydroxylase were upregulated only in Europeans. Given that vitamin D deficiency and decreased 1α,25-(OH)2-D to 25-OH-D ratios are associated with pregnancy complications, our data support a role for high-altitude-induced vitamin D dysregulation impacting reproductive outcomes, particularly in migrants. ## 1. Introduction Residents at high altitude are at higher risk of reproductive challenges including intrauterine growth restriction, stillbirth, preeclampsia (de novo maternal high blood pressure and organ system damage), and perinatal mortality and morbidity [1,2,3,4,5]. Multiple studies have shown multigenerational high-altitude residents, such as Tibetans and Andeans, compared to newcomer residents, such as Han and Hispanics, are less affected by high-altitude chronic hypoxia-induced effects in pregnancy [6,7]. Although physiological adaptations explain some of the reproductive differences between ethnic groups to high-altitude residence [6,8], the molecular and genetic mechanisms remain elusive [9]. Vitamin D has important roles in maternal-fetal tolerance and placental development [10,11]. Vitamin D deficiency has been associated with intrauterine growth restriction, recurrent miscarriages, gestational diabetes, and pre-eclampsia [12,13,14,15]. Furthermore, maternal vitamin D deficiency has been also linked to postnatal infectious diseases, autoimmune diseases (diabetes type I), asthma, and obesity in offspring [16,17,18]. On a global basis, vitamin D deficiency during pregnancy is highly prevalent [19,20,21]. Several developmental (age, pregnancy), lifestyle (diet, exercise, sun exposure), and environmental factors (latitude, altitude, climate, season) are known to affect vitamin D status [22,23]. Most human studies have observed an important correlation between sun exposure and vitamin D levels accounting for more than $50\%$ of changes in vitamin D status [22,23,24]. Ultraviolet B-light exposure is increased at higher altitude; however, several studies have shown significant vitamin D deficiency in humans, including pregnant women, living in high-altitude locations [25,26,27,28,29,30]. This renders Vitamin D deficiency a potentially important mediator of adverse pregnancy outcomes associated with high-altitude residence. Vitamin D metabolism is greatly altered during mammalian pregnancy [31,32,33]. In humans and other mammal species, the maternal plasma levels of calcium and the abundant precursor 25-(OH)-D remain at pre-pregnancy levels while the potent 1α,25-(OH)2-D metabolite levels increase 2–3-fold compared to pre-pregnancy levels as early as in the first trimester [31,33]. The placenta is known to express nearly all vitamin D-related genes and to actively participate in vitamin D homeostasis during pregnancy [32,33]. Vitamin D3 (cholecalciferol) can be endogenously produced from 7-dehydrocholesterol by sun exposure in the maternal skin epidermis or obtained from dietary sources [34,35]. 7-Dehydrocholesterol reductase (DHCR7) can sequester this metabolite to synthesize cholesterol. Vitamin D is endogenously activated by 25-hydroxylase (CYP2R1) producing 25-OH-D, followed by a second hydroxylation by 1α-hydroxylase (CYP27B1) that yields the most active metabolite of vitamin D: 1α,25-(OH)2-D [34,35]. In the blood, vitamin D metabolites are transported bound to albumin and the vitamin D binding protein (VDBP/GC). Megalin (LRP2) and cubulin (CUBN) bind to the VDBP/25-OH-D complex acting as efficient transporters. Inside the nucleus, 1α,25-(OH)2-D binds and activates the vitamin D receptor (VDR) that acts as a transcription factor leading to gene expression changes. A highly sensitive gene upregulated by VDR is the vitamin D-inactivating 24-hydroxylase (CYP24A1), thereby providing a unique negative feedback control mechanism to prevent vitamin D toxicity [34,36]. The aim of the present study is to identify the effect of altitude and genetic ancestry on maternal/fetal vitamin D levels and placental vitamin D metabolism. We hypothesize that high-altitude residence will be associated with vitamin D deficiency, based on prior reports [22,23], and that Andean-ancestry women would demonstrate improved vitamin D homeostasis compared to women of European ancestry. ## 2.1. Demographic and Clinical Characteristics Our cross-sectional mother/neonate dyad study included four groups ($$n = 10$$/group): Andean ancestry at 400 m altitude (An-Low), Andean ancestry at 3600 m altitude (An-High), European ancestry at 400 m altitude (Eu-Low) and European ancestry at 3600 m altitude (Eu-High). The Eu-Low group was significantly younger than the Eu-High and An-High groups, otherwise, the groups had similar maternal age, pre-pregnancy BMI, and neonatal sex distributions (Table 1). There were multiple significant differences in fetal/neonatal outcomes. The Eu-High group delivered at a significantly earlier gestational age than the An-High and An-Low (Table 1), and the Eu-Low group also delivered at a significantly earlier gestational age than the An-Low (Table 1). The Eu-High group had a significantly lower birthweight percentile than the Eu-Low group (31 ± 19.7 vs. 63.6 ± 27.3, Table 1) and significantly lower birthweight than the other three groups (Table 1), although none of the 40 newborns were clinically considered growth restricted. Placental efficiency was estimated indirectly by the birthweight-to-placental weight ratio (BW:PW), with lower ratios indicating lower efficiency, i.e., more placenta is required to support a given fetal size. The BW:PW ratio did not differ between Andean and European groups, but was significantly decreased at high altitude in both ancestry groups (Table 1). The Eu-High group also had a significantly lower ponderal index than the Eu-Low and the An-High group (Table 1). Finally, although there were no differences in head circumference, the Eu-High group of newborns had a significantly lower abdominal circumference than the An-High and An-Low group (Table 1). Two-Way ANOVA suggested that altitude decreased placental efficiency, birthweight, and birthweight percentile, while genetic ancestry had a significant effect on birthweight and abdominal circumference. Ponderal index, birthweight, and birthweight percentile showed a significant interaction between altitude and ancestry. ## 2.2. Vitamin D Status in Bolivian Pregnancies Near-term pregnancies at high altitude were characterized by significantly lower levels of vitamin D metabolites in mothers and neonates (Figure 1). The altitude-associated reduction in 25-OH-D was significant only in An-High mothers and their fetuses (Figure 1A,B). As a result, maternal and fetal/neonatal 25-OH-D plasma levels were significantly lower in the An-High group compared to the other three groups (Figure 1A,B). In contrast, maternal 1α,25-(OH)2-D levels were significantly lower in the Eu-High compared to the Eu-Low while the reduction in this metabolite between An-High and An-Low was not significant (Figure 1C). There were no differences in fetal/neonatal 1α,25-(OH)2-D levels (Figure 1D), or maternal VDBP levels (Figure 1G). Of note, vitamin D deficiency, defined as circulating levels of 25-OH-D less than 25 nM, was present in four mothers and five neonates of the An-High group, one mother and her neonate in the Eu-High group, one neonate in the An-low group and zero subjects of the Eu-low group. The effect of altitude across the four groups is better shown in the altered ratio of the two metabolites; 1α,25-(OH)2-D:25-OH-D. This ratio was similar at low altitude in the two ancestry groups, but it was significantly higher in An-High compared to Eu-High in both mothers (Figure 1E) and neonates (Figure 1F). Analysis of the effect of altitude and ancestry by two-way ANOVA revealed that only altitude had a significant effect on vitamin D levels ($p \leq 0.05$). However, it was ethnicity and its interaction with altitude that significantly influenced the ratio of the two metabolites ($p \leq 0.05$). Multivariable linear regression confirmed that altitude was the most significant determinant of maternal and fetal/neonatal 25-OH-D and 1α,25-(OH)2-D levels (Table 2). In addition, placental efficiency (BW:PW) was a second independent factor associated with fetal/neonatal 25-OH-D as well as maternal and fetal/neonatal 1α,25-(OH)2-D (Table 2). Linear regression analysis showed a significant correlation between maternal and fetal/neonatal blood levels of 25-OH-D in all four groups ($p \leq 0.001$, Supplementary Figure S1A–D). In contrast, the correlation between maternal and fetal/neonatal blood levels of 1α,25-(OH)2-D was only significant in the An-High group ($r = 0.8519$, $p \leq 0.001$, Supplementary Figure S1F). Furthermore, there was no significant correlation between 25-OH-D and 1α,25-(OH)2-D levels, except in maternal and fetal/neonatal samples of the An-High group (Supplementary Figure S2B,F) and fetal-neonatal samples of An-Low and Eu-Low (Supplementary Figure S2E,G). ## 2.3. Placental Expression of Vitamin D-Related Genes Altitude had a significant effect on the placental expression of vitamin D metabolic enzymes (Figure 2). The cholesterol-synthesizing enzyme DHCR7 mRNA levels were significantly higher in An-High and Eu-High than in their respective low-altitude groups (Figure 2A). There were no significant differences among the four groups in the expression of CYP2R1 and CYP27B1 that synthesize 25-OH-D and 1α,25-(OH)2-D, respectively (Figure 2B,E). The vitamin D metabolite transporter LRP2 mRNA was significantly higher in Eu-High than Eu-Low samples (Figure 2C), but there were no significant differences in CUBN transporter expression levels among the four groups (Figure 2D). The vitamin D inactivating enzyme CYP24A1 expression was significantly higher in the Eu-High than the An-High group (Figure 2F). Placental VDR mRNA levels were also significantly higher in the high-altitude groups than in the corresponding low-altitude groups (Figure 2G). Multivariable linear regression analyses did not yield a significant model of placental gene determinants of vitamin D levels for all 40 subjects. However, stratification by altitude uncovered significant placental vitamin D-relevant genes as determinants of vitamin D levels (Table 3). For low-altitude subjects, $26.5\%$ of the maternal 25-OH-D variation could be explained by placental CYP27B1 mRNA levels, while $22\%$ of the fetal/neonatal 25-OH-D variation could be explained by placental CYP2R1 mRNA levels (Table 3). Stronger correlations were found in the high-altitude group. Both maternal and fetal/neonatal 25-OH-D levels were significantly determined by placental CYP24A1 and CYP2R1 expression by $50.5\%$ (maternal) and $35\%$ (fetal/neonatal) (Table 3). Furthermore, $29.7\%$ of maternal 1α,25-(OH)2-D levels were determined by placental LRP2 and CYP24A1 expression while there was no significant model of placental gene expression as determinants of fetal/neonatal 1α,25-(OH)2-D levels (Table 3). The correlation between vitamin D levels and individual placental gene expression (Figure 3) illustrates the differential gene determinants of vitamin D levels according to altitude shown in Table 3. First, placental CYP2R1 and CYP27B1 were differentially regulated by altitude: at low altitude, there was a significant negative correlation with maternal 25-OH-D levels (Figure 3A,C) while at high altitude, there was a positive correlation, although it was only significant for CYP2R1 (Figure 3B) not CYP27B1 (Figure 3D). In addition, there was a significant positive correlation between maternal/fetal 25-OH-D and 1α,25-(OH)2-D levels with placental CYP24A1 expression in high-altitude samples but no correlation in low-altitude samples (shown for maternal 25-OH-D in Figure 3E,F). Finally, high-altitude, but not low-altitude, samples showed a significant positive correlation between maternal 1α,25-(OH)2-D levels and placental LRP2 mRNA levels (Figure 3G,H). ## 3. Discussion The present study has provided novel evidence of associations and potential mechanisms of high-altitude effects on maternal and fetal/neonatal vitamin D status and placental vitamin D metabolism. We confirmed our hypothesis that high-altitude residence is associated with decreased vitamin D levels in both mother and fetus with Andean-ancestry subjects having decreased 25-OH-D and European-ancestry subjects having decreased 1α,25-(OH)2-D. Our data support prior reports showing decreased 25-OH-D levels in indigenous South American populations living at high compared to low altitude [28,29,30]. Other studies report decreased 25-OH-D levels in Tibetans at high altitude and Turkish pregnant women at moderate altitude, but without comparisons of the same population at lower altitudes [25,26,27]. A study of European alpinists showed a significant decrease in 25-OH-D levels after a two-week climb [37]. It is thus consistent that both our multiple linear regression and two-way ANOVA identified high altitude as the most relevant factor affecting term pregnancy vitamin D levels. We identified placental efficiency as a secondary factor determining fetal 25-OH-D, and maternal/fetal 1α,25-(OH)2-D levels, which led us to evaluate placental gene expression of the vitamin D system. These data point to a pivotal role of the placenta in regulating the circulating levels of vitamin D metabolites in both the mother and fetus. Placental impacts on vitamin D status during pregnancy are reflected in the sudden drop in 1α,25-(OH)2-D levels shortly after delivery, and the correlation of placental gene expression with maternal/fetal vitamin D levels [31,32,33,38]. Placental efficiency was significantly decreased by high altitude in both genetic ancestry groups (Table 1). It is important to note that the original study revealed that for any given placental weight, Andean neonates at high altitude were >200 g heavier than their European counterparts, indicating a superior placental adaptation to high altitude in the Andean population [39]. Based on the beneficial effects of both 25-OH-D and 1α,25-(OH)2-D in various aspects of mammalian pregnancy [12,13], decreases in vitamin D status are likely an intermediary of the negative effects of high altitude in human pregnancy outcomes. Of interest, we found that placental gene expression had a stronger correlation with maternal/fetal vitamin D levels at high compared to low altitude. Previous studies have shown positive correlations between maternal 25-OH-D levels and placental LRP2, CUBN, CYP27B1, and CYP24A1 expression [38,40]. However, we did not find similar correlations in the low-altitude subjects, where, in fact, we observed negative correlations of maternal and fetal 25-OH-D levels with the expression of main metabolic genes CYP2R1 and CYP27B1. These are novel findings that we believe could be due to the higher levels of both vitamin D metabolites present in this study’s low-altitude residents, which lead to compensatory downregulation of the two main synthesizing enzymes. The Bolivian low-altitude residing mothers had significantly higher 25-OH-D by $23\%$ and 1α,25-(OH)2-D levels by $41\%$ than in our California cohort of healthy mothers [41]. Furthermore, other than CYP2R1 and CYP27B1, there were no other placental genes that correlated with vitamin D levels in low-altitude pregnancies. In contrast, mothers living at high altitude showed strong correlations between vitamin D metabolite levels and placental vitamin D-related gene expression, with the exception of CYP27B1 (in contrast to low altitude). Multivariable linear regression confirmed that, at high altitude, CYP24A1 and CYP2R1 were strong determinants of maternal and fetal 25-OH-D levels, while CYP24A1 and LRP2 were determinants of maternal 1α,25-(OH)2-D. We hypothesize that mothers exposed to high-altitude environmental stressors (hypoxia, cold) experience a more stringent regulation of vitamin D status by the placenta. The decreased vitamin D levels at high altitude can be partially explained by the changes in placental vitamin D-related gene expression. High-altitude residence was associated with the upregulation of placental DHCR7 and VDR mRNA levels in both Andean- and European-ancestry groups. Placental LRP2, while elevated by altitude in both ancestry groups, was significantly increased only for the Eu-High group. Increased placental DHCR7 expression could partially explain decreased 25-OH-D, since it sequesters the main precursor of vitamin D metabolites, 7-dehydrocholesterol, for cholesterol synthesis. Indeed, it has been reported that high altitude increases total serum cholesterol and other lipids [42,43]. LRP2 is an important transporter of nutrients and lipids, including vitamin D and cholesterol, which is essential for fetal growth [44,45]. Thereby, LRP2 upregulation at high altitude is a beneficial response; however, this is the first study to show upregulation of this gene in high-altitude placentas. In silico analyses of DHCR7 and LRP2 identified multiple (>6) hypoxia-inducible factor alpha (HIF1α) consensus DNA response sites (A/GCGTG) in their gene promoters. In contrast, promoters of CYP24A1, VDR, CUBN, and CYP2R1 had one or two HIF1α sites and CYP27B1 had none. Therefore, it is possible that high altitude upregulates DHCR7 and LRP2 expression via this transcription factor, although HIF1α chromatin immunoprecipitation studies would be required to confirm this. In support of this hypothesis, we have previously shown that the placentas of high-altitude pregnancies show increased levels of HIF1α, its downstream gene targets, and correlations with birth outcomes [46]. Another interesting finding was the upregulation of placental VDR at high altitude in both genetic ancestry groups. Multiple studies on placental vitamin D metabolism in pregnancy complications demonstrated reduced placental VDR expression, in conjunction with vitamin D deficiency, in preeclampsia, intrauterine growth restriction, and preterm birth [47]. Therefore, high-altitude-induced upregulation of placental VDR could be a beneficial compensatory effect against decreased vitamin D status. This finding distinguishes altitude-associated growth restriction from pathological IUGR mechanisms [47]. We also uncovered an interaction of genetic ancestry with high altitude in the regulation of the vitamin D system, since vitamin D levels and placental metabolism were similar between the two genetic ancestry groups at sea level but different at high altitude. This suggests differential genetic adaptations to high-altitude stressors such as hypoxia and cold. Large cross-sectional studies in China [48] and Ecuador [28] revealed that Tibetans and Ecuadorian indigenous ethnicities showed the lowest 25-OH-D levels among all ethnicities studied. However, the interaction of altitude with ethnicity and alterations of vitamin D metabolites like 1α,25-(OH)2-D were not addressed in either study. We found vitamin D metabolism differences according to genetic ancestry in the ratio of 1α,25-(OH)2-D to 25-OH-D, in correlations between vitamin D metabolites, and in placental vitamin D-related gene expression. The 1α,25-(OH)2-D to 25-OH-D ratio was lower in both mother and fetus of European ancestry compared to Andean ancestry at high altitude. Recent studies have shown a stronger association between a low ratio of 1α,25-(OH)2-D:25-OH-D than just low 25-OH-D levels with cardiometabolic disease [49,50] and pregnancy disorders such as preeclampsia, intrauterine growth restriction, gestational diabetes, and preterm birth [41,51]. Our linear regression assays demonstrated that Andeans at high altitude had the highest correlation between 25-OH-D and 1α,25-(OH)2-D levels (both maternal and fetal) followed by the Andean-ancestry and European-ancestry groups (fetal only) residing at low altitude. In contrast, the European-ancestry group at high altitude showed the least correlation between the two metabolites suggesting higher variability and disturbed homeostasis in vitamin D metabolism within this group. Altogether, European-ancestry pregnancies are associated with adaptive changes in vitamin D metabolism more often associated with pregnancy complications and cardiovascular disease. The differences in vitamin D metabolism between genetic ancestry groups at high altitude can be partly explained by the differential placental regulation of CYP24A1. This mitochondrial enzyme inactivates both 25-OH-D and 1α,25-(OH)2-D and was upregulated in placentas of European ancestry, but not those of Andean ancestry, at high altitude. Upregulation of CYP24A1 has been observed in diseases where hypoxia is a key pathogenic component (e.g., cancer). Several studies have identified placental CYP24A1 upregulation in preeclampsia, gestational diabetes, and fetal growth restriction in women residing at low altitude [52,53,54]. In healthy mammalian pregnancy, the increase in maternal 1α,25-(OH)2-D becomes uncoupled from VDR-mediated upregulation of CYP24A1 in maternal renal and placental tissues, thereby preventing the inactivation of vitamin D metabolites [55,56,57]. Potential mechanisms of the differential effects of high altitude in Andeans and Europeans on the placental VDR/CYP24A1 axis include crosstalk with pregnancy-specific hormones such as estrogen. Indeed, a recent study showed that, although high altitude increased estrogen and cortisol levels in near-term pregnant women, Andean women had higher estrogen and lower cortisol levels than European women [58]. Estrogen has been shown to correlate positively with vitamin D levels in human cohorts [59,60] and to upregulate VDR in neuronal tissues in vivo [61]. Therefore, future studies should identify the role of estrogen metabolites and other pregnancy hormones on differential vitamin D metabolism at high altitude according to genetic ancestry. Altogether, we have identified differential vitamin D metabolism in Andeans compared to European Bolivians in response to high-altitude stress, where the placenta plays a pivotal role. Because multiple pregnancy disorders with decreased vitamin D showed placental CYP24A1 upregulation, future studies on placental CYP24A1 regulation by hypoxia/high altitude are warranted. Previous metabolic studies have led us to hypothesize that high-altitude hypoxia alters placental utilization of oxygen in the mitochondria [62]. Since many vitamin D metabolizing enzymes are localized in the mitochondria, placental programming of mitochondrial activity/function could be partly responsible for the alterations in vitamin D related CYP450 expression, subcellular localization, and function. Strengths of this study include the analysis of the vitamin D system in maternal/fetal dyads of different genetic-ancestry and altitude residences. Another important strength was the evaluation of vitamin D levels in correlation with placental mRNA expression of vitamin D-relevant genes that allowed us to uncover novel biological determinants of vitamin D homeostasis at high altitude, and that differed by genetic ancestry. The present study also has important limitations, for instance, skin color, dietary habits, and vitamin D supplementation, which usually have a strong effect on vitamin D status, were not measured. A second important limitation is that the small sample size prevented analysis of known gene polymorphisms in VDR, CYP2R1, CYP24A1, and other vitamin D-relevant genes in association with vitamin D levels. Multiple studies have revealed an association between these polymorphisms and vitamin D levels and disease [63,64,65]. Therefore, larger studies, including genetic polymorphisms are needed to validate the current data. A third limitation is the evaluation of vitamin D metabolites by ELISA instead of using the gold standard LC/MS. We have previously validated our commercial ELISAs using LC/MS in a subset of our samples and found that 1α,25-(OH)2-D levels are overestimated by ELISA as compared to LC/MS. However, the differences between groups remained the same [55,57]. Finally, all of our subjects were healthy-term pregnant women and their singleton neonates. It would be of great interest to investigate the effect of altitude on vitamin D homeostasis and metabolism in pregnancies complicated with fetal growth restriction or other pregnancy disorders. ## 4.1. Study Design This study involved 40 maternal/fetal dyads that were part of a larger, cross-sectional study evaluating the effects of altitude and genetic ancestry on oxygen delivery/consumption, uterine/umbilical blood flows, placental efficiency, and pregnancy outcomes [39]. Genetic ancestry was determined with 133 single nucleotide polymorphisms as previously described [39]. This study was approved by the Bolivian National Bioethics Committee, and local and the US Institutional Review boards, and all the participants gave written and informed consent. All study participants were screened for chronic health conditions and general health 0–10 days prior to scheduled Cesarean delivery [37]. Women were excluded for drug, alcohol, or tobacco use, complications of pregnancy, a positive oral GTT test, pre-term labor, or premature rupture of the membranes. The European group was composed of women of sea level ancestry with no known Andean ancestors and living at ≤400 m for three generations, similar women who migrated to 3600 m, and whose Ancestry-Informative Markers (AIMS, [37]) indicated predominantly European ancestry. The Andean group was comprised of a similar group. Women at high altitude, with three generations of ancestors born and raised at high altitude who either remained in their native environment or migrated to ≤400 m as children or adults, and whose ancestry was predominantly Native American. Maternal blood was collected from the antecubital vein after consent was obtained, during the screening visit 4 ± 1 day prior to delivery (scheduled Cesarean). Umbilical cord blood (fetal/neonatal) was collected from the doubly-clamped umbilical cord vein immediately after delivery. Blood samples were placed in heparinized tubes and processed to collect plasma aliquots that were frozen at −70 °C until further analysis. Placental samples were also obtained from four locations (one from each of the quadrants bisected by the cord insertion), snap-frozen in liquid nitrogen, and stored at −70 °C until further analysis. ## 4.2. Vitamin D Status and Vitamin D Binding Protein ELISA Assays Plasma levels of 25-OH-D and 1α,25-(OH)2-D were analyzed using commercially available EIA kits that determine total (bound and free) levels of vitamin D metabolites (Immunodiagnostic Systems, Ltd., Scottsdale, Arizona). The sensitivity of the 25-OH-D ELISA kit is 5 nM, the intra- and inter-assay variability is ≤$8.7\%$ coefficient of variation (CV), and the specificity is $100\%$ for 25-OH-D3, $75\%$ for 25-OH-D2, $100\%$ for 24,25-(OH)2-D, and less than $0.3\%$ for the remaining vitamin D metabolites. The sensitivity of the 1α,25-(OH)2-D ELISA kit is 6 pM with a specificity of $100\%$ for 1α,25-(OH)2-D3, $39\%$ for 1α,25-(OH)2-D2, less than $0.05\%$ for the remaining vitamin D metabolites, and the intra- and inter-assay variability is estimated at ≤$20\%$ CV. Vitamin D-binding protein (VDBP) was determined with a commercial ELISA kit (Aviva Systems Biology LLC, San Diego, CA, USA) using the manufacturer’s instructions. ## 4.3. Placental Vitamin D-Related Gene Expression Analysis Placental biopsies were ground to obtain a homogenous representative sample. RNA isolation, RT reaction, and real-time PCR were performed as previously described [41]. Samples were analyzed on the CFX Connect™ system (Bio-Rad, Hercules, CA, USA) using the QuantiTect® Probe PCR kit (Qiagen, Hilden, Germany), and Taqman primer sets (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. Negative controls were included in each run. Gene expression was determined for DHCR7 (Hs01023087_m1), CYP2R1 (Hs01379776_m1), LRP2 (Hs00189742_m1), CUBN (Hs00153607_m1), GC (Hs00167096_m1), CYP27B1 (Hs01096154_m1), CYP24A1 (Hs00167999_m1), and VDR (Hs01045840_m1). The housekeeping gene β-actin (ACTB, Hs01060665_g1) was used to normalize target mRNA levels. Quantitative analysis was performed with the aid of standard curves as previously described [41]. ## 4.4. Statistical Analysis Quantitative data are shown as mean and standard error. One-way ANOVA with LSD posthoc analysis was used to compare the four groups on quantitative characteristics of mother and baby, vitamin D status, and mRNA/protein expression variables. Non-parametric data were analyzed with Kruskall–Wallis One-way ANOVA. Two-way ANOVA was used to determine the significance of ethnicity, altitude, and the interaction between them. Chi-square was used for comparisons of categorical characteristics (i.e., fetal sex). Simple linear regression models were developed to examine the relationships between maternal and umbilical cord blood vitamin D levels, and between stable precursor (25-OH-D) and most active (1α,25-(OH)2-D) vitamin D metabolites for each of the four groups. To determine potential predictors of vitamin D levels, multivariable linear regression analysis was performed, and the model with the highest adjusted r2 while showing significance ($p \leq 0.05$) was chosen. IBM SPSS statistics version 26 was used. ## References 1. Herrera E.A., Krause B., Ebensperger G., Reyes R.V., Casanello P., Parra-Cordero M., Llanos A.J.. The Placental Pursuit for an Adequate Oxidant Balance between the Mother and the Fetus. Front. Pharmacol. (2014) 5 149. DOI: 10.3389/fphar.2014.00149 2. Gonzalez-Candia A., Herrera E.A.. High Altitude Pregnancies and Vascular Dysfunction: Observations from Latin American Studies. Front. Physiol. (2021) 12 786038. DOI: 10.3389/fphys.2021.786038 3. Yang L., Helbich-Poschacher V., Cao C., Klebermass-Schrehof K., Waldhoer T.. Maternal Altitude and Risk of Low Birthweight: A Systematic Review and Meta-Analyses. Placenta (2020) 101 124-131. DOI: 10.1016/j.placenta.2020.09.010 4. Julian C.G.. High Altitude during Pregnancy. Clin. Chest Med. (2011) 32 21-31. DOI: 10.1016/j.ccm.2010.10.008 5. van Patot M.C.T., Ebensperger G., Gassmann M., Llanos A.J.. The Hypoxic Placenta. High Alt. Med. Biol. (2012) 13 176-184. DOI: 10.1089/ham.2012.1046 6. Moore L.G., Niermeyer S., Zamudio S.. Human Adaptation to High Altitude: Regional and Life-Cycle Perspectives. Am. J. Phys. Anthropol. (1998) 107 25-64. DOI: 10.1002/(SICI)1096-8644(1998)107:27+<25::AID-AJPA3>3.0.CO;2-L 7. Moore L.G., Shriver M., Bemis L., Hickler B., Wilson M., Brutsaert T., Parra E., Vargas E.. Maternal Adaptation to High-Altitude Pregnancy: An Experiment of Nature—A Review. Placenta (2004) 25 S60-S71. DOI: 10.1016/j.placenta.2004.01.008 8. Postigo L., Heredia G., Illsley N.P., Torricos T., Dolan C., Echalar L., Tellez W., Maldonado I., Brimacombe M., Balanza E.. Where the O. J. Physiol. (2009) 587 693-708. DOI: 10.1113/jphysiol.2008.163634 9. Storz J.F.. High-Altitude Adaptation: Mechanistic Insights from Integrated Genomics and Physiology. Mol. Biol. Evol. (2021) 38 2677-2691. DOI: 10.1093/molbev/msab064 10. Arshad R., Sameen A., Murtaza M.A., Sharif H.R., Iahtisham-Ul-Haq D.S., Ahmed Z., Nemat A., Manzoor M.F.. Impact of Vitamin D on Maternal and Fetal Health: A Review. Food Sci. Nutr. (2022) 10 3230-3240. DOI: 10.1002/fsn3.2948 11. Ganguly A., Tamblyn J.A., Finn-Sell S., Chan S.-Y., Westwood M., Gupta J., Kilby M.D., Gross S.R., Hewison M.. Vitamin D, the Placenta and Early Pregnancy: Effects on Trophoblast Function. J. Endocrinol. (2018) 236 R93-R103. DOI: 10.1530/JOE-17-0491 12. Wei S.-Q., Qi H.-P., Luo Z.-C., Fraser W.D.. Maternal Vitamin D Status and Adverse Pregnancy Outcomes: A Systematic Review and Meta-Analysis. J. Matern.-Fetal Neonatal Med. (2013) 26 889-899. DOI: 10.3109/14767058.2013.765849 13. Weinert L.S., Silveiro S.P.. Maternal-Fetal Impact of Vitamin D Deficiency: A Critical Review. Matern. Child Health J. (2015) 19 94-101. DOI: 10.1007/s10995-014-1499-7 14. van der Pligt P., Willcox J., Szymlek-Gay E., Murray E., Worsley A., Daly R.. Associations of Maternal Vitamin D Deficiency with Pregnancy and Neonatal Complications in Developing Countries: A Systematic Review. Nutrients (2018) 10. DOI: 10.3390/nu10050640 15. Mulligan M.L., Felton S.K., Riek A.E., Bernal-Mizrachi C.. Implications of Vitamin D Deficiency in Pregnancy and Lactation. Am. J. Obstet. Gynecol. (2010) 202 429.e1-429.e9. DOI: 10.1016/j.ajog.2009.09.002 16. Lucas R.M., Ponsonby A.-L., Pasco J.A., Morley R.. Future Health Implications of Prenatal and Early-Life Vitamin D Status. Nutr. Rev. (2008) 66 710-720. DOI: 10.1111/j.1753-4887.2008.00126.x 17. Erkkola M., Nwaru B.I., Viljakainen H.T.. Maternal Vitamin D during Pregnancy and Its Relation to Immune-Mediated Diseases in the Offspring. Vitam. Horm. (2011) 86 239-260. DOI: 10.1016/B978-0-12-386960-9.00010-1 18. Weiss S.T., Litonjua A.A.. The In Utero Effects of Maternal Vitamin D Deficiency: How It Results in Asthma and Other Chronic Diseases. Am. J. Respir. Crit. Care Med. (2011) 183 1286-1287. DOI: 10.1164/rccm.201101-0160ED 19. Rosen C.J., Abrams S.A., Aloia J.F., Brannon P.M., Clinton S.K., Durazo-Arvizu R.A., Gallagher J.C., Gallo R.L., Jones G., Kovacs C.S.. IOM Committee Members Respond to Endocrine Society Vitamin D Guideline. J. Clin. Endocrinol. Metab. (2012) 97 1146-1152. DOI: 10.1210/jc.2011-2218 20. Li H., Ma J., Huang R., Wen Y., Liu G., Xuan M., Yang L., Yang J., Song L.. Prevalence of Vitamin D Deficiency in the Pregnant Women: An Observational Study in Shanghai, China. Arch. Public Health (2020) 78 31. DOI: 10.1186/s13690-020-00414-1 21. Vandevijvere S., Amsalkhir S., Van Oyen H., Moreno-Reyes R.. High Prevalence of Vitamin D Deficiency in Pregnant Women: A National Cross-Sectional Survey. PLoS ONE (2012) 7. DOI: 10.1371/journal.pone.0043868 22. Wang T.J., Zhang F., Richards J.B., Kestenbaum B., Van Meurs J.B., Berry D., Kiel D.P., Streeten E.A., Ohlsson C., Koller D.L.. Common Genetic Determinants of Vitamin D Insufficiency: A Genome-Wide Association Study. Lancet (2010) 376 180-188. DOI: 10.1016/S0140-6736(10)60588-0 23. Neville J.J., Palmieri T., Young A.R.. Physical Determinants of Vitamin D Photosynthesis: A Review. JBMR Plus (2021) 5 e10460. DOI: 10.1002/jbm4.10460 24. Snellman G., Melhus H., Gedeborg R., Olofsson S., Wolk A., Pedersen N.L., Michaëlsson K.. Seasonal Genetic Influence on Serum 25-Hydroxyvitamin D Levels: A Twin Study. PLoS ONE (2009) 4. DOI: 10.1371/journal.pone.0007747 25. Alp H., Tekgündüz K.Ş., Akkar M.K.. Maternal and Cord Blood Vitamin D Status in High-Altitude Pregnancy. J. Matern.-Fetal Neonatal Med. (2016) 29 571-575. DOI: 10.3109/14767058.2015.1011119 26. Gelsor N., Ma L., Duan J., Wangmu T., Gelsor N.. Human Vitamin D Deficiency in Tibet. Food Nutr. Sci. (2017) 08 1127-1136. DOI: 10.4236/fns.2017.812084 27. Norsang G., Ma L., Dahlback A., Zhuoma C., Tsoja W., Porojnicu A., Lagunova Z., Moan J.. The Vitamin D Status among Tibetans. Photochem. Photobiol. (2009) 85 1028-1031. DOI: 10.1111/j.1751-1097.2009.00552.x 28. Orces C.H.. Vitamin D Status among Older Adults Residing in the Littoral and Andes Mountains in Ecuador. Sci. World J. (2015) 2015 545297. DOI: 10.1155/2015/545297 29. Hirschler V., Molinari C., Maccallini G., Intersimone P., Gonzalez C.D.. Vitamin D Levels and Cardiometabolic Markers in Indigenous Argentinean Children Living at Different Altitudes. Glob. Pediatr. Health (2019) 6 1882194. DOI: 10.1177/2333794X18821942 30. Rottenberg M.E., Nylén S., Rodriguez C., Terán G., Cuna W., Brañez F., Persson K.E.M.. Differences in Nutritional and Health Status in School Children from the Highlands and Lowlands of Bolivia. Am. J. Trop. Med. Hyg. (2018) 98 326-333. DOI: 10.4269/ajtmh.17-0143 31. Gray T.K., Lowe W., Lester G.E.. Vitamin D and Pregnancy: The Maternal-Fetal Metabolism of Vitamin D. Endocr. Rev. (1981) 2 264-274. DOI: 10.1210/edrv-2-3-264 32. Paulson S.K., Deluca H.F.. Review Article: Vitamin D Metabolism during Pregnancy. Bone (1986) 7 331-336. DOI: 10.1016/8756-3282(86)90252-8 33. Karras S.N., Wagner C.L., Castracane V.D.. Understanding Vitamin D Metabolism in Pregnancy: From Physiology to Pathophysiology and Clinical Outcomes. Metabolism (2018) 86 112-123. DOI: 10.1016/j.metabol.2017.10.001 34. Dusso A.S., Brown A.J., Slatopolsky E.. Vitamin D. Am. J. Physiol. Renal Physiol. (2005) 289 F8-F28. DOI: 10.1152/ajprenal.00336.2004 35. Norman A.W.. From Vitamin D to Hormone D: Fundamentals of the Vitamin D Endocrine System Essential for Good Health. Am. J. Clin. Nutr. (2008) 88 491S-499S. DOI: 10.1093/ajcn/88.2.491S 36. Haussler M.R., Haussler C.A., Bartik L., Whitfield G.K., Hsieh J.-C., Slater S., Jurutka P.W.. Vitamin D Receptor: Molecular Signaling and Actions of Nutritional Ligands in Disease Prevention. Nutr. Rev. (2008) 66 S98-S112. DOI: 10.1111/j.1753-4887.2008.00093.x 37. Kasprzak Z., Śliwicka E., Hennig K., Pilaczyńska-Szcześniak Ł., Huta-Osiecka A., Nowak A.. Vitamin D, Iron Metabolism, and Diet in Alpinists During a 2-Week High-Altitude Climb. High Alt. Med. Biol. (2015) 16 230-235. DOI: 10.1089/ham.2015.0008 38. Ashley B., Simner C., Manousopoulou A., Jenkinson C., Hey F., Frost J.M., Rezwan F.I., White C.H., Lofthouse E.M., Hyde E.. Placental Uptake and Metabolism of 25(OH)Vitamin D Determine Its Activity within the Fetoplacental Unit. eLife (2022) 11 e71094. DOI: 10.7554/eLife.71094 39. Zamudio S., Postigo L., Illsley N.P., Rodriguez C., Heredia G., Brimacombe M., Echalar L., Torricos T., Tellez W., Maldonado I.. Maternal Oxygen Delivery Is Not Related to Altitude- and Ancestry-Associated Differences in Human Fetal Growth: Maternal O. J. Physiol. (2007) 582 883-895. DOI: 10.1113/jphysiol.2007.130708 40. Park H., Wood M.R., Malysheva O.V., Jones S., Mehta S., Brannon P.M., Caudill M.A.. Placental Vitamin D Metabolism and Its Associations with Circulating Vitamin D Metabolites in Pregnant Women. Am. J. Clin. Nutr. (2017) 106 1439-1448. DOI: 10.3945/ajcn.117.153429 41. Martin C.B., Oshiro B.T., Sands L.D., Kabir S., Thorpe D., Clark T.C., Yao R., Mata-Greenwood E.. Vitamin-D Dysregulation in Early- and Late-Onset Preeclampsia: A Gestational-Age Matched Study. J. Steroid Biochem. Mol. Biol. (2020) 203 105729. DOI: 10.1016/j.jsbmb.2020.105729 42. Temte J.L.. Elevation of Serum Cholesterol at High Altitude and Its Relationship to Hematocrit. Wilderness Environ. Med. (1996) 7 216-224. DOI: 10.1580/1080-6032(1996)007[0216:EOSCAH]2.3.CO;2 43. Gonzales G.F., Tapia V.. Association of High Altitude-Induced Hypoxemia to Lipid Profile and Glycemia in Men and Women Living at 4100m in the Peruvian Central Andes. Endocrinol. Nutr. (2013) 60 79-86. DOI: 10.1016/j.endonu.2012.06.002 44. Baardman M.E., Kerstjens-Frederikse W.S., Berger R.M.F., Bakker M.K., Hofstra R.M.W., Plösch T.. The Role of Maternal-Fetal Cholesterol Transport in Early Fetal Life: Current Insights1. Biol. Reprod. (2013) 88 24. DOI: 10.1095/biolreprod.112.102442 45. Storm T., Christensen E.I., Christensen J.N., Kjaergaard T., Uldbjerg N., Larsen A., Honoré B., Madsen M.. Megalin Is Predominantly Observed in Vesicular Structures in First and Third Trimester Cytotrophoblasts of the Human Placenta. J. Histochem. Cytochem. (2016) 64 769-784. DOI: 10.1369/0022155416672210 46. Zamudio S., Wu Y., Ietta F., Rolfo A., Cross A., Wheeler T., Post M., Illsley N.P., Caniggia I.. Human Placental Hypoxia-Inducible Factor-1α Expression Correlates with Clinical Outcomes in Chronic Hypoxia in Vivo. Am. J. Pathol. (2007) 170 2171-2179. DOI: 10.2353/ajpath.2007.061185 47. Knabl J., Vattai A., Ye Y., Jueckstock J., Hutter S., Kainer F., Mahner S., Jeschke U.. Role of Placental VDR Expression and Function in Common Late Pregnancy Disorders. Int. J. Mol. Sci. (2017) 18. DOI: 10.3390/ijms18112340 48. Li L., Li K., Li J., Luo Y., Cheng Y., Jian M., Xie C., Ji C., Chuan L., Wang Z.. Ethnic, Geographic, and Seasonal Differences of Vitamin D Status among Adults in South-west China. J. Clin. Lab. Anal. (2020) 34 e23532. DOI: 10.1002/jcla.23532 49. Ahmed L.H.M., Butler A.E., Dargham S.R., Latif A., Chidiac O.M., Atkin S.L., Abi Khalil C.. Vitamin D3 Metabolite Ratio as an Indicator of Vitamin D Status and Its Association with Diabetes Complications. BMC Endocr. Disord. (2020) 20. DOI: 10.1186/s12902-020-00641-1 50. Castillo-Peinado L.d.L.S., Calderón-Santiago M., Herrera-Martínez A.D., León-Idougourram S., Gálvez-Moreno M.Á., Sánchez-Cano R.L., Bouillon R., Quesada-Gómez J.M., Priego-Capote F.. Measuring Vitamin D3 Metabolic Status, Comparison between Vitamin D Deficient and Sufficient Individuals. Separations (2022) 9. DOI: 10.3390/separations9060141 51. Albahlol I.A., Almaeen A.H., Alduraywish A.A., Dar U.F., El-Metwally T.H.. Vitamin D Status and Pregnancy Complications: Serum 1,25-Di-Hydroxyl-Vitamin D and Its Ratio to 25-Hydroxy-Vitamin D Are Superior Biomarkers than 25-Hydroxy-Vitamin D. Int. J. Med. Sci. (2020) 17 3039-3048. DOI: 10.7150/ijms.47807 52. Ma R., Gu Y., Zhao S., Sun J., Groome L.J., Wang Y.. Expressions of Vitamin D Metabolic Components VDBP, CYP2R1, CYP27B1, CYP24A1, and VDR in Placentas from Normal and Preeclamptic Pregnancies. Am. J. Physiol.-Endocrinol. Metab. (2012) 303 E928-E935. DOI: 10.1152/ajpendo.00279.2012 53. Wang Y., Wang T., Huo Y., Liu L., Liu S., Yin X., Wang R., Gao X.. Placenta Expression of Vitamin D and Related Genes in Pregnant Women with Gestational Diabetes Mellitus. J. Steroid Biochem. Mol. Biol. (2020) 204 105754. DOI: 10.1016/j.jsbmb.2020.105754 54. O’Brien K.O., Li S., Cao C., Kent T., Young B.V., Queenan R.A., Pressman E.K., Cooper E.M.. Placental CYP27B1 and CYP24A1 Expression in Human Placental Tissue and Their Association with Maternal and Neonatal Calcitropic Hormones. J. Clin. Endocrinol. Metab. (2014) 99 1348-1356. DOI: 10.1210/jc.2013-1366 55. Goyal R., Zhang L., Blood A.B., Baylink D.J., Longo L.D., Oshiro B., Mata-Greenwood E.. Characterization of an Animal Model of Pregnancy-Induced Vitamin D Deficiency Due to Metabolic Gene Dysregulation. Am. J. Physiol. Endocrinol. Metab. (2014) 306 E256-E266. DOI: 10.1152/ajpendo.00528.2013 56. Goyal R., Billings T.L., Mansour T., Martin C., Baylink D.J., Longo L.D., Pearce W.J., Mata-Greenwood E.. Vitamin D Status and Metabolism in an Ovine Pregnancy Model: Effect of Long-Term, High-Altitude Hypoxia. Am. J. Physiol.-Endocrinol. Metab. (2016) 310 E1062-E1071. DOI: 10.1152/ajpendo.00494.2015 57. Mata-Greenwood E., Huber H.F., Li C., Nathanielsz P.W.. Role of Pregnancy and Obesity on Vitamin D Status, Transport, and Metabolism in Baboons. Am. J. Physiol.-Endocrinol. Metab. (2019) 316 E63-E72. DOI: 10.1152/ajpendo.00208.2018 58. Charles S.M., Julian C.G., Vargas E., Moore L.G.. Higher Estrogen Levels During Pregnancy in Andean Than European Residents of High Altitude Suggest Differences in Aromatase Activity. J. Clin. Endocrinol. Metab. (2014) 99 2908-2916. DOI: 10.1210/jc.2013-4102 59. Harmon Q.E., Umbach D.M., Baird D.D.. Use of Estrogen-Containing Contraception Is Associated with Increased Concentrations of 25-Hydroxy Vitamin D. J. Clin. Endocrinol. Metab. (2016) 101 3370-3377. DOI: 10.1210/jc.2016-1658 60. Zhao D., Ouyang P., de Boer I.H., Lutsey P.L., Farag Y.M.K., Guallar E., Siscovick D.S., Post W.S., Kalyani R.R., Billups K.L.. Serum Vitamin D and Sex Hormones Levels in Men and Women: The Multi-Ethnic Study of Atherosclerosis (MESA). Maturitas (2017) 96 95-102. DOI: 10.1016/j.maturitas.2016.11.017 61. Nashold F.E., Spach K.M., Spanier J.A., Hayes C.E.. Estrogen Controls Vitamin D3-Mediated Resistance to Experimental Autoimmune Encephalomyelitis by Controlling Vitamin D3 Metabolism and Receptor Expression. J. Immunol. (2009) 183 3672-3681. DOI: 10.4049/jimmunol.0901351 62. Illsley N.P., Caniggia I., Zamudio S.. Placental Metabolic Reprogramming: Do Changes in the Mix of Energy-Generating Substrates Modulate Fetal Growth?. Int. J. Dev. Biol. (2010) 54 409-419. DOI: 10.1387/ijdb.082798ni 63. Krasniqi E., Boshnjaku A., Wagner K.-H., Wessner B.. Association between Polymorphisms in Vitamin D Pathway-Related Genes, Vitamin D Status, Muscle Mass and Function: A Systematic Review. Nutrients (2021) 13. DOI: 10.3390/nu13093109 64. Ruiz-Ballesteros A.I., Meza-Meza M.R., Vizmanos-Lamotte B., Parra-Rojas I., de la Cruz-Mosso U.. Association of Vitamin D Metabolism Gene Polymorphisms with Autoimmunity: Evidence in Population Genetic Studies. Int. J. Mol. Sci. (2020) 21. DOI: 10.3390/ijms21249626 65. Galvão A.A., de Araújo Sena F., de Andrade Belitardo E.M.M., de Santana M.B.R., de Oliveira Costa G.N., Cruz Á.A., Barreto M.L., dos Santos Costa R., Alcantara-Neves N.M., Figueiredo C.A.. Genetic Polymorphisms in Vitamin D Pathway Influence 25(OH)D Levels and Are Associated with Atopy and Asthma. Allergy Asthma Clin. Immunol. (2020) 16 62. DOI: 10.1186/s13223-020-00460-y
--- title: Phospholipase D1 Attenuation Therapeutics Promotes Resilience against Synaptotoxicity in 12-Month-Old 3xTg-AD Mouse Model of Progressive Neurodegeneration authors: - Chandramouli Natarajan - Charles Cook - Karthik Ramaswamy - Balaji Krishnan journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9967100 doi: 10.3390/ijms24043372 license: CC BY 4.0 --- # Phospholipase D1 Attenuation Therapeutics Promotes Resilience against Synaptotoxicity in 12-Month-Old 3xTg-AD Mouse Model of Progressive Neurodegeneration ## Abstract Abrogating synaptotoxicity in age-related neurodegenerative disorders is an extremely promising area of research with significant neurotherapeutic implications in tauopathies including Alzheimer’s disease (AD). Our studies using human clinical samples and mouse models demonstrated that aberrantly elevated phospholipase D1 (PLD1) is associated with amyloid beta (Aβ) and tau-driven synaptic dysfunction and underlying memory deficits. While knocking out the lipolytic PLD1 gene is not detrimental to survival across species, elevated expression is implicated in cancer, cardiovascular conditions and neuropathologies, leading to the successful development of well-tolerated mammalian PLD isoform-specific small molecule inhibitors. Here, we address the importance of PLD1 attenuation, achieved using repeated 1 mg/kg of VU0155069 (VU01) intraperitoneally every alternate day for a month in 3xTg-AD mice beginning only from ~11 months of age (with greater influence of tau-driven insults) compared to age-matched vehicle ($0.9\%$ saline)-injected siblings. A multimodal approach involving behavior, electrophysiology and biochemistry corroborate the impact of this pre-clinical therapeutic intervention. VU01 proved efficacious in preventing in later stage AD-like cognitive decline affecting perirhinal cortex-, hippocampal- and amygdala-dependent behaviors. Glutamate-dependent HFS-LTP and LFS-LTD improved. Dendritic spine morphology showed the preservation of mushroom and filamentous spine characteristics. Differential PLD1 immunofluorescence and co-localization with Aβ were noted. ## 1. Introduction Alzheimer’s disease (AD) is the most common cause of dementia and the sixth leading cause of death that cannot be prevented, cured or even slowed [1]. In addition, of the 55 million AD patients worldwide, two thirds are women [2,3,4,5,6]. Therefore, there is not only an urgent need to find a cure, but also an emerging need to investigate therapeutics in a gender-specific manner. Recent therapeutic approaches target the molecular mechanisms causing synaptic dysfunction (an early event leading to memory deficits) driven by low molecular weight aggregates of Aβ (AβO) and tau (TauO) [7,8,9,10,11,12,13,14,15]. Robust evidence from earlier studies [16,17,18] reported elevated phospholipase activity in post-mortem AD brains. Clinical trials using phospholipase substrates, phosphatidyl choline (PC) or lecithin, demonstrated a temporary improvement in memory in AD patients [19]. These data suggest that decreasing such activity could be important in the therapeutic attenuation of cognitive decline. Mammalian phosphatidyl-choline phospholipase D (PC-PLD or PLD) is a family of lipolytic enzymes that can affect membrane curvature, exocytosis, endocytosis, vesicle release and neurite outgrowth, all of which are important for maintaining synaptic function [20,21,22,23,24,25,26,27]. While knocking out the gene-encoding PLD is not detrimental to survival across species [28,29,30], the elevated expression of PLD is implicated in cancer, cardiovascular disease and neuropathologies [31,32]. Thus, there has been great interest in developing small molecule inhibitors for mammalian PLD isoforms [33,34,35,36,37,38]. More importantly, upregulated PLD1 expression and activity was reported in reactive astroglial cells from AD brains, where a physical interaction with amyloid precursor protein (APP) was noted [39]. PLD1 activation was also implicated in the mitochondrial dysfunction in the brains of scrapie-infected mice [40] as well as in AD brains [41], thus increasing the importance of downregulating PLD1 levels as a potential therapy. Our study reinforced the human clinical relevance of the involved isoforms in neuronal function [15]. We investigated whether reducing PLD1 levels using a specific small molecule inhibitor in a 3xTg-AD mouse model improved synaptic function underlying AD-like neuropathology [42]. This model contains AD-related human variant genes for amyloid beta precursor protein (APPswe), presenilin 1 (PS1M146V) and microtubule-associated protein tau (tauP301L). The 3xTg-AD is one of the two models widely used to study interventions against human amyloid and tau pathology simultaneously. Recent studies have also outlined sex-specific differences at later ages in 3xTg-AD cognitive responses, making our approach timely and necessary [43,44,45,46,47,48,49,50]. The presence of Aβ plaques were reported in 3-month-old 3xTg-AD mice along with extracellular plaques in the neocortex and hippocampus at six months [51]. Six-month-old 3xTg-AD mice showed decreased hippocampal synaptic physiology with behavioral changes [43] that mirrored the human Aβ-related AD pathology [52]. However, tau expression and underlying pathological changes were prominently observed in the 3xTg-AD mice hippocampal regions only from 9 months onward. Therefore, understanding tau-related effects necessitates the use of mice aged 9 months and above. We were the first to observe that the inducible isoform, PLD1, but not the constitutive isoform, PLD2, shows elevated synaptosomal expression (thus complementing the earlier studies that established that PLD1 was also aberrantly elevated in astrocytes and mitochondrial fractions) in the post-mortem hippocampi of AD brains compared to age-matched controls. Moreover, our systematic functional studies established that increased PLD1 promotes synaptic vulnerability and can exacerbate oligomeric (AβO- and TauO-) amyloidogenic cognitive decline. Our follow-up study investigated the scientific premise of progressive Aβ accumulation in 6-month-old 3xTg-AD mice expressing elevated PLD1 [42], and more importantly, established PLD1 inhibition as a viable therapeutic target in mediating cognitive decline against Aβ. We were the first to report that prolonged administration of VU 0155069 (also referred to as VU01) was sufficient to preserve dendritic spine integrity and preserve hippocampal synaptic function and prevent underlying memory deficits in 3xTg-AD mice. Since Aβ and tau accumulation continue increasing in 3xTg-AD with age, and our data suggesting crude synaptosomal PLD1 expression also increases from 6 months to 18 months [15], we asked whether PLD1 inhibition was effective at 12 months, when the levels of tau are significantly higher, than at 6 months. Further, at 12 months, 3xTg-AD mice display many senescence markers, including neuroinflammation [53]. By performing systematic studies at two different ages, 6 months (~20–30 years old equivalent for human age) and 12 months (~38–47 years old equivalent for human age), we addressed the scientific premise and rationale for the suitability of using PLD1 inhibition (with VU01) as a therapeutic agent over an appropriate age range in terms of human age equivalence. More importantly, this study is among the few to address whether therapeutic intervention with the same dosage of VU01 after the full-blown effects of Aβ and tau is efficacious, a premise that would be extremely important when advancing to clinical trials to estimate whether this intervention can be applied at early stages or even when it is late stage. Recent observations put more emphasis on tau-dependent signaling events over Aβ effects due to its predominance in driving memory deficits and neurodegeneration in the later stages of progressive cognitive decline [54,55,56]. Emerging studies also propose that reducing tau can ameliorate Aβ effects and confer significantly greater improvements in the performance of neuronal circuits, regardless of Aβ and/or tau clearance [56,57,58,59,60]. Thus, in this study, we addressed whether chronic PLD1 inhibition in the later stages of AD-like neuropathology, that has a greater dependence on tau, is still sufficient to provide resistance to the progression of cognitive decline. We [1] looked at whether the dosage at 1mg/kg once every 2 days for 30 days was sufficient to reduce PLD1 expression (using immunofluorescence); [2] studied the effect on dendritic spine morphology (using Golgi–Cox staining) as the underlying mechanism of action that is preserved by inhibiting PLD1; [3] observed for improved hippocampal synaptic function (excitatory and inhibitory) in the Schaffer collateral synapses (electrophysiology); and [4] investigated the rescue of cognitive function in novel object recognition (NOR) and amygdala-associated cued and contextual memory (FC). ## 2.1. Increased PLD1 Levels and Increased Co-Localization with Amyloidogenic Proteins in Post-Mortem Human AD Hippocampal Slices Compared to Age-Matched Control In our previous study, we reported that crude synaptosomal PLD1 expression (using Western blots) was increased in post-mortem AD brains compared to age-matched controls [15]. In the present study, we extend the observation further by quantifying both the increased expression of PLD1 as well as the specific co-staining between Aβ or tau with PLD1 in post-mortem AD brains (Figure 1). We observed that the overall expression of PLD1 is significantly increased in the AD group compared to the control in both the experimental slides used for Aβ (Figure 1B, $$p \leq 0.0022$$) and tau (Figure 1F, $$p \leq 0.0022$$) co-staining with PLD1. Moreover, we report a significantly increased co-localization between PLD1-Aβ (Figure 1C, ** $$p \leq 0.0022$$) and PLD-tau (Figure 1G, * $$p \leq 0.0260$$) in AD compared to age-matched control hippocampi. Thus, the increased levels of PLD1 show an increased association with the amyloidogenic proteins implicated in the progression of AD. This is in agreement with previous studies that show astroglial PLD1 physically interacting and co-localizing with APP and caveolin-3 via the pleckstrin homology domain of PLD1 [39] and cellular studies in the neuronal N2a developmental role of PLD1 assists with the trafficking of an amyloid-precursor form of Aβ [27,61]. ## 2.2. Chronic One-Month Treatment with PLD1 Inhibitor (VU0155069 or VU01 at 1 mg/kg/2 Days i.p.) in 12-Month-Old 3xTg-AD Mice Attenuates PLD1, PLD1-Aβ and PLD1-tau Co-Staining Expression Differentially in Hippocampal Subregions We assessed whether repeated VU01 treatment was effective in reducing hippocampal PLD1 expression with attention to different subregions such as DG, CA1 and CA3 (Figure 2—PLD1-Aβ co-staining and Figure 3—PLD1-tau co-staining). PLD1 levels were significantly decreased in the CA1 (Figure 2J, * $$p \leq 0.0205$$; Figure 3J, * $$p \leq 0.0426$$) and DG (Figure 2B, * $$p \leq 0.0140$$; Figure 3B, $$p \leq 0.0027$$). The reduction in PLD1 levels did not reach significance in the CA3 (Figure 2F, $$p \leq 0.3357$$, ns; Figure 3F, $$p \leq 0.1079$$, ns). Thus, there is an overall reduction in PLD1 expression except in the CA3 region that needs further exploration. Interestingly, treatment with VU01 shows exclusive Aβ level reduction in DG (Figure 2C, * $$p \leq 0.0003$$), but not CA1 (Figure 2K, $$p \leq 0.9551$$, ns) or CA3 (Figure 2G, $$p \leq 0.3357$$, ns). Tau expression levels remain unchanged between VU01 and saline-treated cohorts in all three regions (CA1, Figure 3K, $$p \leq 0.8518$$, ns; CA3, Figure 3G, $$p \leq 0.9497$$, ns; DG, Figure 3C, $$p \leq 0.7546$$, ns). Then we calculated the integrated density of PLD1-Aβ (Figure 2D,H,L) and PLD1-tau (Figure 3D,H,L) co-localization. We observed a significant decrease in the DG co-localization of PLD1-Aβ (Figure 2D, * $$p \leq 0.0289$$) and PLD1-tau (Figure 3D, $$p \leq 0.0087$$), but the other two regions CA3 and CA1 did not show significant differences between the treated and untreated cohorts. The reduction in Aβ in DG, one of the two regions associated with neurogenesis [62,63], due to VU01 could be attributed to the activation of neuroinflammation-specific cytokines that are beneficial, since VU01 has been reported to affect inflammation in somatic tissues [64]. However, further studies will be needed to understand the specific mechanisms affected. Collectively, these results provide evidence of the effectiveness of PLD1 inhibition with attention to regional changes using our treatment regimen. Next, we explored whether such changes were sufficient to preserve the dendritic spine integrity. ## 2.3. VU01 Dependent Chronic Inhibition of PLD1 Preserves Specific Dendritic Spine Morphologies in 12-Month-Old 3xTg-AD Mice We previously reported in our study in 6-month-old 3xTg-AD mice treated with VU01 [42] that the mechanism of resilience involves the preservation of dendritic spine integrity. The role of Aβ on dendritic spines is well established, both using neuronal cultures [65,66] and by the loss of dendritic spines located around plaques in mouse models [67,68]. Taken together, we speculate that the repeated VU01 administration in 6-month-old 3xTg-AD mice may have protected against the Aβ-driven dendritic spine dystrophy. While the role of tau in dendritic spine dystrophy is still emerging, there is enough evidence for its detrimental impacts resulting in clustered dendritic spine loss [9,69,70]. Therefore, we investigated how effective the repeated VU01 regimen was for preserving dendritic spine integrity at the later timeframe of 12 months in the 3xTg-AD mice (with increased tau-related insults—Figure 4). As described in our previous study, we utilized the Golgi–Cox impregnation technique [42], because of its ability to analyze spine morphology through the visualization of a low percentage of neurons (see schematic—Figure 4A). The VU01-treated cohort showed significant differences (increased numbers) in individual spine area averages ( $$p \leq 0.0048$$; Figure 4C), spine perimeter ( $$p \leq 0.0061$$; Figure 4D), filamentous spine length ( $$p \leq 0.0026$$; Figure 4E) and mushroom spine parameters of neck length (** $p \leq 0.0001$; Figure 4F), neck width ( $p \leq 0.0001$; Figure 4G), head length ( $$p \leq 0.0074$$; Figure 4H) and head width (* $$p \leq 0.0101$$; Figure 4I) in the CA1 region of the hippocampus compared to the saline-treated 12-month-old 3xTg-AD female mice. No significant differences were observed in spine density per 10 μm ($$p \leq 0.3593$$; Supplementary Figure S2A), specific spine type per 10 μm (Stubby ($$p \leq 0.2762$$; Supplementary Figure S2B); Filamentous ($$p \leq 0.0663$$; Supplementary Figure S2C); Mushroom ($$p \leq 0.0663$$; Supplementary Figure S2D)), number of dendritic spines ($$p \leq 0.3784$$; Supplementary Figure S2E) or total dendritic area measured in μm2 ($$p \leq 0.7607$$; Supplementary Figure S2F). Thus, repeated treatment attenuating PLD1 is effective against Aβ and tau-driven clustered spine loss via the preservation of mushroom spine morphology and increased filamentous spine lengths in 12-month-old 3xTg-AD mice. ## 2.4. Restoration of Hippocampal Synaptic Function by Chronic PLD1 Inhibition Regimen in 12-Month-Old 3xTg-AD Mice Electrophysiological assessments of high frequency stimulation-dependent long-term potentiation (HFS-LTP) and low frequency stimulation-dependent long-term depression (LFS-LTD) of hippocampal slices from the cohorts were conducted as described in the methods. In our previous studies, we demonstrated the robust recovery of the HFS-LTP by PLD1 inhibition in the hippocampal Schaffer collateral modeled using AD pathology-dependent amyloidogenic proteins (AβO or TauO) acutely in wildtype mice [15] and chronically in 6-month-old 3xTg-AD mouse models [42]. Traditionally, Aβ-associated synaptic dysfunction has been routinely assessed by measuring LTP dysfunction as a readout of functional neurodegeneration [7,12,13,14,15,42,71,72,73,74,75,76,77,78] with only relatively recent work exploring the LTD effects [75]. However, studies have clearly documented an important role for tau in LTD [79,80]. Since the 12-month-old 3xTg-AD mice are reported to show both Aβ and tau pathologies [45,46,48,50], we rationalized that the assessment of both LTP and LTD at this age will be needed to completely understand the therapeutic potential of PLD1 chronic inhibition. Furthermore, very strong correlation has been reported between clustered dendritic spine loss in AD leading to homeostatic imbalance in synaptic function that occurs via the dysregulation of synaptic depression and a corresponding compensatory excitation in brain circuits [81,82,83,84]. This could result in altered synaptic responses of long-term potentiation and long-term depression, justifying our approach to assess both LTP (Figure 5) and LTD (Figure 6) in our VU01-treated 3xTg-AD cohorts at 12 months of age. HFS-LTP decrement observed in the 12-month-old saline-treated 3xTg-AD siblings was prevented in the VU01-treated cohort (**** $p \leq 0.0001$; Figure 5A,B) despite the progressive load of Aβ and tau-driven synaptic toxicity. The improvement in synaptic function was also observed when the data were separated into males (* $p \leq 0.0166$; $$n = 5$$–6 mice, Figure 5C,D) and females ( $p \leq 0.0015$; $$n = 4$$–6 mice, Figure 5E,F) without any sex-specific difference in the inhibition effect ($p \leq 0.9999$ for saline (male vs. female) and $$p \leq 0.3491$$ for inhibitor (male vs. female), Figure 5G). We also measured the input–output relationship before and after HFS-LTP (Supplementary Figure S3). The input—a measure of the presynaptic fiber volley (FV)—did not show any differences when assessed for before and after stimulation (Supplementary Figure S3A) or between male (Supplementary Figure S3B) and female (Supplementary Figure S3C) responses (also see Supplementary Table S1). The output—the slope of the field excitatory post-synaptic potential (fEPSP)—which is also a measure of the viability of the tissue, showed no difference in the stimulation profile between the treatment groups (Supplementary Figure S3D) or between males (Supplementary Figure S3E) or females (Supplementary Figure S3F, also see Supplementary Table S1). It is important to note here that the slope of the response increased in both treatment groups following HFS, but the extent of the increase was pronounced in the inhibitor-treated compared to the saline-treated group (at maximum response, this was 2.3-fold for the saline-treated vs. 5.8-fold for the inhibitor-treated cohort). Qualitative trends in the post-HFS responses of female mice suggest that the efficacy of PLD1 inhibition could be better in females (Supplementary Figure S3F), warranting further electrophysiological assessments in future studies. The fEPSP slope as a function of FV did not reach statistical significance between the treatment groups (Supplementary Figure S3G) or when separated into males (Supplementary Figure S3H) and females (Supplementary Figure S3I), perhaps due to the increased variability within the saline-treated group. We measured the paired-pulse facilitation at four different intervals (200, 100, 50 and 25 ms). We did not observe statistically significant differences between the treatment groups (Supplementary Figure S4A) or when the data were separated into males (Supplementary Figure S4B) and females (Supplementary Figure S4C) (also see Supplementary Table S2). In the present study, we also looked at Schaffer collateral LTD (Figure 6). We report a significant improvement in the response of the VU01-treated 12-month-old 3xTg-AD cohort compared to the saline-treated age-matched control (** $p \leq 0.0001$; $$n = 8$$–9 mice; Figure 6A,B). When the data were separated into males (* $$p \leq 0.0318$$; $$n = 4$$–6 mice, Figure 6C,D) and females (* $$p \leq 0.0138$$; $$n = 3$$–4 mice, Figure 6E,F), we did not observe any sex-specific differences ($$p \leq 0.2626$$ for saline (male vs. female) and $$p \leq 0.6744$$ for inhibitor (male vs. female); Figure 6G). The input–output (Supplementary Figure S5 and Supplementary Table S3) and paired-pulse measurements (Supplementary Figure S6 and Supplementary Table S4) for LFS-LTD were similar in profile to what we observed with HFS-LTP. Thus, an overall beneficial effect reflecting the homeostatic restoration of glutamatergic neurotransmission functionality is observed following our repeated VU01-associated PLD1 inhibition in 3xTg-AD mice at 12 months. ## 2.5. Chronic One-Month Treatment with PLD1 Inhibitor Ameliorates Memory Deficits in 12-Month-Old 3xTg-AD Mice The 3xTg-AD mice injected with either inhibitor (VU01, 1mg/kg) or saline i.p. were subjected to NOR (Figure 7, see schematic in Figure 7A,B). We observed that repeated treatment with PLD1 inhibitor increased the ability of the animals to spend greater time with the novel object compared to their saline-treated sibling (Figure 7C), both at 2 h (*** $$p \leq 0.0003$$, $$n = 12$$–13 mice) and at 24 h (* $$p \leq 0.0211$$, $$n = 12$$–13 mice) after training. When the results were separated based on sex, PLD1 inhibition was effective even when the data was separated into males at 2 h (** $$p \leq 0.0014$$) and at 24 h (* $$p \leq 0.0404$$, $$n = 6$$–7 mice, Figure 7 D) or females at 2 h (* $$p \leq 0.0275$$) and at 24 h (* $$p \leq 0.0412$$, $$n = 6$$ mice, Figure 7E), suggesting equal effect of VU01 on both sexes at this age. We did not observe any changes in the distance travelled (Supplementary Figure S7A) or the time that the animals were moving (Supplementary Figure S7B), suggesting that non-specific locomotor effects were not associated with VU01 administration. Further analysis of male-specific (Supplementary Figure S7C,D) and female-specific (Supplementary Figure S7E,F) responses also did not uncover any significant sex-specific differences. In order to corroborate these behavioral effects and extend the observations, we performed FC behavior (Figure 8, see schematic in Figure 8A,B), where we assessed hippocampal (contextual)- and amygdala (cued)-dependent aversive associative memory [85]. The hippocampal contextual memory showed improvement in the PLD1 inhibitor-injected mice compared to the saline-treated cohort (** $p \leq 0.0001$; $$n = 12$$–13 mice; Figure 8C). The improvement in response remained significant even when the cohorts were separated into males ( $$p \leq 0.0012$$; $$n = 6$$–7 mice, Figure 8F) and females (* $$p \leq 0.0152$$; $$n = 6$$ mice, Figure 8H). The amygdala-dependent cued response showed a significant difference between the two treatment groups (* $$p \leq 0.0005$$; $$n = 12$$–13 mice, Figure 8D). Importantly, the pre-cue freezing was not different between the two treatments ($$p \leq 0.1638$$; Figure 8D) and was reflected even after the separation of the data into males ($$p \leq 0.0906$$; $$n = 6$$–7 mice, Figure 8G) and females ($$p \leq 0.9990$$; $$n = 6$$ mice, Figure 8I). The cued responses showed differences between the sexes (Figure 8G,I). Female mice treated with PLD1 inhibitor showed significant cued freezing compared to their age-matched cohort injected with saline ( $$p \leq 0.0036$$; $$n = 6$$ mice, Figure 8I), but the difference did not reach significance in the male group ($$p \leq 0.0571$$; Figure 8G). However, it is important to note that the difference is not because the inhibitor-treated male mice had a lower response, rather it was because the saline-treated male group showed a higher variability (see distribution of individual animals in Figure 8G and epoch analysis in Supplementary Figure S8E). When we analyzed the training sessions, we did not observe any sex-specific changes (Supplementary Figure S8). Analysis of the epochs for training (Supplementary Figure S8B,C), cued (Supplementary Figure S8E,F) and contextual (Supplementary Figure S8H,I) responses corroborated the averaged responses (Supplementary Figure S8A,D,G). The cohorts did not show any differences in learning the behavior (Figure 8E and Supplementary Figure S8A,B,C). ## 3. Discussion In the present study, we first established the scientific premise of PLD1 as a key element associated with progressive cognitive decline using human clinical samples. We demonstrated that increased PLD1 levels show increased association with Aβ and with tau—two neuropathological elements associated with AD (Figure 1). Then, we used 12-month-old 3xTg-AD mice and performed functional studies to validate whether decreasing the aberrant PLD1 was beneficial to synaptic function and underlying AD-like memory deficits. Most investigations of AD therapies have been routinely conducted only using female 3xTg-AD mice citing the prevalence of AD in females [86,87,88,89,90,91]. We have attempted to study the responses in both sexes to better interpret the underlying mechanism and the therapeutic applicability targeting the PLD1 signalosome toward the amelioration of synaptic deficits in aberrant neurological states. There is emerging literature that the physiological levels of circulating Aβ and tau maintain normal synaptic neurotransmission in healthy states [92]. Such an observation suggests that therapeutics affecting Aβ and tau levels in diseased states should regulate, but not eliminate, Aβ and tau signaling that is important in maintaining normal cognitive functionality. Thus, we assessed how chronic VU01 treatment affects PLD1, Aβ and tau levels using immunofluorescence studies (Figure 2 and Figure 3). We report specific hippocampal regional differences in the reduction in PLD1 levels as well as PLD1-Aβ and PLD1-tau co-localization that warrants further exploration. The specificities of reduction could suggest a possible feedback loop between PLD1, Aβ and tau that results in the preservation of dendritic spines, restoration of synaptic function and prevention of cognitive decline. Additionally, there is literature that supports a role for direct interaction between PLD1 and APP [27,39,40,41,61,93] and, therefore, provides evidence for the reduction in Aβ along with PLD1 reduction by VU01 treatment is perhaps more the case rather than just being a case of “sticky” amyloidgenic protein expression loss. More elaborate studies using proximity ligation assay will be required to quantify further the importance and relevance of this association from a therapeutic point of view. Since PLD1 levels were elevated in the astrocytic and mitochondrial compartments in the previously reported study, our future studies will explore mechanisms in which PLD1 has been implicated and plays an important role in the progression of neurodegenerative states including autophagy [29,94,95,96], neuroinflammation [35,64,97,98,99,100,101,102], reactive oxygen species [103,104,105,106] and infection [35,107,108,109] and how these contribute to the synaptic dysfunction leading to cognitive decline in our studies. It is documented that Aβ and tau progressively accumulate to toxic levels at excitatory synapses [110,111] and result in progressive dendritic spine dystrophy [112,113]. In the present study, we report that chronic PLD1 inhibition demonstrates the ability to preserve key aspects of dendritic spine morphology such as mature mushroom spine integrity and filamentous spine numbers (Figure 4), which are properties that translated to improved synaptic function (HFS-LTP and LFS-LTD) as well as underlying behaviors (NOR and FC). These findings suggest that PLD1 signaling affecting dendritic spine integrity may prove to be a key mechanism that can improve cognitive resilience at different stages in the progression of ADRD. We assessed LTP [7,114] and LTD [115] as functional readouts of synaptic strength and plasticity (HFS-LTP—Figure 5; LFS-LTD—Figure 6) to corroborate whether the preservation of dendritic spine integrity was reflected at the synaptic level by improved functional outcomes in both these measures. We observed robust improvement without any sex-specific differences at the Schaffer collateral synapses. Though LTP involves molecular components that affect both pre- and post-synaptic compartments, the AD literature has more reports on the Aβ and tau-related post-synaptic insults [12,92,116,117,118,119,120,121,122,123]. Many of the studies have reported the molecular changes involved in AD pathology downstream to membrane receptors in the post-synaptic compartment such as protein kinase C (PKC) [124], protein kinase M zeta (PKMζ) [125], calcium/calmodulin-dependent protein kinase II (CaMKII) [126], calcineurin (a phosphatase) [127,128] and cyclic adenosine monophosphate response element binding protein (CREB) [129], all of which affect the maintenance of dendritic spine integrity. On the other hand, glutamate neurotransmission associated LTD that is affected in AD by either Aβ [130] or tau [80], has opposite effects on the above-mentioned signaling cascade and could, therefore, be instrumental in dendritic spine retraction. Thus, it is proposed that LTD in healthy individuals works (along with LTP) to regulate the dendritic spine morphology and enable “forgetting” [131]. Therefore, in theory, Aβ and tau could hijack and compromise LTP, exacerbate LTD and in turn cause “forgetting” via the reduction in dendritic spines [132]. Perhaps repeated PLD1 inhibition at 1 mg/kg may be optimal in modulating glutamatergic neurotransmission to preserve synaptic functionality, thereby altering the homeostasis enough to resist cognitive decline without affecting the physiological functionality associated with Aβ and tau. Taken together, our observations support a possibility that VU01 seems to have a much larger impact on synaptic function and plasticity than on spine morphometric properties (i.e., for most synapse morphometric properties, the magnitude of difference between saline and inhibitor is fairly small, despite significant statistical comparisons). Thus, VU01-mediated inhibition may also affect the biochemical mechanisms of synapse function/plasticity (e.g., glutamate receptor function, balance of protein kinase/phosphatase activities independently of spine morphology). Alternatively, it is possible that there are extrasynaptic effects of VU01 that are responsible for the multimodal improvement that we report here as we have elaborated earlier in the discussion section. Of greater importance to this particular study, where we have attempted to address therapeutic efficacy, we chose to compare the effects of VU01 in the pathological background rather than age-matched wildtype control because we were not addressing the mechanistic changes at this level of study, which would require more extensive and exhaustive analysis that is beyond the scope of this study, but would be addressed in a future study. The 3xTg-AD triple transgenic model of AD-like neuropathology considers [1] age-dependent increase and [2] sex-dependent effects with progressively accumulating Aβ and tau, thus mirroring the human pathological profile [43,44,52]. Familial early-onset Alzheimer’s disease (FEOAD) or early-onset AD (EOAD) associated with dominant mutations in APP, PSEN1 and PSEN2 have a high penetrance for displaying many of the clinical symptoms associated with AD. Thus, characterized in 2003, the 3xTg-AD mouse model is a popular pre-clinical AD model [52], with notable AD-related pathologies which are faithfully replicated. Aβ plaques are first detected at 6 months in the hippocampus that accumulate progressively with age, such as happens in humans. Moreover, cortical Aβ plaques appear at 12 months and also exhibit an age-dependent progression and this was reported initially [52] and replicated recently [47]. Most important, for our current study, tau phosphorylation was observed from 12 months in the hippocampus, particularly in the pyramidal neurons of the CA1 region [47,52]. Thus, while it is important to note here that this extensively used mouse model of AD is unique among many to recapitulate Aβ and tau-driven neurotoxicity found in human tissue or imaging experiments, and vice versa, successful therapies developed and tested have been universally unsuccessful in human clinical trials, prompting a reassessment of the development, use and interpretations of the data acquired from such models [133]. One critical factor that is often disregarded in therapeutic assessment studies at the pre-clinical level is the age. Our present systematic study at 12 months addresses exactly that component of therapeutic importance that is missing in other studies using this remarkable transgenic model that shows the progressive advancement of synaptic dysfunction and associated cognitive decline. Thus, our study is novel in that it emphasizes the observation that a repeated dose of the therapeutic even later in the progression remains efficacious, an observation that is more unique than common and needs to be a gold standard to study the effects of the drug, using multimodal approaches at more than one age to illustrate the robustness of the intervention before proceeding to human clinical trials. Another important novelty that emerges from our study is the potential of PLD1 therapeutics in late-onset AD (LOAD). The search for an “ideal” transgenic model of late-onset AD (LOAD) is a key requirement for the NIH/NIA-established Model Organism Development and Evaluation of Late-onset Alzheimer’s Disease (MODEL-AD) [134]. Under this premise, the LaFerla group conducted a systematic assessment and revisited the adequacy for the 3xTg-AD model recently [133]. They concluded that the slow pathology development in an age-specific and sex-specific manner will be advantageous to study the interaction of amyloidogenic insults on synaptic dysfunction associated with LOAD. Importantly, LOAD is linked to a number of variants with the most well-known occurring in the ε4 allele of apolipoprotein E (APOE4) that result in the disruption of lipid homeostasis [135,136]. APOE occurs in three isoforms—E2, E3 and E4—which differ in a single amino acid change. E4 allele increases the risk of AD in a dose-dependent manner compared to E3 allele, while E2 is reportedly protective with its biological role remaining understudied. It facilitates the binding and uptake of lipoprotein complexes via low-density lipoprotein receptors (LDLRs), or lipoprotein receptor 1 (LRP1). Multiple studies point to the importance of APOE as a universal biological variable along with age and gender in the progression of AD. APOE4 is the major genetic risk factor for developing sporadic AD (sAD), where 1 copy confers a 3-fold increase while 2 copies increase the risk to 12-fold. Female APOE4 carriers have an increased AD risk compared to male APOE4 carriers. Recent transcriptomic studies of human APOE3 and APOE4 knockin mice (APOE3-KI and APOE4-KI resp.) from 5 to 20 months of age show prominent changes in neuronal calcium signaling and synaptic function, implicating the phospholipase D (PLD) pathway [137]. APOE is the most abundant brain lipoprotein chaperones expressed mainly in astrocytes. Combined with the observation that aberrant upregulated PLD1 expression and activity was reported in reactive astroglial cells from AD brains, where a physical interaction with amyloid precursor protein (APP) [39] alters the mitochondrial function in the brains of scrapie-infected mice [40] as well as in AD brains [41], we speculate that attenuating PLD1 using VU01 in 12-month-old 3xTg-AD may impinge on APOE-related dyslipidemia that may be another key mechanism that needs to be explored to realize the full therapeutic potential of VU01 as a potential LOAD intervention. Heavier Aβ burden in female 3xTg-AD [138,139,140,141] correlated with a greater impairment in spatial reorientation [139], compared to male mice. Higher levels of tau are observed in older (18–20 month old) females compared to males [48]. It is important to note here that we have separated the male and female responses for understanding the sex-specific effects of PLD1 inhibition. However, we will further explore these aspects in the future with more numbers for a better powered study. Nevertheless, our stratified approach and a multimodal assessment provide a degree of confidence in conducting such sex-specific assessments that is important for properly addressing the therapeutics and mechanisms in neurodegenerative states that are known to have gender-specific effects. In our previous study [42], no sex-specific differences were observed within the saline- or inhibitor-treated or between the two treatment groups in the amygdala-dependent aversive cued-fear memory response despite detectable levels of intraneuronal and extracellular Aβ in the basolateral/lateral amygdala reported at this age [44]. Robust tau expression in the amygdala of 3xTg-AD mice has been reported at later ages beginning from 9 months through 26 months [51]. Therefore, we speculated that the combined effect of Aβ and tau in 12-month-olds would be more detrimental than the effect of Aβ alone at 6 months in impairing the robust amygdala-dependent cued (fear-conditioned) memory. Indeed, our results (Figure 8) indicate that the 12-month-old 3xTg-AD saline-treated animals have a poor cued memory that is rescued by treatment with the inhibitor. Our sex-stratified approach also uncovered that the inhibitor-treated male cohort failed to reach significance when compared to the saline-treated male group for the cued response. Surprisingly, no intra-treatment differences were noted between the male and female inhibitor-treated cohort for the cued response (Figure 8). This lack of statistical significance may be due to the variability in the saline-treated male cohort (see Figure 8G). An earlier study had reported increased anxiety in 12-month and 15-month-old 3xTg-AD male mice [142] compared to females. Other differences between 2 months to 15 months are attributed to neuroendocrine differences [48] that result in males showing more errors and more latency in memory acquisition [46,143]. Taken together, these results indicate that aging-related, sex-specific endocrine signaling contribute to the variation in male behavioral response seen in our saline-treated male cohort. It is, therefore, remarkable, and pertinent that regardless of such variations, the efficacy of PLD1 inhibition was not different in the male and female mice within the repeated VU01-treated group. Therefore, we speculate that the PLD1 inhibition (by VU01 at 1mg/kg every alternate day for a month regimen) preserves amygdala memory mechanisms by possibly regulating the hormonal effects on synaptic function in males. In the NOR study (Figure 7), no sex-specific difference was observed within the saline-treated or the inhibitor-treated cohorts. It has been reported in human clinical studies that the volume of the total perirhinal cortex in AD patients is decreased compared to control subjects, but without any gender differences [144]. Short-term storage of object memory occurs in the perirhinal cortex before being transferred to the hippocampus for consolidation [145]. Thus, we speculate that the lack of sex differences in our NOR study is because NOR heavily relies on perirhinal cortex function. Thus, in this study, we find that chronic PLD1 inhibition (1 mg/kg VU01 once every 2 days i.p. for 30 days) at later stages of AD-like neuropathology, that has a greater dependence on tau, (with additional effects on neuroinflammation as well as astroglial mechanisms impinging on APOE-related dyslipidemia that need to be explored further) was sufficient to reduce PLD1 expression and association with amyloidogenic proteins, preserve dendritic spine morphology, improve hippocampal synaptic function and rescue hippocampal and amygdala-associative memory deficits. ## 4.1. Drugs PLD1 inhibitor (VU0155069) was obtained from Tocris Bioscience (Bio-Techne, Minneapolis, NE, USA). ## 4.2. Human Samples Post-mortem brain tissues were a gracious gift from Dr. Giulio Taglialatela, PhD, available to us as a member of the Mitchell Center for Neurodegenerative Diseases through his capacity as the Director for the Center. The source of the tissues is the Oregon Brain Bank at Oregon Health and Science University (OHSU; Portland, OR, USA). Briefly, donor subjects of either sex were enrolled and clinically evaluated in studies at the National Institutes of Health-sponsored C. Rex and Ruth H. Layton Aging and Alzheimer’s Disease Center (ADC) at OHSU, via the OHSU Institutional Review Board (IRB). Informed consent was obtained from all participants before their enrollment in the studies at the ADC in brain-aging and for receiving annual neurologic and neuropsychological evaluations. Following neuropathological assessments post-mortem, a neuropathologist-scored database was generated for each brain tissue, according to standardized CERAD (Consortium to Establish a Registry for Alzheimer’s Disease) criteria and Braak staging. The cases used in this study are described in Table 1. To ensure that the variations in post-mortem interval (PMI) did not affect any measurements, we have provided a correlation analysis between PMI values and the immunofluorescence (IF) performed for PLD1. No correlation was found (Supplementary Figure S1), and, therefore, observed differences could not be attributed to differences in non-specific post-mortem tissue degradation. It is nonetheless important to appreciate that five of the brains obtained following 10 h PMI might not necessarily fully reflect freshly obtained brain tissue. ## 4.3. Animals This study was conducted in accordance with the National Research Council’s “Guide for the Care and Use of Laboratory Animals (8th Edition)” in the animal care facility at The University of Texas Medical Branch at Galveston (UTMB) which is accredited by the Association for Assessment and Accreditation of Laboratory Animal Care (AALAS), International. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC) and were performed according to the National Institutes of Health (NIH) Guidelines [146] on the use of laboratory animals. Male and female 3xTg-AD transgenic mice were purchased from Jackson Labs (Bar Harbor, ME, USA) and maintained through a breeding program at UTMB. Mice were housed—five per cage in their filter-top cages in a temperature-controlled environment at 22 °C, humidity $40\%$ and a 12:12 h light–dark cycle, with regular chow provided ad libitum. We had to utilize three cohorts of the 3xTg-AD mice for the experiments described in this study [Round 1: 6 females and 7 males (VU01 injected) and 6 females and 6 males ($0.9\%$ saline); Round 2: 5 females (VU01 injected) and 5 females ($0.9\%$ saline)], separated by few months to complete the experiments described here. Round 3: 4 males and 4 females (2 of each sex injected with VU01 or saline) separated by 7 months from second round. Thus, we had $$n = 20$$ animals total for VU01-injected and $$n = 21$$ animals total for saline- or vehicle-injected to conduct the behavioral, electrophysiological and dendritic spine studies. Animals (aged ~11 months) received a single injection intraperitoneally (i.p.) of 1 mg/kg of VU01 diluted in $0.9\%$ saline solution (inhibitor-treated cohort) or an equivalent amount of $0.9\%$ saline (saline-treated cohort) and returned to their cage. This was repeated every alternate day (alternating the injection side to reduce incidence of repetitive injection-related inflammation), for a period of one month. Based on previous reports, the results were also analyzed separately for males and females. All behavioral testing was performed within the 12 h light cycle for return to home cages prior to the 12 h dark cycle. After completion of the behaviors, the animals were processed as described under field electrophysiological recordings section. The second cohort of animals injected for Golgi staining were also deeply anesthetized with isoflurane, and immediately the brain was extracted from the skull, washed with phosphate buffered saline pH 7.4 (ThermoFisher Scientific, Waltham, MA, USA) and processed as described under Tissue Processing and Golgi Staining section. ## 4.4. Field Electrophysiological Recordings Our standard protocol was used as previously described [13,14,15,42,76,77,147,148]. Briefly, mice were deeply anesthetized with isoflurane and transcardially perfused with ~30 mL of room temperature carbogenated ($95\%$ O2 and $5\%$ CO2 gas mixture) NMDG-artificial cerebrospinal fluid (aCSF) (in mM-93 N-Methyl-D-Gluconate, 2.5 KCl, 1.2 NaH2PO4, 30 NaHCO3, 20 C8H18N2O4S, 25 C6H12O6, 5 C6H7O6Na, 2 CH4N2S, 3 C3H3NaO3, 10 MgSO4,7H2O, 0.5 CaCl2,2H2O, 12 C5H9NO3S, pH 7.4) and sliced using Compresstome VF-300 (Precisionary Instruments, Greenville, NC, USA) in carbogenated NMDG-aCSF to obtain 350 μm transverse brain sections. Slices were allowed to recover for 10 min in carbogenated NMDG-aCSF at 33 °C. Slices were then maintained at room temperature in a modified carbogenated HEPES holding aCSF solution (in mM-92 NaCl, 2.5 KCl, 1.2 NaH2PO4, 30 NaHCO3, 20 C8H18N2O4S, 25 C6H12O6, 5 C6H7O6Na, 2 CH4N2S, 3 C3H3NaO3, 2 MgSO4,7H20, 2 CaCl2,2H20, 12 C5H9NO3S, pH 7.4). Slices were recorded in carbogenated standard recording naCSF (in mM-124 NaCl, 2.5 KCl, 1.2 NaH2PO4, 24 NaHCO3, 5 C8H18N2O4S, 13 C6H12O6, 2 MgSO4,7H20, 2 CaCl2,2H20, pH 7.4). Evoked field excitatory post-synaptic potential (fEPSP) recordings were performed by stimulating the Schaffer collateral pathway (located in stratum radiatum) using a stimulating electrode of ~22 kΩ resistance placed in the CA3 region and glass recording electrodes in the CA1 region. Current stimulation was delivered through a digital stimulus isolation amplifier (A.M.P.I., ISRAEL) and set to elicit a fEPSP approximately $30\%$ of maximum for synaptic potentiation experiments using platinum–iridium-tipped concentric bipolar stimulating electrodes (FHC Inc., Bowdoin, ME, USA). The use of platinum–iridium wire and diphasic pulses can help minimize electrode polarization [149]. Using a horizontal P-97 Flaming/Brown Micropipette puller (Sutter Instruments, Novato, CA, USA), borosilicate glass capillaries were used to pull recording electrodes and filled with naCSF to obtain a resistance of 1–2 MΩ. Field potentials were recorded in CA1 stratum radiatum using a Ag/AgCl bridge with CV7B headstage (Molecular Devices, Sunnyvale, CA, USA) located ~1–2 mm from the stimulating electrode. LTP was induced using a high frequency stimulation protocol (3 × 100 Hz, 20 s) as previously described [13,14,15,42,76,77,147]. LTD was induced using a low frequency stimulation protocol (900 × 1 Hz). Recordings were digitized with Digidata 1550B (Molecular Devices, Sunnyvale, CA, USA), amplified 100× and digitized at 6 kHz using an Axon MultiClamp 700B differential amplifier (Molecular Devices) and analyzed using Clampex 10.7 software (Molecular Devices) as previously described [148]. To assess basal synaptic strength, 250 μs stimulus pulses were given at 10 intensity levels (range, 100–1000 μA) at a rate of 0.1 Hz. Three field potentials at each level were averaged, and measurements of fiber volley (FV) amplitude (in millivolts) and fEPSP slope (millivolts per millisecond) were performed using Clampfit 10.7 software. Synaptic strength curves were constructed by plotting fEPSP slope values against FV amplitudes for each stimulus level. Baseline recordings were obtained for 10 min by delivering single pulse stimulations at 20 s intervals. All data are represented as a percentage change from the initial average baseline fEPSP slope obtained for the 10 min prior to HFS. Maximum of two slices were recorded per animal and averaged to give the response per animal. ## 4.5.1. Novel Object Recognition NOR was performed as described previously [13,14,15,42,76]. Briefly, animals were habituated for two consecutive days and assessed for normal locomotion and acclimation to the test environment (see schematic in Figure 7). After placement in the open field box for two 10 min test sessions that were 24 h apart, the AnyMaze (Stoelting Co., Wood Dale, IL, USA) video tracking software was used to quantify various locomotor parameters: total distance traveled, time spent moving >50mm/s, number of rears, number of entries into and time spent in the center $\frac{1}{9}$th of the locomotor arena. Twenty-four hours after the last habituation session, animals were subjected to training in a 10 min session of exposure to two identical, non-toxic objects in the open field box. The time spent exploring each object was recorded using an area 2 cm2 surrounding the object and was defined such that nose entries within 2 cm of the object were recorded as time exploring the object. After the training session, the animal was returned to its home cage. After a retention interval of 2 h and subsequently 24 h, the animal was returned to the arena in which two objects, one identical to the familiar object but previously unused (to prevent olfactory cues and prevent the necessity to clean objects during experimentation) and one novel object. The animal was allowed to explore for 10 min, during which the amount of time exploring each object was recorded. Objects were randomized and counterbalanced across animals. The animals were returned to their home cages with food/water ad libitum for 24 h minimum. After the rest period of minimum three days, the animals were tested for fear conditioning as described below. For novel object recognition tests, the percentage time exploring each object (familiar versus novel) is reported as an object discrimination index (ODI). An index above 0.5 is indicative of novelty associated with the object. Each mouse was tested at 2 h and at 24 h with the intention of assessing the shorter and longer time frames in memory recall. Different novel objects (color and shape) were used in the 24 h test compared to the 2 h test, to avoid performance deficits. ## 4.5.2. Fear Conditioned Response Contextual and cued fear-conditioned responses were assessed using our standard two-pairings fear-conditioning training protocol as previously described [42,150], utilizing the UTMB Rodent In Vivo Assessment Core Facility. Briefly, the standard protocol consisted of a training phase when the mice were placed in a particular environment—a standard mouse fear-conditioning chamber (Med Associates, Fairfax, VT, USA, a training chamber with lighting, geometry, odor that constitutes the context conditioned stimulus, CS) and allowed to explore for 3 min. An auditory CS (80 dB white noise) was then presented for 30 s and one footshock (0.8 mA, 2 s duration; the unconditioned stimulus, US) delivered during the last 2 s of the auditory CS. A second presentation of the auditory CS and the US was delivered at the 5 min mark and the animals then left in the cage for another 2 min. Twenty-four hours later, the mice were returned to the same training chamber and the context test for fear learning performed. The amount of freezing the mice exhibited during five minutes in the training chamber was measured. Between two to four hours later, the cued test was performed in a completely novel context. The animals were placed in the testing chamber and freezing was measured for three minutes before the auditory CS was represented and freezing quantified over the next three minutes. Freezing was quantified using FreezeFrame automated video capture and software analysis (Coulbourn Instruments, Whitehall, PA, USA) and evaluated as percentage freezing in 30 s (training) or 60 s (contextual, cued) bins. Epochs were averaged to provide the data as number of animals per group. ## 4.6. Tissue Processing and Golgi–Cox Staining As previously reported, brain hemispheres, obtained as described in the animals section, were stained using the FD Rapid Golgi Stain Kit (PK401, FD Neurotechnologies, Columbia, MD, USA) and according to the manufacturer’s instructions [42]. Tissue slices were impregnated in chromate mixture of Solution A (potassium dichromate and mercuric chloride) and Solution B (potassium chromate). After 24 h, chromate solution was replaced gently without disturbing the tissue, and then left in dark for 15 days. Subsequently, tissue slices were immersed in Solution C for 24 h. After 24 h, Solution C was replaced, according to the manufacturer’s instructions. These brains were sliced in 30 μm sections, mounted three per slide on gelatin-coated slides, sequentially for each animal by FD Neurotechnologies. For further microscopic assessments, the slides were shipped back to our lab after being processed by FD Neurotechnologies. Slides were stored in darkness. ## 4.7. Dendrite Imaging Previously published criteria and standards for dendritic imaging were used [151,152]. A blinded experimenter performed imaging and further analysis of the slides. An All-in-One Fluorescence Microscope BZ-X800E (Keyence Corporation of America, Itasca, IL, USA) was used to image Golgi-stained neurons at high magnification (100X oil-immersion objective) using the brightfield options available in the microscope. It was possible to determine the morphology of individual spines by magnifying using the 3X optical zoom option and subsequently quantified. To cover the full depth of the dendritic arbors (20–30 μm), Z-stack images were collected at 0.3 μm intervals and then compressed into a single TIFF image using the BZ-H4A software. ImageJ software was utilized (Open Source from National Institutes of Health, Bethesda, MD, USA) on these TIFF stacks for subsequent quantitative analysis. For each animal, three hemispheric slices per slide with the best representation of the Schaffer collateral were chosen for quantification. From each tissue slice, five distinct areas were imaged and analyzed. The following criteria were used to select areas for imaging: [1] located centrally within the tissue sample depth, [2] not obscured by large staining debris and [3] fully impregnated. If the areas met the criteria, a single dendritic length was imaged per area. For dendrite selection, the following criteria were used: [1] unobstructed/isolated/not overlapping other dendrites, [2] length > 30 µm and [3] diameter approximately 1 µm. If more than two dendrites fulfilled the criteria from a single cell, the first dendrite clockwise was the only dendrite selected. Each tissue slice was initially imaged under low 20× magnification to establish the regions of interest and to determine the five distinct areas for dendrite selection. ## 4.8. Immunofluorescence Fresh frozen tissue blocks from human ($$n = 6$$ per group) or mice ($$n = 4$$ per group, 2 males and 2 females) were removed from storage at −80 °C, equilibrated at −20 °C, embedded in O.C.T. (optimal cutting temperature) compound (Tissue-tek). Sections as thick as 12 µm were captured onto Superfrost Plus slides (ThermoFisher Scientific). Prepared slides were stored at −20 °C until use. Slides were fixed in $4\%$ paraformaldehyde in 0.1 M PBS, pH 7.4 for 30 min at room temperature (RT). Non-specific binding sites in the sections were blocked with $5\%$ bovine serum albumin (BSA, ThermoFisher Scientific), $10\%$ normal goat serum (NGS, Millipore Sigma, Burlington, MA, USA). For permeabilizing the sections, $0.5\%$ Triton X-100 and $0.05\%$ Tween-20 for 1 h were used at RT. Slides were incubated with primary antibodies diluted in PBS containing $1.5\%$ NGS with $0.25\%$ Triton X-100 overnight at 4 °C. Following primary antibodies were used: mouse monoclonal Aβ (1:100, 4G8 or Purified anti-β-Amyloid, 17–24 Antibody (Previously Covance catalog #SIG-39220) #800712, Biolegend, RRID:AB_2734548), rabbit anti-PLD1 (1:200, catalog #ab50695, Abcam, RRID, AB_2237051), mouse monoclonal anti-tau 5 for human tissues (1:2000, catalog #806404, Biolegend, RRID: AB_2715857) or chicken polyclonal anti-tau (1:200, catalog #ab75714, RRID: AB_1310734). Before incubation with the Alexa Fluor-conjugated secondary antibodies, slides were washed in PBS 3 times, 10 min each. Following secondary antibodies were used: goat anti-rabbit Alexa Fluor 488 (1:400, catalog #A-11034, RRID:AB_2576217), goat anti-mouse Alexa Fluor 594 (1:400, catalog #A-11032, RRID:AB_2534091), goat anti-chicken Alexa Fluor 594 (1:400, catalog #A11039, RRID:AB_2534099) diluted in PBS containing $1.5\%$ NGS with $0.25\%$ Triton X-100 for 1 h at RT. Finally, slides were washed in PBS 3 times, 10 min each, then treated with $0.3\%$ Sudan Black in $70\%$ EtOH for 10 min to block autofluorescence caused by lipofuscin, washed again with deionized water and coverslipped using fluoromount-G-containing 4′,6′-diamidino-2-phenylindole dihydrochloride (DAPI, SouthernBiotech, Birmingham, AL, USA). ## 4.9. Quantitative Microscopy Images were acquired either with Keyence BZ-X800 microscope, by using 10× and 60× (oil immersion) objectives or IX83 Confocal microscope (Olympus Corporation, Tokyo, Japan) using 40× and 60× (both oil immersion) objectives for all the immunoreacted sections. For each subject, 2 sections were analyzed and 5 images per section were captured (for the mouse slices, 5 images per region of CA1, CA3 and DG were captured). Resolution was kept at 1920 × 1440 pixel with a Z-step of 2 at 12 µm thickness. All layers from a single image stack were projected on a single slice (stack/Z projection) to increase the confidence of co-localization and quantification. Quantitative analysis was performed using ImageJ software (downloaded from NIH; http://imagej.nih.gov/ij), and the intensity of fluorescence for each marker was analyzed as integrated density to account for overall distribution. The co-localization between two markers was evaluated and quantified using the Pearson’s correlation coefficient and Mander’s correlation coefficient in ImageJ. Representative images were composed in an Adobe Photoshop CC2020 format. ## 4.10. Statistics All data are reported as mean ± SEM. Statistical significance was calculated using GraphPad Prism 9.2 (San Diego, CA, USA). All statistical tests were two-tailed, with the threshold for statistical significance set at 0.05. To account for non-normal distribution of data, either non-parametric t-tests (Mann–Whitney U or Wilcoxon rank sum) or one-way ANOVA (Kruskal–Wallis test) followed by Geisser–Greenhouse correction for mixed effects analysis, Holm–Sidak multiple comparisons test with individual variances computed for each comparison or uncorrected Fisher’s LSD with individual variances computed was applied as appropriate to account for variability of differences. Double blinding was performed—this was achieved by one scientist performing the dilutions and providing it to the experimenter conducting the injections and subsequent experiments with a code denoting the different treatments. Once the analysis was completed, the code was broken. ## 5. Conclusions We previously identified a key mechanism that links increased synaptosomal PLD1 to dendritic spine dystrophy resulting in compromised synaptic function. We also showed that the cognitive decline (driven predominantly by Aβ) in 6-month-old 3xTg-AD was attenuated by a repeated VU01 dose. Here, we increased the thorough and multidisciplinary approach to validate the ability of the therapeutic design and we attempted to address the consistency of outcomes at progressive stages of neurodegenerative insult. Our results suggest that there is efficacy of the same regimen in preventing combined progressive insults of Aβ and tau-driven synaptic dysfunction. We also paid attention to treatment-related sex-specific differences, an important consideration in advancing the therapeutics to human clinical stages. Taken together, we find that repeated PLD1 inhibition is a viable therapeutic approach not only at early but also later stages in the progression of ADRD-like synaptic dysfunction and underlying memory deficits since it confers resilience to dendritic spine dystrophy. Additionally, we report, for the first time using immunofluorescence approaches, that there is a distinct decrease in PLD1 expression (increasing our confidence in the brain-penetrant effects of VU01 functionally) that follows a hippocampal subregion preponderance with effects on Aβ and tau with PLD1 co-localization that warrants deeper assessment of protein–protein interaction in future studies that may provide critical insights regarding the success of our therapeutic approaches. Thus, a continued systematic approach incorporating the above aspects will be critical in addressing the complementary nature of PLD1 therapeutic intervention in complementing the immunosuppression therapies against amyloidogenic proteins in effectively preventing the progression of neurodegenerative states at any clinical stage of the disease. ## References 1. Gaugler J., James B., Johnson T., Reimer J., Solis M., Weuve J., Buckley R.F., Hohman T.J.. 2022 Alzheimer’s Disease Facts and Figures (2022.0) 1-122 2. Vina J., Lloret A.. Why women have more Alzheimer’s disease than men: Gender and mitochondrial toxicity of amyloid-beta peptide. J. Alzheimers Dis. (2010.0) 20 S527-S533. DOI: 10.3233/JAD-2010-100501 3. Rocca W.A., Mielke M.M., Vemuri P., Miller V.M.. Sex and gender differences in the causes of dementia: A narrative review. Maturitas (2014.0) 79 196-201. DOI: 10.1016/j.maturitas.2014.05.008 4. Dubal D.B.. Sex difference in Alzheimer’s disease: An updated, balanced and emerging perspective on differing vulnerabilities. Handb. Clin. Neurol (2020.0) 175 261-273. DOI: 10.1016/b978-0-444-64123-6.00018-7 5. Yan Y., Dominguez S., Fisher D.W., Dong H.. Sex differences in chronic stress responses and Alzheimer’s disease. Neurobiol. Stress (2018.0) 8 120-126. DOI: 10.1016/j.ynstr.2018.03.002 6. Fisher D.W., Bennett D.A., Dong H.. Sexual dimorphism in predisposition to Alzheimer’s disease. Neurobiol. Aging (2018.0) 70 308-324. DOI: 10.1016/j.neurobiolaging.2018.04.004 7. Walsh D.M., Klyubin I., Fadeeva J.V., Cullen W.K., Anwyl R., Wolfe M.S., Rowan M.J., Selkoe D.J.. Naturally secreted oligomers of amyloid beta protein potently inhibit hippocampal long-term potentiation in vivo. Nature (2002.0) 416 535-539. DOI: 10.1038/416535a 8. Reese L.C., Zhang W., Dineley K.T., Kayed R., Taglialatela G.. Selective induction of calcineurin activity and signaling by oligomeric amyloid beta. Aging Cell (2008.0) 7 824-835. DOI: 10.1111/j.1474-9726.2008.00434.x 9. Liao D., Miller E.C., Teravskis P.J.. Tau acts as a mediator for Alzheimer’s disease-related synaptic deficits. Eur. J. Neurosci. (2014.0) 39 1202-1213. DOI: 10.1111/ejn.12504 10. Miller E.C., Teravskis P.J., Dummer B.W., Zhao X., Huganir R.L., Liao D.. Tau phosphorylation and tau mislocalization mediate soluble Abeta oligomer-induced AMPA glutamate receptor signaling deficits. Eur. J. Neurosci. (2014.0) 39 1214-1224. DOI: 10.1111/ejn.12507 11. Bilousova T., Miller C.A., Poon W.W., Vinters H.V., Corrada M., Kawas C., Hayden E.Y., Teplow D.B., Glabe C., Albay R.. Synaptic Amyloid-beta Oligomers Precede p-Tau and Differentiate High Pathology Control Cases. Am. J. Pathol. (2016.0) 186 185-198. DOI: 10.1016/j.ajpath.2015.09.018 12. Fa M., Puzzo D., Piacentini R., Staniszewski A., Zhang H., Baltrons M.A., Li Puma D.D., Chatterjee I., Li J., Saeed F.. Extracellular Tau Oligomers Produce An Immediate Impairment of LTP and Memory. Sci. Rep. (2016.0) 6 19393. DOI: 10.1038/srep19393 13. Comerota M.M., Krishnan B., Taglialatela G.. Near infrared light decreases synaptic vulnerability to amyloid beta oligomers. Sci. Rep. (2017.0) 7 15012. DOI: 10.1038/s41598-017-15357-x 14. Comerota M.M., Tumurbaatar B., Krishnan B., Kayed R., Taglialatela G.. Near Infrared Light Treatment Reduces Synaptic Levels of Toxic Tau Oligomers in Two Transgenic Mouse Models of Human Tauopathies. Mol. Neurobiol. Mol. Neurobiol. (2018.0) 56 3341-3355. DOI: 10.1007/s12035-018-1248-9 15. Krishnan B., Kayed R., Taglialatela G.. Elevated phospholipase D isoform 1 in Alzheimer’s disease patients’ hippocampus: Relevance to synaptic dysfunction and memory deficits. Alzheimers Dement. (2018.0) 4 89-102. DOI: 10.1016/j.trci.2018.01.002 16. Kanfer J.N., McCartney D.G.. Phosphatase and phospholipase activity in Alzheimer brain tissues. J. Neural Transmission. Suppl. (1987.0) 24 183-188 17. Kanfer J.N., Pettegrew J.W., Moossy J., McCartney D.G., Bin Q., Strychor S., McKeag D.W.. Possible enzymatic causes for elevated phosphomonoesters in Alzheimer’s disease brain tissue. Ann. N. Y. Acad. Sci. (1991.0) 640 118-121. DOI: 10.1111/j.1749-6632.1991.tb00202.x 18. Kanfer J.N., Pettegrew J.W., Moossy J., McCartney D.G.. Alterations of selected enzymes of phospholipid metabolism in Alzheimer’s disease brain tissue as compared to non-Alzheimer’s demented controls. Neurochem. Res. (1993.0) 18 331-334. DOI: 10.1007/BF00969091 19. Higgins J.P.T., Flicker L.. Lecithin for dementia and cognitive impairment. Cochrane Database Syst. Rev. (2000.0). DOI: 10.1002/14651858.CD001015 20. Vitale N., Caumont A.S., Chasserot-Golaz S., Du G., Wu S., Sciorra V.A., Morris A.J., Frohman M.A., Bader M.F.. Phospholipase D1: A key factor for the exocytotic machinery in neuroendocrine cells. EMBO J. (2001.0) 20 2424-2434. DOI: 10.1093/emboj/20.10.2424 21. Humeau Y., Vitale N., Chasserot-Golaz S., Dupont J.L., Du G., Frohman M.A., Bader M.F., Poulain B.. A role for phospholipase D1 in neurotransmitter release. Proc. Natl. Acad. Sci. USA (2001.0) 98 15300-15305. DOI: 10.1073/pnas.261358698 22. Ammar M.R., Thahouly T., Hanauer A., Stegner D., Nieswandt B., Vitale N.. PLD1 participates in BDNF-induced signalling in cortical neurons. Sci. Rep. (2015.0) 5 14778. DOI: 10.1038/srep14778 23. Zhang Y., Kanaho Y., Frohman M.A., Tsirka S.E.. Phospholipase D1-promoted release of tissue plasminogen activator facilitates neurite outgrowth. J. Neurosci. (2005.0) 25 1797-1805. DOI: 10.1523/JNEUROSCI.4850-04.2005 24. Lu X.C., Zheng J.Y., Tang L.J., Huang B.S., Li K., Tao Y., Yu W., Zhu R.L., Li S., Li L.X.. MiR-133b Promotes neurite outgrowth by targeting RhoA expression. Cell Physiol. Biochem. (2015.0) 35 246-258. DOI: 10.1159/000369692 25. Koshimizu H., Kiyosue K., Hara T., Hazama S., Suzuki S., Uegaki K., Nagappan G., Zaitsev E., Hirokawa T., Tatsu Y.. Multiple functions of precursor BDNF to CNS neurons: Negative regulation of neurite growth, spine formation and cell survival. Mol. Brain (2009.0) 2 27. DOI: 10.1186/1756-6606-2-27 26. Kanaho Y., Funakoshi Y., Hasegawa H.. Phospholipase D signalling and its involvement in neurite outgrowth. Biochim. Biophys. Acta (2009.0) 1791 898-904. DOI: 10.1016/j.bbalip.2009.03.010 27. Cai D., Zhong M., Wang R., Netzer W.J., Shields D., Zheng H., Sisodia S.S., Foster D.A., Gorelick F.S., Xu H.. Phospholipase D1 corrects impaired betaAPP trafficking and neurite outgrowth in familial Alzheimer’s disease-linked presenilin-1 mutant neurons. Proc. Natl. Acad. Sci. USA (2006.0) 103 1936-1940. DOI: 10.1073/pnas.0510710103 28. Burkhardt U., Stegner D., Hattingen E., Beyer S., Nieswandt B., Klein J.. Impaired brain development and reduced cognitive function in phospholipase D-deficient mice. Neurosci. Lett. (2014.0) 572 48-52. DOI: 10.1016/j.neulet.2014.04.052 29. Dall’Armi C., Hurtado-Lorenzo A., Tian H., Morel E., Nezu A., Chan R.B., Yu W.H., Robinson K.S., Yeku O., Small S.A.. The phospholipase D1 pathway modulates macroautophagy. Nat. Commun. (2010.0) 1 142. DOI: 10.1038/ncomms1144 30. Thakur R., Panda A., Coessens E., Raj N., Yadav S., Balakrishnan S., Zhang Q., Georgiev P., Basak B., Pasricha R.. Phospholipase D activity couples plasma membrane endocytosis with retromer dependent recycling. eLife (2016.0) 5 e18515. DOI: 10.7554/eLife.18515 31. Peng X., Frohman M.A.. Mammalian phospholipase D physiological and pathological roles. Acta Physiol. (2012.0) 204 219-226. DOI: 10.1111/j.1748-1716.2011.02298.x 32. Nelson R.K., Frohman M.A.. Physiological and pathophysiological roles for phospholipase D. J. Lipid Res. (2015.0) 56 2229-2237. DOI: 10.1194/jlr.R059220 33. Frohman M.A.. The phospholipase D superfamily as therapeutic targets. Trends Pharmacol. Sci. (2015.0) 36 137-144. DOI: 10.1016/j.tips.2015.01.001 34. Su W., Chen Q., Frohman M.A.. Targeting phospholipase D with small-molecule inhibitors as a potential therapeutic approach for cancer metastasis. Future Oncol. (2009.0) 5 1477-1486. DOI: 10.2217/fon.09.110 35. Brown H.A., Thomas P.G., Lindsley C.W.. Targeting phospholipase D in cancer, infection and neurodegenerative disorders. Nat. Rev. Drug Discov. (2017.0) 16 351-367. DOI: 10.1038/nrd.2016.252 36. Lavieri R.R., Scott S.A., Selvy P.E., Kim K., Jadhav S., Morrison R.D., Daniels J.S., Brown H.A., Lindsley C.W.. Design, synthesis, and biological evaluation of halogenated N-(2-(4-oxo-1-phenyl-1,3,8-triazaspiro[4.5]decan-8-yl)ethyl)benzamides: Discovery of an isoform-selective small molecule phospholipase D2 inhibitor. J. Med. Chem. (2010.0) 53 6706-6719. DOI: 10.1021/jm100814g 37. Lewis J.A., Scott S.A., Lavieri R., Buck J.R., Selvy P.E., Stoops S.L., Armstrong M.D., Brown H.A., Lindsley C.W.. Design and synthesis of isoform-selective phospholipase D (PLD) inhibitors. Part I: Impact of alternative halogenated privileged structures for PLD1 specificity. Bioorganic Med. Chem. Lett. (2009.0) 19 1916-1920. DOI: 10.1016/j.bmcl.2009.02.057 38. Waterson A.G., Scott S.A., Kett N.R., Blobaum A.L., Alex Brown H., Lindsley C.W.. Isoform selective PLD inhibition by novel, chiral 2,8-diazaspirodecan-1-one derivatives. Bioorganic Med. Chem. Lett. (2018.0) 28 3670-3673. DOI: 10.1016/j.bmcl.2018.10.033 39. Jin J.K., Ahn B.H., Na Y.J., Kim J.I., Kim Y.S., Choi E.K., Ko Y.G., Chung K.C., Kozlowski P.B., Min do S.. Phospholipase D1 is associated with amyloid precursor protein in Alzheimer’s disease. Neurobiol. Aging (2007.0) 28 1015-1027. DOI: 10.1016/j.neurobiolaging.2006.05.022 40. Jin J.K., Kim N.H., Min D.S., Kim J.I., Choi J.K., Jeong B.H., Choi S.I., Choi E.K., Carp R.I., Kim Y.S.. Increased expression of phospholipase D1 in the brains of scrapie-infected mice. J. Neurochem. (2005.0) 92 452-461. DOI: 10.1111/j.1471-4159.2004.02881.x 41. Jin J.K., Kim N.H., Lee Y.J., Kim Y.S., Choi E.K., Kozlowski P.B., Park M.H., Kim H.S., Min do S.. Phospholipase D1 is up-regulated in the mitochondrial fraction from the brains of Alzheimer’s disease patients. Neurosci. Lett. (2006.0) 407 263-267. DOI: 10.1016/j.neulet.2006.08.062 42. Bourne K.Z., Natarajan C., Perez C.X.M., Tumurbaatar B., Taglialatela G., Krishnan B.. Suppressing aberrant phospholipase D1 signaling in 3xTg Alzheimer’s disease mouse model promotes synaptic resilience. Sci. Rep. (2019.0) 9 18342. DOI: 10.1038/s41598-019-54974-6 43. Oddo S., Caccamo A., Shepherd J.D., Murphy M.P., Golde T.E., Kayed R., Metherate R., Mattson M.P., Akbari Y., LaFerla F.M.. Triple-transgenic model of Alzheimer’s disease with plaques and tangles: Intracellular Abeta and synaptic dysfunction. Neuron (2003.0) 39 409-421. DOI: 10.1016/S0896-6273(03)00434-3 44. Billings L.M., Oddo S., Green K.N., McGaugh J.L., LaFerla F.M.. Intraneuronal Abeta causes the onset of early Alzheimer’s disease-related cognitive deficits in transgenic mice. Neuron (2005.0) 45 675-688. DOI: 10.1016/j.neuron.2005.01.040 45. Clinton L.K., Billings L.M., Green K.N., Caccamo A., Ngo J., Oddo S., McGaugh J.L., LaFerla F.M.. Age-dependent sexual dimorphism in cognition and stress response in the 3xTg-AD mice. Neurobiol. Dis. (2007.0) 28 76-82. DOI: 10.1016/j.nbd.2007.06.013 46. Stover K.R., Campbell M.A., Van Winssen C.M., Brown R.E.. Early detection of cognitive deficits in the 3xTg-AD mouse model of Alzheimer’s disease. Behav. Brain Res. (2015.0) 289 29-38. DOI: 10.1016/j.bbr.2015.04.012 47. Belfiore R., Rodin A., Ferreira E., Velazquez R., Branca C., Caccamo A., Oddo S.. Temporal and regional progression of Alzheimer’s disease-like pathology in 3xTg-AD mice. Aging Cell (2019.0) 18 e12873. DOI: 10.1111/acel.12873 48. Giménez-Llort L., García Y., Buccieri K., Revilla S., Suñol C., Cristofol R., Sanfeliu C.. Gender-Specific Neuroimmunoendocrine Response to Treadmill Exercise in 3xTg-AD Mice. Int. J. Alzheimers Dis. (2010.0) 2010 128354. DOI: 10.4061/2010/128354 49. Dennison J.L., Ricciardi N.R., Lohse I., Volmar C.H., Wahlestedt C.. Sexual Dimorphism in the 3xTg-AD Mouse Model and Its Impact on Pre-Clinical Research. J. Alzheimers Dis. (2021.0) 80 41-52. DOI: 10.3233/JAD-201014 50. Fertan E., Wong A.A., Vienneau N.A., Brown R.E.. Age and sex differences in motivation and spatial working memory in 3xTg-AD mice in the Hebb-Williams maze. Behav. Brain Res. (2019.0) 370 111937. DOI: 10.1016/j.bbr.2019.111937 51. Mastrangelo M.A., Bowers W.J.. Detailed immunohistochemical characterization of temporal and spatial progression of Alzheimer’s disease-related pathologies in male triple-transgenic mice. BMC Neurosci. (2008.0) 9. DOI: 10.1186/1471-2202-9-81 52. Oddo S., Caccamo A., Kitazawa M., Tseng B.P., LaFerla F.M.. Amyloid deposition precedes tangle formation in a triple transgenic model of Alzheimer’s disease. Neurobiol. Aging (2003.0) 24 1063-1070. DOI: 10.1016/j.neurobiolaging.2003.08.012 53. Duncan M.J., Guerriero L.E., Kohler K., Beechem L.E., Gillis B.D., Salisbury F., Wessel C., Wang J., Sunderam S., Bachstetter A.D.. Chronic Fragmentation of the Daily Sleep-Wake Rhythm Increases Amyloid-beta Levels and Neuroinflammation in the 3xTg-AD Mouse Model of Alzheimer’s Disease. Neuroscience (2022.0) 481 111-122. DOI: 10.1016/j.neuroscience.2021.11.042 54. Mroczko B., Groblewska M., Litman-Zawadzka A., Kornhuber J., Lewczuk P.. Cellular Receptors of Amyloid β Oligomers (AβOs) in Alzheimer’s Disease. Int. J. Mol. Sci. (2018.0) 19. DOI: 10.3390/ijms19071884 55. Shekari A., Fahnestock M.. Cholinergic neurodegeneration in Alzheimer disease mouse models. Handb. Clin. Neurol. (2021.0) 182 191-209. DOI: 10.1016/B978-0-12-819973-2.00013-7 56. Angulo S.L., Orman R., Neymotin S.A., Liu L., Buitrago L., Cepeda-Prado E., Stefanov D., Lytton W.W., Stewart M., Small S.A.. Tau and amyloid-related pathologies in the entorhinal cortex have divergent effects in the hippocampal circuit. Neurobiol. Dis. (2017.0) 108 261-276. DOI: 10.1016/j.nbd.2017.08.015 57. Busche M.A., Wegmann S., Dujardin S., Commins C., Schiantarelli J., Klickstein N., Kamath T.V., Carlson G.A., Nelken I., Hyman B.T.. Tau impairs neural circuits, dominating amyloid-β effects, in Alzheimer models in vivo. Nature Neurosci. (2019.0) 22 57-64. DOI: 10.1038/s41593-018-0289-8 58. Roberson E.D., Scearce-Levie K., Palop J.J., Yan F., Cheng I.H., Wu T., Gerstein H., Yu G.-Q., Mucke L.. Reducing Endogenous Tau Ameliorates Amyloid ß-Induced Deficits in an Alzheimer’s Disease Mouse Model. Science (2007.0) 316 750-754. DOI: 10.1126/science.1141736 59. Ittner L.M., Ke Y.D., Delerue F., Bi M., Gladbach A., van Eersel J., Wölfing H., Chieng B.C., Christie M.J., Napier I.A.. Dendritic Function of Tau Mediates Amyloid-β Toxicity in Alzheimer’s Disease Mouse Models. Cell (2010.0) 142 387-397. DOI: 10.1016/j.cell.2010.06.036 60. DeVos S.L., Goncharoff D.K., Chen G., Kebodeaux C.S., Yamada K., Stewart F.R., Schuler D.R., Maloney S.E., Wozniak D.F., Rigo F.. Antisense Reduction of Tau in Adult Mice Protects against Seizures. J. Neurosci. (2013.0) 33 12887-12897. DOI: 10.1523/JNEUROSCI.2107-13.2013 61. Cai D., Netzer W.J., Zhong M., Lin Y., Du G., Frohman M., Foster D.A., Sisodia S.S., Xu H., Gorelick F.S.. Presenilin-1 uses phospholipase D1 as a negative regulator of beta-amyloid formation. Proc. Natl. Acad. Sci. USA (2006.0) 103 1941-1946. DOI: 10.1073/pnas.0510708103 62. Winner B., Winkler J.. Adult neurogenesis in neurodegenerative diseases. Cold Spring Harb. Perspect. Biol. (2015.0) 7 a021287. DOI: 10.1101/cshperspect.a021287 63. Hollands C., Bartolotti N., Lazarov O.. Alzheimer’s Disease and Hippocampal Adult Neurogenesis; Exploring Shared Mechanisms. Front. Neurosci. (2016.0) 10 178. DOI: 10.3389/fnins.2016.00178 64. Lee S.K., Kim Y.S., Bae G.H., Lee H.Y., Bae Y.-S.. VU0155069 inhibits inflammasome activation independent of phospholipase D1 activity. Sci. Rep. (2019.0) 9 14349. DOI: 10.1038/s41598-019-50806-9 65. Klein W.L.. Synaptic targeting by A beta oligomers (ADDLS) as a basis for memory loss in early Alzheimer’s disease. Alzheimers Dement. (2006.0) 2 43-55. DOI: 10.1016/j.jalz.2005.11.003 66. Lambert M.P., Barlow A.K., Chromy B.A., Edwards C., Freed R., Liosatos M., Morgan T.E., Rozovsky I., Trommer B., Viola K.L.. Diffusible, nonfibrillar ligands derived from Abeta1-42 are potent central nervous system neurotoxins. Proc. Natl. Acad. Sci. USA (1998.0) 95 6448-6453. DOI: 10.1073/pnas.95.11.6448 67. Spires T.L., Meyer-Luehmann M., Stern E.A., McLean P.J., Skoch J., Nguyen P.T., Bacskai B.J., Hyman B.T.. Dendritic spine abnormalities in amyloid precursor protein transgenic mice demonstrated by gene transfer and intravital multiphoton microscopy. J. Neurosci. (2005.0) 25 7278-7287. DOI: 10.1523/JNEUROSCI.1879-05.2005 68. Spires-Jones T.L., Meyer-Luehmann M., Osetek J.D., Jones P.B., Stern E.A., Bacskai B.J., Hyman B.T.. Impaired spine stability underlies plaque-related spine loss in an Alzheimer’s disease mouse model. Am. J. Pathol. (2007.0) 171 1304-1311. DOI: 10.2353/ajpath.2007.070055 69. Yu W., Polepalli J., Wagh D., Rajadas J., Malenka R., Lu B.. A critical role for the PAR-1/MARK-tau axis in mediating the toxic effects of Abeta on synapses and dendritic spines. Hum. Mol. Genet. (2012.0) 21 1384-1390. DOI: 10.1093/hmg/ddr576 70. Dickstein D.L., Brautigam H., Stockton S.D., Schmeidler J., Hof P.R.. Changes in dendritic complexity and spine morphology in transgenic mice expressing human wild-type tau. Brain Struct. Funct. (2010.0) 214 161-179. DOI: 10.1007/s00429-010-0245-1 71. Sydow A., Van der Jeugd A., Zheng F., Ahmed T., Balschun D., Petrova O., Drexler D., Zhou L., Rune G., Mandelkow E.. Tau-induced defects in synaptic plasticity, learning, and memory are reversible in transgenic mice after switching off the toxic Tau mutant. J. Neurosci. (2011.0) 31 2511-2525. DOI: 10.1523/JNEUROSCI.5245-10.2011 72. Dubal D.B.. The Way of Tau: Secretion and Synaptic Dysfunction. Trends Mol. Med. (2018.0) 24 595-597. DOI: 10.1016/j.molmed.2018.05.006 73. Li K., Wei Q., Liu F.F., Hu F., Xie A.J., Zhu L.Q., Liu D.. Synaptic Dysfunction in Alzheimer’s Disease: Abeta, Tau, and Epigenetic Alterations. Mol. Neurobiol. (2018.0) 55 3021-3032. DOI: 10.1007/s12035-017-0533-3 74. Tracy T.E., Gan L.. Tau-mediated synaptic and neuronal dysfunction in neurodegenerative disease. Curr. Opin. Neurobiol. (2018.0) 51 134-138. DOI: 10.1016/j.conb.2018.04.027 75. Shankar G.M., Li S., Mehta T.H., Garcia-Munoz A., Shepardson N.E., Smith I., Brett F.M., Farrell M.A., Rowan M.J., Lemere C.A.. Amyloid-beta protein dimers isolated directly from Alzheimer’s brains impair synaptic plasticity and memory. Nat. Med. (2008.0) 14 837-842. DOI: 10.1038/nm1782 76. Micci M.A., Krishnan B., Bishop E., Zhang W.R., Guptarak J., Grant A., Zolochevska O., Tumurbaatar B., Franklin W., Marino C.. Hippocampal stem cells promotes synaptic resistance to the dysfunctional impact of amyloid beta oligomers via secreted exosomes. Mol. Neurodegener (2019.0) 14 25. DOI: 10.1186/s13024-019-0322-8 77. Marino C., Krishnan B., Cappello F., Taglialatela G.. Hsp60 Protects against Amyloid beta Oligomer Synaptic Toxicity via Modification of Toxic Oligomer Conformation. ACS Chem. Neurosci. (2019.0) 10 2858-2867. DOI: 10.1021/acschemneuro.9b00086 78. Franklin W., Krishnan B., Taglialatela G.. Chronic synaptic insulin resistance after traumatic brain injury abolishes insulin protection from amyloid beta and tau oligomer-induced synaptic dysfunction. Sci. Rep. (2019.0) 9 8228. DOI: 10.1038/s41598-019-44635-z 79. Kimura T., Whitcomb D.J., Jo J., Regan P., Piers T., Heo S., Brown C., Hashikawa T., Murayama M., Seok H.. Microtubule-associated protein tau is essential for long-term depression in the hippocampus. Philos. Trans. R. Soc. Lond. Biol. Sci. (2014.0) 369 20130144. DOI: 10.1098/rstb.2013.0144 80. Regan P., Piers T., Yi J.-H., Kim D.-H., Huh S., Park S.J., Ryu J.H., Whitcomb D.J., Cho K.. Tau phosphorylation at serine 396 residue is required for hippocampal LTD. J. Neurosci. Off. J. Soc. Neurosci. (2015.0) 35 4804-4812. DOI: 10.1523/JNEUROSCI.2842-14.2015 81. Palop J.J., Chin J., Mucke L.. A network dysfunction perspective on neurodegenerative diseases. Nature (2006.0) 443 768-773. DOI: 10.1038/nature05289 82. Palop J.J., Mucke L.. Synaptic depression and aberrant excitatory network activity in Alzheimer’s disease: Two faces of the same coin?. Neuromol. Med. (2010.0) 12 48-55. DOI: 10.1007/s12017-009-8097-7 83. Falke E., Nissanov J., Mitchell T.W., Bennett D.A., Trojanowski J.Q., Arnold S.E.. Subicular dendritic arborization in Alzheimer’s disease correlates with neurofibrillary tangle density. Am. J. Pathol. (2003.0) 163 1615-1621. DOI: 10.1016/s0002-9440(10)63518-3 84. Mijalkov M., Volpe G., Fernaud-Espinosa I., DeFelipe J., Pereira J.B., Merino-Serrais P.. Dendritic spines are lost in clusters in Alzheimer’s disease. Sci. Rep. (2021.0) 11 12350. DOI: 10.1038/s41598-021-91726-x 85. Izquierdo I., Furini C.R., Myskiw J.C.. Fear Memory. Physiol. Rev. (2016.0) 96 695-750. DOI: 10.1152/physrev.00018.2015 86. Rasool S., Martinez-Coria H., Wu J.W., LaFerla F., Glabe C.G.. Systemic vaccination with anti-oligomeric monoclonal antibodies improves cognitive function by reducing Aβ deposition and tau pathology in 3xTg-AD mice. J. Neurochem. (2013.0) 126 473-482. DOI: 10.1111/jnc.12305 87. Montoliu-Gaya L., Esquerda-Canals G., Bronsoms S., Villegas S.. Production of an anti-Aβ antibody fragment in Pichia pastoris and in vitro and in vivo validation of its therapeutic effect. PLoS ONE (2017.0) 12. DOI: 10.1371/journal.pone.0181480 88. Herline K., Prelli F., Mehta P., MacMurray C., Goñi F., Wisniewski T.. Immunotherapy to improve cognition and reduce pathological species in an Alzheimer’s disease mouse model. Alzheimers Res. Ther. (2018.0) 10 54. DOI: 10.1186/s13195-018-0384-9 89. Montoliu-Gaya L., Güell-Bosch J., Esquerda-Canals G., Roda A.R., Serra-Mir G., Lope-Piedrafita S., Sánchez-Quesada J.L., Villegas S.. Differential effects of apoE and apoJ mimetic peptides on the action of an anti-Aβ scFv in 3xTg-AD mice. Biochem. Pharmacol. (2018.0) 155 380-392. DOI: 10.1016/j.bcp.2018.07.012 90. Rosenberg R.N., Fu M., Lambracht-Washington D.. Active full-length DNA Aβ(42) immunization in 3xTg-AD mice reduces not only amyloid deposition but also tau pathology. Alzheimers Res. Ther. (2018.0) 10 115. DOI: 10.1186/s13195-018-0441-4 91. Acero G., Garay C., Venegas D., Ortega E., Gevorkian G.. Novel monoclonal antibody 3B8 specifically recognizes pyroglutamate-modified amyloid β 3-42 peptide in brain of AD patients and 3xTg-AD transgenic mice. Neurosci. Lett. (2020.0) 724 134876. DOI: 10.1016/j.neulet.2020.134876 92. Kent S.A., Spires-Jones T.L., Durrant C.S.. The physiological roles of tau and Aβ: Implications for Alzheimer’s disease pathology and therapeutics. Acta Neuropathol. (2020.0) 140 417-447. DOI: 10.1007/s00401-020-02196-w 93. Jin J.K., Jang B., Jin H.T., Choi E.K., Jung C.G., Akatsu H., Kim J.I., Carp R.I., Kim Y.S.. Phosphatidylinositol-glycan-phospholipase D is involved in neurodegeneration in prion disease. PLoS ONE (2015.0) 10. DOI: 10.1371/journal.pone.0122120 94. Bae E.J., Lee H.J., Jang Y.H., Michael S., Masliah E., Min D.S., Lee S.J.. Phospholipase D1 regulates autophagic flux and clearance of alpha-synuclein aggregates. Cell Death Differ. (2014.0) 21 1132-1141. DOI: 10.1038/cdd.2014.30 95. Nakagawa H., Hazama K., Ishida K., Komori M., Nishimura K., Matsuo S.. Inhibition of PLD1 activity causes ER stress via regulation of COPII vesicle formation. Biochem. Biophys. Res. Commun. (2017.0) 490 895-900. DOI: 10.1016/j.bbrc.2017.06.137 96. Soreng K., Knaevelsrud H., Holland P., Simonsen A.. HS1BP3 inhibits autophagy by regulation of PLD1. Autophagy (2017.0) 13 985-986. DOI: 10.1080/15548627.2017.1291483 97. Brandenburg L.O., Pufe T., Koch T.. Role of phospholipase d in g-protein coupled receptor function. Membranes (2014.0) 4 302-318. DOI: 10.3390/membranes4030302 98. Gavin A.L., Huang D., Huber C., Mårtensson A., Tardif V., Skog P.D., Blane T.R., Thinnes T.C., Osborn K., Chong H.S.. PLD3 and PLD4 are single-stranded acid exonucleases that regulate endosomal nucleic-acid sensing. Nat. Immunol. (2018.0) 19 942-953. DOI: 10.1038/s41590-018-0179-y 99. Hwang W.C., Seo S.H., Kang M., Kang R.H., Di Paolo G., Choi K.Y., Min D.S.. PLD1 and PLD2 differentially regulate the balance of macrophage polarization in inflammation and tissue injury. J. Cell. Physiol. (2021.0) 236 5193-5211. DOI: 10.1002/jcp.30224 100. Lee S.M., Choi H.J., Oh C.H., Oh J.W., Han J.S.. Leptin increases TNF-alpha expression and production through phospholipase D1 in Raw 264.7 cells. PLoS ONE (2014.0) 9. DOI: 10.1371/journal.pone.0102373 101. Lee Y.Y., Lee S.Y., Park S.-Y., Choi H.-J., Kim E.-G., Han J.-S.. Therapeutic potential of a phospholipase D1 inhibitory peptide fused with a cell-penetrating peptide as a novel anti-asthmatic drug in a Der f 2-induced airway inflammation model. Exp. Mol. Med. (2018.0) 50 1-11. DOI: 10.1038/s12276-018-0083-4 102. Urbahn M.A., Kaup S.C., Reusswig F., Kruger I., Spelleken M., Jurk K., Klier M., Lang P.A., Elvers M.. Phospholipase D1 regulation of TNF-alpha protects against responses to LPS. Sci. Rep. (2018.0) 8 10006. DOI: 10.1038/s41598-018-28331-y 103. Cai M., Wang Z., Luu T.T.T., Zhang D., Finke B., He J., Tay L.W.R., Di Paolo G., Du G.. PLD1 promotes reactive oxygen species production in vascular smooth muscle cells and injury-induced neointima formation. Biochim. Biophys. Acta Mol. Cell Biol. Lipids (2022.0) 1867 159062. DOI: 10.1016/j.bbalip.2021.159062 104. Usatyuk P.V., Kotha S.R., Parinandi N.L., Natarajan V.. Phospholipase D signaling mediates reactive oxygen species-induced lung endothelial barrier dysfunction. Pulm Circ. (2013.0) 3 108-115. DOI: 10.4103/2045-8932.109925 105. Suryadevara V., Huang L., Kim S.J., Cheresh P., Shaaya M., Bandela M., Fu P., Feghali-Bostwick C., Di Paolo G., Kamp D.W.. Role of phospholipase D in bleomycin-induced mitochondrial reactive oxygen species generation, mitochondrial DNA damage, and pulmonary fibrosis. Am. J. Physiol. Lung Cell Mol. Physiol. (2019.0) 317 L175-L187. DOI: 10.1152/ajplung.00320.2018 106. Kim J., Min G., Bae Y.S., Min D.S.. Phospholipase D is involved in oxidative stress-induced migration of vascular smooth muscle cells via tyrosine phosphorylation and protein kinase C. Exp. Mol. Med. (2004.0) 36 103-109. DOI: 10.1038/emm.2004.15 107. Oguin T.H., Sharma S., Stuart A.D., Duan S., Scott S.A., Jones C.K., Daniels J.S., Lindsley C.W., Thomas P.G., Brown H.A.. Phospholipase D facilitates efficient entry of influenza virus, allowing escape from innate immune inhibition. J. Biol. Chem. (2014.0) 289 25405-25417. DOI: 10.1074/jbc.M114.558817 108. Ghossoub R., Lembo F., Rubio A., Gaillard C.B., Bouchet J., Vitale N., Slavik J., Machala M., Zimmermann P.. Syntenin-ALIX exosome biogenesis and budding into multivesicular bodies are controlled by ARF6 and PLD2. Nat. Commun. (2014.0) 5 3477. DOI: 10.1038/ncomms4477 109. Han X., Yu R., Ji L., Zhen D., Tao S., Li S., Sun Y., Huang L., Feng Z., Li X.. InlB-mediated Listeria monocytogenes internalization requires a balanced phospholipase D activity maintained through phospho-cofilin. Mol. Microbiol. (2011.0) 81 860-880. DOI: 10.1111/j.1365-2958.2011.07726.x 110. Koffie R.M., Meyer-Luehmann M., Hashimoto T., Adams K.W., Mielke M.L., Garcia-Alloza M., Micheva K.D., Smith S.J., Kim M.L., Lee V.M.. Oligomeric amyloid beta associates with postsynaptic densities and correlates with excitatory synapse loss near senile plaques. Proc. Natl. Acad. Sci. USA (2009.0) 106 4012-4017. DOI: 10.1073/pnas.0811698106 111. Tai H.C., Serrano-Pozo A., Hashimoto T., Frosch M.P., Spires-Jones T.L., Hyman B.T.. The synaptic accumulation of hyperphosphorylated tau oligomers in Alzheimer disease is associated with dysfunction of the ubiquitin-proteasome system. Am. J. Pathol. (2012.0) 181 1426-1435. DOI: 10.1016/j.ajpath.2012.06.033 112. Wei W., Nguyen L.N., Kessels H.W., Hagiwara H., Sisodia S., Malinow R.. Amyloid beta from axons and dendrites reduces local spine number and plasticity. Nat. Neurosci. (2010.0) 13 190-196. DOI: 10.1038/nn.2476 113. Jaworski T., Lechat B., Demedts D., Gielis L., Devijver H., Borghgraef P., Duimel H., Verheyen F., Kügler S., Van Leuven F.. Dendritic degeneration, neurovascular defects, and inflammation precede neuronal loss in a mouse model for tau-mediated neurodegeneration. Am. J. Pathol. (2011.0) 179 2001-2015. DOI: 10.1016/j.ajpath.2011.06.025 114. Bliss T.V., Collingridge G.L.. A synaptic model of memory: Long-term potentiation in the hippocampus. Nature (1993.0) 361 31-39. DOI: 10.1038/361031a0 115. Bliss T.V., Cooke S.F.. Long-term potentiation and long-term depression: A clinical perspective. Clinics (2011.0) 66 3-17. DOI: 10.1590/s1807-59322011001300002 116. Vitolo O.V., Sant’Angelo A., Costanzo V., Battaglia F., Arancio O., Shelanski M.. Amyloid beta -peptide inhibition of the PKA/CREB pathway and long-term potentiation: Reversibility by drugs that enhance cAMP signaling. Proc. Natl. Acad. Sci. USA (2002.0) 99 13217-13221. DOI: 10.1073/pnas.172504199 117. Puzzo D., Privitera L., Leznik E., Fà M., Staniszewski A., Palmeri A., Arancio O.. Picomolar Amyloid-β Positively Modulates Synaptic Plasticity and Memory in Hippocampus. J. Neurosci. (2008.0) 28 14537-14545. DOI: 10.1523/JNEUROSCI.2692-08.2008 118. Puzzo D., Privitera L., Fa M., Staniszewski A., Hashimoto G., Aziz F., Sakurai M., Ribe E.M., Troy C.M., Mercken M.. Endogenous amyloid-β is necessary for hippocampal synaptic plasticity and memory. Ann. Neurol. (2011.0) 69 819-830. DOI: 10.1002/ana.22313 119. McIntire L.B., Berman D.E., Myaeng J., Staniszewski A., Arancio O., Di Paolo G., Kim T.W.. Reduction of synaptojanin 1 ameliorates synaptic and behavioral impairments in a mouse model of Alzheimer’s disease. J. Neurosci. (2012.0) 32 15271-15276. DOI: 10.1523/JNEUROSCI.2034-12.2012 120. Walsh D.M., Selkoe D.J.. Deciphering the molecular basis of memory failure in Alzheimer’s disease. Neuron (2004.0) 44 181-193. DOI: 10.1016/j.neuron.2004.09.010 121. Lei M., Xu H., Li Z., Wang Z., O’Malley T.T., Zhang D., Walsh D.M., Xu P., Selkoe D.J., Li S.. Soluble Abeta oligomers impair hippocampal LTP by disrupting glutamatergic/GABAergic balance. Neurobiol. Dis. (2016.0) 85 111-121. DOI: 10.1016/j.nbd.2015.10.019 122. Shankar G.M., Walsh D.M.. Alzheimer’s disease: Synaptic dysfunction and Abeta. Mol. Neurodegener. (2009.0) 4 48. DOI: 10.1186/1750-1326-4-48 123. Henstridge C.M., Pickett E., Spires-Jones T.L.. Synaptic pathology: A shared mechanism in neurological disease. Ageing Res. Rev. (2016.0) 28 72-84. DOI: 10.1016/j.arr.2016.04.005 124. Pasinelli P., Ramakers G.M., Urban I.J., Hens J.J., Oestreicher A.B., de Graan P.N., Gispen W.H.. Long-term potentiation and synaptic protein phosphorylation. Behav. Brain Res. (1995.0) 66 53-59. DOI: 10.1016/0166-4328(94)00124-x 125. Patel H., Zamani R.. The role of PKMζ in the maintenance of long-term memory: A review. Rev. Neurosci. (2021.0) 32 481-494. DOI: 10.1515/revneuro-2020-0105 126. Sanhueza M., Lisman J.. The CaMKII/NMDAR complex as a molecular memory. Mol. Brain (2013.0) 6 10. DOI: 10.1186/1756-6606-6-10 127. Norris C.M., Blalock E.M., Chen K.C., Porter N.M., Landfield P.W.. Calcineurin enhances L-type Ca(2+) channel activity in hippocampal neurons: Increased effect with age in culture. Neuroscience (2002.0) 110 213-225. DOI: 10.1016/S0306-4522(01)00574-7 128. Martin Z.S., Neugebauer V., Dineley K.T., Kayed R., Zhang W., Reese L.C., Taglialatela G.. alpha-Synuclein oligomers oppose long-term potentiation and impair memory through a calcineurin-dependent mechanism: Relevance to human synucleopathic diseases. J. Neurochem. (2012.0) 120 440-452. DOI: 10.1111/j.1471-4159.2011.07576.x 129. Yamin G.. NMDA receptor-dependent signaling pathways that underlie amyloid beta-protein disruption of LTP in the hippocampus. J. Neurosci. Res. (2009.0) 87 1729-1736. DOI: 10.1002/jnr.21998 130. Li S., Hong S., Shepardson N.E., Walsh D.M., Shankar G.M., Selkoe D.. Soluble oligomers of amyloid Beta protein facilitate hippocampal long-term depression by disrupting neuronal glutamate uptake. Neuron (2009.0) 62 788-801. DOI: 10.1016/j.neuron.2009.05.012 131. Collingridge G.L., Peineau S., Howland J.G., Wang Y.T.. Long-term depression in the CNS. Nat. Rev. Neurosci. (2010.0) 11 459-473. DOI: 10.1038/nrn2867 132. Spires-Jones T.L., Hyman B.T.. The intersection of amyloid beta and tau at synapses in Alzheimer’s disease. Neuron (2014.0) 82 756-771. DOI: 10.1016/j.neuron.2014.05.004 133. Javonillo D.I., Tran K.M., Phan J., Hingco E., Kramár E.A., da Cunha C., Forner S., Kawauchi S., Milinkeviciute G., Gomez-Arboledas A.. Systematic Phenotyping and Characterization of the 3xTg-AD Mouse Model of Alzheimer’s Disease. Front. Neurosci. (2022.0) 15 785276. DOI: 10.3389/fnins.2021.785276 134. Oblak A.L., Forner S., Territo P.R., Sasner M., Carter G.W., Howell G.R., Sukoff-Rizzo S.J., Logsdon B.A., Mangravite L.M., Mortazavi A.. Model organism development and evaluation for late-onset Alzheimer’s disease: MODEL-AD. Alzheimers Dement. (2020.0) 6 e12110. DOI: 10.1002/trc2.12110 135. Parhizkar S., Holtzman D.M.. APOE mediated neuroinflammation and neurodegeneration in Alzheimer’s disease. Semin. Immunol. (2022.0) 59 101594. DOI: 10.1016/j.smim.2022.101594 136. Sienski G., Narayan P., Bonner J.M., Kory N., Boland S., Arczewska A.A., Ralvenius W.T., Akay L., Lockshin E., He L.. APOE4 disrupts intracellular lipid homeostasis in human iPSC-derived glia. Sci. Transl. Med. (2021.0) 13 aaz4564. DOI: 10.1126/scitranslmed.aaz4564 137. Grone B., Zalocusky K., Hao Y., Yoon S.Y., Arriola P., Huang Y.. Early and lifelong effects of APOE4 on neuronal gene expression networks relevant to Alzheimer’s disease. bioRxiv 2022. DOI: 10.1101/2022.06.16.496371 138. Hirata-Fukae C., Li H.F., Hoe H.S., Gray A.J., Minami S.S., Hamada K., Niikura T., Hua F., Tsukagoshi-Nagai H., Horikoshi-Sakuraba Y.. Females exhibit more extensive amyloid, but not tau, pathology in an Alzheimer transgenic model. Brain Res. (2008.0) 1216 92-103. DOI: 10.1016/j.brainres.2008.03.079 139. Stimmell A.C., Baglietto-Vargas D., Moseley S.C., Lapointe V., Thompson L.M., LaFerla F.M., McNaughton B.L., Wilber A.A.. Impaired Spatial Reorientation in the 3xTg-AD Mouse Model of Alzheimer’s Disease. Sci. Rep. (2019.0) 9 1311. DOI: 10.1038/s41598-018-37151-z 140. Creighton S.D., Mendell A.L., Palmer D., Kalisch B.E., MacLusky N.J., Prado V.F., Prado M.A.M., Winters B.D.. Dissociable cognitive impairments in two strains of transgenic Alzheimer’s disease mice revealed by a battery of object-based tests. Sci. Rep. (2019.0) 9 57. DOI: 10.1038/s41598-018-37312-0 141. Omori C., Motodate R., Shiraki Y., Chiba K., Sobu Y., Kimura A., Nakaya T., Kondo H., Kurumiya S., Tanaka T.. Facilitation of brain mitochondrial activity by 5-aminolevulinic acid in a mouse model of Alzheimer’s disease. Nutr. Neurosci. (2017.0) 20 538-546. DOI: 10.1080/1028415X.2016.1199114 142. Blázquez G., Cañete T., Tobeña A., Giménez-Llort L., Fernández-Teruel A.. Cognitive and emotional profiles of aged Alzheimer’s disease (3×TgAD) mice: Effects of environmental enrichment and sexual dimorphism. Behav. Brain Res. (2014.0) 268 185-201. DOI: 10.1016/j.bbr.2014.04.008 143. Stevens L.M., Brown R.E.. Reference and working memory deficits in the 3xTg-AD mouse between 2 and 15-months of age: A cross-sectional study. Behav. Brain Res. (2015.0) 278 496-505. DOI: 10.1016/j.bbr.2014.10.033 144. Juottonen K., Laakso M.P., Insausti R., Lehtovirta M., Pitkänen A., Partanen K., Soininen H.. Volumes of the entorhinal and perirhinal cortices in Alzheimer’s disease. Neurobiol. Aging (1998.0) 19 15-22. DOI: 10.1016/S0197-4580(98)00007-4 145. Cohen S.J., Stackman R.W.. Assessing rodent hippocampal involvement in the novel object recognition task. A review. Behav. Brain Res. (2015.0) 285 105-117. DOI: 10.1016/j.bbr.2014.08.002 146. Garber J.C.. Guide for the Care and Use of Laboratory Animals (2011.0). DOI: 10.17226/12910 147. Krishnan B., Sallam H.S., Tumurbataar B., Saieva S., Baymon D.M., Tuvdendorj D., Micci M.A., Abate N., Taglialatela G.. Amelioration of hippocampal dysfunction by adipose tissue-targeted stem cell transplantation in a mouse model of type 2 diabetes. J. Neurochem. (2020.0) 153 51-62. DOI: 10.1111/jnc.14915 148. Krishnan B., Natarajan C., Bourne K.Z., Alikhani L., Wang J., Sowa A., Groen K., Perry B., Dickstein D.L., Baulch J.E.. Chronic Low Dose Neutron Exposure Results in Altered Neurotransmission Properties of the Hippocampus-Prefrontal Cortex Axis in Both Mice and Rats. Int. J. Mol. Sci. (2021.0) 22 3668. PMID: 33915974 149. Mathis D.M., Furman J.L., Norris C.M.. Preparation of acute hippocampal slices from rats and transgenic mice for the study of synaptic alterations during aging and amyloid pathology. J. Vis. Exp. (2011.0) 46 e2330. DOI: 10.3791/2330 150. Dineley K.T., Kayed R., Neugebauer V., Fu Y., Zhang W., Reese L.C., Taglialatela G.. Amyloid-beta oligomers impair fear conditioned memory in a calcineurin-dependent fashion in mice. J. Neurosci. Res. (2010.0) 88 2923-2932. DOI: 10.1002/jnr.22445 151. Smith D.L., Pozueta J., Gong B., Arancio O., Shelanski M.. Reversal of long-term dendritic spine alterations in Alzheimer disease models. Proc. Natl. Acad. Sci. USA (2009.0) 106 16877-16882. DOI: 10.1073/pnas.0908706106 152. Boros B.D., Greathouse K.M., Gentry E.G., Curtis K.A., Birchall E.L., Gearing M., Herskowitz J.H.. Dendritic spines provide cognitive resilience against Alzheimer’s disease. Ann. Neurol. (2017.0) 82 602-614. DOI: 10.1002/ana.25049
--- title: 'Tergitol Based Decellularization Protocol Improves the Prerequisites for Pulmonary Xenografts: Characterization and Biocompatibility Assessment' authors: - Susanna Tondato - Arianna Moro - Salman Butt - Martina Todesco - Deborah Sandrin - Giulia Borile - Massimo Marchesan - Assunta Fabozzo - Andrea Bagno - Filippo Romanato - Saima Jalil Imran - Gino Gerosa journal: Polymers year: 2023 pmcid: PMC9967102 doi: 10.3390/polym15040819 license: CC BY 4.0 --- # Tergitol Based Decellularization Protocol Improves the Prerequisites for Pulmonary Xenografts: Characterization and Biocompatibility Assessment ## Abstract Right ventricle outflow tract obstruction (RVOTO) is a congenital pathological condition that contributes to about $15\%$ of congenital heart diseases. In most cases, the replacement of the right ventricle outflow in pediatric age requires subsequent pulmonary valve replacement in adulthood. The aim of this study was to investigate the extracellular matrix scaffold obtained by decellularization of the porcine pulmonary valve using a new detergent (Tergitol) instead of Triton X-100. The decellularized scaffold was evaluated for the integrity of its extracellular matrix (ECM) structure by testing for its biochemical and mechanical properties, and the cytotoxicity/cytocompatibility of decellularized tissue was assessed using bone marrow-derived mesenchymal stem cells. We concluded that Tergitol could remove the nuclear material efficiently while preserving the structural proteins of the matrix, but without an efficient removal of the alpha-gal antigenic epitope. Therefore, Tergitol can be used as an alternative detergent to replace the Triton X-100. ## 1. Introduction Congenital heart diseases (CHDs) are among the most life-threatening factors for the pediatric population worldwide. Approximately $0.8\%$ to $1.2\%$ of live births are affected by CHDs, among which Right Ventricle Outflow Tract Obstruction (RVOTO) contributes to $15\%$ of total congenital heart diseases [1,2,3,4,5]. With the remarkable progress in surgical approaches, the percentage of CHD patients who reach adulthood is increasing over time [6]. Due to the lack of an ideal valve substitute, the search for a more efficient replacement intervention with higher durability is not yet completed. Bioprostheses are the most frequently used devices for the replacement of pulmonary valves, albeit their limited durability and residual immunogenicity cause mid to long-term degeneration [7,8]. On the other hand, mechanical prostheses alter the coagulative status, requiring the administration of permanent anticoagulation therapy [9,10]. Recently, nanofibers development has made a major contribution to the scope of scaffolds for fabrication that can potentially meet this challenge in combination with an acellular scaffold. Regardless of their fabrication criteria, nanofibers have been used as a scaffold for several applications, e.g., musculoskeletal, heart valve, skin, neural, and vascular tissue engineering [11,12,13]. Tranquillo et al. reported for the first time a tubular heart valve fabricated from two decellularized, engineered tissue tubes connected with absorbable sutures, which met this need, in principle [14]. Tissue-engineered biological heart valve substitutes, with the lowest or no immunogenicity, without the concomitant medical treatment with vast availability and durability, are the ideal to achieve this purpose [15]. Based on the clinical data, homograft is the most favorable, and unfortunately, they have limited availability [16,17]. The substitute, which comes closest to the ideal characteristics, is the decellularized pulmonary xenograft [18]. In general, the decellularized valve should be similar to a scaffold, which maintains the functionality of the native valve and with higher potential for remodeling [19,20]. To obtain an optimal tissue-engineered substitute, the first aim is to find the ideal decellularization protocol that can assure low toxicity and very low damage to tissue structure and composition. So far, several decellularization procedures have been reported in the literature [21,22], but most of them are characterized by several limitations, for example, incomplete cell removal, the toxicity of residual detergents, persisting host immune response to the xenograft, thus resulting in a mid-long-term degeneration and calcification [23,24]. The most common detergents applied in the previously reported protocols were SDS, Triton X-100, and Sodium Cholate. Indeed, Triton X-100 has been reported as a hazardous substance for the human endocrine system within the list published by the European Chemical Agency (ECHA) since 4 January 2021 [25]. The specific aim of the present study is to assess the preservation of the ECM scaffold of the decellularized porcine pulmonary valve by using a new eco-friendly detergent, i.e., Tergitol. Our group has reported Tergitol application for the decellularization of the aortic valve [26]. The decellularization protocol includes several treatments, including protease inhibitors’ treatment and the application of Tergitol and sodium cholate detergents in various steps of washings with hyper and hypotonic solutions. The protocol was applied to porcine pulmonary valves. Subsequently, the processed tissues were characterized with regard to their structural and mechanical behavior. Acellular matrices are prone to cell attachment: this proves the biocompatibility of the decellularized scaffold and the possibility for tissue integration and remodeling in vivo. A well-preserved structure can provide a favorable environment for cell repopulation: the crosstalk between ECM and cells can also promote tissue structural organization, directing cell functions. Understanding the process underlying the crosstalk between cells and ECM will lead to the design of the remodeling process [27,28]. ## 2.1. Samples Processing Fresh porcine hearts ($$n = 26$$) were obtained from a local slaughterhouse (F.lli Guerriero, Villafranca Padovana, Padova, Italy) following the protocols consistent with EC regulations $\frac{1099}{2009}$ regarding animal wealth and protection. After removal from pigs (adults 6–8 months old), hearts were brought to the laboratory within 1–2 h, and then pulmonary valves were dissected and cleaned using isotonic saline. The average 3 cm length of the pulmonary wall was obtained with the leaflets. Samples obtained were frozen at −80 °C. ## 2.2. Decellularization Protocol The Tergitol-based decellularization protocol was modified from the TRICOL protocol previously reported [19]. Decellularization protocol was performed in an agitation system: treatment with protease inhibitors cocktail ($1\%$ v/v) and DMSO (supplier) ($10\%$) at 4 °C (8 h) was followed by washing with a hypotonic solution (12 h). Subsequently, a second phase with protease inhibitors in combination with $1\%$ Tergitol (12 h) (Sigma Aldrich, Saint Louis, MO, USA) was carried out at room temperature. After further washing, tissues were treated with Tergitol ($0.1\%$) in a hypertonic solution (24 h in two cycles). Thereafter, they were treated for 20 h with sodium cholate (Sigma Aldrich, $4\%$ v/v). Finally, tissue samples were treated with peracetic acid (Sigma Aldrich, $0.1\%$) and ethanol $4\%$ (Carlo Erba, Cornaredo, MI, Italy) solutions (90′) for bioburden removal and primary decontamination. Valves were cut into 8 mm patches with a biopsy puncher and treated with endonuclease enzyme (Benzonase 25 k U, Sigma Aldrich, E1014) at 37 °C for 48 h to complete the removal of nucleic acid residues. Several washing cycles with Phosphate Buffered Saline (PBS, Sigma Aldrich) were performed to remove residues of the enzyme. Samples (Native pulmonary wall (PW Native), Leaflets (LL Native) and Myocardia (MYO Native), Decellularized Pulmonary wall (PW DC), Leaflets (LL DC) and Myocardia (MYO DC))were then embedded in OCT (Tissue-Tek, 4583, Sakura Finetek, Torrance, CA, USA) and stored at −80 °C for histological and IF analysis, while the rest of the tissues were lyophilized for DNA and biochemical analysis. A subgroup of the decellularized valves ($$n = 3$$) was used for biomechanical tests. ## 2.3. DNA Analysis Lyophilization of the native ($$n = 4$$) and decellularized tissues ($$n = 4$$) was performed with Speed vac SPD130DLX (Thermo Fisher Scientific, Waltham, MA, USA). An amount of 10–12 mg was taken from the pulmonary wall, leaflet, and myocardium tissues and evaluated in duplicate. DNA was extracted according to the manual instructions (DNeasy Blood & Tissue Kit, Qiagen®, Redwood City, CA, USA). Quantification of residual nuclear material was performed directly with Nanodrop (Thermo Fischer Scientific). Afterward, extracted samples were analyzed with a Qubit fluorometer by using Qubit™ dsDNA HS Assay Kit (Thermo Fisher Scientific) according to the manufacturer’s instruction. ## 2.4. Histology OCT-embedded samples from native ($$n = 3$$) and decellularized tissues ($$n = 3$$) frozen, as previously described, were frozen and cut (7.0 µm thickness) using a cryostat (NX 70 HOMVPD Cryostar). Haematoxylin and Eosin (Bio-Optica, 04-061010), Masson Trichrome (Bio-Optica, 04-01082), Mallory Trichrome (Bio-Optica 04-020802), Wiegert Van Gieson (Bio-Optica, 04-051802) staining were performed according to the manufacturer’s instructions. Images of stained tissues were obtained at 10× and 20× magnification using an optical microscope EVOSTM XL Core Cell Imaging System (Thermo Fisher Scientific, Waltham, MA, USA). ## 2.5. Immunofluorescence Analysis The antibodies were applied for the immunological analysis using antibodies indicated in Table S1 to detect structural properties. The primary markers used were Collagen I, Collagen IV, Elastin, and Laminin. Nuclear staining was performed using DAPI (NuBlue Fixed Cell Strain Ready probes reagent, R37606, Thermo Fisher Scientific). Briefly, the cryosections (7–8 μm thick) were rehydrated with PBS at RT and then fixed in PFA ($4\%$ w/v). After treatment with blocking solution ($1\%$ w/v of bovine serum albumin, BSA, Sigma), primary antibodies were added to the samples and incubated overnight at 4 °C. Secondary antibodies were added to the sample and incubated at room temperature for 90 min. After secondary antibodies, the sections were incubated with DAPI for 30 min at RT. Slides were then mounted with Mowiol (Mowiol® 4-88 Sigma-Aldrich). The images were acquired at 10× and 20× magnification with a Leica CTR 6500 fluorescence microscope, and further processing was performed with the LAS AF offline software (Leica Micro-Systems, Wetzlar, Germany). Images of α-Gal epitope were acquired using the confocal microscope Axio Observer LSM 800. ## 2.6. Two-Photon Microscopy Native and decellularized samples frozen blocks in OCT were cut using a cryostat, measuring 10.0 µm thick cryosections. Tissues were compared using DAPI and Phalloidin 647 staining to evaluate the presence of nuclei. For each sample, 5 different areas were recorded in the two-photon microscope with the same acquisition settings, therefore, directly comparable to each other. Briefly, an incident wavelength of 800 nm was adopted, and the images were acquired at fixed magnification through the Olympus 25× water immersion objective with 1.05 numerical aperture (1024 × 1024 pixels), averaged over 70 consecutive frames, with a pixel size of 0.43 μm [29]. Qualitative information about the presence of collagen and its distribution in the tissues was obtained with SHG intensity and coherency analysis methods. ## 2.7. Biochemical Analysis The biochemical analyses included the quantification of hydroxyproline, elastin, and glycosaminoglycans. The protein amount of the decellularized samples was compared to the native tissues. ## 2.7.1. Hydroxyproline Native ($$n = 3$$) and decellularized ($$n = 3$$) samples of the pulmonary wall, leaflet, and myocardium (3–5 mg each) were used for this quantification. Hydroxyproline extraction was performed using a hydroxyproline assay kit (MAK008, Sigma-Aldrich) [30]. Absorbance was measured at 560 nm with a spectrophotometer reader (SPARK). ## 2.7.2. Elastin Elastin quantification was performed with the colorimetric analysis of The Fastin™ Elastin Assay protocol (Biocolor F4000, county Antrim, UK). Elastin content was measured in the pulmonary wall, leaflet, and myocardium of native ($$n = 3$$) and decellularized ($$n = 3$$) pulmonary valves (3–5 mg each). Absorbance was measured with a SPARK spectrophotometer at a wavelength of 513 nm. ## 2.7.3. Glycosaminoglycans GAGs quantification was performed by colorimetric analysis using the Blyscan™ Glycosaminoglycan Assay (Biocolor, B1000, County Antrim, UK) protocol [31]. Native ($$n = 5$$) and decellularized ($$n = 5$$) valves were evaluated in all the components: pulmonary wall, leaflet, and myocardium (3–5 mg each). Absorbance was measured at 656 nm with the SPARK spectrophotometer. ## 2.8. Biomechanical Tests Uniaxial tensile tests were performed on the pulmonary arterial wall and leaflets to biomechanically assess the effects of the decellularization procedure. Three sets of porcine pulmonary heart valves, native and decellularized, were analyzed. The pulmonary valve was opened with a cutter, and 9 dog-bone-shaped samples were taken from each valve. All specimens were cut with a custom-made cutter with a gauge length of 5 mm and a width of 2 mm in the central section. The leaflets were isolated and cut, and one sample was obtained from each leaflet. Six samples were obtained from the pulmonary wall, three in the circumferential direction and three in the longitudinal direction. The sample thus obtained from both were uniaxially tested but can lead to assessment anisotropy [32]. Primarily, the thickness of each tissue sample was measured using Mitutoyo digital Caliber model ID-C112XB (Aurora, IL, USA), and later, the test for the uniaxial tensile property was performed with a customized apparatus (IRS, Padova, Italy) aided by a LabVIEW software (National Instruments, Austin, TX, USA). Tests were carried out as previously described [26,27]. Briefly, the samples were preloaded to 0.1 N at room temperature, then stretched until ruptured at the rate of 0.2 mm/s. Displacement and load were recorded with a sampling frequency of 1000 Hz. Acquired data were analyzed using an in-house developed Matlab® script: for each sample, engineering stress σ (MPa) and strain ε (%) was calculated as the applied load divided by the initial cross-sectional area and the current length divided by the initial length, respectively. Ultimate Tensile Strength (UTS) and Failure Strain (FS) were also calculated in terms of the maximum strength and elongation of each sample. In order to better characterize the non-linear stress-strain curves obtained from each sample, two tensile moduli, E1 and E2, were calculated as the slope of the linear portion of the curve between 0–$10\%$ (elastin region) and 60–$100\%$ deformation (collagen region), respectively [26,28]. A t-test (GraphPad Prism Software, San Diego, CA, USA) was applied to calculate the statistical analysis. Significance was set at $p \leq 0.05.$ ## 2.9. Scanning Electron Microscopy Analysis To evaluate the effect of decellularization on pulmonary valves, a scanning electron microscope (SEM) was used [33]. One sample from the pulmonary wall was collected at each step of the procedure (after Tergitol, sodium cholate, Benzonase® and one after cell seeding) and compared to the native tissue. Photographs of each patch were taken at 200×, 3000×, and 8000× magnification. ## 2.10. Sterility Assessment The sterilization of decellularized tissues was performed following the guidelines of the European Pharmacopoeia 2019 [34]. The decontamination of decellularized samples was performed in two phases: first, samples were treated with $70\%$ ethanol for 30 min at RT. Afterward, they were treated with a cocktail of antibiotics and antimycotics (including vancomycin hydrochloride (50 mg/L, SBR00001, Merck), gentamicin (8 mg/L, G1397, Merck), cefoxitin (240 mg/L, C0688000, Merck) and amphotericin B (25 mg/L, A9528, Merck)) at 37 °C for 24 hr. For the sterility assessment, a thioglycolate medium, (T9032, Sigma-Aldrich) and soya-bean casein digest medium, (22092, Sigma-Aldrich) were used, and turbidity was observed over different time intervals from Day1–14. The patches (8 mm) from native, decellularized and decellularised, and decontaminated samples of the pulmonary wall, leaflets, and myocardium (each in duplicate) were immersed into media to detect the growth of aerobic/anaerobic bacteria and fungal growth. Samples in Thioglycolate medium were incubated at 35 °C in the oven, and those immersed in Soya-bean casein digest medium were kept at RT for 14 days. Images were taken during incubation at time intervals of 48 hrs. ## 2.11. Cytotoxicity/Cytocompatibility hMSC-BM (12974, PromoCell) were cultured in ready-to-use specific media (Mesenchymal Stem Cell Growth medium, C-28009, PromoCell). Cells were expanded and passaged up in an incubator with $5\%$ CO2 at 37 °C and $95\%$ humidity. Sterilized tissue patches were equilibrated with the MSC media for 24 hrs. prior to cell culture. Cells were harvested and resuspended in media with a density of 1 × 106 cells/mL aseptically. Approximately 30,000 cells were on the top of each tissue. The proliferation/toxicity analysis was performed at each time point (24 h, day 7, day 14 after seeding). The analysis was evaluated by the WST-1 assay, Live/dead staining, and DNA quantification. ## 2.11.1. Live and Dead Staining LIVE and DEAD® Viability/Cytotoxicity Kit (L3224 Thermo Fisher Scientific) was used. Staining was performed according to the manufacturer’s instructions. ## 2.11.2. DNA Quantification Quantification of DNA was performed as previously described in the characterization of the decellularized valve: Nanodrop (Quick-Start Protocol DNeasy Blood & Tissue Kit, QIAGEN) and Qubit assay (Qubit™ dsDNA HS Assay Kit, Thermo Fisher Scientific) were used. DNA content was expressed in ng/mg of wet tissue. ## 2.11.3. WST-1 Assay WST-1 solution was made of WST-1 ($5\%$) and mesenchymal cells media., and a volume of 300 µL of this solution was added to each patch and incubated at 37 °C for two hours, according to the supplier’s instructions. Absorbance was measured at 450 nm. ## 2.11.4. SEM Analysis Decellularized and decontaminated pulmonary tissue patches ($$n = 3$$) were imaged using SEM after 14 days of hMSC-BM culture. ## 2.12. Statistical Analysis All data were expressed as mean ± SD. One-way ANOVA was performed with GraphPad Prism 8 to compare the groups of experiments, and multiple comparisons with each control group were performed. Differences were considered statistically significant when $p \leq 0.05.$ ## 3.1. Morphology The decellularized porcine valves appeared white, whereas the myocardium exhibited a light brown compared to the native state (Figure 1b,c). Lyophilized samples were used for DNA quantification and biochemical assays. Freshly decellularized tissues were used along with native aortic valves for biomechanical tests. OCT-embedded tissue punches (8 mm) were frozen in liquid nitrogen for histology and immunological analysis. ## 3.2. DNA Quantification DNA quantification showed a significant reduction in the content of double-strand DNA in treated tissues both with Nanodrop and Qubit assays. Nanodrop showed a decrease in DNA content to $3.54\%$ for the pulmonary wall, $1.47\%$ for the pulmonary leaflet, and $4.37\%$ for the myocardium. Qubit showed a similar reduction in terms of percentage, with a residual DNA of $2.77\%$ for the pulmonary wall, $3.53\%$ for the leaflet, and $1.26\%$ for the myocardium (Figure 1d,e). ## 3.3. Histological Analysis Cell nuclei are very well represented in native tissues, while they are completely absent in decellularized samples. Figure 2a illustrates the results of Haematoxylin and Eosin staining. Collagen seemed well preserved and maintained its distribution in all decellularized tissues as it is likely to observe in three different connective stainings using Masson trichrome (Figure 2b), Mallory trichrome (Figure 2c), Weigert Van Gieson (Figure 2d). Mallory trichrome staining (Figure 2c) and Wiegert Van Gieson staining (Figure 2d) showed a reduction in elastin content. The three-layered organization of the pulmonary wall seemed preserved in all histological samples. ## 3.4. Immunofluorescence DAPI nuclear staining confirmed the effective removal of cells in the decellularized leaflet, pulmonary wall, and myocardium with immunofluorescence and two-photon microscopy (Figure 3). Immunofluorescence showed great preservation of the major structural proteins (Elastin, Collagen I and IV, and Laminin) in the extracellular matrix. The structural architecture was well preserved in the leaflet (Figure 3b), myocardium (Figure 3c), and in all three components of the pulmonary wall: adventitia, media, and intima (Figure 3a). α-Gal persisted in decellularized tissues, with only a slight reduction in the corresponding signal (Figure S1). ## 3.5. Two-Photon Microscopies To deepen the investigation of structural proteins, with a peculiar interest in collagen, we evaluated Collagen I in a semi-quantitative approach with SHG imaging. Compared to immunofluorescence staining, which requires the processing of the tissues, SHG offers a label-free approach to acquire Collagen I signal proportional to the protein content. Moreover, samples were stained with DAPI. In Figure 4a, representative images of native and decellularized tissues from the pulmonary wall, leaflet, and myocardium are shown. DAPI signal is completely absent in decellularized portions. This observation further confirms the effective nuclei removal after the decellularization protocols. First-order analysis of the SHG signal is focused on the average intensity that is proportional to protein content. We observed a decrease in signal intensity in the pulmonary wall and leaflets after decellularization, while no changes appeared in myocardium samples (Figure 4b). In line with this, a re-arrangement of the collagen fibers resulted from coherency analysis for the pulmonary wall and leaflets. Again, no differences have been observed in the myocardium comparing native and decellularized tissue (Figure 4c). ## 3.6. Biochemical Analysis The amount of hydroxyproline in the decellularized valves was higher than the native ones for the pulmonary wall, leaflet, and myocardium (Figure 5a). However, statistically, the quantity of collagen is within the non-significant range except in myocardial tissue. Elastin was shown to be sustained in the pulmonary wall and leaflet but significantly decreased in the myocardium (Figure 5b). GAGs amount seemed to be significantly reduced in all three tissues of the pulmonary scaffold: pulmonary wall, leaflet, and myocardium (Figure 5c). ## 3.7. Biomechanical Results Examples of the stress-strain curves obtained from the uniaxial tensile tests performed on pulmonary valves, along both the circumferential and longitudinal directions, and leaflets are depicted in Figure 6a. It is likely to check that the leaflets exhibit a response to load different from the valve wall, whereas the decellularization procedure seems not to cause appreciable changes in the mechanical behavior of the investigated tissues. Decellularized samples are less thick than native ones in both pulmonary walls and leaflets (Figure 6b) but without statistically significant differences. With respect to UTS and FS values (Figure 6b), the decellularization causes a significant increase in the mechanical strength of the leaflets ($$p \leq 0.003$$) and the pulmonary wall along the circumferential direction ($$p \leq 0.03$$); there is a non-significant decrease in UTS values along the longitudinal direction in decellularized samples with respect to the native ones. The maximum elongation, FS, significantly increases in the circumferential direction ($$p \leq 0.02$$) while it decreases in the longitudinal direction and the leaflets. With regard to Young’s moduli E1 and E2 (Figure 6b), which are characteristic of the elastin and collagen parts of the stress-strain curves, the first one reaches lower values than the second, in accordance with the literature [30]. After decellularization, E1 significantly decreases along the circumferential direction of the valve wall ($$p \leq 0.002$$), while it increases in the longitudinal direction ($$p \leq 0.02$$) and leaflets ($$p \leq 0.02$$). The same trend is observed for E2: there is a significant decrease in stiffness along the circumferential direction ($$p \leq 0.0009$$), while there is an increase in the longitudinal direction (but non-significant) and leaflets ($$p \leq 0.004$$). ## 3.8. Scanning Electron Microscopy (SEM) SEM analysis showed a slight reduction in thickness in decellularized tissue compared to the native one. With the loss of the quaternary structure of the pulmonary wall, the surface appears smoother and more homogeneous, but the overall structure is preserved. Samples from the pulmonary root were analyzed at different steps along the decellularization process: native, after treatment with Tergitol, sodium cholate, and Benzonase, and after 14 days from cell seeding with hMSCs-BM (Figure 7). Images of native tissue showed the presence of cells and the typical structure of collagen fibers inside the matrix. After Tergitol treatment, cells appear in a lower amount or are completely absent, while collagen structure is slightly disrupted. After sodium cholate treatment, pictures show the presence of cellular debris while collagen fibers are maintained: less tissue disruption is caused during this step. After Benzonase, nuclear components are completely absent. In native tissue samples (Figure S2), high turbidity is clearly visible within 24–72 hrs. Decellularized samples were shown to be turbid on day 7. In contrast, sterilized tissue patches were shown to be without turbidity until the end of the test, i.e., day 14. ## 3.9. Cytotoxicity/Cytocompatibility Tests Regarding cytocompatibility, pulmonary wall, and leaflet exhibited different results. From the beginning, WST-1 test demonstrated that the pulmonary wall showed lower values of formazan optical density than the leaflet, and after 72 h, these values have been progressively decreasing (Figure 8). Similar results were obtained by DNA quantification (Figure 8) and confirmed by LIVE/DEAD staining (Figure 9), where the number of live cells decreased progressively from day 7 to day 14. In the leaflet, WST-1 assay showed high optical density from 24 h, and these values grew progressively until stabilization of the growth curve between day 7 and day 14. Similarly, the DNA amounts increased over time: these results were confirmed by live and dead staining. ## 4. Discussion In the present study, a newly reported detergent (Tergitol) was used to decellularize the porcine pulmonary valve. The Tergitol-based protocol proved to be effective in removing cellular components and debris from the native tissue. Histological evaluation allowed for assessing the complete removal of nuclei and the preservation of the extracellular matrix as demonstrated by three different stainings that are specific for the connective tissue (e.g., Massons’ trichrome, Mallory trichrome, Weigert Van Gieson). DAPI confirmed the absence of nuclei in the decellularized tissues both with immunofluorescence and with two-photon microscopy. Immunofluorescence showed no difference in the expression of collagen I, collagen IV, laminin, and elastin. Although the wall thickness of the pulmonary artery was slightly reduced after decellularization, SEM analysis revealed that its surface was not altered by the detergents used. The quaternary structure of matrix proteins, however, was damaged, with a visible reduction in the organization of the fibers and improved smoothness of the pulmonary wall surface. Detection of α-Gal epitope was performed in two ways: a signal derived from the anti-α-Gal epitope antibody and a signal derived from isolectin binding to the antigen were measured. Our results showed that the Tergitol-based decellularization procedure is able to reduce the amount of α-Gal epitope but not to remove it completely. Ongoing investigations are aimed at exploring other strategies for the effective removal of this antigenic epitope before applying the scaffold for implantation in the animal model [33]. Uniaxial tensile tests showed no substantial differences in decellularized and native pulmonary valve tissues, so it is possible to hypothesize that the decellularized scaffold maintains its mechanical properties and functionality. DNA quantification was performed using two different assays, Nanodrop and Qubit: in both cases, the proposed decellularization protocol was demonstrated to be effective in removing cells nuclei and nucleic acids: DNA amount was lower than 50 ng/mg in the decellularized pulmonary wall, leaflet, and myocardium. The higher level of hydroxyproline detected in decellularized tissues compared with native ones can be explained by two hypotheses: first, collagen can be masked by the cells in native tissue so that the assay cannot detect it. SEM analysis showed that the decellularized tissue possesses more disrupted loose fibers, while native tissue is more compact: the compactness of the native tissue can make it difficult to digest, resulting in a lower collagen extraction. The second hypothesis suggests that, with the same weight, the decellularized tissue has a much higher concentration of the ECM than the tissue-containing cells. However, this test quantitatively showed that collagen is well preserved after decellularization, confirming qualitative results. Elastin quantification proved a slight decrease in the amount of this protein, but it is not statistically significant in pulmonary walls and leaflets. In the myocardium, the loss of elastin is significant, but it can be neglected since the myocardium is not relevant to the functionality of the pulmonary valve. Decellularization induced a decrease in the level of GAGs, which are the binding sites of a large number of growth factors and chemokines. GAGs localize the mediators to specific sites in tissues and influence their stability and function [34]. For this reason, the loss of these molecules can reduce the ability of cells to migrate and colonize the scaffold upon implantation. However, GAGs are also deemed to play a role in inflammation and immune response: thus, their reduction in decellularized pulmonary tissues can be considered advantageous to prevent calcification and immune response [34,35]. Biocompatibility tests showed excellent cell growth on the leaflets but not on the pulmonary wall. Biocompatibility was tested qualitatively with WST-1 proliferation index assay and live and dead staining and quantitatively using DNA quantification. All tests showed that mesenchymal cells did not grow on the pulmonary wall, and their amount decreased over time. Indeed, mesenchymal cells did not find a favorable environment in the pulmonary wall: it can be due to inadequate growth stimulation or the absence of specific growth factors for these cells. Furthermore, the thickness of the wall can cause some detergent residues to remain trapped in the tissue: these residues could be toxic to mesenchymal cells. However, previous studies demonstrated that decellularized allografts implanted in the animal model showed an increase in DNA content in the wall comparable to the original root 15 months after implantation [17]. The leaflets showed a progressive growth of mesenchymal cells from both a qualitative and quantitative point of view. The leaflet is much thinner than the pulmonary wall, and this may partly contribute to cell growth since potentially cytotoxic elements were more easily removed. ## 5. Conclusions In our study, we observed that, with the application of Tergitol, decellularization protocol proved to be efficient in removing cells and debris from the porcine pulmonary root, leaving the extracellular matrix well maintained in the anatomical and histological organization. Uniaxial tensile tests let us hypothesize that functionality of the valve is maintained too. The Tergitol modified protocol, therefore, allows obtaining a pulmonary valve scaffold that could mirror the characteristics of an ideal valve substitute. However, the persistence of αGal antigen in tissues requires some specific treatments, as it is not sufficient to remove these xenoantigens completely. Cytotoxic and cytocompatibility tests showed that leaflets might be an optimal substrate for the growth and proliferation of bone marrow mesenchymal cells; however, limited growth in the pulmonary wall requires further in-depth studies to assess the possible reasons involved in this cellular behavior, and if needed, further, surface modifications in the tissue can possibly improve the cell tissue interaction. From the present results, it can be suggested to implant the decellularized valve in an animal model to evaluate both the in vivo functionality and hemodynamic properties of the scaffold. ## References 1. van der Linde D., Konings E.E.M., Slager M.A., Witsenburg M., Helbing W.A., Takkenberg J.J.M., Roos-Hesselink J.W.. Birth Prevalence of Congenital Heart Disease Worldwide: A Systematic Review and Meta-Analysis. J. Am. Coll. Cardiol. (2011) 58 2241-2247. DOI: 10.1016/j.jacc.2011.08.025 2. Chen Z.-Y., Mao S.-F., Guo L.-H., Qin J., Yang L.-X., Liu Y.. Effect of maternal pregestational diabetes mellitus on congenital heart diseases. World J. Pediatr. (2022). DOI: 10.1007/s12519-022-00582-w 3. Luo M., Wang T., Huang P., Zhang S., Song X., Sun M., Liu Y., Wei J., Shu J., Zhong T.. Association of Maternal Betaine-Homocysteine Methyltransferase (BHMT) and BHMT2 Genes Polymorphisms with Congenital Heart Disease in Offspring. Reprod. Sci. (2022) 30 309-325. DOI: 10.1007/s43032-022-01029-3 4. Cao Y., Huang R., Kong R., Li H., Zhang H., Li Y., Liang L., Xiong D., Han S., Zhou L.. Prevalence and risk factors for congenital heart defects among children in the Multi-Ethnic Yunnan Region of China. Transl. Pediatr. (2022) 11 813-824. DOI: 10.21037/tp-21-371 5. Liu Y., Chen S., Zühlke L., Black G., Choy M.-K., Li N., Keavney B.D.. Global birth prevalence of congenital heart defects 1970–2017: Updated systematic review and meta-analysis of 260 studies. Int. J. Epidemiol. (2019) 48 455-463. DOI: 10.1093/ije/dyz009 6. Brida M., Gatzoulis M.A.. Adult congenital heart disease: Past, present and future. Acta Paediatr. (2019) 108 1757-1764. DOI: 10.1111/apa.14921 7. Lee C., Park C.S., Lee C.-H., Kwak J.G., Kim S.-J., Shim W.-S., Song J.Y., Choi E.Y., Lee S.Y.. Durability of bioprosthetic valves in the pulmonary position: Long-term follow-up of 181 implants in patients with congenital heart disease. J. Thorac. Cardiovasc. Surg. (2011) 142 351-358. DOI: 10.1016/j.jtcvs.2010.12.020 8. Jain A., Rajan S.K., Patel K., Garg P., Agrawal V., Kakkar D., Gajjar T., Mishra A., Patel S., Doshi C.. Concomitant pulmonary valve replacement with intracardiac repair for adult tetralogy of fallot. Ann. Pediatr. Cardiol. (2021) 14 323. DOI: 10.4103/apc.apc_125_20 9. Henaine R., Roubertie F., Vergnat M., Ninet J.. Valve replacement in children: A challenge for a whole life. Arch. Cardiovasc. Dis. (2012) 105 517-528. DOI: 10.1016/j.acvd.2012.02.013 10. Ciolacu D.E., Nicu R., Ciolacu F.. Natural Polymers in Heart Valve Tissue Engineering: Strategies, Advances and Challenges. Biomedicines (2022) 10. DOI: 10.3390/biomedicines10051095 11. Atala A., Lanza R., Lanza R.P.. Methods of Tissue Engineering (2001) 12. Wang B., Qinglai T., Yang Q., Li M., Zeng S., Yang X., Xiao Z., Tong X., Lei L., Li S.. Functional acellular matrix for tissue repair. Mater. Today Bio. (2022) 18 100530. DOI: 10.1016/j.mtbio.2022.100530 13. Zhang X., Chen X., Hong H., Hu R., Liu J., Liu C.. Decellularized extracellular matrix scaffolds: Recent trends and emerging strategies in tissue engineering. Bioact. Mater. (2021) 10 15-31. DOI: 10.1016/j.bioactmat.2021.09.014 14. Reimer J.M., Syedain Z.H., Haynie B.H., Tranquillo R.T.. Pediatric tubular pulmonary heart valve from decellularized engineered tissue tubes. Biomaterials (2015) 62 88-94. DOI: 10.1016/j.biomaterials.2015.05.009 15. Cocomello L., Meloni M., Rapetto F., Baquedano M., Ordoñez M.V., Biglino G., Bucciarelli-Ducci C., Parry A., Stoica S., Caputo M.. Long-Term Comparison Between Pulmonary Homograft Versus Bioprosthesis for Pulmonary Valve Replacement in Tetralogy of Fallot. J. Am. Hear. Assoc. (2019) 8 e013654. DOI: 10.1161/JAHA.119.013654 16. Abbas J.R., Hoschtitzky J.A.. Which is the best tissue valve used in the pulmonary position, late after previous repair of tetralogy of Fallot?. Interact. Cardiovasc. Thorac. Surg. (2013) 17 854-860. DOI: 10.1093/icvts/ivt332 17. Haupt J., Lutter G., Gorb S.N., Simionescu D.T., Frank D., Seiler J., Paur A., Haben I.. Detergent-based decellularization strategy preserves macro- and microstructure of heart valves. Interact. Cardiovasc. Thorac. Surg. (2018) 26 230-236. DOI: 10.1093/icvts/ivx316 18. Ramm R., Goecke T., Theodoridis K., Hoeffler K., Sarikouch S., Findeisen K., Ciubotaru A., Cebotari S., Tudorache I., Haverich A.. Decellularization combined with enzymatic removal of N-linked glycans and residual DNA reduces inflammatory response and improves performance of porcine xenogeneic pulmonary heart valves in an ovine in vivo model. Xenotransplantation (2019) 27 e12571. DOI: 10.1111/xen.12571 19. Gallo M., Naso F., Poser H., Rossi A., Franci P., Bianco R., Micciolo M., Zanella F., Cucchini U., Aresu L.. Physiological Performance of a Detergent Decellularized Heart Valve Implanted for 15 Months in Vietnamese Pigs: Surgical Procedure, Follow-up, and Explant Inspection. Artif. Organs (2012) 36 E138-E150. DOI: 10.1111/j.1525-1594.2012.01447.x 20. Yu B.-T., Li W.-T., Song B.-Q., Wu Y.-L.. Comparative study of the Triton X-100-sodium deoxycholate method and detergent-enzymatic digestion method for decellularization of porcine aortic valves. Eur. Rev. Med. Pharmacol. Sci. (2013) 17 2179-2184. PMID: 23893184 21. Malone J.M., Brendel K., Duhamel R.C., Reinert R.L.. Detergent-extracted small-diameter vascular prostheses. J. Vasc. Surg. (1984) 1 181-191. DOI: 10.1016/0741-5214(84)90197-6 22. Grauss R.W., Hazekamp M.G., Oppenhuizen F., Van Munsterena C.J., Groot A.C.G.-D., DeRuiter M.C.. Histological evaluation of decellularised porcine aortic valves: Matrix changes due to different decellularisation methods. Eur. J. Cardio-Thoracic Surg. (2005) 27 566-571. DOI: 10.1016/j.ejcts.2004.12.052 23. Liu X., Li N., Gong D., Xia C., Xu Z.. Comparison of detergent-based decellularization protocols for the removal of antigenic cellular components in porcine aortic valve. Xenotransplantation (2018) 25 e12380. DOI: 10.1111/xen.12380 24. Sandrin M., McKenzie I.F.C.. Galalpha(1,3)Gal, the Major Xenoantigen(s) Recognised in Pigs by Human Natural Antibodies. Immunol. Rev. (1994) 141 169-190. DOI: 10.1111/j.1600-065X.1994.tb00877.x 25. Amending Annex XIV to Regulation (EC) No 1907/2006 of the European Parliament and of the Council Concerning the Registration, Evaluation, Authorization and Restriction of Chemicals (REACH). (2017) 26. Faggioli M., Moro A., Butt S., Todesco M., Sandrin D., Borile G., Bagno A., Fabozzo A., Romanato F., Marchesan M.. A New Decellularization Protocol of Porcine Aortic Valves Using Tergitol to Characterize the Scaffold with the Biocompatibility Profile Using Human Bone Marrow Mesenchymal Stem Cells. Polymers (2022) 14. DOI: 10.3390/polym14061226 27. Li N., Li Y., Gong D., Xia C., Liu X., Xu Z.. Efficient decellularization for bovine pericardium with extracellular matrix preservation and good biocompatibility. Interact. Cardiovasc. Thorac. Surg. (2018) 26 768-776. DOI: 10.1093/icvts/ivx416 28. Burkert J., Kochová P., Tonar Z., Cimrman R., Blassová T., Jashari R., Fiala R., Špatenka J.. The time has come to extend the expiration limit of cryopreserved allograft heart valves. Cell Tissue Bank. (2020) 22 161-184. DOI: 10.1007/s10561-020-09843-2 29. Borile G., Sandrin D., Filippi A., Anderson K., Romanato F.. Label-Free Multiphoton Microscopy: Much More Than Fancy Images. Int. J. Mol. Sci. (2021) 22. DOI: 10.3390/ijms22052657 30. Neuman R.E., Logan M.A.. The determination of hydroxyproline. J. Biol. Chem. (1950) 184 299-306. DOI: 10.1016/S0021-9258(19)51149-8 31. Barbosa I., Garcia S., Barbier-Chassefière V., Caruelle J.-P., Martelly I., Papy-García D.. Improved and simple micro assay for sulfated glycosaminoglycans quantification in biological extracts and its use in skin and muscle tissue studies. Glycobiology (2003) 13 647-653. DOI: 10.1093/glycob/cwg082 32. Soares A., van Geemen D., Bogaerdt A.V.D., Oomens C., Bouten C., Baaijens F.. Mechanics of the pulmonary valve in the aortic position. J. Mech. Behav. Biomed. Mater. (2014) 29 557-567. DOI: 10.1016/j.jmbbm.2013.07.009 33. Nelson A.C., Schoen F.J., Levy R.J.. Scanning electron microscopy methodology for study of the pathophysiology of calcification in bioprosthetic heart valves. Scanning Electron. Microsc. (1985) 1 209-213 34. 2.6.1. Sterility. Eur. Pharm. (2005) 5 145-149 35. Macher B.A., Galili U.. The Galα1,3Galβ1,4GlcNAc-R (α-Gal) epitope: A carbohydrate of unique evolution and clinical relevance. Biochim. Et Biophys. Acta (BBA) Gen. Subj. (2008) 1780 75-88. DOI: 10.1016/j.bbagen.2007.11.003
--- title: Serum OPG and RANKL Levels as Risk Factors for the Development of Cardiovascular Calcifications in End-Stage Renal Disease Patients in Hemodialysis authors: - Michalis Spartalis - Efstratios Kasimatis - Eleni Liakou - Erasmia Sampani - Georgios Lioulios - Michalis Christodoulou - Stamatia Stai - Eleni Moysidou - George Efstratiadis - Aikaterini Papagianni journal: Life year: 2023 pmcid: PMC9967106 doi: 10.3390/life13020454 license: CC BY 4.0 --- # Serum OPG and RANKL Levels as Risk Factors for the Development of Cardiovascular Calcifications in End-Stage Renal Disease Patients in Hemodialysis ## Abstract Cardiovascular calcifications (CVC) are frequently observed in chronic kidney disease (CKD) patients and contribute to their cardiovascular mortality. The aim of the present study was to investigate the impact of osteoprotegerin (OPG)/Receptor Activator of NF-κΒ (RANK)/RANK ligand (RANKL) pathway in the development and evolution of CVCs in hemodialysis patients. In total, 80 hemodialysis patients were assessed for the presence of vascular (abdominal aorta and muscular arteries) calcifications and results were correlated to serum OPG and RANKL levels and the OPG/RANKL ratio. Traditional cardiovascular risk factors and mineral bone disease parameters were also estimated. The presence of VCs was also evaluated 5 years after the initiation of the study, and results were correlated to the initial serum OPG levels. Age, diabetes mellitus, coronary artery disease and OPG levels ($p \leq 0.001$) were associated with VCs, whereas RANKL levels were not. Multivariate analysis though revealed that only OPG levels were significantly associated with abdominal aorta calcifications ($$p \leq 0.026$$), but they were not correlated with the progression of VCs. Serum OPG levels are positively and independently associated with VCs in HD patients, but not with their progression. RANKL levels did not show any associations, whereas further studies are needed to establish the significance of OPG/RANKL ratio. ## 1. Background Cardiovascular complications are the leading cause of death in patients with chronic kidney disease (CKD), especially in those who are under renal replacement treatment (RRT) [1]. Cardiovascular mortality is 10–20 times higher in patients under hemodialysis (HD) compared to the general population, even after adjustment for age, sex and race. Cardiovascular calcifications are an almost universal finding in CKD patients and have a major influence in the development of cardiovascular disease. They appear prematurely and evolve quickly, and their extent and type are actual predictors of the subsequent cardiovascular mortality [2,3]. Both traditional and non-traditional risk factors are involved in the pathogenesis of these calcifications. The former ones are frequently observed in patients with CKD, while the non-traditional ones, such as chronic inflammation and oxidative stress, are usually related to the uremic milieu of these patients [4,5]. There are two major types of vascular calcifications, distinguished by their location. The intimal or atherosclerotic type is mostly seen in larger arteries such as the aorta, and it is characterized by calcification of the intimal layer of the vascular wall and gradually leads to the occlusion of the arteries. The second type is the medial artery calcification, otherwise known as Möckenberg sclerosis, which is characterized by an amorphous mineral deposition within the medial layer, which eventually leads to vessel wall stiffness and loss of its elasticity. This type is more prevalent in CKD patients [6,7]. It is now well known that vascular calcification is not just a passive process resulting from calcium and phosphate deposition. It is an active procedure and is highly regulated by complex enzymatic and cellular pathways, resulting in osteogenesis of the vascular wall, where vascular smooth muscle cells (VSMCs), peripheral lymphocytes and macrophages play a central role. During this process, VSMCs, as well as pericytes are differentiated into osteoblast-like cells, thereby regulating the calcification of the vascular wall [5,6,7,8]. Several co-existing parameters, such as diabetes mellitus, older age, oxidative stress and chronic inflammation, together with calcium and phosphate abnormalities, seem to participate in the osteoblastic transformation of VSMCs and the development of cardiovascular calcifications, mainly by dysregulation of the balance between inhibitors/inducers of vascular calcification [5,9,10]. The OPG/RANK/RANKL system is an important part of this active process and plays a significant role in its regulation [9,11,12]. OPG and RANKL are part of the tumor necrosis factor (TNF) superfamily, and they were originally studied as factors involved in the physiology of bone turnover and the immune system [13]. Recent studies, however, revealed that they have a more complicated and remarkable activity, also being the connective link between bone tissue metabolism and vascular wall morphology [11,12,13]. RANKL is a transmembrane protein expressed by T-cells in lymphoid tissue and osteoblasts in areas of bone remodeling, as well as by endothelial and VSMCs in areas of calcifications. After its secretion, RANKL binds to RANK, a transmembrane receptor expressed on osteoclasts and dendritic cells, and regulates the function and survival of these cells, thus increasing bone resorption [13,14]. It also promotes, through activation of the NF-κΒ pathway, the pathological differentiation of healthy VSMCs into osteoblast-like cells, which in turn leads to osteogenesis of the vascular wall [15]. OPG is a dimeric glycoprotein, mainly produced by osteoblasts, immune cells and cardiovascular cells [13,14], and it acts as a decoy receptor to RANKL, thereby preventing the interaction between RANK and RANKL and reducing bone resorption and the calcification process [5,8,9]. Its generation is upregulated by inflammatory modulators and stimuli on VSMCs and endothelial cells, and these elevated OPG levels may exert an anti-calcific effect on the vascular wall and reflect a status of endothelial dysfunction [16,17]. OPG also binds and deactivates the TNF-related apoptosis-inducing ligand (TRAIL), a molecule expressed on many cells, including T-cells and VSMCs and implicated in ectopic mineralization, thereby neutralizing its apoptotic actions [16]. Recent evidence suggests a vasoprotective role for TRAIL, which seems to counteract RANKL’s pro-calcific actions [18], thus implicating another mechanism through which OPG could regulate the calcification process. It appears that RANKL and OPG exert the opposite effects on the vascular wall to the ones exerted during bone remodeling. Genetic animal studies have supported this theory, showing that OPG-deficient mice developed severe osteoporosis with the simultaneous appearance of vascular calcifications as well [19,20]. In another study, OPGˉʹ¯/ApoEˉʹˉ VSMCs developed increased calcification after RANKL treatment, whereas OPG⁺ʹ⁺/ApoEˉʹˉ cells did not exhibit this result, pointing out a protective role of OPG [21]. Increased levels of serum OPG in the general population, particularly in the elderly and diabetics and patients with ischemic heart disease, are accompanied by increased cardiovascular disease and mortality [22,23,24]. The same association has also been observed in CKD patients and especially in those under RRT, where serum OPG levels are significantly increased [25,26,27], thus generating the question of its true nature and actual action, despite the ones suggested by animal studies. Results concerning the implication and prognostic value of soluble RANKL in the appearance and progression of vascular calcifications are scarce and so far, controversial, whereas the use of OPG/RANKL ratio as a prognostic biomarker of VCs has not been extensively or thoroughly studied with conflicting data to date [12,28]. The aim of the present study was to investigate possible associations of serum OPG and RANKL levels, and their ratio, with the presence and progression of vascular calcifications in CKD patients under HD. ## 2.1. Patients and Study Design Eighty end-stage renal disease patients (ESRD) (42 male—$52.5\%$, and 38 female—$47.5\%$) receiving HD in the Department of Nephrology of Aristotle University, ‘Hippokration’ General Hospital in Thessaloniki, Greece, were enrolled in this study during the first semester of 2015. The study protocol was approved by the Ethics Committee of the School of Medicine, Aristotle University of Thessaloniki, and all protocol procedures were conducted in accordance with the Declaration of Helsinki (2008 Amendment). All patients provided informed written consent prior to enrollment in the study. Patients included in the study were older than 18 years and had been stable on HD for at least 3 months prior to their enrollment. Exclusion criteria were recent (<3 months) or acute infection, chronic inflammation, active autoimmune disease, previous or active malignancy and finally, treatment with antibiotics, steroids or immunosuppressants for at least 3 months prior to the enrollment. All information regarding anthropometric and clinical parameters, such as age, sex, weight, dialysis-related parameters, etiology of CKD, co-morbidities and medication at time of enrollment, was gathered by reviewing of their medical records. All patients were receiving dialysis three times per week, for at least 4 h, with a standard bicarbonate containing dialysate. ## 2.2. Laboratory Measurements Blood samples were collected before a midweek HD session after a 12 h fasting period. Serum levels of glucose, urea, creatinine, total protein, albumin, ALT, AST, total cholesterol, LDL, HDL and triglycerides were determined by routine techniques, using an automated analyzer (Olympus AU560, Hamburg, Germany) at the central laboratory of “Hippokration” General Hospital in Thessaloniki, and they were time averaged for the past 6 months prior to the recruitment to the study. C-Reactive Protein (CRP) was measured by nephelometry, and it was also time averaged for the last 6 months. Serum calcium, phosphorus, alkaline phosphatase (ALP) and intact parathyroid hormone (iPTH) levels (mineral bone disease markers) were also determined, and they were time averaged for the past 12 months before inclusion in the study. Serum levels of iPTH were measured by radioimmunoassay - RIA (Immunotech, Marseille, France) at the B Internal Medicine Training Clinic laboratory of Aristotle University of Thessaloniki. Table 1 depicts the demographic characteristics, primary cause of renal failure, comorbid conditions and the dialysis-related parameters of the study population, whereas Table 2 and Table 3 show the participants’ routine laboratory data and their medication, respectively. ## 2.3. Measurement of Osteoprotegerin and sRANKL Blood samples were drawn from a peripheral vein under fasting conditions in the morning of a midweek routine dialysis session. Serum samples were separated from clotted blood by immediate centrifugation (1500× g for 10 min), aliquoted and stored at −70 °C until assay. Serum levels of osteoprotegerin and sRANKL were measured by an enzyme-linked immunosorbent assay (ELISA) using commercially available standard kits (human osteoprotegerin and human sRANKL (total), respectively, BioVendor, Czech Republic). Serum from patients was diluted 1:3 and 1:100, respectively, for the quantitation of osteoprotegerin and sRANKL. The concentrations of these proteins were calculated by reference to standard curves, performed with the corresponding recombinant molecule. All samples were tested in duplicate. The sensitivity of the ELISA system for osteoprotegerin and sRANKL was 0.03 pmol/L and 0.4 pmol/L, respectively. ## 2.4. Clinical Variables Hypertension was defined as systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg and/or the use of antihypertensive drugs. Diabetes mellitus (DM) was considered present if the patient was on antihyperglycemic medication or had fasting glucose levels > 126mg/dL. Coronary artery disease (CAD) was defined either as at least one documented episode of angina pectoris or a history of myocardial infraction or a coronary stenosis >$75\%$, evidenced by coronary angiography. Cardiovascular disease (CVD) was considered if a patient had a history of CAD and/or atrial fibrillation (AF) and/or peripheral artery disease (PAD) and/or cerebrovascular accident (CVA). Body mass index (BMI) was defined as the post HD body weight (kilograms) divided by height squared (meters). ## 2.5. Vascular Calcifications Vascular calcifications (VCs) were assessed using the Adragao and Kauppila scores, whose methods of estimation and correlation with VCs and cardiovascular outcome has been previously prescribed [29,30]. Adragao score was evaluated with the use of pelvis and hands X-rays which revealed calcifications of the iliac, femoral, radial and digital arteries (muscular arteries calcifications (MACs), and Kauppila score with the use of lateral lumbar spine calcifications, which showed abdominal aorta calcifications (AACs). The assessment of the calcifications was performed by a radiologist blinded to the patients’ clinical and laboratory characteristics. According to their radiological scores, the patients were divided into the following categories: [1] Adragao1: patients with Adragao score 0–2; [2] Adragao2: patients with Adragao score 3–8; [3] Kauppila1: patients with Kauppila score 0–4; [4] Kauppila2: patients with Kauppila score 5–24. Patients with either Adragao2 or Kauppila2 scores were considered as having severe VCs and a higher cardiovascular risk; therefore, based on the above measurements, we further categorized our patients into two groups. The first one was the low calcification score and low cardiovascular risk group, which consisted of patients with both Adragao1 and Kauppila1 scores, whereas the second group—high calcification score and high cardiovascular risk—consisted of patients with at least one high radiological score, either Adragao 3–8 or/and Kauppila 5–24. Five years after the initial enrollment in the study, 47 patients had new pelvis, hand and lateral lumbar spine X-rays which were evaluated for the severity of VCs, and patients were again divided into the above-mentioned groups. Cox regression analysis was then performed to evaluate the association of the clinical and biochemical biomarkers that the patients had upon enrolment, with the progression of their VCs. ## 2.6. Statistical Analysis Statistical analysis was performed using the IBM Statistical Package for Social Sciences (SPSS) Statistics v26 for windows. The Shapiro–Wilk or the Kolmogorov–Smirnov tests were applied to examine the normality of the distribution for continuous variables. Data from normally distributed and non-normally distributed variables were expressed as Mean ± Standard Deviation and Median and Interquartile Range, respectively. Similarly, differences between groups were estimated using Student’s t test for independent samples or Mann–Whitney U test, respectively. Pearson’s and Spearman’s coefficients were used for the correlation between normally and non-normally distributed variables. Odds Ratio (OR) and Receiver Operating Characteristics (ROC) curves were applied to estimate the incidence of serum OPG levels in the presence of vascular calcifications. Multivariate analysis was performed to evaluate serum OPG levels and other independent parameters contributing to the presence of vascular calcifications. Finally, Cox regression analysis was performed to estimate the contribution of OPG serum levels to the progression of vascular calcifications 5 years after the enrollment of the patients. Values of $p \leq 0.05$ (two-tailed) were considered statistically significant for all comparisons. ## 3. Results Eighty patients, M/F $52.5\%$/$47.5\%$, mean age of 57.2 ± 15 years, on standard hemodialysis treatment for a mean time of 64 ± 58 months, were included in the study. The mean value of OPG was 47.6 ± 23.8 pmol/L, of RANKL 455.34 ± 1002.02 pmol/L and of OPG/RANKL ratio 0.40 ± 0.44. ## 3.1. The Severity of Cardiovascular Calcifications and Correlations with Clinical and Laboratory Parameters Based on plain X-rays, $\frac{29}{80}$ ($36.3\%$) patients had a high Adragao score (3–8), and $\frac{39}{80}$ ($48.8\%$) had a Kaupilla score of 5–24. Thirty-five patients ($43.7\%$) were considered to have a low calcification score and low cardiovascular risk, while 45 patients ($56.3\%$) had either an Adragao score of 3–8 or/and a Kauppila score of 5–24 and were considered as having a high calcification score and a high cardiovascular risk. Correlations between the severity of vascular calcifications and clinical and biochemical parameters are shown in Table 4. Age, presence of DM, CAD and CVD were associated with both categories of VCs and a higher calcification score. The use of statins and anti-platelet agents was also associated with the existence of VCs, which most probably reflects the cardiovascular risk of these patients. Serum phosphate levels and elevated BMI had positive correlations with MACs and a high calcification score, whereas hypertension was correlated with AACs. On the other hand, gender, smoking, dyslipidemia, serum albumin, ALP, CRP and iPTH levels did not show statistically important associations with the presence of any VCs in the study population. Serum OPG levels were strongly and positively associated with both categories of VCs ($p \leq 0.001$) and a high calcification score (Table 5). OPG/RANKL ratio was positively associated with AACs and calcification score, whereas RANKL levels were not associated with the presence of VCs. Figure 1 shows differences in OPG levels according to the severity of cardiovascular risk and VCs, whereas the ROC curves depicted in Figure 2 show the importance of OPG serum levels in the severity of VCs as this was estimated by the Kauppila score (AUC 0.793, CI 0.694–0.892, $p \leq 0.001$), Adragao score (AUC 0.751, CI 0.639–0.863, $p \leq 0.001$) and cardiovascular risk (AUC 0.791, CI 0.692–0.891, $p \leq 0.001$). ## 3.2. Multivariate Analysis In a logistic regression modeled analysis, age, DM, the presence of CVD, hypertension and serum OPG levels were evaluated as independent variables, predicting the presence of AACs. A second modeled analysis was also performed for the presence of MACs, where the independent variables examined were age, DM, CVD, BMI and serum OPG levels. Since serum RANKL levels and OPG/RANKL ratio did not have statistically significant correlations with VCs at the univariate analysis, we only used serum OPG levels for the multiple logistic regression analysis. The results of both models are depicted in Table 6. Serum OPG was the only variable that retained its statistical significance at the end of the regression analysis for the presence of AACs ($$p \leq 0.026$$), whereas age failed to do so. Concerning muscular arteries calcifications, BMI was actually the only variable that was positively correlated with them after the multivariate analysis ($$p \leq 0.020$$). ## 3.3. Correlations of Clinical Parameters and Initial Serum OPG Levels with the Progression of VCs Five years after the initial enrolment in the study, 47 patients had new X-rays done. Out of the remaining 33 patients, 22 patients died before the completion of the 5-year follow-up, 1 patient had a transplantation, and 10 patients were lost to follow-up, mostly because of transfer to other dialysis centers. In total, 8 out of the 47 patients ($17\%$) showed a significant progression of AACs and were categorized in the Kauppila2 group, whereas upon enrollment they were in Kauppila1 group. Concerning MACs, this percentage was lower with $\frac{5}{47}$ patients ($10.6\%$) changing their group categorization to Adragao2. The change in group categorization—and hence the progression of VCs—was used as a variable in the Cox regression analysis model. Age, serum OPG levels and hypertension were included as independent variables predicting the progression of AACs, whereas BMI was used instead of hypertension in the regression analysis of MACs. Besides the change in group categorization, we also used the actual increase in VCs as a variable, since it also represents their progression. Patients who presented a doubling—at the least—of their Kauppila score and/or also patients who increased their Adragao score by 1.5 were included in another Cox regression analysis model, using the above-mentioned independent variables. In total, $\frac{9}{47}$ patients had a doubling of their Kauppila score ($19.1\%$), and $\frac{6}{47}$ increased their Adragao score by 1.5 ($12.8\%$). Despite the strong association that serum OPG levels had with the presence of VCs upon enrollment, they were not associated with their progression. In contrast, age showed a positive association with the progression of AACs ($$p \leq 0.022$$ in the group change, and $$p \leq 0.024$$ in the doubling of the Kauppila score), whereas BMI was positively associated ($$p \leq 0.001$$) with the change in Adragao group. ## 4. Discussion The association of OPG, RANKL and their ratio with cardiovascular calcifications in ESRD patients has been investigated in previous studies but the exact nature of this association remains unclear, as well as its prognostic significance. The aim of this study was to verify whether these molecules are associated with the presence and progression of vascular calcifications, as well as with the cardiovascular risk in hemodialysis patients. Evidence acquired over the years suggests that the RANK/RANKL/OPG pathway has a connective role between bone remodeling and vascular calcification and simultaneously acts on osteoblasts and osteoclasts as well as on endothelial cells and VSMCs [12,31,32]. This connecting role is further supported by the fact that OPG is constitutively expressed on the normal vascular wall—contributing most probably to the maintenance of its morphology—and on the surface of endothelial cells and VSMCs. It is rapidly secreted in response to inflammatory stimuli, inhibits the osteoclast activity, acting as a decoy receptor for RANKL, and at the same time, it promotes endothelial cell survival through its anti-apoptotic actions (TRAIL system) [12,31]. In contrast, RANKL and RANK are mostly undetectable on the normal vascular wall, but their expression is significantly increased in calcified areas [16]. Immunohistochemistry performed on calcified areas of the arterial wall of CKD patients revealed that the calcification process is strongly regulated by immunological factors and that the degree of vascular calcification was positively correlated with the intensity of OPG expression, whereas intima media thickness was associated with the degree of RANKL expression [33]. Several studies suggest an association between bone mineral metabolism and vascular calcifications in CKD patients and an impact of bone turnover on their development [34]. There is also evidence of the reduced progression of VCs after the improvement in bone status [35]. In HD patients, additional inhibition of bone resorption by elevated serum OPG levels could result in the inability to accumulate calcium and phosphorus to the bone and metastatic calcification of the vascular tree. Since OPG and RANKL have opposing effects on bone resorption, the OPG/RANKL ratio could be used as a marker of bone turnover—with a high OPG/RANKL value representing a state of low bone turnover—and as a biomarker for the presence of cardiovascular calcifications too. Age, DM and the presence of CAD and CVD were all positively associated with VCs in our study. Hypertension was correlated with AACs, whereas serum phosphate levels and an elevated BMI were positively associated with MACs. These results are in accordance with previous studies, showing the implication of traditional risk factors in the appearance of cardiovascular calcifications [36,37]. Univariate analysis also revealed a strong and positive association of serum OPG levels with VCs, in line with results published from other studies and suggesting its possible role in the pathogenesis and appearance of VCs [38,39]. At the end of the multivariate analysis, serum OPG levels were the only variable that was independently associated with the presence of AACs. Previous studies carried out on pre-dialysis CKD patients have also indicated serum OPG concentration as an independent predictor of VCs, one of them also pointing out a cut-off value of plasma OPG level as a prognostic biomarker for the presence of coronary artery calcifications (CACs) [38,40]. To our knowledge, our study is the first to reveal such a strong and independent association of OPG with AACs in hemodialysis patients, even after adjustment to traditional risk factors including age, thus highlighting the prognostic importance of OPG in younger patients in dialysis. A study by Avila et al. also revealed OPG to be the strongest risk factor associated with arterial calcifications; however, this was performed on peritoneal dialysis patients [41]. The results concerning the presence of MACs were different. After the multivariate analysis, BMI was the only variable that kept its statistical and positive significance. This result comes in accordance with other studies which have associated BMI with VCs, performed both on the general population [42,43] and CKD patients [36]. This is the first study though to reveal BMI as a strong and independent predictor of MACs in hemodialysis patients, affirming its implication and importance. An elevated BMI reflects an excess of adipose tissue, which in turn acts as an inflammatory stimulus that leads to a cascade of actions and eventually vessel atheromatosis, calcification and peripheral artery disease [44]. Our study enhances the assumption that traditional risk factors alone cannot explain the increased incidence of VCs and cardiovascular disease in HD patients, and uremic-related factors are also implicated. OPG remained a strong and independent factor and seems to act as a very potent prognostic biomarker, compared to other traditional risk factors such as age and DM. Its exact function though is still not known and the exact mechanism behind this association is still controversial. Increased OPG levels might either be a result of the vascular damage and the endothelial malfunction observed in CKD patients, or they might have a crucial role in the calcification process itself, or they rise as a compensatory protective mechanism, to counteract the appearance and the progression of vascular calcifications. Results of the Cox regression analysis that we performed suggest that serum OPG levels are not associated with the progression of vascular calcifications, even though they were associated with their presence at the beginning of the study. Previous studies have so far given conflicting information regarding the association of serum OPG with the progression of VCs. Kurnatowska et al. have supported the use of plasma OPG as a marker of the progression of calcification in HD patients [45], and in a study by Ozkok et al., baseline OPG levels were correlated with the progression of CACs at the end of the one-year follow-up. However, after the linear regression analysis, only baseline CAC score and the difference in OPG levels were associated with the progression of CACs and not baseline serum OPG levels [46]. Animal studies have also supported the role of OPG as an inhibitor and a marker of calcification and not as a mediator of atheromatosis [39], and the results obtained by Moldovan et al. support the increase in serum OPG levels as a response and defense to vascular injury [47]. Another study performed on pre-dialysis and hemodialyzed patients did not find any association between OPG levels and the progression of VCs after a 4-year follow-up [48]. These studies support the result obtained from our regression analysis, suggesting that increased serum OPG levels reflect a status of endothelial damage and a compensatory mechanism to protect the vascular wall. On the other hand, age and BMI were both associated with the progression of vascular calcifications in our study. This points to the fact that even though the pathogenesis of VCs is complex in CKD patients, the importance of older age and obesity remains unquestionable. As far as soluble RANKL is concerned, we did not find any significant correlations at all neither with the clinical and laboratory parameters of the study population nor with the presence of vascular calcifications. Despite the growing evidence showing the implication of RANKL in the osteoblastic differentiation of VSMCs [49], the current data concerning the relationship of sRANKL and VCs are controversial. Ozkok et al. showed a significant negative correlation between sRANKL values and CACs at baseline and at one-year follow-up in HD patients [46] and another study by Wei et al. demonstrated a positive association between cardiovascular events and low serum RANKL levels in HD patients [50]. However, other studies performed in CKD patients and general population did not show any association of soluble RANKL with calcification [51,52]. Apparently, RANKL plays an important role in the pathogenesis of VCs, but the exact nature and the significance of the relationship between sRANKL levels and VCs still remains inconclusive. The results concerning the OPG/RANKL ratio are more promising. It was positively correlated with the presence of AACs and a high calcification score. This result is also supported by Ozkok et al. [ 43], where OPG/RANKL ratio values were higher both at baseline and after 1 year of follow-up, in the group of patients who showed a progression of CACs, compared to the non-progressive group. Its use though as a prognostic biomarker of VCs has not been evaluated, and since in our study sRANKL did not show any associations with VCs or any of the other variables, we could attribute the results regarding the association of OPG/RANKL with VCs to the presence of OPG in the equation and not to the ratio itself. Further studies should be carried out to establish the use of the OPG/RANKL ratio as an indicator and regulator of bone turnover and as a prognostic biomarker for the appearance of cardiovascular calcifications. Our study definitely has some limitations. It was a cross sectional study, and the number of patients was relatively small, which did not let us include more variables in our modeled analysis. We should also mention that only $11.3\%$ of our study population had DM, whereas diabetic patients usually comprise about 30–$40\%$ of CKD patients on dialysis, and DM is known to be associated with CVCs. Our study population was also relatively young, with a mean age of 57 years, which may be the reason why age did not retain its significance in the multivariate analysis. However, this does not eliminate the importance of OPG and its association with VCs. It also emphasizes the fact that in younger patients on dialysis, OPG might be a stronger and more important biomarker than age and other clinical parameters concerning the presence of VCs and could help detect patients with increased cardiovascular risk. Another limitation is the fact that only 47 patients had new X-rays done at the 5-year follow-up, which diminishes the strength of the Cox regression analysis. It would also be better if instead of BMI, we used actual visceral adiposity measurements to evaluate the association of adipose tissue with VCs, since BMI is not always representative of it. Visceral adiposity measurements are more complex though and not always available, so simple BMI measurements can still be used as another tool and lead to relatively safe results. ## 5. Conclusions The OPG/RANKL/RANK pathway is considered to play a significant role in the emergence of vascular calcifications in CKD patients. Even though the results are still conflicting, our study reveals that serum OPG levels are strongly associated with vascular calcifications and can be safely used as an independent prognostic marker for them, and this association seems to be of particular interest to younger patients on hemodialysis. Our study also showed that OPG levels are not related to the progression of VCs. Its exact role in the process of vascular calcification is still obscure, as its multiple and conflicting effects remain unspecified and will need further elucidation. On the other hand, RANKL levels do not seem to be associated with the presence of CVCs, whereas more information is needed concerning the use of OPG/RANKL ratio as a prognostic biomarker. ## References 1. Foley R.N., Parfrey P.S., Sarnak M.J.. Clinical epidemiology of cardiovascular disease in chronic renal disease. Am. J. Kidney Dis. (1998) 32 S112-S119. DOI: 10.1053/ajkd.1998.v32.pm9820470 2. Go A.S., Chertow G.M., Fan D., McCulloch C.E., Hsu C.-Y.. Chronic Kidney Disease and the Risks of Death, Cardiovascular Events, and Hospitalization. N. Engl. J. Med. (2004) 351 1296-1305. DOI: 10.1056/NEJMoa041031 3. Moe S.M., Chen N.X.. Mechanisms of vascular calcification in chronic kidney disease. J. Am. Soc. Nephrol. (2008) 19 213-216. DOI: 10.1681/ASN.2007080854 4. Paloian N.J., Giachelli C.M.. A current understanding of vascular calcification in CKD. Am. J. Physiol. Physiol. (2014) 307 F891-F900. DOI: 10.1152/ajprenal.00163.2014 5. Schlieper G., Schurgers L., Brandenburg V., Reutelingsperger C., Floege J.. Vascular calcification in chronic kidney disease: An update. Nephrol. Dial. Transplant. (2015) 31 31-39. DOI: 10.1093/ndt/gfv111 6. Mizobuchi M., Towler D., Slatopolsky E.. Vascular Calcification: The Killer of Patients with Chronic Kidney Disease. J. Am. Soc. Nephrol. (2009) 20 1453-1464. DOI: 10.1681/ASN.2008070692 7. Hénaut L., Sanchez-Nino M.D., Castillo G.A.-E., Sanz A.B., Ortiz A.. Targeting local vascular and systemic consequences of inflammation on vascular and cardiac valve calcification. Expert Opin. Ther. Targets (2015) 20 89-105. DOI: 10.1517/14728222.2015.1081685 8. Harper E., Forde H., Davenport C., Rochfort K.D., Smith D., Cummins P.M.. Vascular calcification in type-2 diabetes and cardiovascular disease: Intergrative roles for OPG, RANKL and TRAIL. Vascul. Pharmacol. (2016) 82 30-40. DOI: 10.1016/j.vph.2016.02.003 9. Schoppet M., Shroff R.C., Hofbauer L.C., Shanahan C.M.. Exploring the biology of vascular calcification in chronic kidney disease: What’s circulating?. Kidney Int. (2008) 73 384-390. DOI: 10.1038/sj.ki.5002696 10. Liu W., Zhang Y., Yu C.-M., Ji Q.-W., Cai M., Zhao Y.-X., Zhou Y.-J.. Current understanding of coronary artery calcification. J. Geriatr. Cardiol. (2015) 12 668-675. DOI: 10.11909/j.issn.1671-5411.2015.06.012 11. Evrard S., Delanaye P., Kamel S., Cristol J.-P., Cavalier E., Arnaud J., Zaoui P., Carlier M., Laville M., Fouque D.. Vascular calcification: From pathophysiology to biomarkers. Clin. Chim. Acta (2015) 438 401-414. DOI: 10.1016/j.cca.2014.08.034 12. Znorko B., Oksztulska-Kolanek E., Michałowska M., Kamiński T., Pawlak K.. Does the OPG/RANKL system contribute to the bone-vascular axis in chronic kidney disease? A systematic review. Adv. Med. Sci. (2017) 62 52-64. DOI: 10.1016/j.advms.2016.08.001 13. Walsh M.C., Choi Y.. Biology of the RANKL—RANK—OPG system in immunity, bone and beyond. Front. Immunol. (2014) 5 511. DOI: 10.3389/fimmu.2014.00511 14. Collin-Osdoby P.. Regulation of Vascular Calcification by Osteoclast Regulatory Factors RANKL and Osteoprotegerin. Circ. Res. (2004) 95 1046-1057. DOI: 10.1161/01.RES.0000149165.99974.12 15. Panizo S., Cardus A., Encinas M., Parisi E., Valcheva P., López-Ongil S., Coll B., Fernandez E., Valdivielso J.M.. RANKL Increases Vascular Smooth Muscle Cell Calcification Through a RANK-BMP4–Dependent Pathway. Circ. Res. (2009) 104 1041-1048. DOI: 10.1161/CIRCRESAHA.108.189001 16. Baud’Huin M., Lamoureux F., Duplomb L., Rédini F., Heymann D.. RANKL, RANK, osteoprotegerin: Key partners of osteoimmunology and vascular diseases. Cell Mol. Life Sci. (2007) 64 2334-2350. DOI: 10.1007/s00018-007-7104-0 17. Higgins C.L., Isbilir S., Basto P., Chen I.Y., Vaduganathan M., Vaduganathan P., Reardon M.J., Lawrie G., Peterson L., Morrisett J.D.. Distribution of Alkaline Phosphatase, Osteopontin, RANK Ligand and Osteoprotegerin in Calcified Human Carotid Atheroma. Protein J. (2015) 34 315-328. DOI: 10.1007/s10930-015-9620-3 18. Forde H., Harper E., Davenport C., Rochfort K.D., Wallace R., Murphy R.P., Smith D., Cummins P.M.. The beneficial pleiotropic effects of tumour necrosis related factor apoptosis ligand (TRAIL) within the vasculature: A review of the literature. Atherosclerosis (2016) 247 87-96. DOI: 10.1016/j.atherosclerosis.2016.02.002 19. Bucay N., Sarosi I., Dunstan C., Morony S., Tarpley J., Capparelli C., Scully S., Tan H.L., Xu W., Lacey D.L.. osteoprotegerin-deficient mice develop early onset osteoporosis and arterial calcification. Genes Dev. (1998) 12 1260-1268. DOI: 10.1101/gad.12.9.1260 20. Bennet B.J., Scatena M., Kirk E.A., Ratazzi M., Varon R.M., Averill M., Schwartz S.M., Giachelli C.M., Rosenfeld M.E.. Osteoprotegerin inactivation accelerates advanced atherosclerotic lesion progression and calcification in older ApoE -/- mice. Arterioscler. Thromb. Vasc. Biol. (2006) 26 2117-2124. DOI: 10.1161/01.ATV.0000236428.91125.e6 21. Callegari A., Coons M.L., Ricks J.L., Rosenfeld M.E., Scatena M.. Increased calcification in osteoprotegerin-deficient smooth muscle cells: Dependence on receptor activator of NF-κB ligand and interleukin 6. J. Vasc. Res. (2014) 51 118-131. DOI: 10.1159/000358920 22. Bjerre M.. Osteoprotegerin (OPG) as a biomarker for diabetic cardiovascular complications. SpringerPLus (2013) 2 658. DOI: 10.1186/2193-1801-2-658 23. Vik A., Mathiesen E.B., Brox J., Wilsgaard T., Njølstad I., Jørgensen L., Hansen J.. Relation between serum osteoprotegerin and carotid intima media thickness in a general population—The Tromsø Study. J. Thromb. Haemost. (2010) 8 2133-2139. DOI: 10.1111/j.1538-7836.2010.03990.x 24. Kiechl S., Schett G., Wenning G., Redlich K., Oberhollenzer M., Mayr A., Santer P., Smolen J., Poewe W., Willeit J.. Osteoprotegerin Is a Risk Factor for Progressive Atherosclerosis and Cardiovascular Disease. Circulation (2004) 109 2175-2180. DOI: 10.1161/01.CIR.0000127957.43874.BB 25. Collado S., Coll E., Nicolau C., Azqueta M., Pons M., Cruzado J.M., De La Torre B., Deulofeu R., Mojal S., Pascual J.. Serum osteoprotegerin in prevalent hemodialysis patients: Associations with mortality, atherosclerosis and cardiac function. BMC Nephrol. (2017) 18. DOI: 10.1186/s12882-017-0701-8 26. Nascimento M., Hayashi S., Riella M., Lindholm B.. Elevated levels of plasma osteoprotegerin are associated with all-cause mortality risk and atherosclerosis in patients with stages 3 to 5 chronic kidney disease. Braz. J. Med. Biol. Res. (2014) 47 995-1002. DOI: 10.1590/1414-431X20144007 27. Yilmaz M.I., Siriopol D., Saglam M., Ünal H.U., Karaman M., Gezer M., Kilinc A., Eyileten T., Guler A.K., Aydin I.. Osteoprotegerin in Chronic Kidney Disease: Associations with Vascular Damage and Cardiovascular Events. Calcif. Tissue Int. (2016) 99 121-130. DOI: 10.1007/s00223-016-0136-4 28. Forde H., Davenport C., Harper E., Cummins P., Smith D.. The role of OPG/RANKL in the pathogenesis of diabetic cardiovascular disease. Cardiovasc. Endocrinol. Metab. (2018) 7 28-33. DOI: 10.1097/XCE.0000000000000144 29. Adragao T., Pires A., Lucas C., Birne R., Magalhaes L., Gonçalves M., Negrao A.P.. A simple vascular calcification score predicts cardiovascular risk in haemodialysis patients. Nephrol. Dial. Transplant. (2004) 19 1480-1488. DOI: 10.1093/ndt/gfh217 30. Wilson P.W.F., Kauppila L.I., O’Donnell C.J., Kiel D., Hannan M., Polak J.M., Cupples L.A.. Abdominal Aortic Calcific Deposits Are an Important Predictor of Vascular Morbidity and Mortality. Circulation (2001) 103 1529-1534. DOI: 10.1161/01.CIR.103.11.1529 31. Klejna K., Naumnik B., Gasowska K., Myśliwiec M.. OPG/RANK/RANKL signaling system and its significance in nephrology. Folia Histochem. Cytobiol. (2009) 47 199-206. DOI: 10.2478/v10042-009-0035-x 32. Pereira L., Frazão J.M.. The bone-vessel axis in chronic kidney disease: An update on biochemical players and its future role in laboratory medicine. Clin. Chim. Acta (2020) 508 221-227. DOI: 10.1016/j.cca.2020.05.023 33. Lysitska A., Galanis N., Skandalos I., Nikolaidou C., Briza S., Fylaktou A., Lioulios G., Mitsoglou Z., Papadopoulou D., Antoniadis N.. Histology and Immunohistochemistry of Radial Arteries Are Suggestive of an Interaction between Calcification and Early Atherosclerotic Lesions in Chronic Kidney Disease. Medicina (2021) 57. DOI: 10.3390/medicina57111156 34. Aaltonen L., Koivuviita N., Seppänen M., Kröger H., Tong X., Löyttyniemi E., Metsärinne K.. Association between bone mineral metabolism and vascular calcification in end-stage renal disease. BMC Nephrol. (2022) 23. DOI: 10.1186/s12882-021-02652-z 35. Barreto D.V., Barreto Fde C., Carvalho A.B., Cuppari L., Draibe S.A., Dalboni M.A., Moyses R.M.A., Neves K.R., Jorgetti V., Miname M.. Association of changes in bone remodeling and coronary calcification in hemodialysis patients: A prospective study. Am. J. Kidney Dis. (2008) 52 1139-1150. DOI: 10.1053/j.ajkd.2008.06.024 36. Barreto D.V., Barreto F.C., Carvalho A.B., Cuppari L., Cendoroglo M., Draibe S.A., Moyses R.M.A., Neves K.R., Jorgetti V., Blair A.. Coronary calcification in hemodialysis patients: The contribution of traditional and uremia-related risk factors. Kidney Int. (2005) 67 1576-1582. DOI: 10.1111/j.1523-1755.2005.00239.x 37. Chen H.-C., Wang W.-T., Hsi C.-N., Chou C.-Y., Lin H.-J., Huang C.-C., Chang C.-T.. Abdominal aortic calcification score can predict future coronary artery disease in hemodialysis patients: A 5-year prospective cohort study. BMC Nephrol. (2018) 19. DOI: 10.1186/s12882-018-1124-x 38. Morena M., Jaussent I., Dupuy A.-M., Bargnoux A.-S., Kuster N., Chenine L., Leray-Moragues H., Klouche K., Vernhet H., Canaud B.. Osteoprotegerin and sclerostin in chronic kidney disease prior to dialysis: Potential partners in vascular calcifications. Nephrol. Dial. Transplant. (2015) 30 1345-1356. DOI: 10.1093/ndt/gfv081 39. Krzanowski M., Krzanowska K., Dumnicka P., Gajda M., Woziwodzka K., Fedak D., Grodzicki T., Litwin J.A., Sułowicz W.. Elevated Circulating Osteoprotegerin Levels in the Plasma of Hemodialyzed Patients with Severe Artery Calcification. Ther. Apher. Dial. (2018) 22 519-529. DOI: 10.1111/1744-9987.12681 40. Morena M., Dupuy A.-M., Jaussent I., Vernhet H., Gahide G., Klouche K., Bargnoux A.-S., Delcourt C., Canaud B., Cristol J.-P.. A cut-off value of plasma osteoprotegerin level may predict the presence of coronary artery calcifications in chronic kidney disease patients. Nephrol. Dial. Transplant. (2009) 24 3389-3397. DOI: 10.1093/ndt/gfp301 41. Avila M., Mora C., Zavala M., Prado M.D.C., Paniagua R.. Osteoprotegerin Is the Strongest Predictor for Progression of Arterial Calcification in Peritoneal Dialysis Patients. Am. J. Nephrol. (2017) 46 39-46. DOI: 10.1159/000477380 42. Powell-Wiley T.M., Poirier P., Burke L.E., Després J.P., Gordon-Larsen P., Lavie C.J., Lear S.A., Ndumele C.E., Neeland I.J., Sanders P.. Obesity and Cardiovascular Disease A Scientific Statement From the American Heart Association. Circulation (2021) 143 e984-e1010. DOI: 10.1161/CIR.0000000000000973 43. Jensen J.C., Dardari Z.A., Blaha M.J., White S., Shaw L.J., Rumberger J., Rozanski A., Berman D.S., Budoff M.J., Nasir K.. Association of Body Mass Index with Coronary Artery Calcium and Subsequent Cardiovascular Mortality. Circ. Cardiovasc. Imaging (2020) 13 e009495. DOI: 10.1161/CIRCIMAGING.119.009495 44. Lai Y., Lin Y., Wang C., Kuo C., Hsu B.. Positive Association of Serum Adipocyte Fatty Acid Binding Protein Level with Peripheral Artery Disease in Hemodialysis Patients. Ther. Apher. Dial. (2019) 24 300-306. DOI: 10.1111/1744-9987.13431 45. Kurnatowska I., Grzelak P., Kaczmarska M., Stefańczyk L., Nowicki M.. Serum Osteoprotegerin is a Predictor of Progression of Atherosclerosis and Coronary Calcification in Hemodialysis Patients. Nephron Clin. Pract. (2010) 117 297-304. DOI: 10.1159/000321169 46. Ozkok A., Caliskan Y., Sakaci T., Erten G., Karahan G., Ozel A., Unsal A., Yildiz A.. Osteoprotegerin/RANKL Axis and Progression of Coronary Artery Calcification in Hemodialysis Patients. Clin. J. Am. Soc. Nephrol. (2012) 7 965-973. DOI: 10.2215/CJN.11191111 47. Morony S., Tintut Y., Zhang Z., Cattley R., Van G., Dwyer D., Stolina M., Kostenuik P., Dimer L.. Osteoprotegerin inhibits vascular calcification without affecting atherosclerosis in ldlr (¯/¯) mice. Circulation (2008) 117 411-420. DOI: 10.1161/CIRCULATIONAHA.107.707380 48. Mesquita M., Demulder A., Wolff F., Melot C., Damry N., Dratwa M., Bergmann P.. Osteoprotegerin and progression of coronary and aortic calcification in chroniv kindey disease. Transpl. Proc. (2010) 42 3444-3449. DOI: 10.1016/j.transproceed.2010.09.129 49. Ono T., Hayashi M., Sasaki F., Nakashima T.. RANKL biology: Bone metabolism, the immune system, and beyond. Inflamm. Regen. (2020) 40 2. DOI: 10.1186/s41232-019-0111-3 50. Wei T., Wang M., Gan L.-Y., Li X.. Relationship of sRANKL Level and Vascular Calcification Score to Cardiovascular Events in Maintenance Hemodialysis Patients. Blood Purif. (2009) 28 342-345. DOI: 10.1159/000232941 51. Lieb W., Gona P., Larson M.G., Massaro J.M., Lipinska I., Keaney J.F., Rong J., Corey D., Hoffmann U., Fox C.S.. Biomarkers of the osteoprotegerin pathway: Clinical correlates, subclinical disease, incident cardiovascular disease, and mortality. Arterioscler. Thromb. Vasc. Biol. (2010) 30 1849-1854. DOI: 10.1161/ATVBAHA.109.199661 52. Shiota J., Izumi N., Kasahara H., Tagawa H., Chiba T., Nihei T.. Relationship between coronary and abdominal calcification score, serum osteoprotegerin (OPG), and serum tartrate—Resistant acid phosphatase (TRACP)—5b in pre-dialysis CKD patients. Nihon Jinzo Gakkai Shi. (2010) 52 1022-1028. PMID: 21254698
--- title: Big Endothelin-1 as a Predictor of Reverse Remodeling and Prognosis in Dilated Cardiomyopathy authors: - Jiayu Feng - Lin Liang - Yuyi Chen - Pengchao Tian - Xuemei Zhao - Boping Huang - Yihang Wu - Jing Wang - Jingyuan Guan - Liyan Huang - Xinqing Li - Yuhui Zhang - Jian Zhang journal: Journal of Clinical Medicine year: 2023 pmcid: PMC9967115 doi: 10.3390/jcm12041363 license: CC BY 4.0 --- # Big Endothelin-1 as a Predictor of Reverse Remodeling and Prognosis in Dilated Cardiomyopathy ## Abstract This study aimed to investigate the predictive value of Big endothelin-1(ET-1) for left ventricular reverse remodeling (LVRR) and prognosis in patients with dilated cardiomyopathy (DCM). Patients with DCM and a left ventricular ejection fraction (LVEF) ≤ $50\%$ from 2008 to 2017 were included. LVRR was defined as the LVEF increased by at least $10\%$ or follow-up LVEF increased to at least $50\%$ with a minimum improvement of $5\%$; meanwhile, the index of left ventricular end-diastolic diameter (LVEDDi) decreased by at least $10\%$ or LVEDDi decreased to ≤33 mm/m2. The composite outcome for prognostic analysis consisted of death and heart transplantations. Of the 375 patients included (median age 47 years, $21.1\%$ female), 135 patients ($36\%$) had LVRR after a median of 14 months of treatment. An independent association was found between Big ET-1 at baseline and LVRR in the multivariate model (OR 0.70, $95\%$ CI 0.55–0.89, $$p \leq 0.003$$, per log increase). Big ET-1, body mass index, systolic blood pressure, diagnosis of type 2 diabetes mellitus (T2DM) and treatment with ACEI/ARB were significant predictors for LVRR after stepwise selection. Adding Big ET-1 to the model improved the discrimination (∆AUC = 0.037, $$p \leq 0.042$$ and reclassification (IDI, $3.29\%$; $$p \leq 0.002$$; NRI, $35\%$; $$p \leq 0.002$$) for identifying patients with LVRR. During a median follow-up of 39 (27–68) months, Big ET-1 was also independently associated with the composite outcome of death and heart transplantations (HR 1.45, $95\%$ CI 1.13–1.85, $$p \leq 0.003$$, per log increase). In conclusion, Big ET-1 was an independent predictor for LVRR and had prognostic implications, which might help to improve the risk stratification of patients with DCM. ## 1. Introduction Dilated cardiomyopathy (DCM) is a heterogeneous disorder defined as dilation and dysfunction of cardiac chambers without ischemic heart disease or loading abnormalities, estimated to affect 1:250 to 1:2500 in the population and cause a 5-year mortality rate of $20\%$ despite therapeutic advancements [1,2]. Left ventricular reverse remodeling (LVRR) is characterized as an increase in LVEF and a decrease in the dimension of the left ventricle, which occurs in 30–$50\%$ of patients with DCM in different cohorts [3,4]. In addition to becoming an essential prognostic indicator in DCM, LVRR also has clinical implications for the appropriate use of device therapy, such as cardiac resynchronization therapy (CRT) and implantable cardioverter-defibrillators (ICDs), as well as the timing for referral to left ventricular assist device (LVAD) or transplantation [5,6]. However, predicting LVRR in clinical practice is still challenging and unclear, particularly in patients with DCM. Big endothelin-1 (Big ET-1) is a 39-amino acid precursor with a much higher concentration and a longer half-life than ET-1 [7]. ET-1 increases collagen production in cardiac fibroblasts, associated with interstitial and vascular fibrosis in cardiac remodeling. ET-1 also has vasoconstriction and inotropic effects on the cardiovascular system [8,9]. Previous studies have shown the association between Big ET-1 and outcome in patients with heart failure (HF), atrial fibrillation (AF), hypertrophic cardiomyopathy (HCM) and left ventricular non-compaction cardiomyopathy (LVNC) [10,11,12,13]. Nevertheless, there is limited evidence regarding Big ET-1’s clinical utility in predicting LVRR for patients with DCM. Therefore, we aim to investigate the ability of Big ET-1 to predict LVRR beyond clinical parameters, as well as its prognostic implications, to improve the current risk stratification of DCM and promote more accurate screening of patients eligible for advanced therapy. ## 2.1. Study Population We retrospectively included patients in this study who were first hospitalized for DCM and had an LVEF ≤ $50\%$ at Fuwai Hospital from 2006 to 2017. DCM was defined as myocardial systolic dysfunction and ventricular dilatation that are not explained by any abnormal loading conditions or coronary artery disease (CAD) [1]. We excluded patients [1] who had significant CAD; [2] primary valvular heart disease; [3] congenital heart disease; [4] cancer or autoimmune disease; [5] viral myocarditis; [6] alcoholic cardiomyopathies; [7] left ventricular non-compaction (LVNC); [8] patients with a CRT or CRT-D implantation; and [9] patients missing echocardiography or follow-up data. All participants signed the informed consent form with the approval of the Ethics Committee of Fuwai Hospital (approval numbers 2014-501). ## 2.2. Follow-Up The participants were re-evaluated and received appropriate medical treatment at Fuwai Hospital as directed by the guideline during follow-up. The follow-up echocardiograms at least 6 months after the baseline examination of patients were searched for in the electronic medical record system. In patients with more than one record, the follow-up echocardiogram was determined based on the examination date closest to 12 months after baseline. In this study, the definition of LVRR was the presence of both during follow-up: [1] LVEF increased by at least $10\%$ or follow-up LVEF increased to at least $50\%$ with a minimum improvement of $5\%$, and [2] LVEDDi decreased by at least $10\%$ or LVEDDi decreased to ≤33 mm/m2. The composite outcome for prognostic analysis consisted of all-cause death and heart transplantations. ## 2.3. Laboratory Measurement A vein blood sample was obtained from patients in the morning following admission. Under standard procedures, all biomarkers were tested at the central laboratory. The measurement of N-terminal Pro Brain natriuretic peptide (NT-Pro-BNP) was performed using a commercial enzyme immunoassay (Biomedica, Vienna, Austria or Bio-Tek, Winooski, VT, USA), and the measurement of Big ET-1 was carried out using a commercial sandwich enzyme immunoassay (BI-2008 2H, Biomedica, Vienna, Austria). ## 2.4. Statistical Analysis We compared the characteristics of patients with or without LVRR using a χ2 test for categorical variables and the Mann–Whitney U-test for continuous variables. Comparison of ∆LVEF/LVEDDi between patients with or without LVRR was tested through the interaction in repeated measures using analysis of variance (ANOVA). We conducted a logistic regression to analyze the relationship between baseline Big ET-1 with other parameters and LVRR. The levels of biomarkers such as Big ET-1, NT-pro-BNP and serum creatine were log2 transformed. Variables that were significantly associated with LVRR ($p \leq 0.1$) in the univariable analysis were included in the multivariable stepwise regression in backward and forward directions guided by Akaike information criteria (AIC). The association between predictors and LVRR was shown as odds ratio (OR) and a $95\%$ confidence interval (CI). We build the best prediction model of LVRR by including predictors with $p \leq 0.05$ in the stepwise model. At last, we measured the discrimination, calibration and reclassification of the model adding Big ET-1 in predicting LVRR using the area under curves (AUC), AIC, integrated discrimination improvement (IDI) and net reclassification improvement (NRI), respectively. The discrimination of the predictive model for LVRR was also tested in three-fold 100-time cross-validation. For sensitivity analysis, we fitted a logistic regression model including clinically significant variables and the interval between two echoes to test the consistency of the relationship between Big ET-1 and LVRR. The Kaplan–Meier curves were plotted to compare the incidence of death or cardiac transplantation in patients with or without LVRR. Cox regressions were conducted to assess the prognostic value of baseline Big ET-1 using stepwise regression with a p-value of 0.1 as a significance level for entry and stay. Then, we plotted a time-dependent receiver operating characteristic (ROC) curve at 1, 3 and 5 years to evaluate the accuracy of the Cox model in predicting the prognosis outcome. Hazard ratio (HR) and a $95\%$ confidence interval were reported, and $p \leq 0.05$ was considered statistically significant. Statistical analysis was performed by the R software 4.1.3. ## 3.1. Baseline Characteristics A total of 375 hospitalized patients diagnosed with DCM were included in this study. A summary of baseline characteristics is shown in Table 1, stratified by the presence of LVRR. The median age of included patients was 47 (34–57) years, and 79 ($21.1\%$) patients were female. Compared to patients without LVRR, patients with LVRR had higher baseline blood pressure, heart rate, body mass index (BMI) and cholesterol but lower levels of Big ET-1 and NT-Pro-BNP. Patients with LVRR were more likely to have a history of T2DM and have non-sustained ventricular tachycardia (NSVT) during hospitalization, and have a higher proportion to use angiotensin-converting enzyme inhibitors (ACEI) or angiotensin II receptor blocker (ARB). Regarding baseline echocardiography parameters, patients with LVRR tended to have smaller left atria and ventricular dilation. During a median interval of 14 (12–24) months between baseline and follow-up echocardiogram, 135 patients ($36\%$) had achieved LVRR. A significant reduction in LVEDDi and a significant improvement in LVEF were observed in patients with LVRR compared with patients without LVRR at follow-up (Figure 1, p for interaction < 0.001). The baseline characteristics stratified using the median of Big ET-1 are shown in Table S1. ## 3.2. Predictors for LVRR In the univariable logistic analysis, the baseline Big-ET 1 was associated with the presence of LVRR (crude OR 0.67, $95\%$ CI 0.55–0.83, $p \leq 0.001$, per log increase). After putting parameters that were significant in univariable regression ($p \leq 0.1$) in the stepwise regression and selecting predictors by AIC, independent baseline predictors of LVRR included Big ET-1 (OR 0.70, $95\%$ CI 0.55–0.89, $$p \leq 0.003$$, per log increase), BMI (OR 1.07, $95\%$ CI 1.01–1.14, $$p \leq 0.026$$), SBP (OR 1.02, $95\%$ CI 1.01–1.04, $$p \leq 0.035$$), diagnosis of T2DM (OR 0.30, $95\%$ CI 0.14–0.65, $$p \leq 0.002$$) and the treatment with ACEI/ARB (OR 2.30, $95\%$ CI 1.02–5.21, $$p \leq 0.045$$). The association between variables and LVRR at univariable and multivariable logistic regression is shown in Table 2. For sensitivity analysis, we fitted a regression model including all clinically significant variables and the interval between two echoes, and the Big ET-1 still showed an independent association with LVRR (Table S2). ## 3.3. Predictive Value of Adding Big ET-1 to the Model for LVRR The performance of the predictive model after adding Big ET-1 is shown in Table 3. In terms of discrimination, adding Big ET-1 improved the AUC from 0.684 to 0.721 ($$p \leq 0.042$$, Figure 2), and improved the mean AUC from 0.653 to 0.682 after cross-validation. The AIC decreased after adding Big ET-1 to the basic model (model 2: 367.25 vs. model 1: 375.94), which showed that including Big ET-1 improved the calibration of the predictive model. Moreover, the model including Big ET-1 (model 2) significantly enhanced the reclassification for patients with LVRR compared with model 1 (IDI, $3.29\%$; $$p \leq 0.002$$; NRI, $35\%$; $$p \leq 0.002$$). ## 3.4. The Outcome of Patients with LVRR and Prognostic Role of Big ET-1 During a median follow-up of 39 (27–68) months, 66 patients died ($17.6\%$) and 19 patients underwent heart transplants ($5.1\%$). DCM patients with LVRR had a better composite outcome than those without LVRR (Figure 3A). After adjusting for confounders, patients with LVRR still had a lower risk of death or heart transplantations (adjusted HR 0.27, $95\%$ CI 0.14–0.52, $p \leq 0.001$). In terms of the relationship between Big ET-1 and long-term prognosis, when the Big ET-1 was dichotomized by the median, the high Big ET-1 group (Big ET-1 > 0.72 pmol/L) was more likely to reach the composite endpoint than the low Big ET-1 group (Figure 3B). In the multivariable regression, baseline Big ET-1 was also independently associated with the outcome as a categorical (adjusted HR 1.94, $95\%$ CI 1.18–3.18, $$p \leq 0.009$$) or a continuous variable (adjusted HR 1.45, $95\%$ CI 1.13–1.85, $$p \leq 0.003$$, per log increase, Table 4). The restricted cubic splines (using four knots) of the relationship between log Big ET-1 and outcome are shown in Figure S1. We screened five significant variables to build a predictive model for all-cause death and heart transplantation through COX stepwise regression, including age, SBP, LVEDD, log NT-Pro-BNP, use of ACEI/ARB and log Big ET-1. Through time ROC analysis, we found that the prognosis prediction model including Big ET-1 can reach AUCs of 0.84, 0.77 and 0.78 for 1-, 3- and 5-year event-free survival (Figure S2). ## 4. Discussion We studied the role of Big ET-1 in predicting LVRR and prognosis in a cohort of DCM patients. We demonstrated that patients with LVRR showed a lower risk of adverse events. Importantly, baseline Big ET-1 proved to be a significant predictor for LVRR as well as the clinical outcome. The predictive model with Big ET-1 for LVRR can improve the current risk stratification of DCM, strengthen the monitoring of high-risk patients and promote more accurate screening of patients eligible for advanced therapy. The reverse remodeling process involves normalizing chamber geometry and function in the failing myocardium. Previous studies have shown that despite reverse remodeling, the myocardium exhibits abnormal gene expression, metabolism pathways, and extracellular matrix structure, indicating that remission is not the same as recovery in myocardial function [14,15]. However, several research results suggest that the earlier the therapy of reverse remodeling is applied, the more apparent cardiac structure and function recovery will be achieved [14,16,17]. As shown in this study, LVRR predicts a better outcome, which is also confirmed by a meta-analysis reporting a relationship between short-term changes in LV structure and function with effects on survival [18]. Therefore, the current challenge is the early stratification to identify patients with a lower probability of LVRR and the most appropriate therapy. Previous studies have confirmed that circulating biomarkers such as NT-Pro-BNP, high-sensitivity troponin, soluble ST2, and galectin-3 may help to identify LVRR [19,20,21]. In this study, our results showed that Big ET-1, a precursor of ET-1, also has a predictive value for LVRR. The potential mechanism of Big ET-1 in predicting LVRR may be related to the biological effects of ET receptor activation on cardiomyocytes or fibroblasts. The ET(A) receptor activates several pathways in HF and cardiac remodeling, including vasoconstriction, inotropy, hypertrophy and fibrosis [8,9]. Additionally, a variety of vasoactive agents (such as NE, angiotensin II, and thrombin) and cytokines (such as TGF-β, TNF, and IL-1) can enhance ET release in vitro [22]. Therefore, the lower level of Big ET-1 may indicate a less severe neuroendocrine activation or cardiac overload, which is more likely to reach myocardial remission after drug treatment. A recent basic study demonstrated the endothelial Forkhead Box transcription factor P1 (EC-Foxp-1) gain of function in protecting from cardiac remodeling and improving cardiac dysfunction by directly downregulating the Transforming Growth Factor-β1 (TGF-β1) gene, which promotes the ET-1 expression [23]. This study supports the relationship between ET-1 and LVRR from the mechanism perspective. In addition, it suggests that a novel therapy targeting the TGF-β1–ET-1 pathway might help relieve or reverse the remodeling of the heart. Several reports have shown that higher Big ET-1 levels correlate with worse prognosis in patients with HF [24,25]. Our study has confirmed this conclusion in patients with HF caused by DCM, and the prognosis prediction models including Big ET-1 showed high accuracy. Previous studies have measured Big ET-1 serially and found that the decline of Big-ET-1 during follow-up after 1 month was associated with less adverse events in patients with ischemic HF [26], demonstrating that the monitoring of Big ET-1 at follow-up may provide additional prognostic value. Conversely, results from the Multi-Ethnic Study of Atherosclerosis Angiogenesis (MEAS) Sub-Study suggested that higher ET-1 levels measured in patients without cardiovascular disease may be associated with a lower risk for incident HF or cardiovascular death [27]. This finding agrees with previous results in pre-clinical animal models that suggest even a modest (~$35\%$) decrease in endothelin-1 gene (Edn1) expression is sufficient to cause cardiac dysfunction [28]. The increase in ET-1 is related to the severity and prognosis of patients with HF, but may have cardioprotective effects on normal hearts. In addition to the Big ET-1, BMI, SBP, diagnosis of T2DM and treatment with ACEI/ARB were also identified as predictors for LVRR. T2DM is proven to be associated with worse LV systolic and diastolic function in DCM patients, and hemoglobin A1C (HbA1c) is related to reduced LV myocardial strain [29]. Our results showed that patients without LVRR had a higher proportion of T2DM and a higher level of HbA1c, despite no significant difference in fast blood glucose (FBG) between patients with LVRR present or absent. Our findings indicated the importance of diabetes management and HbA1c monitoring for patients with DCM. A meta-analysis on sodium–glucose cotransporter-2 (SGLT2) inhibitors and LVRR included 13 trials and a total of 1251 patients with T2DM and/or HF, which found that SGLT2 inhibitors significantly improve LVEF, LV mass index, LVESV index and E-wave deceleration time during follow-up [30]. However, the current study included patients hospitalized from 2006 to 2017; no patients in this study were administrated SGLT2 inhibitors during this period. Nowadays, SGLT2i is being popularized and applied in patients with HF according to new guidelines; further research should analyze the impact of SGLT2i use on LVRR in DCM patients with or without T2DM in real-world situations. Patients with LVRR had a higher prescription of ACEI/ARB, which might be related to the higher blood pressure and better tolerance for drugs of these patients. Therefore, for patients with DCM, avoiding hypotension and titration of standard therapies plays a vital role in the cardiac reverse remodeling. However, this study had several limitations. First, this study was retrospective, and the sample size of patients who had follow-up echocardiograms was relatively small; therefore, selection bias and some potential confounders might not be fully adjusted. Second, the proportion of patients with dynamic monitoring of Big ET-1 during follow-up is low; thus, this study could not analyze the association between the dynamic change and LVRR. Lastly, the prediction model established in this study has not been verified by an external cohort; thus, the generalization of predictive models should be cautious and requires further external validation. ## 5. Conclusions An increased level of serum Big ET-1 was independently associated with LVRR and prognosis in a cohort of Chinese patients with DCM. The predictive model including Big ET-1, BMI, SBP, diagnosis of T2DM and the treatment with ACEI/ARB showed a predictive value for LVRR. The establishment of a predictive model including Big ET-1 helps to improve the risk stratification of patients with DCM and select high-risk patients with a lower probability of LVRR to receive stricter management or device therapy. ## References 1. Richardson P., McKenna W., Bristow M., Maisch B., Mautner B., O’Connell J., Olsen E., Thiene G., Goodwin J., Gyarfas I.. Report of the 1995 World Health Organization/International Society and Federation of Cardiology Task Force on the Definition and Classification of cardiomyopathies. Circulation (1996) 93 841-842. PMID: 8598070 2. Kober L., Thune J.J., Nielsen J.C., Haarbo J., Videbaek L., Korup E., Jensen G., Hildebrandt P., Steffensen F.H., Bruun N.E.. Defibrillator Implantation in Patients with Nonischemic Systolic Heart Failure. N. Engl. J. Med. (2016) 375 1221-1230. DOI: 10.1056/NEJMoa1608029 3. Verdonschot J.A.J., Hazebroek M.R., Wang P., Sanders-van Wijk S., Merken J.J., Adriaansen Y.A., van den Wijngaard A., Krapels I.P.C., Brunner-La Rocca H.-P., Brunner H.G.. Clinical Phenotype and Genotype Associations with Improvement in Left Ventricular Function in Dilated Cardiomyopathy. Circ. Heart Fail. (2018) 11 e005220. DOI: 10.1161/CIRCHEARTFAILURE.118.005220 4. Merlo M., Gobbo M., Stolfo D., Losurdo P., Ramani F., Barbati G., Pivetta A., Di Lenarda A., Anzini M., Gigli M.. The Prognostic Impact of the Evolution of RV Function in Idiopathic DCM. JACC Cardiovasc. Imaging (2016) 9 1034-1042. DOI: 10.1016/j.jcmg.2016.01.027 5. Bristow M.R., Saxon L.A., Boehmer J., Krueger S., Kass D.A., De Marco T., Carson P., DiCarlo L., DeMets D., White B.G.. Cardiac-resynchronization therapy with or without an implantable defibrillator in advanced chronic heart failure. N. Engl. J. Med. (2004) 350 2140-2150. DOI: 10.1056/NEJMoa032423 6. Wilcox J.E., Fang J.C., Margulies K.B., Mann D.L.. Heart Failure with Recovered Left Ventricular Ejection Fraction: JACC Scientific Expert Panel. J. Am. Coll. Cardiol. (2020) 76 719-734. DOI: 10.1016/j.jacc.2020.05.075 7. Hemsén A., Ahlborg G., Ottosson-Seeberger A., Lundberg J.M.. Metabolism of Big endothelin-1 (1-38) and (22-38) in the human circulation in relation to production of endothelin-1 (1-21). Regul. Pept. (1995) 55 287-297. DOI: 10.1016/0167-0115(94)00119-I 8. Leask A.. Getting to the heart of the matter: New insights into cardiac fibrosis. Circ. Res. (2015) 116 1269-1276. DOI: 10.1161/CIRCRESAHA.116.305381 9. Yanagisawa M., Kurihara H., Kimura S., Tomobe Y., Kobayashi M., Mitsui Y., Yazaki Y., Goto K., Masaki T.. A novel potent vasoconstrictor peptide produced by vascular endothelial cells. Nature (1988) 332 411-415. DOI: 10.1038/332411a0 10. Zhang C.L., Xie S., Qiao X., An Y.M., Zhang Y., Li L., Guo X.B., Zhang F.C., Wu L.L.. Plasma endothelin-1-related peptides as the prognostic biomarkers for heart failure: A PRISMA-compliant meta-analysis. Medicine (2017) 96 e9342. DOI: 10.1097/MD.0000000000009342 11. Wu S., Yang Y.M., Zhu J., Ren J.M., Wang J., Zhang H., Shao X.H.. The association between plasma big endothelin-1 levels at admission and long-term outcomes in patients with atrial fibrillation. Atherosclerosis (2018) 272 1-7. DOI: 10.1016/j.atherosclerosis.2018.02.034 12. Fan P., Zhang Y., Lu Y.T., Yang K.Q., Lu P.P., Zhang Q.Y., Luo F., Lin Y.H., Zhou X.L., Tian T.. Prognostic value of plasma big endothelin-1 in left ventricular non-compaction cardiomyopathy. Heart (2021) 107 836-841. DOI: 10.1136/heartjnl-2020-317059 13. Wang Y., Tang Y., Zou Y., Wang D., Zhu L., Tian T., Wang J., Bao J., Hui R., Kang L.. Plasma level of big endothelin-1 predicts the prognosis in patients with hypertrophic cardiomyopathy. Int. J. Cardiol. (2017) 243 283-289. DOI: 10.1016/j.ijcard.2017.03.162 14. Kim G.H., Uriel N., Burkhoff D.. Reverse remodelling and myocardial recovery in heart failure. Nat. Rev. Cardiol. (2018) 15 83-96. DOI: 10.1038/nrcardio.2017.139 15. Tseliou E., Lavine K.J., Wever-Pinzon O., Topkara V.K., Meyns B., Adachi I., Zimpfer D., Birks E.J., Burkhoff D., Drakos S.G.. Biology of myocardial recovery in advanced heart failure with long-term mechanical support. J. Heart Lung Transplant. (2022) 41 1309-1323. DOI: 10.1016/j.healun.2022.07.007 16. Aimo A., Gaggin H.K., Barison A., Emdin M., Januzzi J.L.. Imaging, Biomarker, and Clinical Predictors of Cardiac Remodeling in Heart Failure with Reduced Ejection Fraction. JACC Heart Fail. (2019) 7 782-794. DOI: 10.1016/j.jchf.2019.06.004 17. Aimo A., Vergaro G., Gonzalez A., Barison A., Lupon J., Delgado V., Richards A.M., de Boer R.A., Thum T., Arfsten H.. Cardiac remodelling—Part 2: Clinical, imaging and laboratory findings. A review from the Study Group on Biomarkers of the Heart Failure Association of the European Society of Cardiology. Eur. J. Heart Fail. (2022) 24 944-958. DOI: 10.1002/ejhf.2522 18. Kramer D.G., Trikalinos T.A., Kent D.M., Antonopoulos G.V., Konstam M.A., Udelson J.E.. Quantitative evaluation of drug or device effects on ventricular remodeling as predictors of therapeutic effects on mortality in patients with heart failure and reduced ejection fraction: A meta-analytic approach. J. Am. Coll. Cardiol. (2010) 56 392-406. DOI: 10.1016/j.jacc.2010.05.011 19. Gaggin H.K., Szymonifka J., Bhardwaj A., Belcher A., De Berardinis B., Motiwala S., Wang T.J., Januzzi J.L.. Head-to-head comparison of serial soluble ST2, growth differentiation factor-15, and highly-sensitive troponin T measurements in patients with chronic heart failure. JACC Heart Fail. (2014) 2 65-72. DOI: 10.1016/j.jchf.2013.10.005 20. Lupon J., Gaggin H.K., de Antonio M., Domingo M., Galan A., Zamora E., Vila J., Penafiel J., Urrutia A., Ferrer E.. Biomarker-assist score for reverse remodeling prediction in heart failure: The ST2-R2 score. Int. J. Cardiol. (2015) 184 337-343. DOI: 10.1016/j.ijcard.2015.02.019 21. Daubert M.A., Adams K., Yow E., Barnhart H.X., Douglas P.S., Rimmer S., Norris C., Cooper L., Leifer E., Desvigne-Nickens P.. NT-proBNP Goal Achievement Is Associated with Significant Reverse Remodeling and Improved Clinical Outcomes in HFrEF. JACC Heart Fail. (2019) 7 158-168. DOI: 10.1016/j.jchf.2018.10.014 22. Mayourian J., Ceholski D.K., Gonzalez D.M., Cashman T.J., Sahoo S., Hajjar R.J., Costa K.D.. Physiologic, Pathologic, and Therapeutic Paracrine Modulation of Cardiac Excitation-Contraction Coupling. Circ. Res. (2018) 122 167-183. DOI: 10.1161/CIRCRESAHA.117.311589 23. Liu J., Zhuang T., Pi J., Chen X., Zhang Q., Li Y., Wang H., Shen Y., Tomlinson B., Chan P.. Endothelial Forkhead Box Transcription Factor P1 Regulates Pathological Cardiac Remodeling Through Transforming Growth Factor-beta1-Endothelin-1 Signal Pathway. Circulation (2019) 140 665-680. DOI: 10.1161/CIRCULATIONAHA.119.039767 24. Van Beneden R., Gurne O., Selvais P.L., Ahn S.A., Robert A.R., Ketelslegers J.M., Pouleur H.G., Rousseau M.F.. Superiority of big endothelin-1 and endothelin-1 over natriuretic peptides in predicting survival in severe congestive heart failure: A 7-year follow-up study. J. Card. Fail. (2004) 10 490-495. DOI: 10.1016/j.cardfail.2004.04.001 25. Masson S., Latini R., Anand I.S., Barlera S., Judd D., Salio M., Perticone F., Perini G., Tognoni G., Cohn J.N.. The prognostic value of big endothelin-1 in more than 2300 patients with heart failure enrolled in the Valsartan Heart Failure Trial (Val-HeFT). J. Card. Fail. (2006) 12 375-380. DOI: 10.1016/j.cardfail.2006.02.013 26. Olivier A., Girerd N., Michel J.B., Ketelslegers J.M., Fay R., Vincent J., Bramlage P., Pitt B., Zannad F., Rossignol P.. Combined baseline and one-month changes in big endothelin-1 and brain natriuretic peptide plasma concentrations predict clinical outcomes in patients with left ventricular dysfunction after acute myocardial infarction: Insights from the Eplerenone Post-Acute Myocardial Infarction Heart Failure Efficacy and Survival Study (EPHESUS) study. Int. J. Cardiol. (2017) 241 344-350. DOI: 10.1016/j.ijcard.2017.02.018 27. Leary P.J., Jenny N.S., Bluemke D.A., Kawut S.M., Kronmal R.A., Lima J.A., Maron B.A., Ralph D.D., Rayner S.G., Ryan J.J.. Endothelin-1, cardiac morphology, and heart failure: The MESA angiogenesis study. J. Heart Lung Transplant. (2020) 39 45-52. DOI: 10.1016/j.healun.2019.07.007 28. Hathaway C.K., Grant R., Hagaman J.R., Hiller S., Li F., Xu L., Chang A.S., Madden V.J., Bagnell C.R., Rojas M.. Endothelin-1 critically influences cardiac function via superoxide-MMP9 cascade. Proc. Natl. Acad. Sci. USA (2015) 112 5141-5146. DOI: 10.1073/pnas.1504557112 29. Shen M.T., Li Y., Guo Y.K., Jiang L., Gao Y., Shi R., Yang Z.G.. Impact of type 2 diabetes mellitus on left ventricular deformation in non-ischemic dilated cardiomyopathy patients assessed by cardiac magnetic resonance imaging. Cardiovasc. Diabetol. (2022) 21 94. DOI: 10.1186/s12933-022-01533-5 30. Zhang N., Wang Y., Tse G., Korantzopoulos P., Letsas K.P., Zhang Q., Li G., Lip G.Y.H., Liu T.. Effect of sodium-glucose cotransporter-2 inhibitors on cardiac remodelling: A systematic review and meta-analysis. Eur. J. Prev. Cardiol. (2021) 28 1961-1973. DOI: 10.1093/eurjpc/zwab173
--- title: Genotypes of Hepatitis C Virus and Efficacy of Direct-Acting Antiviral Drugs among Chronic Hepatitis C Patients in a Tertiary Care Hospital authors: - Nahed Mohammed Hawsawi - Tamer Saber - Hussein M. Salama - Walaa S. Fouad - Howaida M. Hagag - Hayaa M. Alhuthali - Emad M. Eed - Taisir Saber - Khadiga A. Ismail - Hesham H. Al Qurashi - Samir Altowairqi - Mohmmad Samaha - Dalia El-Hossary journal: Tropical Medicine and Infectious Disease year: 2023 pmcid: PMC9967136 doi: 10.3390/tropicalmed8020092 license: CC BY 4.0 --- # Genotypes of Hepatitis C Virus and Efficacy of Direct-Acting Antiviral Drugs among Chronic Hepatitis C Patients in a Tertiary Care Hospital ## Abstract Hepatitis C virus (HCV) chronic infection is a major causative factor for several chronic liver diseases, including liver cirrhosis, liver cell failure, and hepatocellular carcinoma. The HCV has seven major genotypes. Genotype 4 is the most prevalent genotype in the Middle East, including Saudi Arabia, followed by genotype 1. The HCV genotype affects the response to different HCV treatments and the progression of liver disease. Currently, combinations of direct-acting antiviral drugs (DAAs) approved for the treatment of HCV achieve high cure rates with minimal adverse effects. Because real-world data from Saudi Arabia about the efficacy of DAAs are still limited, this study was conducted to assess the effectiveness of DAAs in treating patients with chronic hepatitis C and to identify the variables related to a sustained virologic response (SVR) in a real-world setting in Saudi Arabia. This prospective cohort study included 200 Saudi patients with chronic HCV who were 18 years of age or older and had been treated with DAAs at King Abdul-Aziz Specialized Hospital in Taif, Saudi Arabia, between September 2018 and March 2021. The response to treatment was assessed by whether or not an SVR had been achieved at week 12 post treatment (SVR12). An SVR12 was reached in $97.5\%$ of patients. SVR12 rates were comparable for patients of different ages, between men and women, and between patients with and without cirrhosis. In addition, the SVR12 rates did not differ according to the infecting HCV genotype. In this study, the presence of cirrhosis and the patient’s gender were independent predictors of who would not reach an SVR12 (known here as the non-SVR12 group) according to the results of univariate and multivariate binary logistic regression analyses based on the determinants of SVR12. In this population of patients with chronic HCV infection, all DAA regimens achieved very high SVR12 rates. The patients’ gender and the presence of cirrhosis were independent factors of a poor response. ## 1. Introduction Chronic infection with the hepatitis C virus (HCV) is a significant risk factor for developing cirrhosis, liver cell failure, and hepatocellular cancer [1]. In addition, extrahepatic manifestations can be experienced in approximately $40\%$ to $70\%$ of patients with chronic hepatitis C, including chronic kidney disease, metabolic syndrome, autoimmune diseases, lymphoproliferative disorders, and cardiovascular and central nervous system abnormalities [2,3]. Some chronic hepatitis C patients are asymptomatic until they develop severe complications. Delays in the diagnosis of chronic HCV infection can limit treatment options and increase the possibilities of adverse outcomes and the transmission of infection to others [4]. An estimated 58 million people are diagnosed with chronic hepatitis C yearly, and approximately 1.5 million new infections occur. HCV is transmitted parentally and mostly occurs owing to unsafe injection practices, unsafe health care, injection drug use, and sexual practices related to exposure to blood [5,6]. The overall reported prevalence rate of HCV in Saudi Arabia is approximately $1.2\%$ [7]. The World Health Organization (WHO) has set a goal of a $90\%$ decrease in new HCV infections and a $65\%$ decrease in HCV-related deaths by 2030 [8]. HCV has a lipoprotein envelope and a 9.6 kb, single-stranded RNA genome. The genome encodes a large precursor polyprotein, which in turn is cleaved into various nonstructural proteins (p7, NS2, NS3, NS4A, NS4B, NS5A, and NS5B) and structural proteins (core, E1, and E2) [9,10]. The RNA polymerase of HCV has a high propensity for error during replication, leading to high rates of nucleotide substitution in the genome of the progeny virus particles and, consequently, vast genetic variability. Based on these genetic variations, HCV is divided into seven genotypes (GTs) (numbered from GT1 to GT7) and approximately 100 subgenotypes [11,12,13]. Globally, GT1 is the major genotype, followed by GT3, GT4, and GT2. GT 5 and GT6 cause the remaining HCV cases, i.e., approximately $5\%$. Additionally, four immigrants from Congo were diagnosed with GT7 in Canada. However, the distribution of HCV genotypes and subtypes is variable worldwide [14,15,16]. The most prevalent genotype in Saudi Arabia is GT4, followed by GT1 [17]. The HCV genotype affects the response to different HCV treatments, the progression of liver disease, and the overall prognosis of HCV disease [18]. Furthermore, knowledge of the distribution and patterns of HCV genotypes contributes to effective HCV infection control [19]. Therefore, in the context of HCV chronic infection, HCV genotype analysis should be seriously examined [20,21]. A sustained virologic response (SVR), which is defined as the absence of detectable HCV RNA for at least 12 weeks (SVR12) or 24 weeks (SVR24) after completing antiviral treatment, is the primary goal of treating chronic hepatitis C [22,23] Achieving an SVR significantly reduces the HCV-related hepatic and extrahepatic complications, enhances the quality of life, and prevents HCV transmission [24]. In the past, pegylated interferon-based therapy for a long duration (24 to 72 weeks) was the only option for people with chronic hepatitis C. This therapy was not ideal because it achieved an SVR in only a limited percentage of patients and caused many adverse side effects [25,26]. The chronic hepatitis C treatment landscape has changed significantly since 2011 with the development of direct-acting antiviral drugs (DAAs), which target viral replications. In 2014, a highly effective and well-tolerated second-generation DAA was introduced and is now considered the harbinger of a new era in the treatment of chronic hepatitis C [27]. Generally, the approved DAAs can be divided into three classes: NS3/NS4A protease inhibitors, which inhibit the processing of the HCV polyprotein; NS5A complex inhibitors inhibiting viral assembly; and NS5B polymerase inhibitors, which block HCV RNA replication. Each of these DAA classes includes several drugs, and currently, combinations from these approved three classes are recommended in clinical practice. These combinations enable high cure rates with minimal adverse effects, better tolerability, and a shorter treatment duration [22,28]. The efficacy of DAAs has been proven by several clinical trials with “ideal” patients, whose characteristics differed from those of patients using these medications in everyday clinical practice. Comorbidities and/or constitutional characteristics might reduce the effectiveness rates of DAAs reported in controlled clinical studies, suggesting that these findings may not reflect the real world [29,30]. Therefore, evaluations of these drugs in real-life settings from various parts of the world are critical. Because real-world data from Saudi Arabia are still limited, data collected from different ethnic groups are insufficient. This study was carried out to evaluate the efficacy of DAAs in treating chronic hepatitis C patients and to identify the factors associated with an SVR in a real-life experience from a large tertiary care hospital in Taif, Saudi Arabia. The study also aimed to report the frequency of HCV genotypes among the studied patients. ## 2. Materials and Methods This prospective cohort, single-center study included 200 chronic HCV-infected Saudi patients (119 male and 81 female) who were 18 years of age or older and had been treated with DAAs at King Abdul-Aziz Specialized Hospital, a large tertiary care hospital in Taif, Saudi Arabia, between September 2018 and March 2021. Inclusion criteria: This study included both treatment-naïve patients (individuals who had never been treated for HCV before) and treatment-experienced patients (those who had been treated before). All patients were subjected to full history taking and clinical examination, abdominal ultrasonography, and routine laboratory tests, including complete blood count, international normalization ratio (INR), partial thromboplastin time (PTT), serum creatinine, serum albumin, total serum bilirubin, alanine aminotransferase (ALT), and alkaline phosphatase (ALP); gamma-glutamyl transferase (GT) and alfa fetoprotein (AFP), which were taken at the beginning of treatment (basal) and again 12 weeks after treatment ended); and HCV genotype, which was performed only at baseline. Child–Turcotte–Pugh score was calculated for each patient. Assessment of fibrosis: The degree of fibrosis and cirrhosis was assessed using transient elastography (FibroScan): Assessment of liver stiffness using vibration controlled transient elastography is a favored and non-invasive modality. For staging of liver fibrosis, the cut-off values were F1 (>4.8) and (≤7.0 kPa), and F2 ranged between 7.0 and 9.5 kPa, while F3 equal or higher than 9.5 kPa and less than 12.0 kPa and F4 was considered if reading was ≥12.0 kPa. The XL probe was used for the examination of obese patients [31]. Both individuals who had never been treated for HCV before (treatment-naïve) and those who had been treated before (treatment-experienced) participated in the trial. Pregnant or breastfeeding women; patients with significant illnesses such as congestive heart failure, renal failure, respiratory failure, or autoimmune diseases; and patients with concurrent hepatitis B virus and/or HIV infection were excluded from this research. To assess cirrhosis and associated issues, all patients were subjected to full history taking and clinical examination in addition to a baseline abdominal ultrasound and routine laboratory investigation. The grade of liver fibrosis was assessed by using transient elastography (TE) with FibroScan. These lab tests were done at the beginning of treatment and again 12 weeks after treatment ended: complete blood count, international normalization ratio (INR), partial thromboplastin time (PTT), serum creatinine, serum albumin, total serum bilirubin, alanine aminotransferase (ALT), alkaline phosphatase (ALP), and gamma-glutamyl transferase (-GT). HCV genotype was performed only at baseline. Study participants were regularly monitored for 12 weeks after therapy ended. The research Ethics Committee of Taif University and King Abdul-Aziz Specialized Hospital approved this study, and each participating patient signed a written informed consent form. ## 2.1. Screening for HCV Infection According to the manufacturer’s instructions, patients were screened for HCV infection by MonolisaTM HCV Ag-Ab ULTRA V2 (BIO-RAD France). This assay is a qualitative enzyme immunoassay for the detection of anti-HCV antibodies and HCV capsid antigen in human serum or plasma. It uses two recombinant proteins from the non-structural region (NS3 and NS4) and a peptide from the structural region (capsid) of the HCV for the detection of the anti-HCV antibodies. A monoclonal antibody against the HCV capsid is used to detect the HCV capsid antigen. ## 2.2. Confirmation of HCV Infection and Quantification of HCV RNA Identification of HCV infection was confirmed by quantitative HCV- RNA PCR using COBAS® AmpliPrep/COBAS® TaqMan® HCV Quantitative Test, version 2.0 (lower limit of detection, 15 IU/mL), according to the manufacturer’s instructions. This test is used for quantifying HCV RNA genotypes 1 to 6 in human EDTA plasma or serum using the COBAS® AmpliPrep Instrument for automated specimen processing and the COBAS® TaqMan® Analyzer or the COBAS® TaqMan® 48 Analyzer for automated amplification and detection of HCV RNA. ## 2.3. Determination of HCV Genotype HCV genotype was determined by the FDA-approved Abbott RealTime HCV Genotype II assay (Abbott Molecular Inc., Des Plaines, IL, USA), according to the manufacturer’s instructions. These clinical samples are processed by an RT-PCR technique that amplifies the RNA genome of HCV. In addition, at the outset of sample processing, a different RNA sequence than the HCV target sequence is injected into each specimen. For each sample, this unrelated RNA sequence is amplified alongside the target mRNA, using RT-PCR to ensure that the amplification process went well. The assay detects genotypes 1–6 and subtypes 1a and 1b using genotype-specific, fluorescent-labelled oligonucleotide probes. ## 2.4. The Treatment Decision According to the AASLD-IDSA and Saudi Association for the Study of Liver Disease (SASLT) recommendations for hepatitis C, the attending physicians chose the treatment plan [31,32]. Patients were treated with one of the following treatment regimens:Sofosbuvir/Daclatasvir ± Ribavirin (SOF/DCV $\frac{400}{60}$ mg once daily oral dose ± RBV) as a pangenomic regimen for 12 weeks for non-cirrhotic and compensated cirrhotic patients;Sofosbuvir/Ledipasvir ± Ribavirin (SOF/LDV $\frac{400}{90}$ mg once daily dose ± RBV) for GT1 for 8 or 12 weeks according to the baseline HCV RNA and the cirrhosis state;Ombitasvir/Paritaprevir/Ritonavir/Dasabuvir ± Ribavirin (OBV/PTV/Rtv ($\frac{25}{150}$/100 mg once daily orally) plus DSV (250 mg twice daily orally) ± RBV) for 12 weeks for non-cirrhotic GT 1a patients;Elbasvir/Grazoprevir (EBR/GZR) $\frac{50}{100}$ mg single oral daily dose for 12 weeks as a non-pangenomic regimen for patients with GT1b;Glecaprevir/Pibrentasvir/Ribavirin (GLE/PIB $\frac{300}{120}$ mg single daily oral dose ± RBV) as a pangenotypic regimen for 8 weeks. The medication regimen was selected depending on genotypes, treatment history, and the presence or absence of cirrhosis. The dose of RBV for non-cirrhotic or CTP A was weight-based: 1200 mg (for patients ≥ 75 Kg) given orally daily (in two divided doses), while those >75 Kg body weight received 1000 mg orally daily (in two divided doses). For cirrhotic CTP B receiving SOF/LDV, the dose was RBV 600 mg/day, which increased by 200 mg/day every 2 weeks as tolerated. For patients with renal impairment, CrCl 30–50 mL/min 200 mg was given, alternating with 400 mg daily; and for those with CrCl < 50 mL/min and hemodialysis patients, 200 mg was given daily. For patients with baseline Hgb > 12 g, no dose adjustment was made, and RBV was discontinued if Hgb. ## 2.5. Assessment of Treatment Efficacy Sustained virologic response (SVR12) was evaluated by quantitative HCV-RNA PCR at twelve weeks after the end of treatment. ## 2.6. Sample Size Calculation To evaluate the difference between treatment groups (INF and SOF) and subgroups (INF; INF24, INF48, and SOF; SOF/SIM and SOF/DAC), two-way analysis of variance was performed. A total sample size of 181 was deemed sufficient to detect an effect size of 0.295 at a power of 0.95 ($95\%$) at a partial eta squared of 0.08 at a significance level of 0.05. Taking into account a non-response rate of $10.0\%$, the sample size was increased to 199 individuals, i.e., 200 patients were applied. Sample size was calculated using power analysis using Gpower version 3.9.1.6 for Mac OS. ## 2.7. Statistical Analysis All collected data were tabulated and statistically analyzed using the following statistical tests. Descriptive data analysis was in the form of percentages, and mean and data were expressed as mean± standard deviation (SD) or number and percentages (%) as appropriate. The Pearson chi-square test or Fisher’s exact test were used for group comparisons of categorical variables. Two-sample t-tests were used for all independent variables for numerical data. For categorical data, all independent factors were subjected to univariate binary logistic regression analysis, and the odds ratio with $95\%$ confidence intervals was computed for VR evaluation. All statistical analyses were performed using the computer program SPSS software for windows version 26.0 (Statistical Package for Social Science, IBM Corp, Armonk, NY, USA). A two-tailed p-value < 0.05 was considered statistically significant. ## 3.1. Baseline Characteristics of the Studied Population The study included a total of 200 chronic hepatitis C patients with complete SVR12 data. Of these patients, 119 ($59.5\%$) were treatment-naïve, and 81 ($40.5\%$) were treatment-experienced. Sixty-one patients received interferon/ribavirin therapy (24 to 48 weeks), 11 patients received sofosbuvir/simeprevir therapy (12 weeks), and 9 patients received sofosbuvir/daclatasvir therapy (12 weeks). The evaluated patients were 18 to 95 years of age (mean ± SD 53.17 ± 15.71 years). Of the 200 patients studied, 119 ($59.5\%$) were male, and 81 ($40.5\%$) were female. Liver cirrhosis was diagnosed in 70 patients ($35\%$), and all cirrhotic patients were compensated; 44 patients had a Child–Pugh score of B, and 24 patients had a Child–Pugh score of A. The main FibroScan score of the studied patients was 8.7 ± 4.1 kPa. Regarding associated comorbidities, most of our patients were overweight, with a mean body mass index of 28.8 ± 6.2. Approximately one-third of our patients were hypertensive, and approximately $39\%$ had been diagnosed with diabetes; $10\%$ had a history of coronary artery disease, and chronic kidney disease had been diagnosed in 23 patients ($11.5\%$). The baseline viral load of the studied patients was 2,340,000 ± 2,150,000 IU/mL. HCV GT4 was the most frequent genotype detected (109 patients; 54,$5\%$) among the infected patients, followed by GT1a and b (64 patients; $32\%$), GT1b, GT3 (14 patients; $7\%$), and GT 2 (3 patients; $1.5\%$). A mixed genotype (GT4 + GT1a) was detected in 10 patients ($5\%$) (Table 1). No statistically significant difference was found among the different genotypes regarding the study patients’ baseline characteristics and prescribed treatment regimens except for gender ($$p \leq 0.001$$) (Table 2). ## 3.2. Prescribed Treatment Regimens The majority of the study patients (158; $79\%$) received sofosbuvir-based regimens. Sofosbuvir/daclatasvir with or without ribavirin (SOF/DCV ± RBV) was prescribed for 94 patients ($47\%$), and sofosbuvir/ledipasvir with or without ribavirin (SOF/LDV ± RBV) was prescribed for 64 patients ($32\%$). Other prescribed treatment regimens included ombitasvir/paritaprevir/ritonavir/dasabuvir with or without ribavirin (OBV/PTV/Rtv/DSV ± RBV) (28 patients; $14\%$), elbasvir/grazoprevir (EBR/GZR) (10 patients; $5\%$), and glecaprevir/pibrentasvir/ribavirin (GLE/PIB/RBV) (4 patients; $2\%$). Overall, RBV was included in the treatment regimens for 67 patients ($33.5\%$) (see Table 1). A total of 162 patients ($81\%$) received therapy for 12 weeks, and the remaining 38 patients ($19\%$) were treated for 24 weeks. ## 3.3. Treatment Efficacy Overall, an SVR12 was attained in 195 patients ($97.5\%$). As shown in Table 3, in terms of the baseline demographic and clinical characteristics, there was no statistically significant difference between the patients who achieved an SVR12 (SVR12 group) and those who did not achieve an SVR12 (non-SVR12 group) apart from the rate of cirrhosis ($$p \leq 0.032$$) and the distribution of HCV genotypes ($$p \leq 0.001$$). Similarly, the two groups did not differ in terms of the prescribed antiviral drugs ($$p \leq 0.348$$ and $$p \leq 0.755$$ for the antiviral drug combinations and concomitant RBV therapy, respectively). The SVR12 rates were similar among patients in different age groups ($100\%$ for patients younger than 40 years, $95.8\%$ for patients 40 to 60 years, and $98.4\%$ for patients older than 60 years; $$p \leq 0.632$$), among male and female patients ($99.2\%$ versus $95.1\%$, respectively; $$p \leq 0.160$$), or among patients with and without cirrhosis ($94.3\%$ versus $99.2\%$, respectively; $$p \leq 0.121$$). In addition, the SVR12 rates did not differ according to the infecting HCV genotype ($$p \leq 0.847$$); the percentages were $95.5\%$ for GT1b, $96.3\%$ for GT4, and $100\%$ for GT 1a, GT2, and GT3 and the mixed genotype. On the other hand, a statistically significant difference in SVR12 rates was found among treatment-naïve and treatment-experienced patients ($95.8\%$ versus $100\%$, respectively; $$p \leq 0.018$$) (Table 4). Regarding the different drug regimens, there was no statistically significant difference in SVR12 rates achieved by them ($$p \leq 0.348$$); the SVR12 rates for SOF/DCV with or without RBV, SOF/LDV with or without RBV, OBV/PTV/Rtv/DSV with or without RBV, EBR/GZR, and GLE/PIB/RBV were $96.8\%$, $96.9\%$, $100\%$, $100\%$, and $100\%$, respectively (see Table 4 and Figure 1). In patients receiving SOF/DCV with or without RBV, the SVR12 rates were similar among the patients in different age groups ($$p \leq 0.898$$), among male and female patients ($$p \leq 0.553$$), among treatment-naïve and treatment-experienced patients ($$p \leq 0.116$$), among patients with and without cirrhosis ($$p \leq 0.252$$), and among patients infected with different genotypes ($$p \leq 0.660$$). Similarly, the efficacy of SOF/LDV with or without RBV did not differ according to age ($$p \leq 0.878$$), gender ($$p \leq 0.613$$), treatment history ($$p \leq 0.075$$), cirrhosis status ($$p \leq 0.283$$), or infecting HCV genotype ($$p \leq 0.448$$) (see Table 4). ## 3.4. Predictors of SVR12 The patient’s age and gender, treatment history, presence of liver cirrhosis or chronic kidney disease, genotype, and antiviral drug regimen were assessed as potential predictors of the SVR12. Results of the univariate binary logistic regression analysis revealed a nonsignificant effect ($p \leq 0.05$) of all of the explanatory variables on the outcome. Furthermore, the multivariate binary logistic regression analysis showed a nonsignificant effect ($p \leq 0.05$) of all of the explanatory variables on the outcome except for gender and the presence of cirrhosis ($$p \leq 0.045$$ and $$p \leq 0.036$$, respectively) (Table 5). ## 4. Discussion Hepatitis C virus infection is a serious challenge to global health with a significant economic impact. Chronic HCV infection is one of the main causes of liver cirrhosis, liver cell failure, and hepatocellular carcinoma. It is the most common indication for liver transplantation worldwide [33,34]. The prevalence of positive HCV antibodies in Saudi Arabia is* approximately $0.7\%$, and the most prevalent genotype is GT4, followed by GT1 [17,35]. Prior to 2011, pegylated interferon alpha and RBV were the recommended antiviral treatments for 24 to 48 weeks. This regimen resulted in a moderate SVR and was associated with multiple side effects [36]. The introduction of the DAA-based regimens achieved high cure rates with minimal adverse effects, better tolerability, and shorter treatment duration [22,28]. Further, treatment with DAA in patients with chronic HCV infection has a positive effect on the bioelectrical brain activity, with an increase in the amplitude of evoked potentials indicating an improvement in the activity of the cerebral cortex, and this improvement was correlated with the neuroimaging parameters [29,30,31,32,33,34,35]. Long-term clinical outcomes and the health-related quality of life may increase with SVR achievement because of the decreased risk of liver disease progression [37]. In the current investigation, we monitored a total of 200 patients who were treated at a single medical facility and were found to have HCV infections. More than half of our patients had GT4 ($54\%$), followed by GT1($32\%$) including subtypes a and b; this finding was consistent with the overall genotype prevalence among HCV-infected patients in Saudi Arabia [17,38]. Alarfaj et al. [ 39] reported in a study carried out in Riyadh, Saudi Arabia, that the most observed genotype was GT4 ($63.7\%$), followed by GT1 ($24\%$). According to Bawazer et al. [ 17], GT4 was the most prevalent genotype in Saudi Arabia, representing $65\%$ of infections, followed by GT1 in $23\%$ of cases. Younger age groups showed an apparent reduction in the prevalence of GT4 but had an increased rate of GT1. Moreover, several review publications and meta-analyses on the distribution of HCV genotypes reported that GT4 predominated in Middle Eastern nations, notably Egypt, Iraq, Saudi Arabia, and Syria, with rates of $86\%$, $60\%$, $56\%$, and $57\%$, respectively [40,41]. In contrast to our findings, a recent study from Bahrain revealed an increasing tendency of HCV GT1 compared with GT4 in the studied population [29]. Patients in the current study achieved high rates of SVR12 ($97.5\%$), with therapeutic failure occurring only in five cases ($2.5\%$). The high success rates in this study were supported by the results of Alarfaj et al. [ 39], who reported an SVR12 rate of $95.9\%$ in Saudi Arabia, which is consistent with real-world data reported in Middle Eastern countries [41,42,43]. There was no statistically significant difference in SVR12 rates between age groups, genders, and individuals with or without chronic renal disease, liver cirrhosis, or genotypes. However, there was a statistically significant difference in SVR12 rates between treatment-naïve and treatment-experienced individuals ($$p \leq 0.018$$). These findings were consistent with those of Yang et al. [ 44], who reported no statistically significant variations in SVR rates in patients with various genotypes. These results were also in agreement with the findings of Kamal et al. [ 45], who studied the effectiveness of DAA-based regimens in elderly Egyptian patients with chronic HCV infection and found that all of these regimens were well-tolerated, safe, and highly successful even in patients aged 75 years or beyond. They found that age did not influence the effectiveness of DAA treatment. Regarding patients with chronic kidney disease, the results of this study were in concordance with previous studies in the literature [46,47]; they found that treatment with OBV/PTV/R and DSV with or without RBV was safe and effective in HCV-infected patients with chronic renal disease. There was no statistically significant difference in SVR12 rates between patients treated with SOF-based regimens and those treated with other regimens regardless of whether they also used RBV. This was in line with previous research and real-world studies from all over the world, showing similar outcomes with high SVRs across a wide range of regimens, genotypes, and durations of therapy with or without the addition of RBV and in both treatment-naïve and treatment-experienced patients [29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48]. Additionally, the previous use of an antiviral medication by the patient did not affect the SVR rate. Therefore, based on these findings, one can say that the proper DAA-based regimens can yield good curative outcomes. When looking at the patient’s age and gender, whether patients were treatment experienced or treatment-naïve, whether patients had chronic kidney disease or not, and the treatment regimens used and whether RBV was included in them or not, the current study was unable to show a significant difference between the SVR12 group and non-SVR group. However, there were significant differences between the SRV12 and non-SVR12 groups in terms of genotype groups, the presence of liver cirrhosis, and the Child–Pugh score of cirrhotic patients ($$p \leq 0.032$$, $$p \leq 0.02$$, and $$p \leq 0.001$$, respectively). Patients with cirrhosis and a Child–Pugh score of B and GT4 and GT1b had significantly lower SVR12 rates compared with patients without cirrhosis, a Child–Pugh score of A, and viral genotypes other than GT4 and GT1b, according to the results of univariate and multivariate binary logistic regression analyses based on the determinants of the SVR12 ($$p \leq 0.036$$). In our research, the presence of cirrhosis and a Child–Pugh score of B stood out as significant predictors of the failure to achieve an SVR12. SOF with RBV treatment resulted in an SVR in only $71.2\%$ of patients with HCV-related cirrhosis, and more than $5\%$ of the patients discontinued the medication owing to side effects; this was consistent with recent Egyptian research that included a large number of patients [49]. Additionally, the findings indicated a gender difference that was statistically significant ($$p \leq 0.045$$). In general, SVR12-associated factors were inconsistent between clinical trials and real-world investigations, making it difficult to compare the efficacies of different DAA combinations. Thus far, baseline factors (i.e., liver cirrhosis, past treatment experience, infecting HCV genotype, high viral load, increased liver enzymes, and natural polymorphisms in nonstructural HCV genes that limit drug sensitivity) have been linked to poorer SVR rates [50]. The limitation of this study was the limited number of cases (i.e., only 200 cases) compared with the prevalence rate of $1.2\%$ in the overall population of Saudi Arabia. Furthermore, the investigation was carried out at a single center. Additionally, there was no randomization in the patient assignment to treatment [51,52]. This was also confounded by the fact that there was great heterogeneity in the regimens used. Patients younger than 18 years and those with decompensated cirrhosis were not included in the study. ## 5. Conclusions The current study verified that DAAs were successful in treating Saudi HCV patients and achieved an SVR in $97.5\%$ of the patients in a real-world context. The presence of cirrhosis, a Child–Pugh score of B, and viral genotypes GT4 and GT1b were significant predictors of the failure to achieve an SVR12. ## References 1. Mohamoud Y.A., Riome S., Abu-Raddad L.J.. Epidemiology of hepatitis C virus in the Arabian Gulf countries: Systematic review and meta-analysis of prevalence. Int. J. Infect. Dis. (2016.0) 46 116-125. DOI: 10.1016/j.ijid.2016.03.012 2. Jadoul M., Berenguer M.C., Doss W., Fabrizi F., Izopet J., Jha V., Kamar N., Kasiske B.L., Lai C.-L., Morales J.M.. Executive summary of the 2018 KDIGO Hepatitis C in CKD Guideline: Welcoming advances in evaluation and management. Kidney Int. (2018.0) 94 663-673. DOI: 10.1016/j.kint.2018.06.011 3. Pol S., Parlati L., Jadoul M.. Hepatitis C virus and the kidney. Nat. Rev. Nephrol. (2018.0) 15 73-86. DOI: 10.1038/s41581-018-0081-8 4. Jefferies M., Rauff B., Rashid H., Lam T., Rafiq S.. Update on global epidemiology of viral hepatitis and preventive strategies. World J. Clin. Cases (2018.0) 6 589-599. DOI: 10.12998/wjcc.v6.i13.589 5. Global Hepatitis Report 2021 6. Al-Raddadi R.M., Dashash N.A., Alghamdi H.A., Alzahrani H.S., Alsahafi A.J., Algarni A.M., Alraddadi Z.M., Alghamdi M.M., Hakim R.F., Al-Zalabani A.H.. Prevalence and predictors of hepatitis B in Jeddah City, Saudi Arabia: A population-based seroprevalence study. J. Infect. Dev. Ctries. (2016.0) 10 1116-1123. DOI: 10.3855/jidc.7880 7. Sanai F.M., Aljumah A.A., Babatin M., Hashim A., Abaalkhail F., Bassil N., Safwat M.. Hepatitis B care pathway in Saudi Arabia: Current situation, gaps and actions. Saudi J. Gastroenterol. (2019.0) 25 73-80. DOI: 10.4103/sjg.SJG_421_18 8. Global Health Sector Strategy on Viral Hepatitis 2016–2021. Towards Ending Viral Hepatitis 9. Kato N.. Genome of human hepatitis C virus (HCV): Gene organization, sequence diversity, and variation. Microb. Comp. Genom. (2000.0) 5 129-151. DOI: 10.1089/omi.1.2000.5.129 10. Op De Beeck A., Dubuisson J.. Topology of hepatitis C virus envelope glycopro-teins. Rev. Med. Virol. (2003.0) 13 233-241. DOI: 10.1002/rmv.391 11. Simmonds P., Bukh J., Combet C., Deléage G., Enomoto N., Feinstone S., Halfon P., Inchauspé G., Kuiken C., Maertens G.. Consensus proposals fora unified system of nomenclature of hepatitis C virusgenotypes. Hepatology (2005.0) 42 962-973. DOI: 10.1002/hep.20819 12. Tsukiyama-Kohara K., Kohara M.. Hepatitis C virus: Viralquasispecies and genotypes. Int. J. Mol. Sci. (2017.0) 19. DOI: 10.3390/ijms19010023 13. Li G., De Clercq E.. Current therapy for chronic hepatitis C: The role of direct-acting antivirals. Antivir. Res. (2017.0) 142 83-122. DOI: 10.1016/j.antiviral.2017.02.014 14. Abdelhameed R.F.A., Ibrahim A.K., Elfaky M.A., Habib E.S., Mahamed M.I., Mehanna E.T., Darwish K.M., Khodeer D.M., Ahmed S.A., Elhady S.S.. Antioxidant and Anti-Inflammatory Activity of Cynanchum Acutum L. Isolated Flavonoids Using Experimentally Induced Type 2 Diabetes Mellitus: Biological and In Silico Investigation for NF-ΚB Pathway/MiR-146a Expression Modulation. Antioxidants (2021.0) 10. DOI: 10.3390/antiox10111713 15. Petruzziello A., Marigliano S., Loquercio G., Cozzolino A., Cacciapuoti C.. Globalepidemiology of hepatitis C virus infection: An up-date of thedistribution and circulation of hepatitis C virus genotypes. World J. Gastroenterol. (2016.0) 22 7824-7840. DOI: 10.3748/wjg.v22.i34.7824 16. Blach S., Zeuzem S., Manns M., Altraif I., Duberg A.S., Muljono D.H., Waked I., Alavian S.M., Lee M.H., Negro F.. Global prevalence andgenotype distribution of hepatitis C virus infection in 2015, amodelling study. Lancet Gastroenterol. Hepatol. (2017.0) 2 161-176. DOI: 10.1016/S2468-1253(16)30181-9 17. Bawazir A., AlGusheri F., Jradi H., AlBalwi M., Abdel-Gader A.-G.. Hepatitis C virus genotypes in Saudi Arabia: A future prediction and laboratory profile. Virol. J. (2017.0) 14 208. DOI: 10.1186/s12985-017-0873-7 18. Yu M.-L., Tsai W.-L., Chi-Jen Chu C.-J., Kao J.-H.. Characteristics of patients with hepatitis C virus infection and antiviral treatment initiation in Taiwan: The MOSAIC study. Kaohsiung J. Med. Sci. (2020.0) 37 245-252. DOI: 10.1002/kjm2.12317 19. Gower E., Estes C., Blach S., Razavi-Shearer K., Razavi H.. Global epidemiology and genotype distribution of the hepatitis C virus infection. J. Hepatol. (2014.0) 61 S45-S57. DOI: 10.1016/j.jhep.2014.07.027 20. Messina J.P., Humphreys I., Flaxman A., Brown A., Cooke G.S., Pybus O.G., Barnes E.. Global distribution and prevalence of hepatitis C virus genotypes. Hepatology (2015.0) 61 77-87. DOI: 10.1002/hep.27259 21. de Castroa G.L.C., Amoras E.G.S., de Araújo M.S.M., da Silva Conde S.R.S., Vallinoto A.C.R.. Hepatitis C virus genotypes and associated risk factors in the state of Pará, Northern Brazil. Braz. J. Infect. Dis. (2020.0) 24 304-309. DOI: 10.1016/j.bjid.2020.06.010 22. Ellabban M.A., Fattah I.O.A., Kader G.A., Eldeen O.S., Mehana A.E., Khodeer D.M., Hosny H., Elbasiouny M.S., Masadeh S.. The Effects of Sildenafil and/or Nitroglycerin on Random-Pattern Skin Flaps After Nicotine Application in Rats. Sci. Rep. (2020.0) 10 3212. DOI: 10.1038/s41598-020-60128-w 23. EASL 2017 clinical practice guidelines on the management of hepatitis C virus infection. J. Hepatol. (2017.0) 66 153-194. PMID: 27667367 24. Shaheen M.A., Idrees M.. Evidence-based consensus on the diagnosis, prevention and management of hepatitis C virus disease. World J. Hepatol. (2015.0) 7 616-627. DOI: 10.4254/wjh.v7.i3.616 25. Manns M.P., McHutchison J.G., Gordon S.C., Rustgi V.K., Shiffman M., Reindollar R., Goodman Z.D., Koury K., Ling M.-H., Albrecht J.K.. Peginterferon alfa-2b plus ribavirin compared with interferon alfa-2b plus ribavirin for initial treatment of chronic hepatitis C: A randomised trial. Lancet (2001.0) 358 958-965. DOI: 10.1016/S0140-6736(01)06102-5 26. Bertino G., Ardiri A., Proiti M., Rigano G., Frazzetto E., Demma S., Ruggeri M.I., Scuderi L., Malaguarnera G., Bertino N.. Chronic hepatitis C: This and the new era of treatment. World J. Hepatol. (2016.0) 8 92-106. DOI: 10.4254/wjh.v8.i2.92 27. Hong C.M., Liu C.H., Su T.H., Yang H.C., Chen P.J., Chen Y.W., Kao J.H., Liu C.J.. Real-world effectiveness of direct-acting antiviral agents for chronic hepatitis C in Taiwan: Real-world data. J. Microbiol. Immunol. Infect. (2020.0) 53 569-577. DOI: 10.1016/j.jmii.2018.09.005 28. Zeng H., Li L., Hou Z., Zhang Y., Tang Z., Liu S.. Direct-acting Antiviral in the Treatment of Chronic Hepatitis C: Bonuses and Challenges. Int. J. Med. Sci. (2020.0) 17 892-902. DOI: 10.7150/ijms.43079 29. Abdulla M., Ali H., Nass H., Khamis J., Al Qamish J.. Efficacy of direct-acting antiviral therapy for hepatitis C viral infection. Real-life experience in Bahrain. Hepat. Med. (2019.0) 11 69-78. DOI: 10.2147/HMER.S190967 30. Cárdaba-García M.E., Abad-Lecha E., Calleja-Hernández M.Á.. Effectiveness of direct-acting antiviral drugs against hepatitis C virus: Predictive factors of response to the treatment. Libyan J. Med. (2021.0) 16 1949797. DOI: 10.1080/19932820.2021.1949797 31. Chung R.T., Davis G.L., Jensen D.M., Masur H., Saag M.S., Thomas D.L., Aronsohn A.I., Charlton M.R., Feld J.J., Fontana R.J.. Hepatitis C guidance: AASLDIDSA recommendations for testing, managing, and treating adults infected with hepatitis C virus. Hepatology (2015.0) 62 932-954. DOI: 10.1002/hep.27950 32. Alghamdi A.S., Alghamdi M., Sanai F.M., Alghamdi H., Aba-Alkhail F., Alswat K., Babatin M., Alqutub A., Altraif I., Alfaleh F.. SASLT guidelines: Update in treatment of Hepatitis C virus infection. Saudi J. Gastroenterol. (2016.0) 22 S25-S27. DOI: 10.4103/1319-3767.188067 33. Axley P., Ahmed Z., Ravi S., Singal A.K.. Hepatitis C virus and hepatocellular carcinoma: A narrative review. J. Clin. Transl. Hepatol. (2018.0) 6 79-84. DOI: 10.14218/JCTH.2017.00067 34. Bhamidimarri K.R., Satapathy S.K., Martin P.. Hepatitis C virus and liver transplantation. Gastroenterol. Hepatol. (2017.0) 13 214-220 35. Aljumah A., Abaalkhail F., Al-Ashgar H., Assiri A., Babatin M., Al-Faleh F., Alghamdi A., Al-Hakeem R., Hashim A., Alqutub A.. Epidemiology, disease burden, and treatment strategies of chronic hepatitis C virus infections in Saudi Arabia in the new treatment paradigm shift. Saudi J. Gastroenterol. (2016.0) 22 269. DOI: 10.4103/1319-3767.187609 36. Ahmed O.A., Kaisar H.H., Hawash N., Samir H., Shabana S., Fouad H.A., Rizk F., Abd-Elsalam S.. Efficacy of sofosbuvir plus ribavirin with or without peginterferon- alfa interferon in treatment of a cohort of Egyptian patients with hepatitis C virus infection. Infect. Disord. Drug Targets (2017.0) 17 95-100. DOI: 10.2174/1871526517666170417143216 37. Juanbeltz R., Castilla J., Martínez-Baz I., O’Leary A., Sarobe M., San Miguel R.. Health-related quality of life in hepatitis C patients who achieve sustained virological response to direct-acting antivirals: A comparison with the general population. Qual. Life Res. (2019.0) 28 1477-1484. DOI: 10.1007/s11136-019-02111-1 38. Abdel-Moneim A., Bamaga M., Shehab G., Abu Elsaad A., Farahat F.. HCV Infection among Saudi Population: High Prevalence of Genotype 4 and Increased Viral Clearance Rate. PLoS ONE (2012.0) 7. DOI: 10.1371/journal.pone.0029781 39. Alarfaj S.J., Alzahrani A., Alotaibi A., Almutairi M., Hakami M., Alhomaid N., Alharthi N., Ghazwa B.K., Alghamdi A.. The effectiveness and safety of direct-acting antivirals for hepatitis C virus treatment: A single-center experience in Saudi Arabia. Saudi Pharm. J. (2022.0) 30 1448-1453. DOI: 10.1016/j.jsps.2022.07.005 40. Ghaderi-Zefrehi H., Gholami-Fesharaki M., Sharafi H., Sadeghi F., Alavian S.. The Distribution of Hepatitis C Virus Genotypes in Middle Eastern Countries: A Systematic Review and Meta-Analysis. Hepat. Mon. (2016.0) 16 e40357. DOI: 10.5812/hepatmon.40357 41. Hashim A.E.M., Zaky S., Alghannam M.Z., Ibraheem S.A., El-Raey F.M., Abdelaziz A.. Efficacy and safety of daclatasvir-containing regimens in the treatment of chronic hepatitis C patients in Damietta. Al-Azhar Assiut Med. J. (2020.0) 18 159 42. Ahmed O.A., Kaisar H.H., Badawi R., Hawash N., Samir H., Shabana S.S.T., Fouad M.H.A., Rizk F.H., Khodeir S.A., Abd-Elsalam S.. Efficacy and safety of sofosbuvir-ledipasvir for treatment of a cohort of Egyptian patients with chronic hepatitis C genotype 4 infection. Drug Resist. Infect. (2018.0) 11 295-298. DOI: 10.2147/IDR.S153060 43. El Kassas M., Alboraie M., Omar H., El Latif Y.A., Algaber M.A., El Tahan A., El Halwagy H., Afify S., Elserafy M., Elsaeed K.. High success rates for the use of ombitasvir/paritaprevir/ritonavir containing regimens in treatment of naïve and experienced chronic hepatitis C genotype 4: Real world results. J. Med. Virol. (2022.0) 94 667-674. DOI: 10.1002/jmv.25478 44. Yang Y., Wu F.-P., Wang W.-J., Shi J.-J., Li Y.-P., Zhang X., Dang S.S.. Real-life efficacy and safety of direct-acting antiviral therapy for treatment of patients infected with hepatitis C virus genotypes 1, 2 and 3 in northwest China. World J. Gastroenterol. (2019.0) 25 6551. DOI: 10.3748/wjg.v25.i44.6551 45. Kamel S., Elessawy H., Ashraf O., Elbaz A., Dabbous H., El-Sayed M., Ali S., Kamel H.. Effectiveness of Direct-Acting Antivirals in Treatment of Elderly Egyptian Chronic Hepatitis C Patients. Gastroenterol. Insights (2021.0) 12 336-346. DOI: 10.3390/gastroent12030031 46. Iliescu E.L., Mercan-Stanciu A., Toma L.. Safety and efficacy of direct-acting antivirals for chronic hepatitis C in patients with chronic kidney disease. BMC Nephrol. (2020.0) 21. DOI: 10.1186/s12882-020-1687-1 47. Tatar B., Kose S., Ergun N., Turken M., Onlen Y., Yilmaz Y., Akhan S.. Response to direct-acting antiviral agents in chronic hepatitis C patients with end-stage renal disease: A clinical experience. Rev. Assoc. Med. Bras. (2019.0) 65 1470-1475. DOI: 10.1590/1806-9282.65.12.1470 48. Marciano S., Haddad L., Reggiardo M.V., Peralta M., Vistarini C., Marino M., Descalzi V.I., D’Amico C., Escuti S.F., Gaite L.A.. Effectiveness and safety of original and generic sofosbuvir for the treatment of chronic hepatitis C: A real-world study. J. Med. Virol. (2018.0) 90 951-958. DOI: 10.1002/jmv.25033 49. Abd-Elsalam S., Sharaf-Eldin M., Soliman S., Elfert A., Badawi R., Ahmad Y.K.. Efficacy and safety of sofosbuvir plus ribavirin for treatment of cirrhotic patients with genotype 4 hepatitis C virus in real-life clinical practice. Arch. Virol. (2018.0) 163 51-56. DOI: 10.1007/s00705-017-3573-0 50. Buti M., Riveiro-Barciela M., Esteban R.. Management of direct-acting antiviral agent failures. J. Hepatol. (2015.0) 63 1511-1522. DOI: 10.1016/j.jhep.2015.08.010 51. Abdelsameea E., Alsebaey A., Abdel-Razek W., Ehsan N., Morad W., Salama M., Waked I.. Elastography and serum markers of fbrosis versus liver biopsy in 1270 Egyptian patients with hepatitis C. Eur. J. Gastroenterol. Hepatol. (2020.0) 32 1553-1558. DOI: 10.1097/MEG.0000000000001672 52. Prosół M., Bladowskab J., Ejmaa M., Stepniewskac K., Rymer W., Sąsiadek M., Pawłowski T., Małyszczak K., Inglot M., Żelwetro A.. Visual and brainstem auditory evoked potentials in HCV-infected patients before and after interferon-free therapy—A pilot study. Int. J. Infect. Dis. (2019.0) 80 122-128. DOI: 10.1016/j.ijid.2019.01.011
--- title: Intra-Oral Photograph Analysis for Gingivitis Screening in Orthodontic Patients authors: - Han-Na Kim - Kyuseok Kim - Youngjin Lee journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9967138 doi: 10.3390/ijerph20043705 license: CC BY 4.0 --- # Intra-Oral Photograph Analysis for Gingivitis Screening in Orthodontic Patients ## Abstract This study aimed to confirm the presence of gingival inflammation through image analysis of the papillary gingiva using intra-oral photographs (IOPs) before and after orthodontic treatment and to confirm the possibility of using gingival image analysis for gingivitis screening. Five hundred and eighty-eight ($$n = 588$$) gingival sites from the IOPs of 98 patients were included. Twenty-five participants who had completed their orthodontic treatments and were aged between 20 and 37 were included. Six points on the papillary gingiva were selected in the maxillary and mandibular anterior incisors. The red/green (R/G) ratio values were obtained for the selected gingival images and the modified gingival index (GI) was compared. The change in the R/G values during the orthodontic treatment period appeared in the order of before orthodontic treatment (BO), mid-point of orthodontic treatment (MO), three-quarters of the way through orthodontic treatment (TO), and immediately after debonding (IDO), confirming that it was similar to the change in the GI. The R/G value of the gingiva in the image correlated with the GI. Therefore, it could be used as a major index for gingivitis diagnosis using images. ## 1. Introduction The major dental diseases encountered in dentistry include dental caries, gingivitis, and periodontitis. Unfortunately, dentistry focuses on a downstream, patient-centered, curative, and rehabilitative approach to oral diseases [1]. Recently, there has been increased research on potential preventive approaches, with studies being conducted on the early diagnosis and screening of diseases before diagnosis [2]. Periodontal disease can be treated by dividing it into periodontitis and gingivitis. Gingivitis is a condition with no loss of alveolar bone in which the inflammation is limited to the gingiva; therefore, proper treatment can restore a healthy state [3]. Gingivitis is a reversible disease that can be restored to a healthy state, and it is necessary to prevent the disease from progressing to periodontitis through early diagnosis. The diagnosis of gingivitis relies on the identification of signs and symptoms of inflammation resulting from the disease in gingival tissues. It is difficult for patients to recognize periodontitis independently, and patients often expect to recover from gingivitis with time rather than seeking to actively treat it. Invasive indices such as the gingival index (GI) [4] use a periodontal probe that is adapted to gently probe around the gingival sulcus of each tooth in the mouth to enable bleeding and to enable qualitative changes in the marginal and interproximal tissues to be recorded. It is important that patients recognize bleeding and red inflammation as signs of disease and visit a dental institution. In addition to using a dental probe for gingivitis diagnosis, checking the depth of the periodontal pocket [5,6], and checking the level of inflammation through visual inspection, various methods have been attempted [7,8,9]. Many different methods have been introduced, and representative examples include diagnosis using saliva and diagnosis of diseases by X-ray using artificial intelligence technology [10,11]. These studies were performed to enable the diagnosis of diseases at an early stage so that treatment is possible and to enable diagnosis without visiting a dentist. Recently, as economic conditions have improved and interest in aesthetics has increased, the number of patients undergoing orthodontic treatment has increased. Patients undergoing orthodontic treatment find it particularly difficult to maintain satisfactory oral hygiene because of the presence of bands, wires, and ligatures. In particular, after the removal of orthodontic braces, teeth may be in poor oral condition owing to advanced dental caries, gingival recession, and severe periodontitis [12,13,14,15]. Therefore, oral health education for orthodontic patients is emphasized, and dental staff need to relay its importance whenever patients visit for treatment. Orthodontic patients generally have intra-oral photographs (IOPs) taken before, during, and after orthodontic treatment to record the state of orthodontic treatment. In cases of gingivitis or periodontitis, the presence of red inflammation in the gingiva can be confirmed on intra-oral radiographs, and it is necessary to confirm the possibility of diagnosing oral diseases using intra-oral radiographs. To date, there have been various attempts to diagnose diseases using IOPs, including oral cancer detection on an imaging system using a fluorescence enhancement method [16,17]. For the diagnosis of dental caries, an optical imaging system was used to distinguish sound teeth and caries lesions with quantitative values of the reflectance, transmittance, and absorbance [18]. Computer-assisted systems have been actively studied to improve the accuracy of dental health screening. In particular, radiographic and photographic intra-oral images have been intensively discussed when studying alveolar bone loss and periodontal disease [19,20,21]. Image analysis for periodontal diagnosis has mainly interpreted correlation with the disease by extracting features of abnormal symptoms, such as the degree of alveolar bone loss and periodontal redness. Sela et al. [ 22] introduced a method of structural analysis using morphological operators to segment trabeculae in dental X-ray images. This method has a limitation in that it is highly affected by degradation, such as noise. Statistical analysis based on methods including the gray level co-occurrence matrix (GLCM) [23], multichannel GLCM (MGLCM) [24], and domain transform methods [25] was applied to radiographic and photographic images to extract and classify features [26]. Based on the results of the extracted features, gingivitis, periodontitis, and the stage of periodontal disease were predicted. However, it is difficult to present a linear relationship between the extracted features and disease; therefore, there is a limit to deriving highly accurate results. Recently, machine learning-based approaches have shown superior results in the diagnosis of periodontitis [27,28]. Periodontal disease prediction methods based on convolution neural networks predict diseases with high accuracy using multiple radiographs, optical photographs, and patient information [29,30,31]. Chang et al. [ 32] demonstrated that high accuracy and reliability were achieved by training each network suitable for the purpose and providing the derived results by synthesizing them into one panoramic image. Li et al. [ 33,34] attempted to diagnose and predict gingivitis using a particle swarm optimization neural network incorporating contrast-limited adaptive histogram equalization and MGLCM methods. This combination approach derived more accurate and sensitive results compared to state-of-the-art methods. However, data-driven prediction methods are dependent on the data size and acquisition conditions. Therefore, it is necessary to verify the reference data (or label data) with considerable expertise and time. A study that attempted to screen for gingivitis automatically using deep learning technology suggested that the area under the curve for detecting gingivitis, dental calculus, and soft deposits was $87.11\%$, $80.11\%$, and $78.57\%$ using oral photos [31]; however, since the range was wide and oral sites not related to gingival inflammation were included, the reliability of their findings between the actual inflammation occurrence and the target data need to be confirmed. This study aimed to confirm the presence or absence of inflammation through image analysis of the papillary gingiva using IOPs before and after orthodontic treatment and to confirm the possibility of using the gingival image analysis method for periodontal disease screening. ## 2.1. Study Participants Ninety-eight oral photos, including 588 targeted gingivae, were captured from 25 patients admitted at three orthodontic dental clinics in Cheongju city, South Korea, between September 2018 and November 2022. A power analysis was performed using the GPower software, version 3.1.9.2 (Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany), using mean differences, with mean differences of $0.7\%$ for gingivitis from a related study. An error probability of 0.05 and actual power of 0.95 were used. The number of subjects calculated through the GPower software was 23. The number of subjects initially included in the study was 30, but in the process of evaluating the quality of the obtained photographs the final number of subjects was selected as 25 participants. A dental hygienist captured all the photos. The patients were aged between 20 and 39 years. Gingivitis was considered the primary disease in this study. The inclusion criteria were those who had completed orthodontic treatment in their 20s and 30s, had undergone orthodontic treatment for at least six months, had received orthodontic treatment with fixed orthodontic devices, and had been diagnosed with gingivitis by a dentist during orthodontic treatment. Patients were excluded from the study if they were aged 40 years or older, had systemic disease (such as diabetes and hypertension), had dental caries on the buccal smooth tooth surface, had undergone tooth extraction during orthodontic treatment, had noticeable discoloration of their teeth or had severe melanin pigmentation on the gingiva, or had gingiva that was dark red even though there was no inflammation. In addition, patients with abnormal anatomical structures due to periodontal disease were excluded. The details of exclusion criteria were as follows: subjects with a systemic disease within the past 6 months who are receiving continuous treatment, subjects with other diseases that cause inflammatory transformation of the gingiva, including oral cancer, or subjects taking antibiotics due to disease. Among patients aged ≥ 40 years, some patients with advanced periodontitis were included. Advanced periodontitis can cause alveolar bone loss and gingival recession. The size of the gingiva can be different depending on the participants’ oral condition due to the deformation of the gingiva; therefore, the size of the gingiva was not considered. None of the patients in this study had dental caries on smooth surfaces. The IOPs used in the research analysis were received by one data organizer who attached a number to each photograph to prevent exposure to personal information and ensure the patients’ anonymity. Then the IOPs were prepared for analysis. The requirement for informed consent was waived because this was a retrospective study and all data were anonymized. We received an exemption for institutional review board review from the Bioethics Review Committee of Cheongju University (1041107-202212-HR-053-01). ## 2.2. Targeted Gingiva A total of 588 targeted gingiva samples were included. The gingiva selected included six points of papillary gingiva in the maxillary and mandibular anterior incisors (FDI dental numbering system: Nos. 13, 12, 11, 21, 22, 23, 43, 42, 41, 31, 32, and 32) from 25 participants. As the orthodontic treatment progressed, four IOPs were taken before orthodontic treatment (BO, see the Figure 1A), mid-point of orthodontic treatment (BO, see the Figure 1B), three-quarters of the way through orthodontic treatment (TO, see the Figure 1C), and immediately after debonding (IDO, see the Figure 1D). ## 2.3. Proposed Framework to Measure the Redness Figure 2 shows a simplified flowchart of the proposed redness measurement scheme, which involves extracting several areas on the upper and lower gums around the tooth using high-definition imaging. First, oral images are acquired by a standard protocol using an advanced 4 k ultra high-definition (4 k UHD) optical camera (Nikon, CMOS-type, 24.2 Megapixels, AF-S NIKKOR 85 MM F1.4 G telephoto lens), and then the region is selected to measure the redness. The selected region, Cin, is composed of a three-dimensional image (i.e., width, height, and depth), which can be expressed as follows:[1]Cin={CR(x,y)CG(x,y)CB(x,y) 0≤CR,G,B(x,y)≤255, where x and y denote the coordinates of the oral image. Here, when the flash is used to ensure the appropriate brightness of the image, the halo artifact is often included in the periodontal region to obtain an image. We performed gamma correction [35] to emphasize the high intensity of the halo artifacts and the traditional gamma transformation given by [2]Cout=bCinγ where Cout is the result of the gamma correction, which is obtained by applying the two constant parameters of b and γ to control the shape of the transformation curve. In this study, we used $b = 1.0$ and γ=2.2, empirically. The corresponding values are not fixed and can change according to the exposure conditions in oral photography. The region of the artifact generated by the flash was separated using the Otsu method [36]. L is an index map with halo artifacts and Linv is the opposite. Cnew, which is an artifact-removed image, is generated by the element product between Cin and Linv. Here, ∘ is an element-wise multiplication operator. Finally, the mean values of the three-color channels, MeanR,G,B(x,y), were calculated as follows:[3]MeanR,G,B(x,y)=∑$x = 1$max∑$y = 1$maxIR,G,B(x,y)∑$x = 1$max∑$y = 1$maxLinv(x,y) where IR,G,B(x,y) indicate the intensity values of the three-color channels in Cnew(x,y). These processes were repeated in other areas, and datasets for statistical analysis were collected. Based on the above descriptions, we implemented the proposed algorithm using MATLAB TM version 8.3 (MathWorks, MA, USA) programming language and a normal workstation (operating system: Windows 10, CPU: 2.13 GHz, RAM: 64 GB). The obtained image was in JPEG format, and the image dimensions were 3300 × 5000 × 3. ## 2.4. Modified Gingival Index (GI) This study attempted to confirm the association between redness obtained from the images and the GI. The degree of gingivitis, which can be confirmed by visual inspection, was scored using a modified GI [37]. The degree of gingival inflammation, which reflected the same lesion as the red/green (R/G) ratio analysis of the participants in the IOP, was scored as 0–4 points with the modified GI according to the criteria presented by Tobias et al. ( Table 1) [37]. One researcher (HK), a dental hygienist, performed the GI analysis. The degree of gingivitis was confirmed by visually checking the images on the same computer. Duplicate analysis of $10\%$ of all subject photos confirmed that the concordance of the results was more than $95\%$. ## 2.5. Statistical Methods A descriptive statistical analysis of the gingival R/G ratio values at four time points during orthodontic treatment was performed. As a descriptive statistical analysis, the average and standard deviation of the gingival R values were presented for each of the six regions, and a correlation analysis between the modified GI and the highest R/G value was performed using the maxilla and mandible using Spearman’s test. According to orthodontic treatment, the differences in the R/G values were confirmed using Friedman’s and Kendall’s W. All analyses were performed using SPSS version 24.0 (IBM Corp., New York, NY, USA), and the alpha levels were set at 0.05. A value of 0.1 or less as a significance level was set as a tendency. ## 3.1. General Characteristics Twenty-five patients underwent orthodontic treatment. The study included eleven males and fourteen females. Of the participants, $75.7\%$ and $24.3\%$ were in their 20s and 30s, respectively. The average duration of orthodontic treatment was 22.1 months. The orthodontic devices used by the study subjects were mostly ceramic orthodontic devices ($64\%$) (Table 2). ## 3.2. Gingival Index (GI) Table 3 shows the changes in the GI according to the progress of orthodontic treatment. Although there was a difference in the degree according to the six gingival locations, gingival inflammation confirmed by the GI was confirmed to increase IDO and TO. In contrast, there was less inflammation BO and TO ($p \leq 0.05$). ## 3.3. R/G Ratio Table 4 shows changes in the R/G ratio according to the orthodontic treatment. Changes in the R/G values in the IOPs according to the progress of orthodontic treatment were confirmed in each of the six gingivae. The results of the R/G values analysis confirmed that IDO, TO, and MO were high in that order ($p \leq 0.05$). ## 3.4. Correlation between the GI and R/G Values Table 5 shows the results of the correlation analysis between the GI and R/G values. The correlation between the degree of gingival inflammation was visually confirmed, and the degree of redness was confirmed using photographic image analysis. The correlation coefficients in the maxilla were 0.63 for GI_MO and R/G_MO, 0.70 for GI_TO and R/G_TO, and 0.87 for GI_IDO and R/G_IDO, respectively ($p \leq 0.05$). Correlation coefficients in the mandible were 0.60 for GI_TO and R/G_TO and 0.73 for GI_IDO and R/G_IDO, respectively. ( $p \leq 0.05$). ## 4. Discussion Inspection of conditions such as color and contour changes in the gingiva is important in periodontal clinical examinations and disease determination [38,39]. It is important to confirm changes in the gingiva at a preclinical stage during the course of the disease. Since gingivitis is a stage without loss of the alveolar bone, it must be detected through early screening and appropriate measures should be taken. In this study, we compared the GI confirmed by visual inspection with the R/G value in the actual image to check whether the result of the image analysis was similar to the image and confirm the degree of gingival inflammation during orthodontic treatment. In this study, by analyzing photographs of suspected gingivitis using a recently developed image analysis method, it was found that the R/G value of the gingiva in the image can be used as an important indicator for gingivitis diagnosis. Previous studies using images have attempted to confirm periodontal inflammation. However, the R/G ratio was confirmed in a state that included part of the teeth or buccal mucosa beyond the gingival region in the image [33,34,35]. As a result, the method for selecting the target lesion was reconsidered to ensure reliability. In this study, we attempted to improve the precision of the research results by sectioning and analyzing only the gingiva papillary [40,41], where gingival inflammation first begins. Additionally, the results obtained from the images were compared with the GI, a clinical indicator, to confirm the reliability of the actual data, thereby increasing the value of the study. The inflammation of the gingiva confirmed by the GI increased immediately after debonding and three-quarters of the way through orthodontic treatment. In contrast, less inflammation was apparent before orthodontic treatment, at the mid-point of orthodontic treatment, and at the beginning of orthodontic treatment with orthodontic appliances. These findings confirm that there is an increase in gingivitis due to poor oral hygiene management with braces. Relatively little inflammation was observed at the mid-point of orthodontic treatment. However, patients received oral hygiene management education at the beginning of orthodontic treatment, and it was confirmed that the degree of inflammation was the lowest among those with a high interest in oral health. Ozlu et al. [ 42] reported that the Loë–Silness index is frequently used and that the second GI is often used as a secondary outcome to confirm the effectiveness of oral education in patients undergoing orthodontic treatment. However, the GI varies depending on the general condition of patients and the targeting area of the GI does not include the bonded dental surfaces. Bardal et al. emphasized [43] that oral healthcare guidance should be provided before and during treatment. Previous studies have suggested the use of mobile applications to deliver oral hygiene information and increase the efficiency of education [44,45]. During periodontal disease, host inflammatory cells are recruited and inflammatory cytokines such as IL-1β, IL-6, and TNF-α are released from fibroblasts, macrophages, connective tissue, and junctional epithelial cells. In particular, prostaglandins appear and increase microvascular permeability, and the color of the gingiva turns red or dark red [45]. The change in the R/G value during the orthodontic treatment period appeared in the order of immediately after the removal of the orthodontic appliance, three-quarters of the way through the orthodontic treatment, before the orthodontic appliance, and at the beginning of the orthodontic treatment, confirming that it was similar to the change in the GI. In other words, it was confirmed that the R/G value had a similar tendency to the change in the ranking of the GI value. Jönsson et al. [ 46] demonstrated the effectiveness of an individually tailored oral health education program for oral hygiene self-care in patients with chronic periodontitis. The validation of the GI was significantly correlated with histological parameters of inflammation during gingivitis development; specifically, the infiltrated connective tissue volume and its ratio with the volume of non-infiltrated connective tissue increased with increasing GI [47]. In the results of the correlation analysis between the R/G and GI, the highest value of each R/G and GI was selected and analyzed in the upper and lower jaws of the three gingivae. The correlation coefficient between the R/G and GI was not confirmed in the gingiva before orthodontic treatment. However, a coefficient of 0.63–0.87 was confirmed three-quarters of the way through orthodontic treatment and immediately after debonding. These results show that the R/G value has limitations in reflecting the degree of gingival inflammation in a state of low gingival inflammation, but the correlation was confirmed with GI scores of two or three where gingivitis occurred as a result of orthodontic treatment. However, in the image analysis, the gingival volume or edema due to inflammation was not reflected, and only the color change and R/G ratio were obtained. Therefore, some of the variables were not significantly confirmed in the results of the correlation analysis. The proposed method can derive the correlation between gingivitis and periodontitis from oral images without additional devices. However, several limitations need to be addressed to improve the performance. First, it is difficult to maintain the dynamic range under various exposure conditions, and it is necessary to unify the dynamic range of the obtained oral images by using the absolute intensity of the RGB color image. This problem is currently being studied through research and development. Another limitation is the accurate segmentation of the area saturated by the camera flash. The saturated value influences the result, and this problem can be overcome using more complicated methods, including the energy minimization method [48], Gaussian mixture model [49], prior knowledge-based method [50,51], multiscale-based method [52], and random walk method [53]. However, these approaches increase the computation time. Recently, deep-learning-based image segmentation methods have shown promising results [54] and are expected to improve the usefulness of the proposed method. Another limitation of the current study is that we could not include patients over 40 years old because of the possibility of gingival recession in the oral cavity. In patients over 40 years old, the shape of the papillary gingiva on imaging is different for each patient, which could cause a difference in the results. Standardized IOPs using digital cameras, excluding personal cell phones, were used for the analysis, so many photos could not be included in the current study. Future studies should be conducted that include a larger number of samples. The anatomical structure in the oral cavity is three-dimensional, but images are two-dimensional, so gingival edema or morphological changes were not considered in the current study. For this, it is necessary to check whether the volume and shape have changed in a future study using IOPs while expanding the age range of the study subjects. The failure to secure a sufficient number of study subjects is considered to be a limitation of this study. However, since the number of analyzed gingival regions was sufficient, significant results were confirmed during statistical analysis. A study that secures a sufficient number of research subjects is planned for the future Nevertheless, our findings are still valuable. The current study shows that it is important to compare and judge gingival inflammation using images and the GI within the range designated by experts. Furthermore, since the correlation between the two indicators has been confirmed, the analysis method using images can be used. Due to poor oral hygiene management, periodontal disease is accelerated by biofilms and bacteria in interdental spaces or oral environments. The relationship between cardiovascular disease and diabetes has been reported in many studies [55,56]. Moreover, increased carious teeth can be positively associated with the risk of cerebral or myocardial infarction [57]. As oral health management is a major factor in overall health, promoting oral hygiene and reducing oral inflammation must be continuously highlighted to patients. If an image analysis related to periodontal inflammation screening can be developed, patients’ overall health could be greatly improved. ## 5. Conclusions In this study, a recently developed image analysis method was applied to analyze gingivitis using IOPs of orthodontic patients. It was confirmed that the R/G value of the gingiva in the image was correlated with the GI. The change in the R/G value during the orthodontic treatment period appeared in the order of immediately after the removal of the orthodontic appliance, three-quarters of the way through orthodontic treatment, before the orthodontic appliance, and at the beginning of the orthodontic treatment, confirming that it was similar to the change in the GI. In addition, the change in the degree of gingivitis according to the progress of orthodontic treatment was confirmed using the R/G values, showing that it could be used as a major index for gingivitis diagnosis using images. ## References 1. Baelum V.. Dentistry and population approaches for preventing dental diseases. J. Dent. (2011.0) 39 S9-S19. DOI: 10.1016/j.jdent.2011.10.015 2. Landry R.G., Jean M.. Periodontal Screening and Recording (PSR) Index: Precursors, utility and limitations in a clinical setting. Int. Dent. J. (2002.0) 52 35-40. DOI: 10.1111/j.1875-595X.2002.tb00595.x 3. Eke P.I., Wei L., Borgnakke W.S., Thornton-Evans G., Zhang X.Y., Lu H., Mcguire L.C., Genco R.J.. Periodontitis prevalence in adults >= 65 years of age, in the USA. Periodontology 2000 (2016.0) 72 76-95. DOI: 10.1111/prd.12145 4. Loe H., Silness J.. Periodontal disease in pregnancy I. Prevalence and severity. Acta Odontol. Scand. (1963.0) 21 533-551. DOI: 10.3109/00016356309011240 5. Corbet E.F., Ho D.K.L., Lai S.M.L.. Radiographs in periodontal disease diagnosis and management. Aust. Dent. J. (2009.0) 54 S27-S43. DOI: 10.1111/j.1834-7819.2009.01141.x 6. Highfield J.. Diagnosis and classification of periodontal disease. Aust. Dent. J. (2009.0) 54 S11-S26. DOI: 10.1111/j.1834-7819.2009.01140.x 7. Schmalz G., Li S., Burkhardt R., Rinke S., Krause F., Haak R., Ziebolz D.. MicroRNAs as salivary markers for periodontal diseases: A new diagnostic approach?. BioMed Res. Int. (2016.0) 2016 1027525. DOI: 10.1155/2016/1027525 8. Xiang X., Sowa M.G., Lacopino A.M., Maev R.G., Hewko M.D., Man A., Liu K.Z.. An Update on Novel Non-Invasive Approaches for Periodontal Diagnosis. J. Periodontol. (2010.0) 81 186-198. DOI: 10.1902/jop.2009.090419 9. Elashiry M., Meghil M.M., Cutler C.W.. From manual periodontal probing to digital 3-D imaging to endoscopic capillaroscopy: Recent advances in periodontal disease diagnosis. J. Periodontal Res. (2019.0) 54 1-9. DOI: 10.1111/jre.12585 10. Trombelli L., Farina R., Silva C.O., Tatakis D.N.. Plaque-induced gingivitis: Case definition and diagnostic considerations. J. Clin. Periodontol. (2018.0) 45 S44-S67. DOI: 10.1111/jcpe.12939 11. Revilla-León M., Gómez-Polo M., Barmak A.B., Inam W., Kan J.Y., Kois J.C., Akal O.. Artificial intelligence models for diagnosing gingivitis and periodontal disease: A systematic review. J. Prosthetic Dent. (2022.0). DOI: 10.1016/j.prosdent.2022.01.026 12. Chang H.S., Walsh L.J., Freer T.J.. Enamel demineralization during orthodontic treatment. Aetiology and prevention. Aust. Dent. J. (1997.0) 42 322-327. DOI: 10.1111/j.1834-7819.1997.tb00138.x 13. Paolantonio M., D’ercole S., Perinetti G., Tripodi D., Catamo G., Serra E., Bruè C., Piccolomini R.. Clinical and microbiological effects of different restorative materials on the periodontal tissues adjacent to subgingival class V restorations. J. Clin. Periodontol. (2004.0) 31 200-207. DOI: 10.1111/j.0303-6979.2004.00472.x 14. Freitas A.O., Marquezan M., Mda C.N., Alviano D.S., Maia L.C.. The influence of orthodontic fixed appliances on the oral microbiota: A systematic review. Dental Press J. Orthod. (2014.0) 19 46-55. DOI: 10.1590/2176-9451.19.2.046-055.oar 15. Joss-Vassalli I., Grebenstein C., Topouzelis N., Sculean A., Katsaros C.. Orthodontic therapy and gingival recession: A systematic review. Orthod. Craniofacial Res. (2010.0) 13 127-141. DOI: 10.1111/j.1601-6343.2010.01491.x 16. Jiang C.F., Wang C.Y., Chiang C.P.. Comparative study of protoporphyrin IX fluorescence image enhancement methods to improve an optical imaging system for oral cancer detection. J. Biomed Opt. (2011.0) 16 076006. DOI: 10.1117/1.3595860 17. Malik B.H., Jabbour J.M., Cheng S., Cuenca R., Cheng Y.S.L., Wright J.M., Jo J.A., Maitland K.C.. A novel multimodal optical imaging system for early detection of oral cancer. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. (2016.0) 121 290-300. DOI: 10.1016/j.oooo.2015.10.020 18. Abdel Gawad A.L., El-Sharkawy Y.H., El-Sherif A.F.. Classification of human teeth caries using custom non-invasive optical imaging system. Lasers Dent. Sci. (2017.0) 1 73-81. DOI: 10.1007/s41547-017-0008-x 19. Teeuw W.J., Coelho L., Silva A., Van Der Palen C.J.N.M., Lessmann F.G.J.M., Van der Velden U., Loos B.G.. Validation of a dental image analyzer tool to measure alveolar bone loss in periodontitis patients. J. Periodontal Res. (2009.0) 44 94-102. DOI: 10.1111/j.1600-0765.2008.01111.x 20. Moutinho R.P., Coelho L., Silva A., Lobo Pereira J.A., Pinto M., Baptista I.P.. Validation of a dental image-analyzer tool to measure the radiographic defect angle of the intrabony defect in periodontitis patients. J. Periodontal Res. (2012.0) 47 695-700. DOI: 10.1111/j.1600-0765.2012.01483.x 21. Uthoff R.D., Song B., Sunny S., Patrick S., Suresh A., Kolur T., Keerthi G., Spires O., Anbarani A., Wilder-Smith P.. Point-of-care, smartphone-based, dual-modality, dual-view, oral cancer screening device with neural network classification for low-resource communities. PLoS ONE (2018.0) 13. DOI: 10.1371/journal.pone.0207493 22. Sela E.I., Hartati S., Harjoko A., Wardoyo R., Munakhir M.S.. Segmentation on the dental periapical X-ray images for osteoporosis screening. Int. J. Adv. Computer Sci. Appl. (2013.0) 4 147-151 23. Djuričić G.J., Radulovic M., Sopta J.P., Nikitović M., Milošević N.T.. Fractal and Gray Level Cooccurrence Matrix Computational Analysis of Primary Osteosarcoma Magnetic Resonance Images Predicts the Chemotherapy Response. Front. Oncol. (2017.0) 7 246. DOI: 10.3389/fonc.2017.00246 24. Huang X., Liu X., Zhang L.. A multichannel gray level co-occurrence matrix for multi/hyperspectral image texture representation. Remote Sens. (2014.0) 6 8424-8445. DOI: 10.3390/rs6098424 25. Faber T.D., Yoon D.C., Service S.K., White S.C.. Fourier and wavelet analyses of dental radiographs detect trabecular changes in osteoporosis. Bone (2004.0) 35 403-411. DOI: 10.1016/j.bone.2004.03.002 26. Lin P.L., Huang P.W., Huang P.Y., Hsu H.C.. Alveolar bone-loss area localization in periodontitis radiographs based on threshold segmentation with a hybrid feature fused of intensity and the H-value of fractional Brownian motion model. Comput. Methods Programs Biomed (2015.0) 121 117-126. DOI: 10.1016/j.cmpb.2015.05.004 27. Farhadian M., Shokouhi P., Torkzaban P.. A decision support system based on support vector machine for diagnosis of periodontal disease. BMC Res. Notes (2020.0) 13. DOI: 10.1186/s13104-020-05180-5 28. Chang J., Chang M.F., Angelov N., Hsu C.Y., Meng H.W., Sheng S., Ayilavarapu S.. Application of deep machine learning for the radiographic diagnosis of periodontitis. Clin. Oral Investig. (2022.0) 26 6629-6637. DOI: 10.1007/s00784-022-04617-4 29. Lee J.H., Kim D.H., Jeong S.N., Choi S.H.. Detection and diagnosis of dental caries using a deep learning-based convolutional neural network algorithm. J. Dent. (2018.0) 77 106-111. DOI: 10.1016/j.jdent.2018.07.015 30. Krois J., Ekert T., Meinhold L., Golla T., Kharbot B., Wittemeier A., Schwendicke F.. Deep learning for the radiographic detection of periodontal bone loss. Sci. Rep. (2019.0) 9 8495. DOI: 10.1038/s41598-019-44839-3 31. Li W., Liang Y., Zhang X., Liu C., He L., Miao L., Sun W.. A deep learning approach to automatic gingivitis screening based on classification and localization in RGB photos. Sci. Rep. (2021.0) 11 16831. DOI: 10.1038/s41598-021-96091-3 32. Chang H.J., Lee S.J., Yong T.H., Shin N.Y., Jang B.G., Kim J.E., Yi W.J.. Deep learning hybrid method to automatically diagnose periodontal bone loss and stage periodontitis. Sci. Rep. (2020.0) 10 7531. DOI: 10.1038/s41598-020-64509-z 33. Li W., Chen Y., Sun W., Brown M., Zhang X., Wang S., Miao L.. A gingivitis identification method based on contrast-limited adaptive histogram equalization, gray-level co-occurrence matrix, and extreme learning machine. Int. J. Imaging Sys. Tech. (2019.0) 29 77-82. DOI: 10.1002/ima.22298 34. Li W., Jiang X., Sun W., Wang S.H., Liu C., Zhang X., Miao L.. Gingivitis identification via multichannel gray-level co-occurrence matrix and particle swarm optimization neural network. Int. J. Imaging Sys. Tech. (2020.0) 30 401-411. DOI: 10.1002/ima.22385 35. Rahman S., Rahman M.M., Abdullah-Al-Wadud M., Al-Quaderi G.D., Shoyaib M.. An adaptive gamma correction for image enhancement. EURASIP J. Image Video Proc. (2016.0) 2016 35. DOI: 10.1186/s13640-016-0138-1 36. Otsu N.. A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man Cybern. (1979.0) 9 62-66. DOI: 10.1109/TSMC.1979.4310076 37. Tobias G., Spanier A.B.. Modified gingival index (MGI) classification using dental selfies. Appl. Sci. (2020.0) 10. DOI: 10.3390/app10248923 38. Armitage G.C.. The complete periodontal examination. Periodontology 2000 (2004.0) 34 22-33. DOI: 10.1046/j.0906-6713.2002.003422.x 39. Pihlstrom B.L.. Periodontal risk assessment, diagnosis and treatment planning. Periodontology 2000 (2001.0) 25 37-58. DOI: 10.1034/j.1600-0757.2001.22250104.x 40. Bergströum J.. The topography of papillary gingiva in health and early gingivitis. J. Clin. Periodontol. (1984.0) 11 423-431. DOI: 10.1111/j.1600-051X.1984.tb01341.x 41. Rebelo M.A.B., Queiroz A.C.D., Panagakos F.S., Davies R.M.. Gingival indices: State of art. Gingival Diseases-Their Aetiology, Prevention and Treatment (2011.0) 41-54 42. Ozlu F.C., Aktunc E., Yilmaz H., Karadeniz E.I.. Effectiveness of three different types of educational methods on implementation of proper oral hygiene behaviour prior to orthodontic treatment. Dent. Press J. Orthod. (2021.0) 26 e2119248. DOI: 10.1590/2177-6709.26.1.e2119248.oar 43. Bardal P.A.P., Olympio K.P.K., Bastos J.R.D.M., Henriques J.F.C., Buzalaf M.A.R.. Educação e motivação em saúde bucal: Prevenindo doenças e promovendo saúde em pacientes sob tratamento ortodôntico. Dent. Press J. Orthod. (2011.0) 16 95-102. DOI: 10.1590/S2176-94512011000300012 44. Patil S., Hedad I.A., Jafer A.A., Abutaleb G.K., Arishi T.M., Arishi S.A., Raj A.T.. Effectiveness of mobile phone applications in improving oral hygiene care and outcomes in orthodontic patients. J. Oral Bio Craniofacial Res. (2021.0) 11 26-32. DOI: 10.1016/j.jobcr.2020.11.004 45. Meyle J., Chapple I.. Molecular aspects of the pathogenesis of periodontitis. Periodontology 2000 (2015.0) 69 7-17. DOI: 10.1111/prd.12104 46. Jönsson B., Öhrn K., Oscarson N., Lindberg P.. The effectiveness of an individually tailored oral health educational programme on oral hygiene behaviour in patients with periodontal disease: A blinded randomized-controlled clinical trial (one-year follow-up). J. Clin. Periodontol. (2009.0) 36 1025-1034. DOI: 10.1111/j.1600-051X.2009.01453.x 47. Brecx M.C., Schlegel K., Gehr P., Lang N.P.. Comparison between histological and clinical parameters during human experimental gingivitis. J. Periodontal Res. (1987.0) 22 50-57. DOI: 10.1111/j.1600-0765.1987.tb01539.x 48. Rousson M., Brox T., Deriche R.. Active unsupervised texture segmentation on a diffusion based feature space. Proceedings of the 2003 IEEE Computer Society Conference on Computer Vision and Pattern Recognition 49. Riaz F., Rehman S., Ajmal M., Hafiz R., Hassan A., Aljohani N.R., Coimbra M.. Gaussian mixture model based probabilistic modeling of images for medical image segmentation. IEEE Access (2020.0) 8 16846-16856. DOI: 10.1109/ACCESS.2020.2967676 50. Navalpakkam V., Itti L.. An integrated model of top-down and bottom-up attention for optimizing detection speed. Proceedings of the 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06) 51. Liu L., Bian H., Yagi S.I., Yang X.. A prior-knowledge-based threshold segmentation method of forward-looking sonar images for underwater linear object detection. Jpn. J. Appl. Phys. (2016.0) 55 07KG06. DOI: 10.7567/JJAP.55.07KG06 52. Li J., Levine M.D., An X., Xu X., He H.. Visual saliency based on scale-space analysis in the frequency domain. IEEE Trans. Pattern Anal. Mach. Intell. (2012.0) 35 996-1010. DOI: 10.1109/TPAMI.2012.147 53. Craswell N., Szummer M.. Random walks on the click graph. Proceedings of the 30th Annual International ACM SIGIR Conference on Research and Development in Information Retrieval 239-246 54. Minaee S., Boykov Y., Porikli F., Plaza A., Kehtarnavaz N., Terzopoulos D.. Image segmentation using deep learning: A survey. IEEE Trans. Pattern Anal. Mach. Intell. (2022.0) 44 3523-3542. DOI: 10.1109/TPAMI.2021.3059968 55. Fagundes N.C.F., Carvalho A.P.C.P.S., Vilhena K.F.B., Magno M.B., Maia L.C., Lima R.R.. Periodontitis as a risk factor for stroke: A systematic review and meta-analysis. Vasc. Health Risk Manag. (2019.0) 15 519. DOI: 10.2147/VHRM.S204097 56. Song T.J., Kim J.W., Kim J.. Oral health and changes in lipid profile: A nationwide cohort study. J. Clin. Periodontol. (2020.0) 47 1437-1445. DOI: 10.1111/jcpe.13373 57. Song T.J., Jeon J., Kim J.. Cardiovascular risks of periodontitis and oral hygiene indicators in patients with diabetes mellitus. Diab. Metabol. (2021.0) 47 101252. DOI: 10.1016/j.diabet.2021.101252
--- title: The Immune, Inflammatory and Hematological Response in COVID-19 Patients, According to the Severity of the Disease authors: - Felicia Trofin - Eduard-Vasile Nastase - Andrei Vâță - Luminița Smaranda Iancu - Cătălina Luncă - Elena Roxana Buzilă - Mădălina Alexandra Vlad - Olivia Simona Dorneanu journal: Microorganisms year: 2023 pmcid: PMC9967162 doi: 10.3390/microorganisms11020319 license: CC BY 4.0 --- # The Immune, Inflammatory and Hematological Response in COVID-19 Patients, According to the Severity of the Disease ## Abstract Introduction: The aim of this study was to evaluate the immune and inflammatory responses in COVID-19 patients by dosing specific IgM and IgG total antibodies and interleukin 6, correlating them with the hematological and biochemical blood parameters and comparing them by the form of the disease. Materials and methods: One hundred twenty-five patients with polymerase chain reaction-confirmed COVID-19, hospitalized between 15.03.2020 and 1.07.2020 in the Clinical Hospital of Infectious Diseases “Sf. Parascheva” Iaşi, were tested by chemiluminescence for the presence of anti-SARS-CoV-2 IgM and IgG and IL-6 in the serum. The results were correlated with the results of the CBC count and serum biochemical parameters detected on the admission day. The patients presented different forms of the disease (asymptomatic, mild, moderate, severe, and critical) according to World Health Organization (WHO) criteria for the clinical management of COVID-19. Results: The amplitude of the immune response was directly correlated with the form of the disease. In the asymptomatic/mild form patients, the IL-6 and CRP concentrations were significantly higher and eosinophil count was significantly lower compared with the reference interval. In the moderate form, the concentrations of IL-6, CRP, and IgG were significantly higher, compared with the reference interval, while eosinophil count and eGFR were significantly lower. In severe/critical COVID-19 patients, IL-6, CRP, NLR, PLR, glucose, AST, urea, creatinine, and eGFR were significantly higher compared with the reference interval, while eosinophil count was significantly lower. IL-6 boosted in all forms of COVID-19, with a major increase in severe and critical patients. IL-6, neutrophil count, % neutrophils, NLR, PLR, CRP, AST, and urea increased with the severity of the SARS-CoV-2 infection, and the lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, and hemoglobin decreased with the increased severity of COVID-19. Conclusions: The amplitude and the moment of appearance of the immune response depended on the form of the disease. IgM generally occurred in the first 14 days of illness, and IgG appeared beginning with the second week of disease. IgG titer increased rapidly until the fourth week of disease and decreased slowly after 4 weeks. The amplitudes of all the tested inflammatory and serological markers depended on the COVID-19 form, increasing somewhat in the moderate forms and even more in the critical ones. The lymphocyte and eosinophil count are able to predict the risk of severe COVID-19. ## 1. Introduction The importance of coronaviruses (CoVs) was emphasized by doctors and researchers in 2002, with the outbreak of the severe acute respiratory syndrome (SARS) virus. Until then, CoVs had not generally been considered important pathogens for humans, as circulating strains caused only mild infections even in immunocompromised hosts [1]. In December 2019, a new virus, identified as the severe acute respiratory syndrome virus-2 (SARS-CoV-2), reminded us that CoVs are a severe threat to global health [2]. Laboratory blood tests are recommended to monitor the evolution and the prognosis of the coronavirus disease 2019 (COVID-19) [3]. Antibody detection plays an important role in the epidemiology of the infection because they can reveal the populations that had an immune response after vaccination or natural infection [1,4]. Anti-SARS-CoV-2 immunity is given by the sum of cell-mediated immune response and specific antibodies production. Identifiable antibodies in COVID-19 are immunoglobulins (Ig) M, G, and A. IgM occurs in the serum of patients approximately 5–7 days after the onset of the disease, while IgG occurs 10–21 days after the onset and may persist for a long time [1,2,4,5]. Many authors argued that the immune response in SARS-CoV-2 infection targets the receptor-binding domain (RBD) or other parts of the protein subunits [1]. Some authors described an early humoral immune response of IgA in COVID-19 patients, which provides a more rapid diagnosis and better tool for prognosis compared with IgM [6,7,8,9]. The medium seroconversion time for IgA is 4–6 days after the onset of symptoms [6]. The IL-6 concentration is known to predict the development of the severe COVID-19 and of the hypoxemia that will need hospitalization [10,11,12]. The severe form of this infection is also associated with a higher serum concentration of C-reactive protein (CRP) [11], glucose, alanine aminotransferase (ALT), aspartate aminotransferase (AST), creatinine [10,13,14], urea [14], estimated glomerular filtration rate (eGFR) [14], neutrophilia, and low lymphocytes and eosinophils count, respectively [10,13,15], and higher amount of serum ferritin [16]. A neutrophil-to-lymphocyte ratio (NLR) of 2.90 and a platelets-to-lymphocyte ratio (PLR) of 186 have good specificities for the association with severe forms of COVID-19. Reported cut-off values were 3.3 and 180 for NLR and PLR, respectively [15]. There are four different forms of COVID-19, according to the World Health Organization, and they depend on the severity of the disease: mild, moderate, severe, and critical [17]. Additionally, some infected patients did not develop any symptoms at all, having asymptomatic forms of COVID-19 [18]. Many questions revolve around COVID-19′s serological response. It is currently unclear what the antibodies’ titer requires for immune protection. The lifespan of the antibodies is uncertain. There have also been reported infections with no antibody detection, at least not during the study period [19]. Secondly, it is known that all forms of disease, including the mild ones, may be followed by complications, especially the development of the post-COVID-19 syndrome [20]. Some COVID-19 symptoms may persist for up to 3 months in $10\%$ of patients [18]. Usually, they manifest as one to three of the following symptoms: dyspnea, sleep disturbance, fatigue [21], asthenia, cough, or anosmia [22,23]. To the best of our knowledge, a limited number of studies have correlated the serological parameters with the forms of the SARS-CoV-2 infection, but only few brought up the immune and inflammatory responses or the blood parameters of the asymptomatic or mildly infected SARS-CoV-2 patients. The study was motivated by the fact that patients with mild or asymptomatic COVID-19 forms were not studied in detail, and often were even omitted entirely. The mild-form patients are thought to spread the disease more frequently, as they do not seek medical attention or self-implement social restrictions. Therefore, we proposed that the tested parameters should be related to the form of the infection. The aim of our study was to quantify the total anti-SARS-CoV-2 IgM and IgG and IL-6 in human serum from COVID-19 patients and to assess if any of the routine blood tests performed on the first day of hospitalization can predict the extent of the inflammatory response or that of the specific antibody production. ## 2.1. Study Design and Participants We conducted a prospective study that analyzed the antibody titers and IL-6 concentration in SARS-CoV-2-infected patients’ serum samples by the form of the disease. The inclusion criteria were as follows: a positive SARS-CoV-2 RT-PCR test at admission to hospital, aged over 18 years, and the agreement of the patient to participate in the study. The exclusion criteria were as follows: treatment with tocilizumab at the time of or in the days before the collection of the serum sample, insufficient amount of serum sample, incomplete hematological or biochemical data from the blood tests at hospital admission or at patient progression, or the patient not agreeing to participate in the study. We included 125 consecutive patients admitted to the Clinical Hospital of Infectious Diseases “Sf. Parascheva” Iași between 15 March and 1 July, 2020. SARS-CoV-2-positive patients were confirmed by positive RT-PCR through both nasopharyngeal and oropharyngeal swab samples. Depending on the signs and symptoms at the admission, they were diagnosed with different forms of COVID-19 (asymptomatic, mild, moderate, severe, and critical). The total specific IgM and/or IgG antibodies were tested for all patients, but, for economic reasons, we were able to dose IL-6 concentration for only 80 of the included patients. Demographic data (age, sex) and comorbidities were collected for all patients. These data are available in Table S1 in the Supplementary Materials. The results were corroborated with the form of the disease; CBC, NLR, PLR, CRP, ALT, AST, glucose, urea, creatinine, and eGFR on the day of hospital admission; and the length of hospital stay (LOS). ## 2.2. Collection of Samples Depending on LOS, 1, 2, or 3 sera from patients were collected. The number of serum samples collected from each patient was correlated with the number of days of hospitalization—one serum sample for patients who underwent 7–12 hospitalization days, two serum samples for patients with 12–20 hospitalization days, and three serum samples for patients with more than 21 hospitalization days—so that we could observe any significant dynamics. The first serum sample was used for anti-SARS-CoV-2 IgM and/or anti-SARS-CoV-2 IgG and IL-6 detection; the following two sera were tested only for specific IgG. The samples were transferred in Eppendorf tubes, frozen at −20 °C, and stored until analyzed. A total of 251 sera were collected as follows: for all 125 patients, one sample was taken 10 +/− 3 days from the onset of symptoms; for 90 patients, a second one was taken at 15 +/− 3 days; and for 36 patients from the patients who collected the second sample, a third sample was collected at 24 +/− 3 days after the onset of disease. Fifty-four patients had two serum samples collected and thirty-six patients had three serum samples taken. The sera were sampled from 125 consecutive patients who met the inclusion criteria (Figure 1). ## 2.3. Analysis of Samples Luminescence for total anti-SARS-CoV-2 specific antibodies (IgM and IgG, respectively) was detected by a chemiluminescence immunoassay. The light signal, which is measured by a photomultiplier as relative light units, is proportional to the concentration of anti-SARS-CoV-2 IgM or IgG present in the sample. Serum samples were evaluated using the MAGLUMI 2019-nCov IgM (Catalog number: 130219016M) and 2019-nCov IgG (Catalog number: 130219015M) kits (Snibe Diagnostic, Shenzhen, China). The cut off value according to the manufacturer is 1.00 AU/mL for both SARS-CoV-2 immunoglobulins. The manufacturer reported a sensitivity of $78.65\%$ for IgM detection and $91.21\%$ for IgG detection, while the specificities of IgM and IgG were 97.50 and $97.3\%$, respectively. The results were also analyzed individually as the mean of the concentrations obtained from all patients investigated in each 7-day interval from the onset of the infection. The IL-6 concentration was quantified by the same method, using the MAGLUMI IL-6 kits (Snibe Diagnostic, Shenzhen, China, Catalog number: 130616004M). The assay is linear between 1.5 pg/mL and 5000 pg/mL. ## 2.4. Ethical Principles The study complied with the ethical principles stated by the World Medical Association’s Declaration of Helsinki, regarding medical research involving human subjects. The study was approved by the Commission of Ethics of Research from the University of Medicine and Pharmacy “Grigore T. Popa” Iasi, Romania (IRB number: 99), and by the Hospital Ethics Committee (IRB number: 30). ## 2.5. Statistical Analysis We used the IBM SPSS statistical software version 20 for data analysis. The distribution of the variables was verified using the Kolmogorov–Smirnov test. For the normally distributed variables, we used the Pearson correlation test, and for the others we used the Spearman correlation test. For group statistics, we compared the groups using a one-way ANOVA test. The patients’ blood test results were compared with the reference interval of each parameter using a one-sample t-test. We considered a $p \leq 0.05$ as significant, due to the less than $5\%$ chance probability of the event occurring. The level of significance that we used refers to the α-significance, which measures the probability of false positives. The ROC curve and the area under curve (AUC) values were generated in order to assess the sensitivity and specificity of the tested biomarkers in predicting the risk of a severe form of COVID-19. The conclusions of the studies were supported by the results of the statistics tests. The mean concentration, median, standard deviation, variance, and range of concentrations were calculated using the same software. ## 3. Results The serological study was conducted on 125 patients. Among these patients 80 were tested for IL-6: seven ($8.75\%$) asymptomatic, ten ($12.5\%$) mild, twenty-nine ($36.25\%$) moderate, twenty-two ($27.5\%$) severe, and twelve ($15\%$) critical patients. They were merged into three groups: 17 ($21.25\%$) patients were placed in the asymptomatic/mild group, 29 ($36.25\%$) patients in the moderate group, and 34 ($42.5\%$) patients in the severe/critical group. This step aided us in the statistical analysis through batch balancing and more significant group sizes (Figure 1). The patients were aged between 19 and 88 years old. The mean age was 52 years old for the asymptomatic/mild form, 53 years old for the moderate form, and 59 for the severe/critical form of disease. Forty-five ($56.25\%$) of the patients were female. In most patients with the asymptomatic/mild form of the disease, IgM became detectable after the 7th day of illness. IgG was found in most patients with asymptomatic/mild forms, beginning with the second week of illness, with the highest concentration recorded in the fourth week of disease. In patients with moderate forms, the immune response began in the first week for IgM and in the second for IgG. The maximum value of the concentration mean per week was reached in the first week of hospitalization for IgM and in the fourth week for IgG. Almost all patients with severe/critical forms developed both types of antibodies after the seventh day of illness. The maximum mean concentration per week for IgM was recorded in the second week of illness, while for IgG it was recorded in the third week of the disease. IL-6 was increased in $74\%$ patients with asymptomatic/mild disease; the mean value for these patients was 19.57 pg/mL. The mean concentration of anti-SARS-CoV-2 IgG in these patients was 25.82 AU/mL. All patients from the other two groups had higher IL-6 levels. In patients with moderate forms, the mean IL-6 value was 73.26 pg mL and the mean anti-SARS-CoV-2 IgG concentration was 36.81 AU/mL. In severe/critical patients the mean values were 149.36 pg/mL and 64.17 AU/mL for IL-6 and anti-SARS-CoV-2 IgG, respectively. The mean values of certain blood parameters increased accordingly with the severity of the infection: leucocyte count, neutrophil count, % neutrophils, NLR, PLR, CRP, urea, and creatinine. On the other hand, there are some means that decrease depending on COVID-19 severity: lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, hemoglobin, and eGFR (Table 1). The relevant statistical parameters (mean, σ—standard deviation, median, and inter-quartile range, and variance values) for all the investigated parameters are listed in Table 1 and Table 2. We used the Kolmogorov–Smirnov test to assess the distribution of the variables. For normally distributed parameters we used the Pearson test to uncover whether there was any correlation between them. The LOS days did not correlate with % neutrophils, lymphocyte count, % lymphocytes, hemoglobin, or platelet count nor with the PLR. To correlate all the other variables which were not normally distributed ($p \leq 0.05$), we used the Spearman test. Its results underlined a statistically significant correlation between COVID-19 severity and the IL-6, IgG, neutrophil count, % neutrophils, NLR, PLR, CRP, AST, and urea (Table 3). The lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, and hemoglobin were inversely correlated with the form of the disease (Table 3). LOS correlated significantly only with creatinine ($r = 0.316$, $$p \leq 0.007$$) and eGFR (r = −0.342, $$p \leq 0.003$$). The IL-6 concentration correlated with the IgG titer, % neutrophils, NLR, PLR, and CRP and negatively correlated with lymphocyte count, % lymphocytes, hemoglobin, and eGFR (Table 4). The IgG titer was correlated with the leucocyte count, neutrophil count, % neutrophils, NLR, CRP, and ALT but only weakly correlated with the IL-6, PLR, and AST (Table 5). There was a moderate negative correlation between the IgG concentration and the % lymphocytes (Table 5). Using the One-Sample t-Test, we compared the obtained results of the blood tests of our patients with the reference intervals (Table 6). In asymptomatic/mild forms, biomarkers IL-6, CRP, and the IgG were significantly increased (Table 6). Asymptomatic/mild patients displayed significantly lower eosinophil counts and % eosinophils (Table 6). The moderate form of COVID-19 registered significantly higher values of IL-6, CRP, and IgG (Table 6). Patients with moderate forms displayed significantly decreased eosinophil counts and a significantly lower eGFR (Table 6). IgG, IL-6, CRP, NLR, PLR, glucose, AST, urea, creatinine, and eGFR were significantly higher in severe/critical COVID-19 patients. In the same group, eosinophil counts were significantly lower (Table 6). We compared the blood tests results as dependent variables with the sex of the patient as the independent variable using the independent samples t-test. This showed a significantly higher IgG secretion ($p \leq 0.001$) and NLR ($$p \leq 0.038$$) in adult males compared with adult females. We also compared all the blood parameters according to the COVID-19 severity groups. In this case, we used the one-way ANOVA test (Table 7). The form of the disease was considered as an independent variable and all the other parameters as dependent variables. This test highlighted significant differences between the three groups of COVID-19 severity in IL-6 concentrations, CRP, % neutrophils, lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, NLR, PLR, and urea (Table 7). On the other hand, in IgG levels, leucocyte count, neutrophils, hemoglobin, platelet count, LOS, glucose, ALT, AST, creatinine, and eGFR, the registered p was > 0.05. To appraise the sensitivity and specificity of the investigated biomarkers in predicting the severe forms of COVID-19, we performed ROC analysis. The variables that best predicted the risk of severe COVID-19 were the lymphocyte count (AUC = 0.779), lymphocyte percentage (AUC = 0.752), eosinophil count (AUC = 0.767), and eosinophil percentage (AUC = 0.768) (Supplementary Materials, Figure S1). The corresponding cut-off values were 1.275 × 103/µL (sensitivity = 0.733; specificity =_0.339) for lymphocyte count, $27.9\%$ (sensitivity = 0.733; specificity = 0.288) for lymphocyte percentage, 0.025 × 103/µL (sensitivity = 0.733; specificity =_0.322) for eosinophil count, and $1.05\%$ for eosinophil percentage (sensitivity = 0.533; specificity =_0.153). The IL-6 concentration and the severe form of COVID-19 are correlated with the presence of comorbidities (Table 8). Forty-six ($58.75\%$) patients had at least one cardiovascular disorder; thirty-two ($40\%$) had associated cardiovascular, metabolic, renal, or endocrinological conditions; and twenty-two ($27.5\%$) had no comorbidities (Table S1, Supplementary Materials). ## 4. Discussion Until now, most COVID-19 research has been focused on risk factors, such as age [24,25], body mass index [26], sex [27], specific geographical regions [28,29,30], pregnancy [31], ethnicity [32], or specific comorbidities [33,34,35,36,37,38]. However, there has recently been a trend of researchers reorienting themselves towards understanding the possible mechanisms of COVID-19 infection [39,40,41]. To our knowledge, few studies have targeted the evaluation of patients with asymptomatic/mild forms of the disease. In our study, in asymptomatic/mild patients, the mean seroconversion time was day 12 for IgM and day 14 for IgG. In the first week, fewer than $40\%$ of patients developed antibodies; the percentage increased beginning with day 15. Sun et al. [ 2020] showed that in most non-ICU patients IgM reached the peak in the second week after symptom onset [42]. Additionally, they observed that within one week after the symptom onset, the seropositive rates of IgM in non-ICU patients were $41.7\%$ [42]. This is in line with our findings. There are some reported cases in which anti-SARS-CoV-2 antibodies were not detected at all, or at least they did not develop in the time frame proposed for the study [6,7,8,41]. This statement is supported by our results, too. In our study, we observed that IgM became detectable in most of patients ($77\%$) with asymptomatic/mild forms beginning with day 7 of illness, while IgG became detectable at the end of the second week of illness in $95\%$ of the tested patients. The week of illness with the highest mean IgG concentration was approximately the same as the week of seroconversion in other studies. We calculated the peak of the mean concentration for IgG as being in the fourth week. In $22.6\%$ of patients with asymptomatic/mild forms, both immunoglobulins were absent in the tested sera. This can be explained in two ways: either they did not develop specific antibodies or they were not present during the study period. Twelve ($38.7\%$) patients had a negative IgM test result in their first week of illness but had a positive test for IgG in the second or in the third week. This may be due to an earlier onset of disease than was declared. Eight ($25.8\%$) patients developed IgM in the first 10 days, but they did not develop IgG until day 21, possibly due to a later IgG seroconversion. Patients with moderate forms of COVID-19 developed IgM 5 days earlier than the asymptomatic/mild patients. The immune response in their sera was observed during the first week of the infection for IgM, and in the second week for IgG. The maximum mean concentration per week was reached in the first 7 days for IgM, and in the fourth week for IgG. This can be explained by an earlier onset of the suggestive symptoms of disease compared with the patients with the asymptomatic/mild form. Seventeen ($45.9\%$) patients developed both IgM and IgG within the study period. A few patients ($8.1\%$) were positive either only for IgM or only for IgG by the 14th day of hospitalization. The presence of IgM, but not IgG, within two weeks of hospitalization may be explained by IgM decline and IgG tardive production. We recorded six ($13.6\%$) patients with moderate COVID-19 with undetectable antibodies by day 20. Only for some severe/critical patients were IgM detected in the first week of illness and IgG in the second week. The concentration of IgM reached its maximum value within 7 days, while the IgG concentration increased until the third week of the disease. One third of the patients developed both types of antibodies before day 14. Twelve ($37.5\%$) patients were negative for IgM in the second week, but they tested positive for IgG at that point. On the other hand, six ($13\%$) patients tested positive for IgM in the second week, but were negative for IgG until the third week. The results for the severe/critical patients are similar to those of other authors. Guo et al. [ 2020] [1] and Theel et al. [ 2020] [4] concluded that IgM can be detected during the first week after the onset of symptoms but IgG can be detected only at around 14 days. Wang et al. [ 2020] [2] observed that the average time of IgM occurrence was about 5 days, while for IgG it was about 14 days. Many other authors observed that IgM occurred in the serum of patients approximately 5–7 days after the onset of the disease, while IgG occurred after 10–21 days and could persist for a long time [6,7,8,41]. The life span of the anti-SARS-CoV-2 antibodies is uncertain. However, using the existing information on the other coronaviruses, we can assume that the anti-SARS-CoV-2 antibodies decrease over time (12–52 weeks after the onset of symptoms) [9]. Numerous SARS-CoV-2 reinfections have also been documented [9]. It is known that $90\%$ of patients are positive for IgG two years after SARS-CoV infection, and in $50\%$ of cases, these antibodies persist for over 3 years [19]. Zeng et al. [ 2020] stated that, in $80\%$ of cases, IgM persisted until day 49 [6]. We observed an increase in IgG concentration until a peak in the fourth week. The concentration increase is important in the first 4 weeks, while the decrease is slow after the fourth week. Sun et al. [ 2020] [42] observed that the antibody response gradually increased for weeks 1–3 after the onset of the disease and that IgM reached a peak in the second week, while IgG antibodies continued to increase in the third week. Our results are similar to those obtained by Sun et al. [ 2020] [42] for all forms of the disease. The absence of detectable antibodies could be due to the low sensitivity of the method ($74.5\%$ in first 7 days of disease). This may explain the variability in the antibody detection described so far by several authors due to the sensitivity of detection kits and the moment of appearance of the antibodies and their low titer [4,5,6,7,8]. Serological tests seem to be gaining popularity due to their low price, short time to results, use of common equipment, and accessibility of detection methods. The evolution of the infection can be predicted by measuring: IL-6, leucocyte count, neutrophil count, % neutrophils, NLR, PLR, CRP, urea, and creatinine, which tend to increase with COVID-19 severity, or lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, hemoglobin, and eGFR, which decrease in correlation with the severity of the disease. Most of the asymptomatic/mild-form patients had elevated IL-6 concentrations. Hambali et al. [ 2020] also described a non-severe COVID-19 patient with persistently high IL-6 level [43]. Our patients with moderate, severe, and critical forms of disease had higher concentrations of IL-6. Zhang et al. [ 2020] [12] observed similar differences in IL-6 concentrations according to the form of the infection. We also detected certain blood parameters in the asymptomatic/mild form which were significantly increased, such as CRP, or were significantly lower, such as the eosinophil count and % eosinophils. Gu et al. [ 2021] mentioned that the WBC differential count may be abnormal in asymptomatic/mild COVID-19 patients [44]. IL-6 and CRP were significantly increased in the moderate form, while eosinophil count and eGFR were significantly lowered (Table 3, Table 4, Table 5, Table 6 and Table 7). IL-6, CRP, NLR, PLR, glucose, AST, urea, creatinine, and eGFR were significantly higher in our severe/critical COVID-19 patients. On the other hand, eosinophil count is significantly decreased in the same group (Table 3, Table 4, Table 5, Table 6 and Table 7). LOS did not correlate with % neutrophils, lymphocyte count, % lymphocytes, hemoglobin, or platelet count nor with the PLR. It is influenced by the creatinine level (moderate, positive correlation) and eGFR (moderate, negative correlation). IL-6, neutrophil count, % neutrophils, NLR, PLR, CRP, AST, and urea increased with the severity of the infection. The lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, and hemoglobin decreased with the severity of COVID-19 (Table 3, Table 4, Table 5, Table 6 and Table 7). As the IL-6 concentration increased, the CRP, % neutrophils, NLR, and PLR increased and the lymphocyte count, % lymphocytes, hemoglobin, and eGFR decreased. The NLR and PLR rose as CRP and AST increased. NLR and PLR were also positively strongly correlated one with each other, meaning that both increased at the same time. The statistical comparison tests showed a significant difference between males and females only in IgG secretion and NLR: males develop a higher immune response and a higher NLR value compared to women. Boon et al. [ 2021] stated that females develop a weaker immune response after infection and vaccination due to estrogen secretion [45]. We used the one-way ANOVA test to compare the blood parameters according to the severity of COVID-19. According to the f test of the One-Way ANOVA, there were significant differences between the three severity groups of COVID-19 in IL-6, CRP, % neutrophils, lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, NLR, PLR, and urea (Table 7). On the other hand, concerning the leucocyte count, neutrophil count, hemoglobin, platelet count, LOS, glucose, ALT, AST, creatinine, and eGFR, there were no significant differences between the three groups. The test of homogeneity of variances showed that the variances of the IL-6 were not homogeneous. Based on the results, in terms of IL-6, there were significant differences between patients with asymptomatic/mild disease, those with moderate disease, and those with severe/critical disease, respectively, meaning that patients in the asymptomatic/mild category had a significantly lower level of IL-6 compared with those with the moderate or severe/critical form of disease. The statistical analysis of CRP, NLR, and PLR showed a significant difference between the severe/critical and the asymptomatic/mild or moderate forms, respectively. Since the variances of the % neutrophils and % lymphocytes were homogeneous, we applied Tukey’s post hoc test. The results allowed us to conclude that there were significant differences in % neutrophils and % lymphocytes between the severe/critical forms, the moderate forms, and the asymptomatic/mild forms of COVID-19. The variances of the lymphocyte count were not homogeneous. Upon applying the Games–Howell test, we found that there was a significant difference between the asymptomatic/mild forms of the disease and the moderate and severe/critical forms, respectively, regarding the lymphocyte count category. For the result of the homogeneity of variances test for % eosinophils, we used Tukey’s post hoc test. This test showed a significant difference between patients with asymptomatic/mild forms of disease compared to those with moderate and severe/critical forms, respectively. Our results are similar to those of Hou et al. [ 2020] [46]. They concluded that severe COVID-19 disease was associated with significantly increased neutrophils, infection biomarkers (such as CRP), and cytokine levels and decreased lymphocyte counts [46]. Xu et al. [ 2020] showed that the IL-6 concentration was significantly higher in fatal forms of COVID-19 when compared with patients with mild cases [47,48]. Ding et al. [ 2021] compared the blood parameters of the patients with mild/moderate cases to patients with severe COVID-19 [49]. We noticed that no significant differences in WBC counts and hemoglobin were observed between the compared groups of patients. The WBC differential count in the severe/life-threatening cases exhibited an increase in neutrophil count and a decrease in lymphocyte and eosinophil counts. NLR and CRP levels increased more in the severe COVID-19 group compared with the mild/moderate group [49]. The review by Velavan and Meyer [2020] [13] showed that the WBC was within the reference interval in all the tested cases. Mo et al. [ 2021] concluded that platelets count was normal or slightly lower in severe cases [50]. Some authors found that the neutrophil count [51,52], creatinine [53], and glucose [54] were significantly increased in critically ill patients. Other authors found that lymphocyte count [55,56,57] and eosinophil count [52] declined in most cases. The CRP value was described as increasing in most cases [50,57]. Chen et al. [ 53] observed that the IL-6 concentrations increased according to the severity of the disease from the mild to the critical form (mild ˂ severe ˂ critical). Our results are in line with the results described above. Our findings are also similar to those of Hachim et al. [ 2020], who proved that raised urea concentration and decreased lymphocyte count could predict the admission to an intensive care unit [14]. Man et al. [ 2021] argued that NLR and PLR have been proven to be reliable markers in COVID-19 patients and are increasingly correlated with CRP [15]. Similar findings were found in our study. Xia et al. [ 2021] [58] and Xiang et al. [ 2021] [59] established the presence a high proportion of early kidney function injury in COVID-19 patients at admission. The decline of eGFR and the escalation of creatinine on admission were related to poor prognosis. Our results support their conclusions. The absence of an increase in LOS in critically ill patients and the lack of correlation between LOS and this form of the disease is due to the fact that, in several critically ill patients, death occurred after just 14 days. In asymptomatic/mild-form patients, the IL-6 and CRP concentrations were significantly higher and eosinophil count was significantly lower compared with the reference interval. In the moderate form, the concentration of IL-6, CRP, and IgG were significantly higher compared with the reference interval, while eosinophil count and eGFR were significantly lower. In severe/critical COVID-19 patients, IL-6, CRP, NLR, PLR, glucose, AST, urea, creatinine, and eGFR were significantly higher compared with the reference interval, while eosinophil count was significantly lower. As IL-6, CRP, NLR, PLR, lymphocyte count, and eosinophil count are strongly correlated with the form of the disease (Table 3), we can state that they may represent a minimum set of markers to be tested as the best predictors for the aggressive evolution of COVID-19. Additionally, as ROC analyses revealed, lymphocyte count, lymphocyte percentage, eosinophil count, and eosinophil percentage can predict severe forms of COVID-19. Comorbidities are correlated with the severe form of the disease and IL-6 concentration. Patients who had more comorbidities were associated with a higher concentration of IL-6 and developed a more severe form of the disease. Additionally, patients who had diabetes, malignancies, or renal dysfunction had more severe forms compared to the patients with cardiovascular diseases, mostly presenting critical forms that culminated in death. Our results are in line with Sanyaolu et al. [ 2020] who concluded that COVID-19 patients with history of hypertension, diabetes, and cardiovascular disease had bad prognoses. Additionally, chronic kidney disease patients and cancer patients are not only at risk for contracting the virus, but there is a significantly increased risk of death among these groups of patients [60]. The most frequent comorbidity encountered in our patients was hypertension ($58.75\%$ patients). Our results are similar to those of Franki et al. [ 2020], who observed that one of the leading comorbidities among COVID-19 deaths in NY, USA, was hypertension, as seen in $55.4\%$ of their patients [61]. A large group of patients ($40\%$) had two or more of the following comorbidities: cardiovascular, metabolic, renal, or endocrinological conditions. ## 4.1. Strengths of the Study A strong point of our study was the analysis of a statistically significant number of serum samples from patients with different forms of the disease, including asymptomatic patients, thus enabling us to obtain an image of how the body reacts when triggering the necessary physical response depending on each form of COVID-19. Most of the patients had mild or moderate forms of COVID-19. Another strength of the present study was that all the patients included in our cohort were infected at the beginning of the pandemic and all had a prompt hospital admission, so all the serum samples were collected at the onset of the disease, leading to the most correct detection as possible of the onset of the antibody response. In general, most authors have focused on searching for the risk factors or to understand possible mechanisms in severe or critical COVID-19; however, this paper also explored the variations of the blood parameters, not only in moderate or severe COVID-19 cases but also in mild or asymptomatic patients. Our study explored a multitude of factors that may influence the severity of the disease (e.g., sex, age, comorbidities) and correlated a wide range of hematological, inflammatory, biochemical, and serological tests. Although the study was carried in a single hospital, patients from all northeastern Romania were hospitalized there and were included in the study. ## 4.2. Limitations of the Study Our study has several shortcomings. We were not able to assess the antibody titers at longer time intervals for all patients due to their different LOS. We could not follow the evolution of the patients after their hospital discharge due to the lack of availability of all patients; only a small number of patients, statistically insignificant, returned for clinical–biological reevaluation at the hospital. ## 5. Conclusions The amplitude and the time-point of the appearance of a detectable immune response depended on the severity of COVID-19 and the sensitivity of the test that was used—seroconversion being more frequently reported in severe cases. IL-6 was increase in all forms of COVID-19, with a major rise in severe and critical patients. IL-6, neutrophil count, % neutrophils, NLR, PLR, CRP, AST, and urea rose with the increased severity of the SARS-CoV-2 infection, and lymphocyte count, % lymphocytes, eosinophil count, % eosinophils, hemoglobin decreased with the increased severity of COVID-19. These changes allow us to conclude that all COVID-19 patients must be surveilled carefully—patients who develop asymptomatic or mild forms of disease should not be neglected. The negative evolution of COVID-19 can be presumed by measuring the IL-6, CRP, lymphocytes, and NLR, which are strongly correlated. The lymphocyte count and the eosinophil count can predict the severe form of COVID-19. Patients with associated comorbidities developed more severe forms of COVID-19. The LOS depended on the kidney injury in COVID-19 patients. ## References 1. Guo L., Ren L., Yang S., Xiao M., Chang D., Yang F., Dela Cruz C.S., Wang Y., Wu C., Xiao Y.. Profiling Early Humoral Response to Diagnose Novel Coronavirus Disease (COVID-19). Clin. Infect. Dis. (2020.0) 71 778-785. DOI: 10.1093/cid/ciaa310 2. Wang W., Xu Y., Gao R., Lu R., Han K., Wu G., Tan W.. Detection of SARS-CoV-2 in Different Types of Clinical Specimens. JAMA (2020.0) 323 1843-1844. DOI: 10.1001/jama.2020.3786 3. Bawiskar N., Talwar D., Acharya S., Kumar S.. Hematological Manifesta-tions of COVID-19 and Their Prognostic Significance in an Intensive Care Unit: A Cross-Sectional Study. Cureus (2021.0) 13 e19887. DOI: 10.7759/cureus.19887 4. Theel E.S., Slev P., Wheeler S., Couturier M.R., Wong S.J., Kadkhoda K.. The Role of Antibody Testing for SARS-CoV-2: Is There One?. J. Clin. Microbiol. (2020.0) 58 e00797-20. DOI: 10.1128/JCM.00797-20 5. Immune Responses and Immunity to SARS-CoV-2 6. Ma H., Zeng W., He H., Zhao D., Jiang D., Zhou P., Cheng L., Li Y., Ma X., Jin T.. Serum IgA, IgM, and IgG responses in COVID-19. Cel. Molec. Immunol. (2020.0) 17 773-775. DOI: 10.1038/s41423-020-0474-z 7. Wang P.. Significance of IgA antibody testing for early detection of SARS-CoV-2. J. Med. Virol. (2021.0) 93 1888-1889. DOI: 10.1002/jmv.26703 8. Infantino M., Manfredi M., Grossi V., Lari B., Fabbri S., Benucci M., Fortini A., Damiani A., Mobilia E.M., Panciroli M.. Closing the serological gap in the diagnostic testing for COVID-19: The value of anti-SARS-CoV-2 IgA antibodies. J. Med. Virol. (2021.0) 93 1436-1442. DOI: 10.1002/jmv.26422 9. Yu H.Q., Sun B.Q., Fang Z.F., Zhao J.C., Liu X.Y., Li Y.M., Sun X.Z., Liang H.F., Zhong B., Huang Z.F.. Distinct features of SARS-CoV-2-specific IgA response in COVID-19 patients. Europ. Resp. J. (2020.0) 56 2001526. DOI: 10.1183/13993003.01526-2020 10. Sabaka P., Koščálová A., Straka I., Hodosy J., Lipták R., Kmotorková B., Kachlíková M., Kušnírová A.. Role of interleukin 6 as a predictive factor for a severe course of COVID-19: Retrospective data analysis of patients from a long-term care facility during COVID-19 outbreak. BMC Infect. Dis. (2021.0) 21. DOI: 10.1186/s12879-021-05945-8 11. Herold T., Jurinovic V., Arnreich C., Lipworth B.J., Hellmuth J.C., von Bergwelt-Baildon M., Klein M., Weinberger T.. Elevated levels of IL-6 and CRP predict the need for mechanical ventilation in COVID-19. J. Allergy Clin. Immunol. (2020.0) 146 128-136.e4. DOI: 10.1016/j.jaci.2020.05.008 12. Zhang J., Hao Y., Ou W., Ming F., Liang G., Qian Y., Cai Q., Dong S., Hu S., Wang W.. Serum interleukin-6 is an indicator for severity in 901 patients with SARS-CoV-2 infection: A cohort study. J. Transl. Med. (2020.0) 18 406. DOI: 10.1186/s12967-020-02571-x 13. Velavan T.P., Meyer C.G.. Mild versus severe COVID-19: Laboratory markers. Int. J. Infect. Dis. (2020.0) 95 304-307. DOI: 10.1016/j.ijid.2020.04.061 14. Hachim M.Y., Hachim I.Y., Naeem K.B., Hannawi H., Salmi I.A., Hannawi S.. D-dimer, Troponin, and Urea Level at Presentation With COVID-19 can Predict ICU Admission: A Single Centered Study. Front. Med. (2020.0) 7 585003. DOI: 10.3389/fmed.2020.585003 15. Man M.A., Rajnoveanu R.M., Motoc N.S., Bondor C.I., Chis A.F., Lesan A., Puiu R., Lucaciu S.R., Dantes E., Gergely-Domokos B.. Neutrophil-to-lymphocyte ratio, platelets-to-lymphocyte ratio, and eosinophils correlation with high-resolution computer tomography severity score in COVID-19 patients. PloS ONE (2021.0) 16. DOI: 10.1371/journal.pone.0252599 16. Hulkoti V.S., Acharya S., Kumar S., Talwar D., Khanna S., Annadatha A., Madaan S., Verma V., Sagar V.V.S.S.. Association of serum ferritin with COVID-19 in a cross-sectional study of 200 intensive care unit patients in a rural hospital: Is ferritin the forgotten biomarker of mortality in severe COVID-19?. J. Fam. Med. Prim. Care (2022.0) 11 2045-2050. DOI: 10.4103/jfmpc.jfmpc_1921_21 17. Coronavirus Disease (COVID-19) Pandemic 18. Living with COVID-19—Second Review. (2021.0) 19. Petersen L.R., Sami S., Vuong N., Pathela P., Weiss D., Morgenthau B.M., Henseler R.A., Daskalakis D.C., Atas J., Patel A.. Lack of Antibodies to Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) in a Large Cohort of Previously Infected Persons. Clin. Infect.Dis. (2021.0) 73 e3066-e3073. DOI: 10.1093/cid/ciaa1685 20. Malkova A., Kudryavtsev I., Starshinova A., Kudlay D., Zinchenko Y., Glushkova A., Yablonskiy P., Shoenfeld Y.. Post COVID-19 Syndrome in Patients with Asymptomatic/Mild Form. Pathogens (2021.0) 10. DOI: 10.3390/pathogens10111408 21. Carod-Artal F.J.. Post-COVID-19 syndrome: Epidemiology, diagnostic criteria and pathogenic mechanisms involved. Rev. Neurol. (2021.0) 72 384-396. DOI: 10.33588/rn.7211.2021230 22. Ladds E., Rushforth A., Wieringa S., Taylor S., Rayner C., Husain L., Greenhalgh T.. Persistent symptoms after COVID-19: Qualitative study of 114 “long Covid” patients and draft quality principles for services. BMC Health Serv. Res. (2020.0) 20. DOI: 10.1186/s12913-020-06001-y 23. Coronavirus (COVID-19) Infection Survey, UK Statistical Bulletins 24. Covino M., De Matteis G., Polla D.A.D., Santoro M., Burzo M.L., Torelli E., Simeoni B., Russo A., Sandroni C., Gasbarrini A.. Predictors of in-hospital mortality AND death RISK STRATIFICATION among COVID-19 PATIENTS aged ≥ 80 YEARs OLD. Arch. Geront. Geriat. (2021.0) 95 104383. DOI: 10.1016/j.archger.2021.104383 25. Fagard K., Gielen E., Deschodt M., Devriendt E., Flamaing J.. Risk factors for severe COVID-19 disease and death in patients aged 70 and over: A retrospective observational cohort study. Acta Clin. Belg. (2022.0) 77 487-494. DOI: 10.1080/17843286.2021.1890452 26. Kompaniyets L., Goodman A.B., Belay B., Freedman D.S., Sucosky M.S., Lange S.J., Gundlapalli A.V., Boehmer T.K., Blanck H.M.. Body Mass Index and Risk for COVID-19-Related Hospitalization, Intensive Care Unit Admission, Invasive Mechanical Ventilation, and Death–United States, March-December 2020. MMWR. Morb. Mortal. Wkly. Rep. (2021.0) 70 355-361. DOI: 10.15585/mmwr.mm7010e4 27. Peckham H., de Gruijter N.M., Raine C., Radziszewska A., Ciurtin C., Wedderburn L.R., Rosser E.C., Webb K., Deakin C.T.. Male sex identified by global COVID-19 meta-analysis as a risk factor for death and ITU admission. Nat. Commun. (2021.0) 11 6317. DOI: 10.1038/s41467-020-19741-6 28. Silva I., Faria N.C., Ferreira Á.R.S., Anastácio L.R., Ferreira L.G.. Risk factors for critical illness and death among adult Brazilians with COVID-19. Rev. Soc. Bras. Med. Trop. (2021.0) 54 e0014-2021. DOI: 10.1590/0037-8682-0014-2021 29. Telle K.E., Grøsland M., Helgeland J., Håberg S.E.. Factors associated with hospitalization, invasive mechanical ventilation treatment and death among all confirmed COVID-19 cases in Norway: Prospective cohort study. Scand. J. Pubic Health (2021.0) 49 41-47. DOI: 10.1177/1403494820985172 30. Grasselli G., Greco M., Zanella A., Albano G., Antonelli M., Bellani G., Bonanomi E., Cabrini L., Carlesso E.. Risk Factors Associated With Mortality Among Patients With COVID-19 in Intensive Care Units in Lombardy, Italy. JAMA Intern. Med. (2020.0) 180 1345-1355. DOI: 10.1001/jamainternmed.2020.3539 31. Sakowicz A., Ayala A.E., Ukeje C.C., Witting C.S., Grobman W.A., Miller E.S.. Risk factors for severe acute respiratory syndrome coronavirus 2 infection in pregnant women. Am. J. Obstet. Gynecol. MFM (2020.0) 2 100198. DOI: 10.1016/j.ajogmf.2020.100198 32. Mathur R., Rentsch C.T., Morton C.E., Hulme W.J., Schultze A., MacKenna B., Eggo R.M., Bhaskaran K., Wong A.Y.S., Williamson E.J.. Ethnic differences in SARS-CoV-2 infection and COVID-19-related hospitalisation, intensive care unit admission, and death in 17 million adults in England: An observational cohort study using the OpenSAFELY platform. Lancet (2021.0) 397 1711-1724. DOI: 10.1016/S0140-6736(21)00634-6 33. Choi Y.J., Park J.Y., Lee H.S., Suh J., Song J.Y., Byun M.K., Cho J.H., Kim H.J., Lee J.H., Park J.W.. Effect of asthma and asthma medication on the prognosis of patients with COVID-19. Eur. Resp. J. (2021.0) 57 2002226. DOI: 10.1183/13993003.02226-2020 34. Khunti K., Knighton P., Zaccardi F., Bakhai C., Barron E., Holman N., Kar P., Meace C., Sattar N., Sharp S.. Prescription of glucose-lowering therapies and risk of COVID-19 mortality in people with type 2 diabetes: A nationwide observational study in England. Lancet Diabetes Endocrinol. (2021.0) 9 293-303. DOI: 10.1016/S2213-8587(21)00050-4 35. Xia Y., Jin R., Zhao J., Li W., Shen H.. Risk of COVID-19 for patients with cancer. Lancet Oncol. (2020.0) 21 e180. DOI: 10.1016/S1470-2045(20)30150-9 36. Rüthrich M.M., Giessen-Jung C., Borgmann S., Classen A.Y., Dolff S., Grüner B., Hanses F., Isberner N., Köhler P., Lanznaster J.. COVID-19 in cancer patients: Clinical characteristics and outcome-an analysis of the LEOSS registry. Ann. Hematol. (2021.0) 100 383-393. DOI: 10.1007/s00277-020-04328-4 37. Pilgram L., Eberwein L., Wille K., Koehler F.C., Stecher M., Rieg S., Kielstein J.T., Jakob C.E.M., Rüthrich M., Burst V.. Clinical course and predictive risk factors for fatal outcome of SARS-CoV-2 infection in patients with chronic kidney disease. Infection (2021.0) 49 725-737. DOI: 10.1007/s15010-021-01597-7 38. Huber M.K., Raichle C., Lingor P., Synofzik M., Borgmann S., Erber J., Tometten L., Rimili W., Dolff S., Wille K.. Outcomes of SARS-CoV-2 Infections in Patients with Neurodegenerative Diseases in the LEOSS Cohort. Mov. Disord. (2021.0) 36 791-793. DOI: 10.1002/mds.28554 39. Wehbe Z., Hammoud S.H., Yassine H.M., Fardoun M., El-Yazbi A.F., Eid A.H.. Molecular and Biological Mechanisms Underlying Gender Differences in COVID-19 Severity and Mortality. Front. Immunol. (2021.0) 12 659339. DOI: 10.3389/fimmu.2021.659339 40. Wanschel A.C.B.A., Moretti A.I.S., Ouimet M.. Editorial: COVID-19 Mechanisms on Cardio-Vascular Dysfunction: From Membrane Receptors to Immune Response. Front. Cardiovasc. Med. (2021.0) 8 686495. DOI: 10.3389/fcvm.2021.686495 41. Pesce M., Agostoni P., Bøtker H.E., Brundel B., Davidson S.M., Caterina R., Ferdinandy P., Girao H., Gyöngyösi M., Hulot J.S.. COVID-19-related cardiac complications from clinical evidences to basic mechanisms: Opinion paper of the ESC Working Group on Cellular Biology of the Heart. Cardiovasc. Resp. (2021.0) 117 2148-2160. DOI: 10.1093/cvr/cvab201 42. Sun B., Feng Y., Mo X., Zheng P., Wang Q., Li P., Peng P., Liu X., Chen Z., Huang H.. Kinetics of SARS-CoV-2 specific IgM and IgG responses in COVID-19 patients. Emerg. Microbes. Infect. (2020.0) 9 940-948. DOI: 10.1080/22221751.2020.1762515 43. Hambali N.L., Mohd Noh M., Paramasivam S., Chua T.H., Hayati F., Payus A.O., Tee T.Y., Rosli K.T., Abd Rachman Isnadi M.F., Manin B.O.. A Non-severe Coronavirus Disease 2019 Patient With Persistently High Interleukin-6 Level. Front. Public Health (2020.0) 8 584552. DOI: 10.3389/fpubh.2020.584552 44. Gu X., Sha L., Zhang S., Shen D., Zhao W., Yi Y.. Neutrophils and Lymphocytes Can Help Distinguish Asymptomatic COVID-19 From Moderate COVID-19. Front. Cell. Infect Microbiol. (2021.0) 11 654272. DOI: 10.3389/fcimb.2021.654272 45. Li F., Boon A., Michelson A.P., Foraker R.E., Zhan M., Payne P.. Estrogen Hormone Is an Essential Sex Factor Inhibiting Inflammation and Immune Response in COVID-19. Res. Sq. (2021.0) preprint rs.3.rs-936900. DOI: 10.1038/s41598-022-13585-4 46. Hou H., Zhang B., Huang H., Luo Y., Wu S., Tang G., Liu W., Mao L., Mao L., Wang F.. Using IL-2R/lymphocytes for predicting the clinical progression of patients with COVID-19. Clin. Exp. Immunol. (2020.0) 201 76-84. DOI: 10.1111/cei.13450 47. Xu Z.S., Shu T., Kang L., Wu D., Zhou X., Liao B.W., Sun X.L., Zhou X., Wang Y.Y.. Temporal profiling of plasma cytokines, chemokines and growth factors from mild, severe and fatal COVID-19 patients. Signal Transduct Target Ther. (2020.0) 5 100. DOI: 10.1038/s41392-020-0211-1 48. Talwar D., Kumar S., Acharya S., Raisinghani N., Madaan S., Hulkoti V., Akhilesh A., Khanna S., Shah D., Nimkar S.. Interleukin 6 and Its Correlation with COVID-19 in Terms of Outcomes in an Intensive Care Unit of a Rural Hospital:A Cross-sectional Study. Indian J. Crit. Care. Med. (2022.0) 26 39-42. DOI: 10.5005/jp-journals-10071-24075 49. Ding R., Yang Z., Huang D., Wang Y., Li X., Zhou X., Yan L., Lu W., Zhang Z.. Identification of parameters in routine blood and coagulation tests related to the severity of COVID-19. Int. J. Med. Sci. (2021.0) 18 1207-1215. DOI: 10.7150/ijms.47494 50. Mo P., Xing Y., Xiao Y., Deng L., Zhao Q., Wang H., Xiong Y., Cheng Z., Gao S., Liang K.. Clinical Characteristics of Refractory Coronavirus Disease 2019 in Wuhan, China. Clin. Infect. Dis. (2021.0) 73 e4208-e4213. DOI: 10.1093/cid/ciaa270 51. Huang C., Wang Y., Li X., Ren L., Zhao J., Hu Y., Zhang L., Fan G., Xu J., Gu X.. Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China. Lancet (2020.0) 395 497-506. DOI: 10.1016/S0140-6736(20)30183-5 52. Wu C., Chen X., Cai Y., Xia J., Zhou X., Xu S., Huang H., Zhang L., Zhou X., Du C.. Risk Factors Associated With Acute Respiratory Distress Syndrome and Death in Patients With Coronavirus Disease 2019 Pneumonia in Wuhan, China. JAMA Intern. Med. (2020.0) 180 934-943. DOI: 10.1001/jamainternmed.2020.0994 53. Chen L., Liu H.G., Liu W., Liu J., Liu K., Shang J., Deng Y., Wei S.. [Analysis of clinical features of 29 patients with 2019 novel coronavirus pneumonia]. Zhonghua Jie He He Hu Xi Za Zhi Chin. J. Tuberc. Respir. Dis. (2020.0) 43 E005. DOI: 10.3760/cma.j.issn.1001-0939.2020.0005 54. Gao Y., Li T., Han M., Li X., Wu D., Xu Y., Zhu Y., Liu Y., Wang X., Wang L.. Diagnostic utility of clinical laboratory data determinations for patients with the severe COVID-19. J. Med. Virol. (2020.0) 92 791-796. DOI: 10.1002/jmv.25770 55. Zhang J.J., Dong X., Cao Y.Y., Yuan Y.D., Yang Y.B., Yan Y.Q., Akdis C.A., Gao Y.D.. Clinical characteristics of 140 patients infected with SARS-CoV-2 in Wuhan, China. Allergy (2020.0) 75 1730-1741. DOI: 10.1111/all.14238 56. Li L.Q., Huang T., Wang Y.Q., Wang Z.P., Liang Y., Huang T.B., Zhang H.Y., Sun W., Wang Y.. COVID-19 patients’ clinical characteristics, discharge rate, and fatality rate of meta-analysis. J. Med. Virol. (2020.0) 92 577-583. DOI: 10.1002/jmv.25757 57. Li Y.Y., Wang W.N., Lei Y., Zhang B., Yang J., Hu J.W., Ren Y.L., Lu Q.F.. Comparison of the clinical characteristics between RNA positive and negative patients clinically diagnosed with coronavirus disease 2019. Zhonghua Jie He He Hu Xi Za Zhi Chin. J. Tuberc. Respir. Dis. (2020.0) 43 427-430. DOI: 10.3760/cma.j.cn112147-20200214-00095 58. Xia T., Zhang W., Xu Y., Wang B., Yuan Z., Wu N., Xiang Y., Li C., Shan Y., Xie W.. Early kidney injury predicts disease progression in patients with COVID-19: A cohort study. BMC Infect. Dis. (2021.0) 21. DOI: 10.1186/s12879-021-06576-9 59. Xiang H.X., Fei J., Xiang Y., Xu Z., Zheng L., Li X.Y., Fu L., Zhao H.. Renal dysfunction and prognosis of COVID-19 patients: A hospital-based retrospective cohort study. BMC Infect. Dis. (2021.0) 21. DOI: 10.1186/s12879-021-05861-x 60. Sanyaolu A., Okorie C., Marinkovic A., Patidar R., Younis K., Desai P., Hosein Z., Padda I., Mangat J., Altaf M.. Comorbidity and its Impact on Patients with COVID-19. SN Compr. Clin. Med. (2020.0) 2 1069-1076. DOI: 10.1007/s42399-020-00363-4 61. Franki R.. Comorbidities the Rule in New York’s COVID-19 Deaths. Hospitalist. (2020.0)
--- title: Clinicopathological Characteristics and Prognostic Profiles of Breast Carcinoma with Neuroendocrine Features authors: - Yue Qiu - Yongjing Dai - Li Zhu - Xiaopeng Hao - Liping Zhang - Baoshi Bao - Yuhui Chen - Jiandong Wang journal: Life year: 2023 pmcid: PMC9967167 doi: 10.3390/life13020532 license: CC BY 4.0 --- # Clinicopathological Characteristics and Prognostic Profiles of Breast Carcinoma with Neuroendocrine Features ## Abstract Background: Breast carcinoma with neuroendocrine features includes neuroendocrine neoplasm of the breast and invasive breast cancer with neuroendocrine differentiation. This study aimed to investigate the clinicopathological features and prognosis of this disease according to the fifth edition of the World Health Organization classification of breast tumors. Materials and Methods: A total of 87 patients with breast carcinoma with neuroendocrine features treated in the First Medical Center, Chinese PLA General Hospital from January 2001 to January 2022 were retrospectively enrolled in this study. Results: More than half of the patients were postmenopausal patients, especially those with neuroendocrine neoplasm ($62.96\%$). There were more patients with human epidermal growth factor receptor 2 negative and hormone receptor positive tumors, and most of them were Luminal B type ($71.26\%$). The multivariate analysis showed that diabetes and stage IV disease were related to the progression-free survival of breast carcinoma with neuroendocrine features patients ($$p \leq 0.004$$ and $p \leq 0.001$, respectively). Conclusion: Breast carcinoma with neuroendocrine features tended to be human epidermal growth factor receptor 2 negative and hormone receptor positive tumors, most of them were Luminal B type, and the related factors of progression-free survival were diabetes and stage IV disease. ## 1. Introduction Breast carcinoma with neuroendocrine features consists of a group of diseases with high heterogeneity. It was reported that the incidence rate of breast carcinoma with neuroendocrine features ranged from $0.1\%$ to $20\%$ [1,2,3]. Breast carcinoma with neuroendocrine features included neuroendocrine neoplasm (NEN) of the breast and invasive breast cancer (IBC) with neuroendocrine differentiation. According to the latest World Health Organization (WHO) classification of breast tumors, NEN was divided into neuroendocrine tumor (NET) and neuroendocrine carcinoma (NEC) based on the degree of differentiation [3]. IBC with neuroendocrine differentiation was classed into breast carcinoma of no special type and breast carcinoma of special types, such as solid papillary carcinoma and the hypercellular subtype of mucinous carcinoma. Since the third edition of the WHO classification of breast tumors was published, the definition and classification of this disease had changed greatly in different editions. As a result, there have been controversies surrounding the definition and classification of breast carcinoma with neuroendocrine features. The diagnostic criteria for subjects included in existing studies were not identical. As well, the research results were not completely consistent or were even contradictory [4,5]. Further, due to the rarity of the disease, few studies had been conducted, and those which have were mainly case reports [6,7,8]. At present, the treatment strategy of IBC of no special type is used directly in breast carcinoma with neuroendocrine features. The general practice guidelines for breast carcinoma with neuroendocrine features are still not formed. The TNM stage of breast carcinoma with neuroendocrine features was defined by the eighth version of the America joint committee on cancer staging systems [9]. According to the guidelines of the Chinese Society of Clinical Oncology published in 2020 [10], the minimum positive threshold of estrogen receptor (ER), progesterone receptor (PR), and Ki-67 are $1\%$, $1\%$, and $14\%$, respectively, and human epidermal growth factor receptor 2 (Her-2) (3+) or ISH positivity meant Her-2 positivity. Breast carcinoma with neuroendocrine features was divided into Luminal A (ER/PR positive, Her-2 negative with low Ki-67 index) disease, Luminal B (ER/PR positive, Her-2 negative with high Ki-67 index, or ER/PR positive, Her-2 positive) disease, Her-2 positive (ER and PR negative, Her-2 positive) disease, and Triple-negative (ER, PR, and Her-2 negative) disease according to molecular subtyping. To investigate the clinicopathological features and prognosis of this disease under the fifth edition of the WHO classification of breast tumors, we designed this study. ## 2.1. Study Groups The data of 87 patients with breast carcinoma with neuroendocrine features treated in the First Medical Center, Chinese PLA General Hospital from January 2001 to January 2022 were retrospectively collected. Patients with breast carcinoma derived from other organs were excluded. Pregnant patients and patients who were breastfeeding were excluded as well. There were no patients without definite pathological diagnosis or without complete medical records. All procedures performed in this study involving human participants were in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by the ethics committee of the Chinese PLA General Hospital (NO.: S2022-746). Individual consent for this retrospective analysis was waived. ## 2.2. Study Variables General information on the patients was collected, such as age at diagnosis, gender, laterality, smoking history, drinking history, body mass index (BMI), family history, menopause status, hypertension, diabetes, hyperlipidemia, T stage, Her-2 status, Ki-67, ER, PR, molecular typing, vessel carcinoma embolus, N stage, skin or chest wall invasion, distant metastasis, and stage. Unique clinical features of the patients were analyzed, including clinical symptoms and history of thyroid diseases. Pathological characteristics of the patients were also described, such as detailed classification, expression of neuroendocrine markers, and ductal carcinoma in situ composition. The treatment strategy of the patients, such as neoadjuvant chemotherapy, surgery, and adjuvant therapy, was discussed. All patients enrolled in this study were followed up. The 5-year overall survival (OS), 5-year progression-free survival (PFS), and 5-year disease-specific survival (DSS) of patients in the study group were described. Finally, the factors related to 5-year PFS were analyzed. ## 2.3. Statistical Analysis All statistical analyses of this study were performed using Stata Statistical Software version 15.1 (StataCorp LLC, College Station, TX, USA). The measurement data were described by median (inter-quartile range, IQR). Frequency was used to show the counting data. Comparison of counting data between two groups was conducted by Pearson chi-square test. Kruskal–Wallis H test was used to examine multiple comparisons of ranked counting data between groups. Kaplan–Meier method was used in survival analysis, and Log-rank test was used to compare different survival curves. Univariate and multivariate analysis were performed using Cox model. A two-tailed $p \leq 0.05$ was considered statistically significant, and all confidence intervals (CI) were expressed at $95\%$ confidence level. ## 3.1. General Clinicopathological Characteristics The median age at diagnosis of the patients in the study group was 53 (42–64) years old. More than half of the patients were postmenopausal patients, especially those with neuroendocrine neoplasm ($62.96\%$). The proportion of patients with Her-2 negative and hormone receptor (HR) positive tumors was high, and most of them were Luminal B type ($71.26\%$ vs. $28.74\%$). Around $49.43\%$ of the patients had stage II disease. There was no difference between the NEN group and IBC with neuroendocrine differentiation group except in diabetes when analyzing general characteristics ($$p \leq 0.039$$) (Table 1). ## 3.2. Special Clinicopathological Features of Breast Carcinoma with Neuroendocrine Features Patients A total of $22.99\%$ of patients had clinical symptoms such as pain, nipple discharge, or both. About $21.84\%$ of patients were complicated, with thyroid diseases such as thyroid nodule, diffuse thyroid disease, and thyroid cancer. A total of 34 patients with lymph node metastasis all had axillary lymph node metastasis, two cases also had supraclavicular lymph node metastasis, and one case had subclavian lymph node metastasis at the same time (Table 2). A total of 87 cases with breast carcinoma with neuroendocrine features were classified into 29 cases of NEN and 58 cases of IBC with neuroendocrine differentiation (Table 3). After H and E staining and immunohistochemical staining, some tumor cells showed typical positive synaptophysin (Syn) and Chromogranin (CgA) staining (Figure 1). There were also some cases that had CD56+ tumors and neuron-specific enolase (NSE)+ tumors (Table 3). ## 3.3. Treatment and Follow-Up of Breast Carcinoma with Neuroendocrine Features Patients Only 7 out of 87 patients received neoadjuvant chemotherapy. All patients received surgical treatment of the breast and/or axillary lymph node. There were 75 cases that underwent mastectomy, while the rest of the patients underwent breast-conserving surgery or nipple-areola complex-sparing mastectomy. As for adjuvant therapy, $81.61\%$ of patients received chemotherapy, $31.03\%$ of patients received radiotherapy, $73.56\%$ of patients received endocrine therapy, and only $9.20\%$ of patients received targeted therapy after surgery (Table 4). A total of 87 patients were followed up, and 2 patients dropped out. The median follow-up time was 57 (25–74) months. During the follow-up period, local recurrence involving the breast, axilla, or chest wall occurred in five cases. There were 12 cases recorded of distant metastasis including bone, lung, liver, brain, retroperitoneal lymph node, and contralateral axillary lymph node. Eight cases died, six of them died of breast cancer and two died of other causes (Table 5). ## 3.4. Kaplan-Meier Survival Analysis of Breast Carcinoma with Neuroendocrine Features Patients The five-year PFS, five-year DSS, and five-year OS of all of the patients were $81.19\%$ ($95\%$CI: 0.6964–0.8869), $91.53\%$ ($95\%$CI: 0.8022–0.9651), and $90.25\%$ ($95\%$CI: 0.7901–0.9564), respectively. No significant differences were found in the five-year PFS, five-year DSS, and five-year OS between NEN and IBC with neuroendocrine differentiation ($p \leq 0.05$) (Figure 2, Figure 3 and Figure 4). ## 3.5. Univariate Analysis of PFS in Breast Carcinoma with Neuroendocrine Features Patients The results of the univariate analysis of PFS showed that smoking history, diabetes, Her-2 positive disease, N3 stage disease, distant metastasis, and targeted therapy were related to the progression-free survival of breast carcinoma in neuroendocrine features patients ($$p \leq 0.038$$, $$p \leq 0.008$$, $$p \leq 0.003$$, $p \leq 0.001$, $p \leq 0.001$; $$p \leq 0.004$$, respectively). The results of Her-2 status and family history slightly missed the margin of significance ($$p \leq 0.097$$ and $$p \leq 0.092$$, respectively) (Table 6). ## 3.6. Multivariate Analysis of PFS in Breast Carcinoma with Neuroendocrine Features Patients The results of the multivariate analysis of PFS showed that diabetes and stage IV disease were related to the progression-free survival of breast carcinoma in neuroendocrine features patients ($$p \leq 0.004$$ and $p \leq 0.001$, respectively). The result of targeted therapy showed a barely detectable statistical significance ($$p \leq 0.051$$) (Table 7). ## 4. Discussion Breast carcinoma with neuroendocrine features is a group of heterogeneous tumors. Its definitions and diagnostic criteria have varied with the revisions of the WHO classification of breast tumors. As a result, the results of some studies on the clinicopathological characteristics of this disease have been controversial for a long time. Breast carcinoma with neuroendocrine features is a group of tumors that exhibit morphological features similar to those of neuroendocrine tumors of the gastrointestinal tract and lung [11]. Before 2003, there were no criteria for the definition and diagnosis of this disease. With the further study of breast carcinoma, the consensus on breast carcinoma with neuroendocrine features has gradually been formed. The third version of the WHO criteria of breast tumors defined it as >$50\%$ tumor cells with neuroendocrine differentiation confirmed by immunohistochemical staining [12]. From then on, it was recognized as single breast carcinoma entities named “neuroendocrine breast carcinomas”. In 2012, the WHO classification used the category of “carcinoma with neuroendocrine features” and described this disease as tumors expressing neuroendocrine markers to any extent [13]. It included well-differentiated NET, poorly differentiated NEC, and carcinoma with neuroendocrine differentiation. In this version, small cell neuroendocrine carcinoma (SCNEC) was brought into the NEC group. The current WHO classification adopted the term “NEN”, including well-differentiated (NET) and poorly-differentiated (NEC) tumors with predominant neuroendocrine differentiation [3]. The main distinction between the latest classification and the past version is that carcinoma with neuroendocrine differentiation without distinct or uniform enough neuroendocrine histological features and neuroendocrine marker expression is no longer classified as NET or NEC. In this version, large cell neuroendocrine carcinoma (LCNEC) was classified into NEC as well. All the criteria mentioned above are used for the classification of primary breast carcinoma with neuroendocrine features; however, before a diagnosis of primary NEN is made, the possibility of metastasis from other organs should be carefully ruled out. Immunohistochemistry staining is conducive to distinction between NEN derived from other organs from invasive mammary carcinoma with neuroendocrine features [14]. This study only discussed primary breast carcinoma with neuroendocrine features. The results of the previous studies in different periods were different or even contradictory. Previous studies found that most patients were 50 years old or older [15,16], and that the clinical symptoms of this disease were mainly bloody nipple discharge [16], which was consistent with the results of this study. NEN can be divided into functional and non-functional tumors according to whether the tumor has hormone activity, and most NENs are non-functional. Functional NEN produces excessive hormones, leading to clinical symptoms such as diarrhea and facial flushing. Non-functional NENs do not produce enough hormones to cause these symptoms [17]. Paraneoplastic endocrine syndrome may occur in breast cancer with or without neuroendocrine differentiation [18]. There were few reports on paraneoplastic endocrine syndrome related to breast neuroendocrine tumors. One case with hyperprolactinemia was reported in the previous literature [19]. Studies have illustrated that patients often had ER/PR positive and Her-2 negative tumors [5,20], supporting this study. Another study found that neuroendocrine differentiation was more common in Luminal B breast cancer [21], which is consistent with the results of this study. However, among them, Her-2 positive patients were rare, only occasionally seen in case reports [22]. Research showed that breast carcinoma with neuroendocrine features was highly aggressive, with a high rate of local recurrence and distant metastasis, and a poor prognosis [5]. A study has also shown that its general clinical characteristics are not different from other breast cancers, and its biological behavior was not aggressive; on the contrary, it tended to be an independent good-prognosis subgroup [20]. Morphologically, the typical features of the lung/gastrointestinal tract NET, such as ribbons, cords, and rosettes, are not prominent in the breast NET; histological and immunohistochemical features of breast NEC are sometimes difficult to distinguish from lung NEC features [23]. CgA, Syn, NSE, and CD56 were neuroendocrine differentiation markers for breast carcinoma with neuroendocrine features [4]. It has been reported that only $23\%$ of patients were detected with Syn+ and CgA+ at the same time [15]. The expression level of neuroendocrine markers in tumor tissues of patients in this study was also not high. There were significant differences in cytological characteristics between focal and diffuse neuroendocrine differentiated breast carcinomas [24]. However, when breast cancer of no special type with focal neuroendocrine differentiation was regarded as a separate entity, focal neuroendocrine differentiation had no obvious significance for its prognosis [15,25]. In this study, there was no difference between NEN and IBC with neuroendocrine differentiation in five-year OS, five-year DSS, and five-year PFS. SCNEC was classified into NEC in 2012 and LCNEC was brought into NEC in 2019. A study indicated that approximately half of these patients had triple-negative breast cancer, with a $61.6\%$ five-year DSS rate and $47.7\%$ five-year OS rate [26]. Chemotherapy, surgery, and stage were predictive factors of prognosis. In this study, there was only one case of pure SCNEC and pure LCNEC, respectively, accounting for a relatively low proportion. The special type of breast cancer with neuroendocrine differentiation was rare. According to one study, half of the invasive solid papillary carcinomas were accompanied by neuroendocrine differentiation [7]. In this study, there were four cases of a special type of breast cancer with neuroendocrine differentiation, including two cases of solid papillary carcinoma, one case of invasive papillary carcinoma, and one case of type B mucinous carcinoma. A study reported that the five-year OS and disease-free survival rates of HR positive/Her-2 negative breast cancer were $93.0\%$ and $92.6\%$, respectively [27]. The PFS of breast carcinoma with neuroendocrine features was lower than that of the same molecular typing of breast cancer of no special type [28], which was consistent with the results of this study. The overall five-year PFS of patients in this study was $82.37\%$ ($95\%$CI: 0.7084–0.8966), five-year DSS was $91.53\%$ ($95\%$CI: 0.8022–0.9651), and five-year OS was $90.25\%$ ($95\%$CI: 0.7901–0.9564). The study of neoadjuvant therapy for this disease has been limited. In this study, there were seven patients receiving neoadjuvant chemotherapy, and none of them reached a pathological complete response. A study found that endocrine therapy or radiotherapy might improve the prognosis [5]. It was suggested that HR positive NEN patients receive endocrine therapy, especially those with SCNEC with recurrence and metastasis [29]. Endocrine therapy was also found to be effective for liver metastasis of breast cancer with neuroendocrine differentiation [6]. However, surgery was still the main treatment method, and the effect of chemotherapy on prognosis was still uncertain [16]. This study also analyzed the related factors of PFS. A study found that the prognostic factors of NEN of the breast were similar to those of gastrointestinal tract tumors [30], among which lymph node metastasis was an adverse factor of OS [5]. A previous study found that histological grade, pathological stage, ER status, and HER2 status were independent prognostic indicators of OS and disease-free survival [31]. The results of this study showed that diabetes and stage IV disease were related to the PFS of breast carcinoma with neuroendocrine features patients. The influence of diabetes on the PFS of this disease may be related to higher BMI, and there is a lack of relevant research at present. ## Limitations This study was a retrospective study, and the sample size was relatively insufficient. We failed to compare this disease with other types of breast cancer because of lacking a control group. In addition, this study did not describe its imaging characteristics. Because of the rarity of this disease, there was a gap between the length of follow-up time of the patients. ## 5. Conclusions Breast carcinoma with neuroendocrine features is relatively rare compared with other types of breast cancer. In this study, it was illustrated that this disease tended to be HR+/Her-2- tumor. In addition, diabetes and stage IV were related to the PFS of patients. These results may provide evidence for the treatment and prognosis prediction of breast carcinoma with neuroendocrine features. Further studies such as large-sample randomized clinical trials are needed to validate the theoretical value and practical significance of these findings and improve understanding of this disease. ## References 1. Zhang Y., Geng C., Xu B.. Controversies and research progress on the diagnosis and treatment of breast neuroendocrine neoplasm. Chin. J. Clin. Oncol. (2022) 49 968-972. DOI: 10.12354/j.issn.1000-8179.2022.20220593 2. Wang J., Wei B., Albarracin C.T., Hu J., Abraham S.C., Wu Y.. Invasive neuroendocrine carcinoma of the breast: A population-based study from the surveillance, epidemiology and end results (SEER) database. BMC Cancer (2014) 14. DOI: 10.1186/1471-2407-14-147 3. Breast Tumours. WHO Classification of Tumors (2019) 4. Rovera F., Lavazza M., La Rosa S., Fachinetti A., Chiappa C., Marelli M., Sessa F., Giardina G., Gueli R., Dionigi G.. Neuroendocrine breast cancer: Retrospective analysis of 96 patients and review of literature. Int. J. Surg. (2013) 11 S79-S83. DOI: 10.1016/S1743-9191(13)60023-0 5. Wei B., Ding T., Xing Y., Wei W., Tian Z., Tang F., Abraham S.C., Nayeemuddin K.M., Hunt K.K., Wu Y.. Invasive neuroendocrine carcinoma of the breast: A distinctive subtype of aggressive mammary carcinoma. Cancer (2010) 116 4463-4473. DOI: 10.1002/cncr.25352 6. Rischke H.C., Staib-Sebler E., Mose S., Adams S.W., Herrmann G., Böttcher H.D., Lorenz M.. Metastasiertes Mamma-Karzinom neuroendokriner Differenzierung--kombinierte Therapie mit Tamoxifen und dem Somatostatin-Analogon Octreotid [Metastatic breast carcinoma with neuroendocrine differentiation--its combined therapy with tamoxifen and the somatostatin analog octreotide]. DMW-Dtsch. Med. Wochenschr. (1999) 124 182-186. DOI: 10.1055/s-2007-1024271 7. Lin X., Matsumoto Y., Nakakimura T., Ono K., Umeoka S., Torii M., Yoshibayashi H., Toi M.. Invasive solid papillary carcinoma with neuroendocrine differentiation of the breast: A case report and literature review. Surg. Case Rep. (2020) 6 143. DOI: 10.1186/s40792-020-00905-x 8. Salemis N.S.. Primary neuroendocrine carcinoma of the breast: A rare presentation and review of the literature. Intractable Rare Dis. Res. (2020) 9 233-246. DOI: 10.5582/irdr.2020.03046 9. Giuliano A.E., Connolly J.L., Edge S.B., Mittendorf E.A., Rugo H.S., Solin L.J., Weaver D.L., Winchester D.J., Hortobagyi G.N.. Breast Cancer-Major changes in the American Joint Committee on Cancer eighth edition cancer staging manual. CA Cancer J. Clin. (2017) 67 290-303. DOI: 10.3322/caac.21393 10. Jiang Z., Song E., Wang X., Wang H., Wang X., Wu J., Yin Y., Zhang Q., Chen J., Che W.. Guidelines of Chinese Society of Clinical Oncology (CSCO) on Diagnosis and Treatment of Breast Cancer (2020 version). Transl. Breast Cancer Res. (2020) 1. DOI: 10.21037/tbcr-2020-2 11. Rindi G., Klimstra D.S., Abedi-Ardekani B., Asa S.L., Bosman F.T., Brambilla E., Busam K.J., De Krijger R.R., Dietel M., El-Naggar A.K.. A common classification framework for neuroendocrine neoplasms: An International Agency for Research on Cancer (IARC) and World Health Organization (WHO) expert consensus proposal. Mod. Pathol. (2018) 31 1770-1786. DOI: 10.1038/s41379-018-0110-y 12. Tavassoli F.A., Devilee P.. World Health Organization Classification of Tumours: Pathology and Genetics of Tumours of the Breast and Female Genital Organs (2003) 13. Lakhani S.R., Ellis I.O., Schnitt S.J., Tan P.H., van de Vijver M.J.. WHO Classification of Tumours of the Breast (2012) 14. Mohanty S.K., Kim S.A., DeLair D.F., Bose S., Laury A.R., Chopra S., Mertens R.B., Dhall D.. Comparison of metastatic neuroendocrine neoplasms to the breast and primary invasive mammary carcinomas with neuroendocrine differentiation. Mod. Pathol. (2016) 29 788-798. DOI: 10.1038/modpathol.2016.69 15. Van Krimpen C., Elferink A., Broodman C., Hop W., Pronk A., Menke M.. The prognostic influence of neuroendocrine differentiation in breast cancer: Results of a long-term follow-up study. Breast (2004) 13 329-333. DOI: 10.1016/j.breast.2003.11.008 16. Li Y., Du F., Zhu W., Xu B.. Neuroendocrine carcinoma of the breast: A review of 126 cases in China. Chin. J. Cancer (2017) 36 1-4. DOI: 10.1186/s40880-017-0211-x 17. Ozaki Y., Miura S., Oki R., Morikawa T., Uchino K.. Neuroendocrine Neoplasms of the Breast: The Latest WHO Classification and Review of the Literature. Cancers (2021) 14. DOI: 10.3390/cancers14010196 18. Dimitriadis G.K., Angelousi A., Weickert M.O., Randeva H.S., Kaltsas G., Grossman A.. Paraneoplastic endocrine syndromes. Endocrine-Related Cancer (2017) 24 R173-R190. DOI: 10.1530/ERC-17-0036 19. Zhang Q., He L., Lv W., Wang N.. Neuroendocrine carcinoma of the breast with hyperprolactinemia: Report of two cases and a minireview. Int. J. Clin. Exp. Pathol. (2020) 13 1457-1462. PMID: 32661484 20. Menendez J.A., Lopez-Bonet E., Alonso-Ruano M., Barraza G., Vázquez-Martín A., Bernadó L.. Solid neuroendocrine breast carcinomas: Incidence, clinico-pathological features and immunohistochemical profiling. Oncol. Rep. (1994) 20 1369-1374. DOI: 10.3892/or_00000154 21. Wachter D.L., Hartmann A., Beckmann M.W., Fasching P.A., Hein A., Bayer C.M., Agaimy A.. Expression of Neuroendocrine Markers in Different Molecular Subtypes of Breast Carcinoma. BioMed Res. Int. (2014) 2014 408459. DOI: 10.1155/2014/408459 22. Marijanović I., Kraljević M., Buhovac T., Križanac D.K.. Rare Human Epidermal Growth Factor Receptor 2 (HER-2)-Positive Neuroendocrine Carcinoma of the Breast: A Case Report with 9-Year Follow-up. Am. J. Case Rep. (2020) 21 e925895. DOI: 10.12659/AJCR.925895 23. Tsang J.Y., Tse G.M.. Breast cancer with neuroendocrine differentiation: An update based on the latest WHO classification. Mod. Pathol. (2021) 34 1062-1073. DOI: 10.1038/s41379-021-00736-7 24. Talu C.K., Guzelbey B., Hacihasanoglu E., Cakir Y., Nazli M.A.. The effect of the extent of neuroendocrine differentiation on cytopathological findings in primary neuroendocrine neoplasms of the breast. J. Cytol. (2021) 38 216. DOI: 10.4103/JOC.JOC_56_21 25. Bogina G., Munari E., Brunelli M., Bortesi L., Marconi M., Sommaggio M., Lunardi G., Gori S., Massocco A., Pegoraro M.C.. Neuroendocrine differentiation in breast carcinoma: Clinicopathological features and outcome. Histopathology (2015) 68 422-432. DOI: 10.1111/his.12766 26. Zhu J., Wu G., Zhao Y., Yang B., Chen Q., Jiang J., Meng Y., Ji S., Gu K.. Epidemiology, Treatment and Prognosis Analysis of Small Cell Breast Carcinoma: A Population-Based Study. Front. Endocrinol. (2022) 13 802339. DOI: 10.3389/fendo.2022.802339 27. Wang Y., Zhang S., Zhang H., Liang L., Xu L., Cheng Y.J., Duan X.N., Liu Y.H., Li T.. Clinicopathological features and prognosis of hormone receptor-positive/hum an epidermal growth factor receptor 2-negative breast cancer. Beijing Da Xue Xue Bao Yi Xue Ban (2022) 54 853-862. PMID: 36241228 28. Lavigne M., Menet E., Tille J.-C., Lae M., Fuhrmann L., Bonneau C., Deniziaut G., Melaabi S., Ng C.C.K., Marchiò C.. Comprehensive clinical and molecular analyses of neuroendocrine carcinomas of the breast. Mod. Pathol. (2018) 31 68-82. DOI: 10.1038/modpathol.2017.107 29. Alkaied H., Harris K., Brenner A., Awasum M., Varma S.. Does Hormonal Therapy Have a Therapeutic Role in Metastatic Primary Small Cell Neuroendocrine Breast Carcinoma? Case Report and Literature Review. Clin. Breast Cancer (2012) 12 226-230. DOI: 10.1016/j.clbc.2012.01.008 30. Roininen N., Takala S., Haapasaari K.-M., Jukkola-Vuorinen A., Mattson J., Heikkilä P., Karihtala P.. Neuroendocrine Breast Carcinomas Share Prognostic Factors with Gastroenteropancreatic Neuroendocrine Tumors: A Putative Prognostic Role of Menin, p27, and SSTR-2A. Oncology (2018) 96 147-155. DOI: 10.1159/000493348 31. Kwon S.Y., Bae Y.K., Gu M.J., Choi J.E., Kang S.H., Lee S.J., Kim A., Jung H.R., Kang S.H., Oh H.K.. Neuroendocrine differentiation correlates with hormone receptor expression and decreased survival in patients with invasive breast carcinoma. Histopathology (2013) 64 647-659. DOI: 10.1111/his.12306
--- title: Bioinformatics and Next-Generation Data Analysis for Identification of Genes and Molecular Pathways Involved in Subjects with Diabetes and Obesity authors: - Prashanth Ganekal - Basavaraj Vastrad - Satish Kavatagimath - Chanabasayya Vastrad - Shivakumar Kotrashetti journal: Medicina year: 2023 pmcid: PMC9967176 doi: 10.3390/medicina59020309 license: CC BY 4.0 --- # Bioinformatics and Next-Generation Data Analysis for Identification of Genes and Molecular Pathways Involved in Subjects with Diabetes and Obesity ## Abstract Background and Objectives: A subject with diabetes and obesity is a class of the metabolic disorder. The current investigation aimed to elucidate the potential biomarker and prognostic targets in subjects with diabetes and obesity. Materials and Methods: The next-generation sequencing (NGS) data of GSE132831 was downloaded from Gene Expression Omnibus (GEO) database. Functional enrichment analysis of DEGs was conducted with ToppGene. The protein–protein interactions network, module analysis, target gene–miRNA regulatory network and target gene–TF regulatory network were constructed and analyzed. Furthermore, hub genes were validated by receiver operating characteristic (ROC) analysis. A total of 872 DEGs, including 439 up-regulated genes and 433 down-regulated genes were observed. Results: Second, functional enrichment analysis showed that these DEGs are mainly involved in the axon guidance, neutrophil degranulation, plasma membrane bounded cell projection organization and cell activation. The top ten hub genes (MYH9, FLNA, DCTN1, CLTC, ERBB2, TCF4, VIM, LRRK2, IFI16 and CAV1) could be utilized as potential diagnostic indicators for subjects with diabetes and obesity. The hub genes were validated in subjects with diabetes and obesity. Conclusion: This investigation found effective and reliable molecular biomarkers for diagnosis and prognosis by integrated bioinformatics analysis, suggesting new and key therapeutic targets for subjects with diabetes and obesity. ## 1. Introduction Diabetes mellitus and obesity are major metabolic or endocrine disorders and are dramatically increasing throughout the globe [1]. The prevalence of obesity and type 2 diabetes mellitus is considerably higher [2]. Diabetes mellitus and obesity are linked with progression of cardiovascular diseases [3], hypertension [4], and neurological and neuropsychiatric disorders [5] and asthma [6]. Till today, there is no cure for diabetes mellitus and obesity, and treatment and mediation tailored to clinical features are endorsed. Genetic and environmental factors are two initial contributors to these disorders [7]. Exploration of the molecular mechanisms of diabetes mellitus and obesity will develop the considerate of its pathogenesis and has key implications for designing new therapy. Molecular mechanisms of subject with diabetes and obesity have been increasingly studied. Previous investigations showed that genes and signaling pathways are associated with diabetes mellitus and obesity. Key genes such as ENPP1 [8] and FTO [9] were responsible for development of diabetes mellitus and obesity. Recent investigations showed that PI3K/AKT pathway [10] and TLR pathway [11] as a potential target for diabetes mellitus and obesity. However, certain key genes and pathways associated with diabetes mellitus and obesity have not been completely investigated. Further studies are necessary to elucidate these essential genes and pathways to provide novel therapeutic targets for the treatment of diabetes and obesity. In recent years, the analysis of biological information, known as bioinformatics, has attracted a great deal of attention and sustained breakthroughs in the search for biomarkers for various diseases [12,13,14]. With the gradual advancement of next-generation sequencing (NGS) technology, bioinformatics has become increasingly essential in molecular pathogenesis, performing a major role in elucidating diseases mechanisms and finding novel targets for diseases treatment and patient prognosis [15]. With the wide function of NGS, a huge amount of data hasbeen generated, and most of the data have been deposited and stored in public databases. NGS data analyses have been carried out on diabetes and obesity in recent years [16], and hundreds of differentially expressed genes (DEGs) have been obtained. Bioinformatics methods combining with NGS techniques will be innovative. Therefore, in this investigation, we downloaded the next-generation sequencing (NGS)data GSE132831, provided by Osinski et al. [ 17], from Gene Expression Omnibus (GEO, http://www.ncbi.nlm.nih.gov/geo/, accessed on 11 June 2020) [18] database to identify the differentially expressed genes (DEGs) between diabetes mellitus and obesity samples and normal control samples. With the identified DEGs, we performed Gene Ontology (GO) and pathway enrichment analyses to investigate the functions and pathways enriched by the DEGs. Additionally, we constructed a protein–protein interaction (PPI) network and modules screened out some important gene nodes to perform clustering analysis. Furthermore, we constructed a target gene–miRNA regulatory network and target gene–TF regulatory network based on these key genes to investigate the potential relationships between genes and subject with diabetes and obesity. Finally, hub genes were validated by using receiver operating characteristic (ROC) curve analysis. The research design of this study was shown in Figure 1. These results might provide novel ideas for future investigation and treatment of diabetes mellitus and obesity by exploring prognostic markers and therapeutic targets in diabetes mellitus and obesity. ## 2.1. RNA Sequencing Data The NGS data GSE132831 was downloaded from the GEO database, which was based on the platform of GPL1857 Illumina NextSeq 500 (Homo sapiens). This dataset, including samples of 104 diabetic obese and samples of 120 normal control, was deposited by Osinski et al. [ 17]. ## 2.2. Identification of DEGs The limma R/Bioconductor software package was used to perform the identification of DEGs between samples of diabetic obese and normal control in R software [19]. The cutoff criteria were \|logFC\| > 1.112 for up-regulated genes, \|logFC\| <−0.64 for down-regulated genes, and a p-value < 0.05. The significance of p value measures how likely it is that any observed difference between two groups (diabetes mellitus and obesity samples and normal control samples). The significance of log FC looks only at genes which vary wildly amongst other genes. ## 2.3. GO and Pathway Enrichment Analyses of DEGs The ToppGene (ToppFun) (https://toppgene.cchmc.org/enrichment.jsp, accessed on 11 June 2020) [20] bioinformatics resources was utilized to distinguish and enrich the biological attributes, such as biological processes (BP), cellular components (CC), molecular functions (MF) and pathways, of identified DEGs (Up- and down-regulated genes separately). Moreover, GO (http://geneontology.org/, accessed on 11 June 2020) [21] and REACTOME (https://reactome.org/, accessed on 11 June 2020) [22] pathway enrichment analyses were used to identify the significant GO terms and pathways. $p \leq 0.05$ was set as the cutoff criterion for significant enrichment. ## 2.4. Protein–Protein Interaction (PPI) Network and Module Analysis The IID interactome (http://iid.ophid.utoronto.ca/, accessed on 11 June 2020) [23] is an online database containing known and predicted PPI networks. In this investigation, a PPI network of identified DEGs in dataset was identified using the IID interactome database (combined score >0.4) and subsequently visualized using Cytoscape (http://www.cytoscape.org/, accessed on 11 June 2020) software (version 3.8.2) [24]. The regulatory relationship between genes were analyzed through topological property of computing network including the node degree [25], betweenness centrality [26], stress centrality [27] and closeness centrality [28] by using the Network Analyzer app within Cytoscape. The PEWCC1 (http://apps.cytoscape.org/apps/PEWCC1, accessed on 11 June 2020) [29] program within Cytoscape was used to detect modules of the PPI network. The GO and pathway enrichment analysis of the identified modules was then performed using the ToppGene database. ## 2.5. Target Gene–miRNARegulatory Network miRNet database (https://www.mirnet.ca/, accessed on 11 June 2020) [30] is a bioinformatics platform for predicting target gene–miRNA pairs. In the present study, the target genes were predicted using 14 miRNA databases: TarBase, miRTarBase, miRecords, miRanda (S mansoni only), miR2Disease, HMDD, PhenomiR, SM2miR, PharmacomiR, EpimiR, starBase, TransmiR, ADmiRE, and TAM 2.0. In this study, miRNAs were considered the targeted miRNAs of hub genes based on these miRNA databases. The target gene–miRNA regulatory network was depicted and visualized using Cytoscape software. ## 2.6. Target Gene–TF Regulatory Network NetworkAnalyst database (https://www.networkanalyst.ca/, accessed on 11 June 2020) [31] is a bioinformatics platform for predicting target gene–TF pairs. In the present study, the target genes were predicted using ChEA TF database. In this study, TFs were considered the targeted TFs of hub genes based on this TF database. The target gene–TF regulatory network was depicted and visualized using Cytoscape software. ## 2.7. Receiver Operating Characteristic (ROC) Analysis A ROC analysis is a technique for visualizing, construct and determining classifiers based on their achievement. A diagnostic test was firstly performed in order to measure the diagnostic value of candidate biomarkers in subject with diabetes and obesity. Sensitivity and specificity of each biomarker in this diagnostic test were determined. ROC curves were retrieved by plotting the sensitivity, against the specificity using the pROC in R software [32]. Area under the ROC curve (AUC) was determined to predict the efficiency of this diagnostic test. A test with AUC bigger than 0.9 is assigned great efficiency, 0.7–0.9, modest efficiency and 0.5–0.7, small efficiency. ## 3.1. Identification of DEGs The DEGs were screened by “limma” package (p-value < 0.05, and \|logFC\| > 1.112 for up-regulated genes and \|logFC\| <−0.64 for down-regulated genes). The GSE132831 dataset contained 872 DEGs, including 439 up-regulated genes and 433 down-regulated genes. DEGs are listed in Table S1. The volcano plot is presented in Figure 2. The heat map DEGs is shown in Figure 3. ## 3.2. GO and Pathway Enrichment Analyses of DEGs To gain in-depth and comprehensive biological characteristics of these DEGs, GO functional annotation and REACTOME pathway enrichment analysis were performed through online analytical tool ToppGene. The BP was mainly enriched in plasma membrane bounded cell projection organization, neurogenesis, cell activation and secretion (Table S2). The CC was mainly enriched in neuron projection, golgi apparatus, secretory granule and secretory vesicle (Table S2). The MF was significantly enriched in drug binding, ribonucleotide binding, signaling receptor binding and molecular transducer activity (Table S2). Result of REACTOME enrichment analysis showed that top pathways were axon guidance, extracellular matrix organization, neutrophil degranulation and innate immune system (Table S3). ## 3.3. Protein–Protein Interaction (PPI) Network and Module Analysis To find the hub genes in the DEGs, Network Analyzer, a plug-in Cytoscape was performed. All the genes and edges were determined. IID interactome mapped 872 DEGs into a PPI network containing 3894 nodes and 7142 edges (Figure 4). Hub genes with the high node degree, betweenness centrality, stress centrality and closeness centrality are listed in Table S4. MYH9 (Degree 231; Betweenness 0.083106; Stress 11909200; Closeness 0.348923), FLNA (Degree 196; Betweenness 0.07999; Stress 10927852; Closeness 0.35285), DCTN1 (Degree 168; Betweenness 0.080808; Stress 7748054; Closeness 0.330668), CLTC (Degree 158; Betweenness 0.071579; Stress 7687192; Closeness 0.351161), ERBB2 (Degree 158; Betweenness 0.069347; Stress 7857692; Closeness 0.327216), TCF4 (Degree 186; Betweenness 0.080443; Stress 7982798; Closeness 0.320102), VIM (Degree 146; Betweenness 0.062238; Stress 9673084; Closeness 0.327078), LRRK2 (Degree 114; Betweenness 0.046449; Stress 5881270; Closeness 0.333305), IFI16 (Degree 91; Betweenness 0.035934; Stress 2890936; Closeness 0.294671) and CAV1 (Degree 74; Betweenness 0.034501; Stress 3642486; Closeness 0.323593). Then, PEWCC1 was used to find clusters in the network. Module1 contained 16 nodes and 39 edges (Figure 5A). Module1 was associated with including axon guidance, signaling by NGF, plasma membrane bounded cell projection organization and neurogenesis. Module2contained 13 nodes and 24 edges (Figure 5B). Module 2 was associated with an innate immune system. ## 3.4. Target Gene–miRNA Regulatory Network The target gene–miRNA regulatory network included 2520 nodes (miRNAs: 2224; gene: 296) and 15485 edges (Figure 6). The nodes with degrees were listed in Table S5. We discovered that MYH9 was targeted by 116 miRNAs (ex; hsa-mir-4329); ERBB2 was targeted by 73miRNAs (ex; hsa-mir-4315); MYO18A was targeted by 71miRNAs (ex; hsa-mir-1299); SEC16A was targeted by 66 miRNAs (ex; hsa-mir-4779); PLCG1 was targeted by 56 miRNAs (ex; hsa-mir-3685); CCNB1 was targeted by 84miRNAs (ex; hsa-mir-6134); CAV1 was targeted by 58miRNAs (ex; hsa-mir-4459); VIM was targeted by 30miRNAs (ex; hhsa-mir-6124); HAP1 was targeted by 22miRNAs (ex; hsa-mir-9500); MAD2L1 was targeted by 17miRNAs (ex; hsa-mir-1297). ## 3.5. Target Gene–TF Regulatory Network The target gene–TF regulatory networkincluded 487 nodes (TFs: 195; gene: 292) and 7094 edges (Figure 7). The nodes with degrees were listed in Table S5. We discovered that that CLTC was targeted by 59 TFs (ex; SMARCA4); MYH9 was targeted by 53 TFs (ex; TCF7); NOTCH1 was targeted by 45 TFs (ex; MYB); SHC1 was targeted by 42 TFs (ex; E2F4); KIFC3 was targeted by 42 TFs (ex; CUX1); TCF4 was targeted by 50 TFs (ex; NANOG); VIM was targeted by 43 TFs (ex; GFI1B); CAV1 was targeted by 36 TFs (ex; GATA4); FBL was targeted by 33 TFs (ex; HIF1A); TUBA1A was targeted by 32 TFs (ex; CLOCK). ## 3.6. Receiver Operating Characteristic (ROC) Analysis To identify new potential biomarkers for diabetes and obesity, ROC curves of data derived from healthy controls and patients with diabetes and obesity was analyzed using the R package. The AUC calculated to assess the discriminatory ability of hub genes (Figure 8). Validated by ROC curves, we found that hub genes had high sensitivity and specificity, including MYH9, FLNA, DCTN1, CLTC, ERBB2, TCF4, VIM, LRRK2, IFI16 and CAV1, and AUC values more than 0.7. This analysis demonstrated that the hub genes had a diagnostic role. ## 4. Discussion A NGS investigation is an ideal way to comprehensively investigate diabetes mellitus and obesity. In this investigation, we collected NGS dataset from the GEO database, and a total of 872 DEGs, including 439 up-regulated genes and 433 down-regulated genes, were found. Altered expression of XIST (X inactive specific transcript) [33] and SELL (selectin L) [34] are associated with prognosis ofdiabetes. S100A9 and S100A8 are associated with the prognosis of diabetes mellitus and obesity [35]. IL1R2 [36] and SPINK5 [37] plays an important role in the diabetes mellitus and obesity. GO term and REACTOME enrichment analyzes were accomplished to examine interactions between the DEGs. The altered expression of genes including ERBB2 [38], DACT1 [39], ARAP1 [40], MYH9 [41], INPPL1 [42], SARM1 [43], NOTCH1 [44], ROBO1 [45], MAPK8IP1 [46], ANK1 [47], SARM1 [43], SREBF2 [48], SIK1 [49], PASK (PAS domain-containing serine/threonine kinase) [50], NOS2 [51], OAS3 [52], KL (klotho) [53], PECAM1 [54], S100A12 [55], S100P [56], BATF3 [57], PLEK (pleckstrin) [58], ALOX5 [59], ARG1 [60], CXCL8 [61], CXCR1 [62], PTAFR (platelet-activating factor receptor) [63], PYGL (glycogen phosphorylase L) [64], TCF4 [65], CAMP (cathelicidin antimicrobial peptide) [66], RUNX2 [67], PLA2G2A [68], GCG (glucagon) [69], RARRES2 [70] and HAP1 [71] in diabetes mellitus was reported to be an independent prognostic factors. ACHE (acetylcholinesterase) [72], FGFR3 [73], VLDLR (very-low-density lipoprotein receptor) [74], SHC1 [75], HDAC6 [76], CHRNA2 [77], CASR (calcium-sensing receptor) [78], ELK1 [79], TYK2 [80], CIITA (class II major histocompatibility complex transactivator) [81], ZAP70 [82], GPT (glutamic-pyruvic transaminase) [83], CHI3L1 [84], AIF1 [85], MMP9 [86], ITGB2 [87], CFD (complement factor D) [88], C3AR1 [89], LGALS1 [90], CD14 [91], TIMP1 [92], TLR2 [93], LTF (lactotransferrin) [94], BRCA2 [95] and IGFBP3 [96] are a potential prognostic markers in obesity. Sun et al. [ 97] reported that TRPM2 was significantly regulated in diabetes and obesity. Findings were implied by Richter et al. [ 98], Suchkova et al. [ 99], Qureshi et al. [ 100], Wang et al. [ 101], Wang et al. [ 102], Aoki-Suzuki et al. [ 103], Ohno et al. [ 104], Richter et al. [ 98], Rahman and Copeland [105], Congiu et al. [ 106], Ji et al. [ 107], Wollmer et al. [ 108], Yamazaki et al. [ 109], Bardien et al. [ 110], Comella Bolla et al. [ 111], Horvath et al. [ 112], Watanabe et al. [ 113], Kushima et al. [ 114], Grünblatt et al. [ 115], and Sato and Kawata [116] when they found that TAOK2, ACAP3, PLXNA3, PLXNA4, DCTN1, NTNG2, LRP4, AGRN (agrin), TAOK2, POLG (DNA polymerase gamma, catalytic subunit), KCNK2, OPRK1, ABCA2, ABCA7, LRRK2, CD200, PAK3, PADI2, EPHB1, CHAT (choline O-acetyltransferase) and SLC18A1 plays a substantial role in the patients with neurological and neuropsychiatric disorders. Studies showed that biomarkers include PLD2 [117], FLNA (filamin A) [118], SMURF1 [119], LINGO1 [120], CACNA1H [121], NLRP6 [122], NLRC3 [123], CXCR2 [124] and C5AR1 [125] plays an important role in progression of hypertension. Sauzeau et al. [ 126], Xu et al. [ 127], Hirota et al. [ 128], Alharatani et al. [ 129], Beitelshees et al. [ 130], Zhu et al. [ 131], Gil-Cayuela et al. [ 132], Liu et al. [ 133], Xie et al. [ 134], Kroupis et al. [ 135], López-Mejías et al. [ 136], Gremmel et al. [ 137] Yamada and Guo [138], Petri et al. [ 139], DeFilippis et al. [ 140], Rocca et al. [ 141] and Tur et al. [ 142] found that genes include VAV2, RASAL1, LIF (LIF interleukin 6 family cytokine), CTNND1, CACNA1C, MAP3K10, NRBP2, TRPM4, LILRB2, FCGR2A, PIK3CG, SELPLG (selectin P ligand), PRDX4, FPR2, PLG (plasminogen), SELENOM (selenoprotein M) and NCAM1 were a diagnostic markers of cardiovascular diseasesand could be used as therapeutic targets. Accumulating evidence shows that ITGB4 [143], SEMA3D [144], FCAR (Fc fragment of IgA receptor) [145], KIT (KIT proto-oncogene, receptor tyrosine kinase) [146], PGLYRP1 [147], IL17RB [148], BIRC5 [149] and PTGS1 [150] are associated with prognosis in asthma. Studies showed that GRK2 [151], ADCY3 [152], FASN (fatty acid synthase) [153], DGKD (diacylglycerol kinase delta) [154], DGKQ (diacylglycerol kinase theta) [154], IP6K1 [155], ANXA1 [156], SUCNR1 [157], PRNP (prion protein) [158], CXCR4 [159], CAV1 [160], LCN2 [161], AQP9 [162], NMU (neuromedin U) [163], NPY1R [164], FFAR2 [165], OSM (oncostatin M) [166] and TREM1 [167] might be a potential markers for diabetes mellitus and obesity. Researchers have shown that UNC13B [168], PFKFB3 [169], FCN1 [170] and SLC11A1 [171] were diagnostic markers for type 1 diabetes. DEGs involved in GO terms and pathways were more likely related to diabetes mellitus and obesity, and DEGs also involved in neurological and neuropsychiatric disorders, hypertension, cardiovascular diseases and asthma. As known, dynamic networks analysis and disease gene association were criteria for progression of various diseases [172,173]. Protein–protein interaction (PPI) network and its module can be regarded as key to the understanding of progression of diabetes mellitus and obesity, and might also lead to novel therapeutic way. MYH9 [41,174,175,176], ERBB2 [38,177,178,179,180], TCF4 [65,181], VIM (vimentin) [182,183], LRRK2 [184,185] and CAV1 [161,186,187,188,189,190,191,192] have been implicated as a principal mediator of diabetes mellitus. VIM (vimentin) binds to insulin-responsive aminopeptidas, a major cargo protein of glucose transporter type 4, and decreases the glucose tolerance [182]. IFI16 [193], ERBB2 [194], VIM (vimentin) [182,195] and CAV1 [160,196,197,198,199] are crucial factors for advancement of obesity. IFI16 showed adipogenesis, an enhanced inflammatory state and damaged insulin-stimulated glucose uptake in adipose tissue [193]. Motor protein MYH9 bindsto actin and producesmechanical force through magnesium-dependent hydrolysis of ATP, and it generatesthe contraction of striated and smooth muscles [200]. ErbB2 is a receptor tyrosine kinase family whose activity in cells depends on dimerization with another ligand-binding ErbB receptor, and associated with progression of various diseases [201]. TCF4 is a member of the basic helix–loop–helix (bHLH) family of transcription factors that have a key role in a various diseases [202]. VIM (vimentin) is an intermediate filament (IF) protein and plays an important role in epithelial–mesenchymal transition (EMT), a process that occurs during the development of various diseases [203]. LRRK2 is an enigmatic protein and has been one of the central molecules in a number of human diseases [204]. CAV1 is a cell surface protein shownto play a key role in insulin resistance [205]. IFI16 is an innate immune sensor for intracellular DNA and is associated with DNA damage in various diseases [206]. We identified novel targets including CLTC (clathrin heavy chain), TNS2, PLCG1 and NIFK (nucleolar protein interacting with the FHA domain of MKI67) for specific therapy of diabetes mellitus and obesity. Further investigation is needed to validate these results and investigate the roles of these biomarkers in diabetes mellitus and obesity. In the present investigation, NGS data analysis revealed that the mechanism of occurrence of diabetes mellitus and obesity might be related to the expression of miRNA and TF. To validate the accuracy of the target genes, miRNAs and TFs identified by target gene–miRNA regulatory network and target gene–TF regulatory network analysis. Yan et al. [ 207], Wang et al. [ 208], Yan et al. [ 209] and Guo et al. [ 210] showed that expression and prognosis of hsa-mir-4329, hsa-mir-3685, hsa-mir-6124, hsa-mir-1297 and SMARCA4 are associated with the risk of cardiovascular diseases. Several studies have shown that biomarkers including hsa-mir-1299 [211], hsa-mir-4779 [212] and hsa-mir-4459 [213] might be predictive biomarkers for the efficacy of diabetes mellitus treatment. TCF7 was revealed and regarded as diagnostic biomarker in type 1 diabetes mellitus [214]. Transcription factor MYB was involved in asthma [215]. MYB might be associated with diabetes and obesity. E2F4 [216] and CLOCK [217] are associated with prognosis in patients with diabetes mellitus and obesity. CUX1 [218], NANOG [219], GATA4 [220] and HIF1A [221] plays a vital role in the patients with obesity. Novel targets include MYO18A, SEC16A, CCNB1, MAD2L1, hsa-mir-4315, hsa-mir-6134, hsa-mir-9500, KIFC3, FBL (fibrillarin), TUBA1A and GFI1B might have crucial biologic functions in the pathogenesis of patients with diabetes mellitus and obesity. This result indicated that our identified biomarkers are involved in the pathological progression of diabetes and obesity, its associated complications beingneurological and neuropsychiatric disorders, hypertension, cardiovascular diseases and asthma, thus warranting further exploration. However, there are some limitations in this investigation. For instance, the NGS data were obtained from the GEO database and were not given by the authors. Therefore, further research should be conducted to verify whether these target genes can be used in the clinical treatment of diabetes mellitus and obesity. ## 5. Conclusions Using a bioinformatics analysis of NGS dataset GSE132831, we identified the genes of diabetes and obesity. We found that DEGs in patients were enriched for pathways mainly involved in the axon guidance, neutrophil degranulation, plasma membrane-bounded cell projection organization, and cell activation. Focusing on the key genes and corresponding pathways involved in diabetes and obesity could provide new insights for diabetes mellitus and obesity treatment. Hub genes including MYH9, FLNA, DCTN1, CLTC, ERBB2, TCF4, VIM, LRRK2, IFI16 and CAV1 were identified as potential novel biomarkers for diabetes and obesity. The validation of hub genes was demonstrated by ROC analysis. Further investigation isurgently demanded to validate the hub genes, and further molecular mechanisms would be uncovered. All the output will lay the foundation for finding a possible therapeutic strategy to treat diabetes mellitus and obesity. ## References 1. Hossain P., Kawar B., El Nahas M.. Obesity and Diabetes in the Developing World—A Growing Challenge. N. Engl. J. Med. (2007.0) 356 213-215. DOI: 10.1056/NEJMp068177 2. Daousi C., Casson I.F., Gill G.V., MacFarlane I.A., Wilding J.P., Pinkney J.H.. Prevalence of obesity in type 2 diabetes in secondary care: Association with cardiovascular risk factors. Postgrad Med. J. (2006.0) 82 280-284. DOI: 10.1136/pmj.2005.039032 3. Scherer P.E., Hill J.A.. Obesity, Diabetes, and Cardiovascular Diseases: A Compendium. Circ. Res. (2016.0) 118 1703-1705. DOI: 10.1161/CIRCRESAHA.116.308999 4. Babu G.R., Murthy G.V.S., Ana Y., Patel P., Deepa R., Neelon S.E.B., Kinra S., Reddy K.S.. Association of obesity with hypertension and type 2 diabetes mellitus in India: A meta-analysis of observational studies. World J. Diabetes (2018.0) 9 40-52. DOI: 10.4239/wjd.v9.i1.40 5. Naseer M., Bibi F., Alqahtani M.H., Chaudhary A.G., Azhar E.I., Kamal M.A., Yasir M.. Role of Gut Microbiota in Obesity, Type 2 Diabetes and Alzheimer’s Disease. CNS Neurol. Disord. Drug Targets (2014.0) 13 305-311. DOI: 10.2174/18715273113126660147 6. Medina-Remón A., Kirwan R., Lamuela-Raventós R.M., Estruch R.. Dietary patterns and the risk of obesity, type 2 diabetes mellitus, cardiovascular diseases, asthma, and neurodegenerative diseases. Crit. Rev. Food Sci. Nutr. (2018.0) 58 262-296. DOI: 10.1080/10408398.2016.1158690 7. Romao I., Roth J.. Genetic and Environmental Interactions in Obesity and Type 2 Diabetes. J. Am. Diet. Assoc. (2008.0) 108 S24-S28. DOI: 10.1016/j.jada.2008.01.022 8. Meyre D., Bouatia-Naji N., Tounian A., Samson C., Lecoeur C., Vatin V., Ghoussaini M., Wachter C., Hercberg S., Charpentier G.. Variants of ENPP1 are associated with childhood and adult obesity and increase the risk of glucose intolerance and type 2 diabetes. Nat. Genet. (2005.0) 37 863-867. DOI: 10.1038/ng1604 9. Ramya K., Radha V., Ghosh S., Majumder P.P., Mohan V.. Genetic Variations in the. Diabetes Technol. Ther. (2011.0) 13 33-42. DOI: 10.1089/dia.2010.0071 10. Huang X., Liu G., Guo J., Su Z.. The PI3K/AKT pathway in obesity and type 2 diabetes. Int. J. Biol. Sci. (2018.0) 14 1483-1496. DOI: 10.7150/ijbs.27173 11. Aamir K., Khan H.U., Sethi G., Hossain M.A., Arya A.. Wnt signaling mediates TLR pathway and promote unrestrained adipogenesis and metaflammation: Therapeutic targets for obesity and type 2 diabetes. Pharmacol. Res. (2020.0) 152 104602. DOI: 10.1016/j.phrs.2019.104602 12. Kumar S.U., Kumar D.T., Siva R., Doss C.G.P., Zayed H.. Integrative Bioinformatics Approaches to Map Potential Novel Genes and Pathways Involved in Ovarian Cancer. Front. Bioeng. Biotechnol. (2019.0) 7 391. DOI: 10.3389/fbioe.2019.00391 13. Udhaya Kumar S., Thirumal Kumar D., Bithia R., Sankar S., Magesh R., Sidenna M., George Priya Doss C., Zayed H.. Analysis of Differentially Expressed Genes and Molecular Pathways in Familial Hypercholesterolemia Involved in Atherosclerosis: A Systematic and Bioinformatics Approach. Front. Genet. (2020.0) 11 734. DOI: 10.3389/fgene.2020.00734 14. Fu D., Zhang B., Yang L., Huang S., Xin W.. Development of an Immune-Related Risk Signature for Predicting Prognosis in Lung Squamous Cell Carcinoma. Front. Genet. (2020.0) 11 978. DOI: 10.3389/fgene.2020.00978 15. Li X., Liao Z., Deng Z., Chen N., Zhao L.. Combining bulk and single-cell RNA-sequencing data to reveal gene expression pattern of chondrocytes in the osteoarthritic knee. Bioengineered (2021.0) 12 997-1007. DOI: 10.1080/21655979.2021.1903207 16. Prashanth G., Vastrad B., Tengli A., Vastrad C., Kotturshetti I.. Investigation of candidate genes and mechanisms underlying obesity associated type 2 diabetes mellitus using bioinformatics analysis and screening of small drug molecules. BMC Endocr. Disord. (2021.0) 21. DOI: 10.1186/s12902-021-00718-5 17. Osinski C., Le Gléau L., Poitou C., de Toro-Martin J., Genser L., Fradet M., Soula H.A., Leturque A., Blugeon C., Jourdren L.. Type 2 diabetes is associated with impaired jejunal enteroendocrine GLP-1 cell lineage in human obesity. Int. J. Obes. (2020.0) 45 170-183. DOI: 10.1038/s41366-020-00694-1 18. Clough E., Barrett T.. The Gene Expression Omnibus Database. Methods Mol. Biol. (2016.0) 1418 93-110. PMID: 27008011 19. Ritchie M.E., Phipson B., Wu D., Hu Y., Law C.W., Shi W., Smyth G.K.. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. (2015.0) 43 e47. DOI: 10.1093/nar/gkv007 20. Chen J., Bardes E.E., Aronow B.J., Jegga A.G.. ToppGene Suite for gene list enrichment analysis and candidate gene prioritization. Nucleic Acids Res. (2009.0) 37 W305-W311. DOI: 10.1093/nar/gkp427 21. Thomas P.D.. The Gene Ontology and the Meaning of Biological Function. Methods Mol. Biol. (2017.0) 1446 15-24. PMID: 27812932 22. Fabregat A., Jupe S., Matthews L., Sidiropoulos K., Gillespie M., Garapati P., Haw R., Jassal B., Korninger F., May B.. The Reactome Pathway Knowledgebase. Nucleic Acids Res. (2018.0) 46 D649-D655. DOI: 10.1093/nar/gkx1132 23. Kotlyar M., Pastrello C., Malik Z., Jurisica I.. IID 2018 update: Context-specific physical protein–protein interactions in human, model organisms and domesticated species. Nucleic Acids Res. (2019.0) 47 D581-D589. DOI: 10.1093/nar/gky1037 24. Shannon P., Markiel A., Ozier O., Baliga N.S., Wang J.T., Ramage D., Amin N., Schwikowski B., Ideker T.. Cytoscape: A software environment for integrated models of Biomolecular Interaction Networks. Genome Res. (2003.0) 13 2498-2504. DOI: 10.1101/gr.1239303 25. Przulj N., Wigle D.A., Jurisica I.. Functional topology in a network of protein interactions. Bioinformatics (2004.0) 20 340-348. DOI: 10.1093/bioinformatics/btg415 26. Nguyen T.-P., Liu W.-C., Jordán F.. Inferring pleiotropy by network analysis: Linked diseases in the human PPI network. BMC Syst. Biol. (2011.0) 5. DOI: 10.1186/1752-0509-5-179 27. Shi Z., Zhang B.. Fast network centrality analysis using GPUs. BMC Bioinform. (2011.0) 12. DOI: 10.1186/1471-2105-12-149 28. Fadhal E., Gamieldien J., Mwambene E.C.. Protein interaction networks as metric spaces: A novel perspective on distribution of hubs. BMC Syst. Biol. (2014.0) 8. DOI: 10.1186/1752-0509-8-6 29. Zaki N., Efimov D., Berengueres J.. Protein complex detection using interaction reliability assessment and weighted clustering coefficient. BMC Bioinform. (2013.0) 14. DOI: 10.1186/1471-2105-14-163 30. Fan Y., Xia J.. miRNet—Functional Analysis and Visual Exploration of miRNA–Target Interactions in a Network Context. Methods Mol. Biol. (2018.0) 1819 215-233. DOI: 10.1007/978-1-4939-8618-7_10 31. Zhou G., Soufan O., Ewald J., Hancock R.E.W., Basu N., Xia J.. NetworkAnalyst 3.0: A visual analytics platform for comprehensive gene expression profiling and meta-analysis. Nucleic Acids Res. (2019.0) 47 W234-W241. DOI: 10.1093/nar/gkz240 32. Robin X., Turck N., Hainard A., Tiberti N., Lisacek F., Sanchez J.-C., Müller M.. pROC: An open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinform. (2011.0) 12. DOI: 10.1186/1471-2105-12-77 33. Sohrabifar N., Ghaderian S.M.H., Parsa S.A., Ghaedi H., Jafari H.. Variation in the expression level of MALAT1, MIAT and XIST lncRNAs in coronary artery disease patients with and without type 2 diabetes mellitus. Arch. Physiol. Biochem. (2020.0) 128 1308-1315. DOI: 10.1080/13813455.2020.1768410 34. Stavarachi M., Panduru N.M., Serafinceanu C., Moţa E., Moţa M., Cimponeriu D., Ion D.A.. Investigation of P213S SELL gene polymorphism in type 2 diabetes mellitus and related end stage renal disease. A case-control study. Romanian J. Morphol. Embryol. (2011.0) 52 995-998 35. Lylloff L., Bathum L., Madsbad S., Grundtvig J.L.G., Nordgaard-Lassen I., Fenger M.. S100A8/A9 (Calprotectin), Interleukin-6, and C-Reactive Protein in Obesity and Diabetes before and after Roux-en-Y Gastric Bypass Surgery. Obes. Facts (2017.0) 10 386-395. DOI: 10.1159/000478097 36. Gagné-Ouellet V., Guay S.-P., Boucher-Lafleur A.-M., Bouchard L., Laprise C.. DNA methylation signature of interleukin 1 receptor type II in asthma. Clin. Epigenet. (2015.0) 7 80. DOI: 10.1186/s13148-015-0114-0 37. Martínez-Aguilar N.E., Del Río-Navarro B.E., Navarro-Olivos E., García-Ortíz H., Orozco L., Jiménez-Morales S.. J. Asthma (2015.0) 52 232-239. DOI: 10.3109/02770903.2014.966913 38. Muhammad I.F., Borné Y., Bao X., Melander O., Orho-Melander M., Nilsson P.M., Nilsson J., Engström G.. Circulating HER2/ErbB2 Levels Are Associated With Increased Incidence of Diabetes: A Population-Based Cohort Study. Diabetes Care (2019.0) 42 1582-1588. DOI: 10.2337/dc18-2556 39. Yu C.-Y., Yang C.-Y., Rui Z.-L.. MicroRNA-125b-5p improves pancreatic β-cell function through inhibiting JNK signaling pathway by targeting DACT1 in mice with type 2 diabetes mellitus. Life Sci. (2019.0) 224 67-75. DOI: 10.1016/j.lfs.2019.01.031 40. Kulzer J.R., Stitzel M.L., Morken M.A., Huyghe J.R., Fuchsberger C., Kuusisto J., Laakso M., Boehnke M., Collins F.S., Mohlke K.L.. A Common Functional Regulatory Variant at a Type 2 Diabetes Locus Upregulates ARAP1 Expression in the Pancreatic Beta Cell. Am. J. Hum. Genet. (2014.0) 94 186-197. DOI: 10.1016/j.ajhg.2013.12.011 41. Freedman B.I., Langefeld C.D., Lu L., Divers J., Comeau M.E., Kopp J., Winkler C.A., Nelson G.W., Johnson R.C., Palmer N.D.. Differential Effects of MYH9 and APOL1 Risk Variants on FRMD3 Association with Diabetic ESRD in African Americans. PLoS Genet. (2011.0) 7. DOI: 10.1371/journal.pgen.1002150 42. Marion E., Kaisaki P.J., Pouillon V., Gueydan C., Levy J.C., Bodson A., Krzentowski G., Daubresse J.-C., Mockel J., Behrends J.. The Gene INPPL1, Encoding the Lipid Phosphatase SHIP2, Is a Candidate for Type 2 Diabetes In Rat and Man. Diabetes (2002.0) 51 2012-2017. DOI: 10.2337/diabetes.51.7.2012 43. Cheng Y., Liu J., Luan Y., Liu Z., Lai H., Zhong W., Yang Y., Yu H., Feng N., Wang H.. Diabetes (2019.0) 68 2120-2130. DOI: 10.2337/db18-1233 44. DiNicolantonio J.J., Mccarty M.. Autophagy-induced degradation of Notch1, achieved through intermittent fasting, may promote beta cell neogenesis: Implications for reversal of type 2 diabetes. Open Heart (2019.0) 6 e001028. DOI: 10.1136/openhrt-2019-001028 45. Liu J., Hou W., Guan T., Tang L., Zhu X., Li Y., Hou S., Zhang J., Chen H., Huang Y.. Slit2/Robo1 signaling is involved in angiogenesis of glomerular endothelial cells exposed to a diabetic-like environment. Angiogenesis (2018.0) 21 237-249. DOI: 10.1007/s10456-017-9592-3 46. Waeber G., Delplanque J., Bonny C., Mooser V., Steinmann M., Widmann C., Maillard A., Miklossy J., Dina C., Hani E.H.. The gene MAPK8IP1, encoding islet-brain-1, is a candidate for type 2 diabetes. Nat. Genet. (2000.0) 24 291-295. DOI: 10.1038/73523 47. Sun L., Zhang X., Wang T., Chen M., Qiao H.. Association of ANK1 variants with new-onset type 2 diabetes in a Han Chinese population from northeast China. Exp. Ther. Med. (2017.0) 14 3184-3190. DOI: 10.3892/etm.2017.4866 48. Galavi H., Noorzehi N., Saravani R., Sargazi S., Mollashahee-Kohkan F., Shahraki H.. Association study of SREBF-2 gene polymorphisms and the risk of type 2 diabetes in a sample of Iranian population. Gene (2018.0) 660 145-150. DOI: 10.1016/j.gene.2018.03.080 49. Song D., Yin L., Wang C., Wen X.. Adenovirus-mediated expression of SIK1 improves hepatic glucose and lipid metabolism in type 2 diabetes mellitus rats. PLoS ONE (2019.0) 14. DOI: 10.1371/journal.pone.0210930 50. Da Silva Xavier G., Farhan H., Kim H., Caxaria S., Johnson P., Hughes S., Bugliani M., Marselli L., Marchetti P., Birzele F.. Per-arnt-sim (PAS) domain-containing protein kinase is downregulated in human islets in type 2 diabetes and regulates glucagon secretion. Diabetologia (2011.0) 54 819-827. DOI: 10.1007/s00125-010-2010-7 51. Chen F., Li Y.-M., Yang L.-Q., Zhong C.-G., Zhuang Z.-X.. Association of NOS2 and NOS3 gene polymorphisms with susceptibility to type 2 diabetes mellitus and diabetic nephropathy in the Chinese Han population. IUBMB Life (2016.0) 68 516-525. DOI: 10.1002/iub.1513 52. Park S., Liu M., Kang S.. Alcohol Intake Interacts with. Alcohol. Clin. Exp. Res. (2018.0) 42 2326-2336. DOI: 10.1111/acer.13888 53. Donate-Correa J., Martín-Núñez E., Ferri C., Hernández-Carballo C., Tagua V.G., Delgado-Molinos A., López-Castillo Á., Rodríguez-Ramos S., Cerro-López P., López-Tarruella V.C.. FGF23 and Klotho Levels are Independently Associated with Diabetic Foot Syndrome in Type 2 Diabetes Mellitus. J. Clin. Med. (2019.0) 8. DOI: 10.3390/jcm8040448 54. Završnik M., Kariž S., Makuc J., Šeruga M., Cilenšek I., Petrovič D.. PECAM-1 Leu125Val (rs688) Polymorphism and Diabetic Nephropathy in Caucasians with Type 2 Diabetes Mellitus. Anal. Cell. Pathol. (2016.0) 2016 3152967. DOI: 10.1155/2016/3152967 55. Dong N., Shi H., Xu B., Cai Y.. Increased Plasma S100A12 Levels Are Associated with Diabetic Retinopathy and Prognostic Biomarkers of Macrovascular Events in Type 2 Diabetic Patients. Investig. Opthalmology Vis. Sci. (2015.0) 56 4177-4185. DOI: 10.1167/iovs.15-16470 56. Afarideh M., Esteghamati V.Z., Ganji M., Heidari B., Esteghamati S., Lavasani S., Ahmadi M., Tafakhori A., Nakhjavani M., Esteghamati A.. Associations of Serum S100B and S100P With the Presence and Classification of Diabetic Peripheral Neuropathy in Adults with Type 2 Diabetes: A Case-Cohort Study. Can. J. Diabetes (2019.0) 43 336-344.e2. DOI: 10.1016/j.jcjd.2019.01.003 57. Ferris S.T., Carrero J.A., Mohan J.F., Calderon B., Murphy K.M., Unanue E.R.. A Minor Subset of Batf3-Dependent Antigen-Presenting Cells in Islets of Langerhans Is Essential for the Development of Autoimmune Diabetes. Immunity (2014.0) 41 657-669. DOI: 10.1016/j.immuni.2014.09.012 58. Ding Y., Kantarci A., Badwey J.A., Hasturk H., Malabanan A., Van Dyke T.E.. Phosphorylation of Pleckstrin Increases Proinflammatory Cytokine Secretion by Mononuclear Phagocytes in Diabetes Mellitus. J. Immunol. (2007.0) 179 647-654. DOI: 10.4049/jimmunol.179.1.647 59. Nejatian N., Häfner A.-K., Shoghi F., Badenhoop K., Penna-Martinez M.. 5-Lipoxygenase (ALOX5): Genetic susceptibility to type 2 diabetes and vitamin D effects on monocytes. J. Steroid Biochem. Mol. Biol. (2019.0) 187 52-57. DOI: 10.1016/j.jsbmb.2018.10.022 60. Shah S.F.A., Iqbal T., Naveed N., Akram S., Rafiq M.A., Hussain S.. ARG1 single nucleotide polymorphisms rs2781666 and rs2781665 confer risk of Type 2 diabetes mellitus. EXCLI J. (2018.0) 17 847-855. DOI: 10.17179/excli2018-1178 61. Da Silva B.R., Cirelli T., Nepomuceno R., Theodoro L.H., Orrico S.R., Cirelli J.A., Barros S.P., Scarel-Caminaga R.M.. Functional haplotype in the Interleukin8 (CXCL8) gene is associated with type 2 Diabetes Mellitus and Periodontitis in Brazilian population. Diabetes Metab. Syndr. (2020.0) 14 1665-1672. DOI: 10.1016/j.dsx.2020.08.036 62. Gond D.P., Singh S., Agrawal N.. Testing an association between TLR4 and CXCR1 gene polymorphisms with susceptibility to urinary tract infection in type 2 diabetes in north Indian population. Gene (2018.0) 641 196-202. DOI: 10.1016/j.gene.2017.10.060 63. Yamaguchi M., Matsui M., Higa R., Yamazaki Y., Ikari A., Miyake M., Miwa M., Ishii S., Sugatani J., Shimizu T.. A platelet-activating factor (PAF) receptor deficiency exacerbates diet-induced obesity but PAF/PAF receptor signaling does not contribute to the development of obesity-induced chronic inflammation. Biochem. Pharmacol. (2015.0) 93 482-495. DOI: 10.1016/j.bcp.2014.12.022 64. Nagy L., Docsa T., Szántó M., Brunyánszki A., Hegedűs C., Márton J., Kónya B., Virág L., Somsák L., Gergely P.. Glycogen Phosphorylase Inhibitor N-(3,5-Dimethyl-Benzoyl)-N’-(β-D-Glucopyranosyl)Urea Improves Glucose Tolerance under Normoglycemic and Diabetic Conditions and Rearranges Hepatic Metabolism. PLoS ONE (2013.0) 8. DOI: 10.1371/journal.pone.0069420 65. Boj S.F., van Es J.H., Huch M., Li V.S., José A., Hatzis P., Mokry M., Haegebarth A., Born M.V.D., Chambon P.. Diabetes Risk Gene and Wnt Effector Tcf7l2/TCF4 Controls Hepatic Response to Perinatal and Adult Metabolic Demand. Cell (2012.0) 151 1595-1607. DOI: 10.1016/j.cell.2012.10.053 66. Patrick C., Wang G.-S., Lefebvre D.E., Crookshank J.A., Sonier B., Eberhard C., Mojibian M., Kennedy C.R., Brooks S.P., Kalmokoff M.L.. Promotion of Autoimmune Diabetes by Cereal Diet in the Presence or Absence of Microbes Associated with Gut Immune Activation, Regulatory Imbalance, and Altered Cathelicidin Antimicrobial Peptide. Diabetes (2013.0) 62 2036-2047. DOI: 10.2337/db12-1243 67. Zhang G., Li H., Zhao W., Li M., Tian L., Ju W., Li X.. miR-205 regulates bone turnover in elderly female patients with type 2 diabetes mellitus through targeted inhibition of Runx2. Exp. Ther. Med. (2020.0) 20 1557-1565. DOI: 10.3892/etm.2020.8867 68. Khajeniazi S., Marjani A., Shakeri R., Hakimi S.. Polymorphism of Secretary PLA2G2A Gene Associated with Its Serum Level in Type2 Diabetes Mellitus Patients in Northern Iran. Endocrine Metab. Immune Disord. Drug Targets (2019.0) 19 1192-1197. DOI: 10.2174/1871530319666190528111225 69. Li L., Gao K., Zhao J., Feng T., Yin L., Wang J., Wang C., Li C., Wang Y., Wang Q.. Glucagon gene polymorphism modifies the effects of smoking and physical activity on risk of type 2 diabetes mellitus in Han Chinese. Gene (2014.0) 534 352-355. DOI: 10.1016/j.gene.2013.09.121 70. Zhao K., Ding W., Zhang Y., Ma K., Wang D., Hu C., Liu J., Zhang X.. Variants in the RARRES2 gene are associated with serum chemerin and increase the risk of diabetic kidney disease in type 2 diabetes. Int. J. Biol. Macromol. (2020.0) 165 1574-1580. DOI: 10.1016/j.ijbiomac.2020.10.030 71. Gong Y.-J., Feng Y., Cao Y.-Y., Zhao J., Wu W., Zheng Y.-Y., Wu J.-R., Li X., Yang G.-Z., Zhou X.. Huntingtin-associated protein 1 plays an essential role in the pathogenesis of type 2 diabetes by regulating the translocation of GLUT4 in mouse adipocytes. BMJ Open Diabetes Res. Care (2020.0) 8 e001199. DOI: 10.1136/bmjdrc-2020-001199 72. Shenhar-Tsarfaty S., Sherf-Dagan S., Berman G., Webb M., Raziel A., Keidar A., Goitein D., Sakran N., Zwang E., Shapira I.. Obesity-related acetylcholinesterase elevation is reversed following laparoscopic sleeve gastrectomy. Int. J. Obes. (2019.0) 43 297-305. DOI: 10.1038/s41366-018-0014-4 73. Saint-Laurent C., Garcia S., Sarrazy V., Dumas K., Authier F., Sore S., Tran A., Gual P., Gennero I., Salles J.-P.. Early postnatal soluble FGFR3 therapy prevents the atypical development of obesity in achondroplasia. PLoS ONE (2018.0) 13. DOI: 10.1371/journal.pone.0195876 74. Kim O.Y., Lee S.-M., Chung J.H., Do H.J., Moon J., Shin M.-J.. Arginase I and the very low-density lipoprotein receptor are associated with phenotypic biomarkers for obesity. Nutrition (2012.0) 28 635-639. DOI: 10.1016/j.nut.2011.09.012 75. Feigelson H.S., Teras L.R., Diver W.R., Tang W., Patel A.V., Stevens V.L., Calle E.E., Thun M.J., Bouzyk M.. Genetic variation in candidate obesity genes ADRB2, ADRB3, GHRL, HSD11B1, IRS1, IRS2, and SHC1 and risk for breast cancer in the Cancer Prevention Study II. Breast Cancer Res. (2008.0) 10 R57. DOI: 10.1186/bcr2114 76. Lieber A.D., Beier U.H., Xiao H., Wilkins B.J., Jiao J., Li X.S., Schugar R.C., Strauch C.M., Wang Z., Brown J.M.. Loss of HDAC6 alters gut microbiota and worsens obesity. FASEB J. (2019.0) 33 1098-1109. DOI: 10.1096/fj.201701586R 77. Kim J.. Association of CHRNA2 polymorphisms with overweight/obesity and clinical characteristics in a Korean population. Clin. Chem. Lab. Med. (2008.0) 46 1085-1089. DOI: 10.1515/CCLM.2008.230 78. Mattar P., Sanhueza S., Yuri G., Briones L., Perez-Leighton C., Rudich A., Lavandero S., Cifuentes M.. Calcium-Sensing Receptor in Adipose Tissue: Possible Association with Obesity-Related Elevated Autophagy. Int. J. Mol. Sci. (2020.0) 21. DOI: 10.3390/ijms21207617 79. Pang L., You L., Ji C.-B., Shi C., Chen L., Yang L., Huang F., Zhou Y., Zhang J., Chen X.. miR-1275 inhibits adipogenesis via ELK1 and its expression decreases in obese subjects. J. Mol. Endocrinol. (2016.0) 57 33-43. DOI: 10.1530/JME-16-0007 80. Derecka M., Gornicka A., Koralov S.B., Szczepanek K., Morgan M., Raje V., Sisler J., Zhang Q., Otero D., Cichy J.. Tyk2 and Stat3 Regulate Brown Adipose Tissue Differentiation and Obesity. Cell Metab. (2012.0) 16 814-824. DOI: 10.1016/j.cmet.2012.11.005 81. Deng T., Lyon C.J., Minze L.J., Lin J., Zou J., Liu J.Z., Ren Y., Yin Z., Hamilton D.J., Reardon P.R.. Class II Major Histocompatibility Complex Plays an Essential Role in Obesity-Induced Adipose Inflammation. Cell Metab. (2013.0) 17 411-422. DOI: 10.1016/j.cmet.2013.02.009 82. Torres K., Torres A., Chrościcki A., Maciejewski R., Radej S., Roliński J., Pietrzyk Ł., Wallner G.. Evaluation of lymphocytes CD4+ and CD8+ and expression of ZAP-70 kinase on CD3+ and CD19+ lymphocytes in obese patients undergoing laparoscopic cholecystectomy. Surg. Endosc. (2013.0) 27 872-879. DOI: 10.1007/s00464-012-2527-6 83. Nakamura S., Takezawa Y., Nakajima Y., Maeda T.. Elevation of glutamic pyruvic transaminase and .GAMMA.-glutamyl transpeptidase in obesity. Tohoku J. Exp. Med. (1980.0) 132 473-478. DOI: 10.1620/tjem.132.473 84. Teitsdottir U.D., Arnardottir E.S., Bjornsdottir E., Gislason T., Petersen P.H.. Obesity modulates the association between sleep apnea treatment and CHI3L1 levels but not CHIT1 activity in moderate to severe OSA: An observational study. Sleep Breath. (2018.0) 22 1101-1109. DOI: 10.1007/s11325-018-1731-6 85. Parikh D., Riascos-Bernal D.F., Egaña-Gorroño L., Jayakumar S., Almonte V., Chinnasamy P., Sibinga N.E.S.. Allograft inflammatory factor-1-like is not essential for age dependent weight gain or HFD-induced obesity and glucose insensitivity. Sci. Rep. (2020.0) 10 3594. DOI: 10.1038/s41598-020-60433-4 86. Allott E.H., Lysaght J., Cathcart M.C., Donohoe C.L., Cummins R., McGarrigle S.A., Kay E., Reynolds J.V., Pidgeon G.P.. MMP9 expression in oesophageal adenocarcinoma is upregulated with visceral obesity and is associated with poor tumour differentiation. Mol. Carcinog. (2013.0) 52 144-154. DOI: 10.1002/mc.21840 87. Awaya T., Yokosaki Y., Yamane K., Usui H., Kohno N., Eboshida A.. Gene-environment Association of an. Obesity (2008.0) 16 1463-1466. DOI: 10.1038/oby.2008.68 88. Mathews J.A., Wurmbrand A.P., Ribeiro L., Neto F.L., Shore S.A.. Induction of IL-17A Precedes Development of Airway Hyperresponsiveness during Diet-Induced Obesity and Correlates with Complement Factor D. Front. Immunol. (2014.0) 5 440. DOI: 10.3389/fimmu.2014.00440 89. Cero C., Razzoli M., Han R., Sahu B.S., Patricelli J., Guo Z., Zaidman N.A., Miles J.M., O’Grady S.M., Bartolomucci A.. The neuropeptide TLQP-21 opposes obesity via C3aR1-mediated enhancement of adrenergic-induced lipolysis. Mol. Metab. (2016.0) 6 148-158. DOI: 10.1016/j.molmet.2016.10.005 90. Mukherjee R., Kim S.W., Park T., Choi M.S., Yun J.W.. Targeted inhibition of galectin 1 by thiodigalactoside dramatically reduces body weight gain in diet-induced obese rats. Int. J. Obes. (2015.0) 39 1349-1358. DOI: 10.1038/ijo.2015.74 91. Leite F., Leite Â., Santos A., Lima M., Barbosa J., Cosentino M., Ribeiro L.. Predictors of Subclinical Inflammatory Obesity: Plasma Levels of Leptin, Very Low-Density Lipoprotein Cholesterol and CD14 Expression of CD16+ Monocytes. Obes. Facts (2017.0) 10 308-322. DOI: 10.1159/000464294 92. Andrade V.L., Petruceli E., Belo V.A., Andrade-Fernandes C.M., Russi C.V.C., Bosco A.A., Tanus-Santos J.E., Sandrim V.C.. Evaluation of plasmatic MMP-8, MMP-9, TIMP-1 and MPO levels in obese and lean women. Clin. Biochem. (2012.0) 45 412-415. DOI: 10.1016/j.clinbiochem.2012.01.008 93. Martínez-García M., Ojeda-Ojeda M., Rodríguez-Martín E., Insenser M., Moncayo S., Álvarez-Blasco F., Luque-Ramírez M., Escobar-Morreale H.F.. TLR2 and TLR4 Surface and Gene Expression in White Blood Cells after Fasting and Oral Glucose, Lipid and Protein Challenges: Influence of Obesity and Sex Hormones. Biomolecules (2020.0) 10. DOI: 10.3390/biom10010111 94. Jamka M., Kaczmarek N., Mądry E., Krzyżanowska-Jankowska P., Bajerska J., Kręgielska-Narożna M., Bogdański P., Walkowiak J.. Metabolic Health in Obese Subjects—Is There a Link to Lactoferrin and Lactoferrin Receptor-Related Gene Polymorphisms?. Nutrients (2020.0) 12. DOI: 10.3390/nu12092843 95. Peplonska B., Bukowska A., Wieczorek E., Przybek M., Zienolddiny S., Reszka E.. Rotating night work, lifestyle factors, obesity and promoter methylation in BRCA1 and BRCA2 genes among nurses and midwives. PLoS ONE (2017.0) 12. DOI: 10.1371/journal.pone.0178792 96. Moreno-Santos I., Castellano-Castillo D., Lara M.F., Fernandez-Garcia J.C., Tinahones F.J., Macias-Gonzalez M.. IGFBP-3 Interacts with the Vitamin D Receptor in Insulin Signaling Associated with Obesity in Visceral Adipose Tissue. Int. J. Mol. Sci. (2017.0) 18. DOI: 10.3390/ijms18112349 97. Sun L., Liu Y.-L., Ye F., Xie J.-W., Zeng J.-W., Qin L., Xue J., Wang Y.-T., Guo K.-M., Ma M.-M.. Free fatty acid-induced H2O2 activates TRPM2 to aggravate endothelial insulin resistance via Ca. Free. Radic. Biol. Med. (2019.0) 143 288-299. DOI: 10.1016/j.freeradbiomed.2019.08.018 98. Richter M., Murtaza N., Scharrenberg R., White S.H., Johanns O., Walker S., Yuen R.K.C., Schwanke B., Bedürftig B., Henis M.. Altered TAOK2 activity causes autism-related neurodevelopmental and cognitive abnormalities through RhoA signaling. Mol. Psychiatry (2019.0) 24 1329-1350. DOI: 10.1038/s41380-018-0025-5 99. Suchkova I.O., Borisova E.V., Patkin E.L.. Length Polymorphism and Methylation Status of UPS29 Minisatellite of the. Int. J. Mol. Sci. (2020.0) 21. DOI: 10.3390/ijms21239206 100. Qureshi M., Hatem M., Alroughani R., Jacob S.P., Al-Temaimi R.A.. NeuroMolecular Med. (2017.0) 19 286-292. DOI: 10.1007/s12017-017-8443-0 101. Wang H., Sun F.-R., Tan L., Zhang W., Wang Z.-X., Jiang T., Yu J.-T., Tan L.. Association study of the PLXNA4 gene with the risk of Alzheimer’s disease. Ann. Transl. Med. (2016.0) 4 108. DOI: 10.21037/atm.2016.03.23 102. Wang N., Ma Q., Peng P., Yu Y., Xu S., Wang G., Ying Z., Wang H.. Autophagy and Ubiquitin-Proteasome System Coordinate to Regulate the Protein Quality Control of Neurodegenerative Disease-Associated DCTN1. Neurotox. Res. (2020.0) 37 48-57. DOI: 10.1007/s12640-019-00113-y 103. Aoki-Suzuki M., Yamada K., Meerabux J., Iwayama-Shigeno Y., Ohba H., Iwamoto K., Takao H., Toyota T., Suto Y., Nakatani N.. A family-based association study and gene expression analyses of netrin-G1 and -G2 genes in schizophrenia. Biol. Psychiatry (2005.0) 57 382-393. DOI: 10.1016/j.biopsych.2004.11.022 104. Ohno K., Ohkawara B., Ito M.. Agrin-LRP4-MuSK signaling as a therapeutic target for myasthenia gravis and other neuromuscular disorders. Expert Opin. Ther. Targets (2017.0) 21 949-958. DOI: 10.1080/14728222.2017.1369960 105. Rahman S., Copeland W.C.. POLG-related disorders and their neurological manifestations. Nat. Rev. Neurol. (2019.0) 15 40-52. DOI: 10.1038/s41582-018-0101-0 106. Congiu C., Minelli A., Bonvicini C., Bortolomasi M., Sartori R., Maj C., Scassellati C., Maina G., Trabucchi L., Segala M.. The role of the potassium channel gene KCNK2 in major depressive disorder. Psychiatry Res. (2015.0) 225 489-492. DOI: 10.1016/j.psychres.2014.11.061 107. Ji H., Wang Y., Liu G., Xu X., Dai D., Chen Z., Zhou D., Zhou X., Han L., Li Y.. OPRK1 promoter hypermethylation increases the risk of Alzheimer’s disease. Neurosci. Lett. (2015.0) 606 24-29. DOI: 10.1016/j.neulet.2015.08.027 108. Wollmer M.A., Kapaki E., Hersberger M., Muntwyler J., Brunner F., Tsolaki M., Akatsu H., Kosaka K., Michikawa M., Molyva D.. Ethnicity-dependent genetic association of. Am. J. Med. Genet. Part B: Neuropsychiatr. Genet. (2006.0) 141 534-536. DOI: 10.1002/ajmg.b.30345 109. Yamazaki K., Yoshino Y., Mori T., Yoshida T., Ozaki Y., Sao T., Mori Y., Ochi S., Iga J.-I., Ueno S.-I.. Gene Expression and Methylation Analysis of ABCA7 in Patients with Alzheimer’s Disease. J. Alzheimer Dis. (2017.0) 57 171-181. DOI: 10.3233/JAD-161195 110. Bardien S., Lesage S., Brice A., Carr J.. Genetic characteristics of leucine-rich repeat kinase 2 (LRRK2) associated Parkinson’s disease. Park. Relat. Disord. (2011.0) 17 501-508. DOI: 10.1016/j.parkreldis.2010.11.008 111. Bolla A.C., Valente T., Miguez A., Brito V., Gines S., Solà C., Straccia M., Canals J.M.. CD200 is up-regulated in R6/1 transgenic mouse model of Huntington’s disease. PLoS ONE (2019.0) 14. DOI: 10.1371/journal.pone.0224901 112. Horvath G.A., Tarailo-Graovac M., Bartel T., Race S., Van Allen M.I., Blydt-Hansen I., Ross C.J., Wasserman W.W., Connolly M.B., van Karnebeek C.D.M.. Improvement of Self-Injury with Dopamine and Serotonin Replacement Therapy in a Patient with a Hemizygous. J. Child Neurol. (2018.0) 33 106-113. DOI: 10.1177/0883073817740443 113. Watanabe Y., Nunokawa A., Kaneko N., Arinami T., Ujike H., Inada T., Iwata N., Kunugi H., Itokawa M., Otowa T.. A two-stage case–control association study of. J. Hum. Genet. (2009.0) 54 430-432. DOI: 10.1038/jhg.2009.52 114. Kushima I., Nakamura Y., Aleksic B., Ikeda M., Ito Y., Shiino T., Okochi T., Fukuo Y., Ujike H., Suzuki M.. Resequencing and Association Analysis of the KALRN and EPHB1 Genes and Their Contribution to Schizophrenia Susceptibility. Schizophr. Bull. (2012.0) 38 552-560. DOI: 10.1093/schbul/sbq118 115. Grünblatt E., Reif A., Jungwirth S., Galimberti D., Weber H., Scarpini E., Sauer C., Wichart I., Rainer M.K., Huber K.. Genetic variation in the choline O-acetyltransferase gene in depression and Alzheimer’s disease: The VITA and Milano studies. J. Psychiatr. Res. (2011.0) 45 1250-1256. DOI: 10.1016/j.jpsychires.2011.03.017 116. Sato D.X., Kawata M.. Positive and balancing selection on. Evol. Lett. (2018.0) 2 499-510. DOI: 10.1002/evl3.81 117. Kundu A., Ramakrishnan P., Rajendran A., Dharwar N.V., Anbarasu A.. Analysis of non-synonymous single-nucleotide polymorphisms and population variability of PLD2 gene associated with hypertension. Int. J. Bioinform. Res. Appl. (2013.0) 9 227-241. DOI: 10.1504/IJBRA.2013.053604 118. Jain M., Mann T.D., Stulić M., Rao S.P., Kirsch A., Pullirsch D., Strobl X., Rath C., Reissig L., Moreth K.. RNA editing of Filamin A pre- mRNA regulates vascular contraction and diastolic blood pressure. EMBO J. (2018.0) 37 e94813. DOI: 10.15252/embj.201694813 119. Baptista R., Marques C., Catarino S., Enguita F.J., Costa M.C., Matafome P., Zuzarte M., Castro G., Reis A., Monteiro P.. MicroRNA-424(322) as a new marker of disease progression in pulmonary arterial hypertension and its role in right ventricular hypertrophy by targeting SMURF1. Cardiovasc. Res. (2018.0) 114 53-64. DOI: 10.1093/cvr/cvx187 120. Fu Q.-L., Hu B., Li X., Shao Z., Shi J.-B., Wu W., So K.-F., Mi S.. LINGO-1 negatively regulates TrkB phosphorylation after ocular hypertension. Eur. J. Neurosci. (2010.0) 31 1091-1097. DOI: 10.1111/j.1460-9568.2010.07127.x 121. Scholl U.I., Stölting G., Nelson-Williams C., Vichot A.A., Choi M., Loring E., Prasad M.L., Goh G., Carling T., Juhlin C.C.. Recurrent gain of function mutation in calcium channel. Elife (2015.0) 4 e06315. DOI: 10.7554/eLife.06315 122. Glorioso N., Herrera V.L., Didishvili T., Ortu M.F., Zaninello R., Fresu G., Argiolas G., Troffa C., Ruiz-Opazo N.. Sex-Specific Effects of NLRP6/AVR and ADM Loci on Susceptibility to Essential Hypertension in a Sardinian Population. PLoS ONE (2013.0) 8. DOI: 10.1371/journal.pone.0077562 123. Zha L., Zhou J., Li T., Luo H., Zhang M., Li S., Yu Z.. NLRC3 inhibits MCT-induced pulmonary hypertension in rats via attenuating PI3K activation. J. Cell. Physiol. (2019.0) 234 15963-15976. DOI: 10.1002/jcp.28255 124. Zhang Y., Teng F., Han X., Li P., Yan X., Guo S., Li H.. Selective blocking of CXCR2 prevents and reverses atrial fibrillation in spontaneously hypertensive rats. J. Cell. Mol. Med. (2020.0) 24 11272-11282. DOI: 10.1111/jcmm.15694 125. Weiss S., Rosendahl A., Czesla D., Meyer-Schwesinger C., Stahl R.A.K., Ehmke H., Kurts C., Zipfel P.F., Köhl J., Wenzel U.O.. The complement receptor C5aR1 contributes to renal damage but protects the heart in angiotensin II-induced hypertension. Am. J. Physiol. Physiol. (2016.0) 310 F1356-F1365. DOI: 10.1152/ajprenal.00040.2016 126. Sauzeau V., Jerkic M., López-Novoa J.M., Bustelo X.R.. Loss of. Mol. Biol. Cell (2007.0) 18 943-952. DOI: 10.1091/mbc.e06-09-0877 127. Xu X., Tan X., Tampe B., Nyamsuren G., Liu X., Maier L.S., Sossalla S., Kalluri R., Zeisberg M., Hasenfuss G.. Epigenetic balance of aberrant Rasal1 promoter methylation and hydroxymethylation regulates cardiac fibrosis. Cardiovasc. Res. (2015.0) 105 279-291. DOI: 10.1093/cvr/cvv015 128. Hirota H., Izumi M., Hamaguchi T., Sugiyama S., Murakami E., Kunisada K., Fujio Y., Oshima Y., Nakaoka Y., Yamauchi-Takihara K.. Circulating interleukin-6 family cytokines and their receptors in patients with congestive heart failure. Heart Vessel. (2004.0) 19 237-241. DOI: 10.1007/s00380-004-0770-z 129. Alharatani R., Ververi A., Beleza-Meireles A., Ji W., Mis E., Patterson Q.T., Griffin J.N., Bhujel N., Chang C.A., Dixit A.. Novel truncating mutations in. Hum. Mol. Genet. (2020.0) 29 1900-1921. DOI: 10.1093/hmg/ddaa050 130. Beitelshees A.L., Navare H., Wang D., Gong Y., Wessel J., Moss J.I., Langaee T.Y., Cooper-DeHoff R.M., Sadee W., Pepine C.J.. Circ. Cardiovasc. Genet. (2009.0) 2 362-370. DOI: 10.1161/CIRCGENETICS.109.857839 131. Zhu J., Chen T., Yang L., Li Z., Wong M.M., Zheng X., Pan X., Zhang L., Yan H.. Regulation of MicroRNA-155 in Atherosclerotic Inflammatory Responses by Targeting MAP3K10. PLoS ONE (2012.0) 7. DOI: 10.1371/journal.pone.0046551 132. Gil-Cayuela C., López A., Martinez-Dolz L., Juanatey J.R.G., Lago F., Roselló-Lletí E., Rivera M., Portolés M.. The altered expression of autophagy-related genes participates in heart failure:. PLoS ONE (2019.0) 14. DOI: 10.1371/journal.pone.0215818 133. Liu H., El Zein L., Kruse M., Guinamard R., Beckmann A., Bozio A., Kurtbay G., Mégarbané A., Ohmert I., Blaysat G.. Gain-of-Function Mutations in. Circ. Cardiovasc. Genet. (2010.0) 3 374-385. DOI: 10.1161/CIRCGENETICS.109.930867 134. Xie M., Hu C., Li D., Li S.. MicroRNA-377 Alleviates Myocardial Injury Induced by Hypoxia/Reoxygenation via Downregulating LILRB2 Expression. Dose-Response (2020.0) 18 1559325820936124. DOI: 10.1177/1559325820936124 135. Kroupis C., Theodorou M., Chaidaroglou A., Dalamaga M., Oliveira S.C., Cokkinos D.V., Degiannis D., Manginas A.. The Association Between a CommonFCGR2APolymorphism and C-Reactive Protein and Coronary Artery Disease Revisited. Genet. Test. Mol. Biomark. (2010.0) 14 839-846. DOI: 10.1089/gtmb.2010.0108 136. López-Mejías R., Genre F., García-Bermúdez M., Ubilla B., Castañeda S., Llorca J., González-Juanatey C., Corrales A., Miranda-Filloy J.A., Pina T.. Lack of Association between. Mediat. Inflamm. (2014.0) 2014 756279. DOI: 10.1155/2014/756279 137. Koppensteiner R., Kaider A., Eichelberger B., Mannhalter C., Panzer S., Gremmel T.. Impact of variables of the P-selectin–P-selectin glycoprotein ligand-1 axis on leukocyte-platelet interactions in cardiovascular disease. Thromb. Haemost. (2015.0) 113 806-812. DOI: 10.1160/TH14-08-0690 138. Yamada S., Guo X.. Peroxiredoxin 4 (PRDX4): Its critical in vivo roles in animal models of metabolic syndrome ranging from atherosclerosis to nonalcoholic fatty liver disease. Pathol. Int. (2018.0) 68 91-101. DOI: 10.1111/pin.12634 139. Petri M.H., Laguna-Fernández A., Gonzalez-Diez M., Paulsson-Berne G., Hansson G.K., Bäck M.. The role of the FPR2/ALX receptor in atherosclerosis development and plaque stability. Cardiovasc. Res. (2015.0) 105 65-74. DOI: 10.1093/cvr/cvu224 140. DeFilippis A.P., Chernyavskiy I., Amraotkar A.R., Trainor P.J., Kothari S., Ismail I., Hargis C.W., Korley F.K., Leibundgut G., Tsimikas S.. Circulating levels of plasminogen and oxidized phospholipids bound to plasminogen distinguish between atherothrombotic and non-atherothrombotic myocardial infarction. J. Thromb. Thrombolysis (2016.0) 42 61-76. DOI: 10.1007/s11239-015-1292-5 141. Rocca C., Pasqua T., Boukhzar L., Anouar Y., Angelone T.. Progress in the emerging role of selenoproteins in cardiovascular disease: Focus on endoplasmic reticulum-resident selenoproteins. Cell. Mol. Life Sci. (2019.0) 76 3969-3985. DOI: 10.1007/s00018-019-03195-1 142. Tur M.K., Etschmann B., Benz A., Leich E., Waller C., Schuh K., Rosenwald A., Ertl G., Kienitz A., Haaf A.T.. The 140-kD Isoform of CD56 (NCAM1) Directs the Molecular Pathogenesis of Ischemic Cardiomyopathy. Am. J. Pathol. (2013.0) 182 1205-1218. DOI: 10.1016/j.ajpath.2012.12.027 143. Yu P.-F., Pang L.-L., Mao Q.-S., Zou S.-C., Shi Y., Lin D.-J.. Dose dependency PM2.5 aggravated airway inflammation in asthmatic mice via down-regulating expression of ITGB4. Eur. Rev. Med. Pharmacol. Sci. (2019.0) 23 1688-1697. PMID: 30840294 144. McGeachie M.J., Wu A.C., Tse S.M., Clemmer G.L., Sordillo J., Himes B.E., Lasky-Su J., Chase R.P., Martinez F.D., Weeke P.. CTNNA3 and SEMA3D: Promising loci for asthma exacerbation identified through multiple genome-wide association studies. J. Allergy Clin. Immunol. (2015.0) 136 1503-1510. DOI: 10.1016/j.jaci.2015.04.039 145. Jasek M., Obojski A., Mańczak M., Wiśniewski A., Winiarska B., Małolepszy J., Jutel M., Łuszczek W., Kuśnierczyk P.. Are Single Nucleotide Polymorphisms of the Immunoglobulin A Fc Receptor Gene Associated with Allergic Asthma?. Int. Arch. Allergy Immunol. (2004.0) 135 325-331. DOI: 10.1159/000082327 146. Cahill K.N., Katz H.R., Cui J., Lai J., Kazani S., Crosby-Thompson A., Garofalo D., Castro M., Jarjour N., DiMango E.. KIT Inhibition by Imatinib in Patients with Severe Refractory Asthma. N. Engl. J. Med. (2017.0) 376 1911-1920. DOI: 10.1056/NEJMoa1613125 147. Banskar S., Detzner A.A., Juarez-Rodriguez M.D., Hozo I., Gupta D., Dziarski R.. The. J. Immunol. (2019.0) 203 3113-3125. DOI: 10.4049/jimmunol.1900711 148. Hunninghake G.M., Chu J.-H., Sharma S.S., Cho M.H., Himes B.E., Rogers A.J., Murphy A., Carey V.J., Raby B.A.. The CD4+ T-cell transcriptome and serum IgE in asthma: IL17RB and the role of sex. BMC Pulm. Med. (2011.0) 11. DOI: 10.1186/1471-2466-11-17 149. Ungvári I., Hadadi E., Virág V., Bikov A., Nagy A., Semsei A.F., Gálffy G., Tamási L., Horvath I., Szalai C.. Implication of BIRC5 in asthma pathogenesis. Int. Immunol. (2012.0) 24 293-301. DOI: 10.1093/intimm/dxs007 150. Yucesoy B., Kashon M.L., Johnson V.J., Lummus Z.L., Fluharty K., Gautrin D., Cartier A., Boulet L.-P., Sastre J., Quirce S.. Genetic variants in. J. Immunotoxicol. (2016.0) 13 119-126. DOI: 10.3109/1547691X.2015.1017061 151. Vila-Bedmar R., Cruces-Sande M., Lucas E., Willemen H.L.D.M., Heijnen C.J., Kavelaars A., Mayor F., Murga C.. Reversal of diet-induced obesity and insulin resistance by inducible genetic ablation of GRK2. Sci. Signal. (2015.0) 8 ra73. DOI: 10.1126/scisignal.aaa4374 152. Grarup N., Moltke I., Andersen M.K., Dalby M., Vitting-Seerup K., Kern T., Mahendran Y., Jørsboe E., Larsen C.V.L., Dahl-Petersen I.. Loss-of-function variants in ADCY3 increase risk of obesity and type 2 diabetes. Nat. Genet. (2018.0) 50 172-174. DOI: 10.1038/s41588-017-0022-7 153. Berndt J., Kovacs P., Ruschke K., Klöting N., Fasshauer M., Schön M.R., Körner A., Stumvoll M., Blüher M.. Fatty acid synthase gene expression in human adipose tissue: Association with obesity and type 2 diabetes. Diabetologia (2007.0) 50 1472-1480. DOI: 10.1007/s00125-007-0689-x 154. Mannerås-Holm L., Kirchner H., Björnholm M., Chibalin A.V., Zierath J.R.. mRNA expression of diacylglycerol kinase isoforms in insulin-sensitive tissues: Effects of obesity and insulin resistance. Physiol. Rep. (2015.0) 3 e12372. DOI: 10.14814/phy2.12372 155. Ghoshal S., Zhu Q., Asteian A., Lin H., Xu H., Ernst G., Barrow J.C., Xu B., Cameron M.D., Kamenecka T.M.. TNP [N2-(m-Trifluorobenzyl), N6-(p-nitrobenzyl)purine] ameliorates diet induced obesity and insulin resistance via inhibition of the IP6K1 pathway. Mol. Metab. (2016.0) 5 903-917. DOI: 10.1016/j.molmet.2016.08.008 156. Pietrani N.T., Ferreira C.N., Rodrigues K.F., Perucci L.O., Carneiro F.S., Bosco A.A., Oliveira M.C., Pereira S.S., Teixeira A.L., Alvarez-Leite J.I.. Proresolving protein Annexin A1: The role in type 2 diabetes mellitus and obesity. Biomed. Pharmacother. (2018.0) 103 482-489. DOI: 10.1016/j.biopha.2018.04.024 157. Van Diepen J.A., Robben J.H., Hooiveld G.J., Carmone C., Alsady M., Boutens L., Bekkenkamp-Grovenstein M., Hijmans A., Engelke U.F.H., Wevers R.A.. SUCNR1-mediated chemotaxis of macrophages aggravates obesity-induced inflammation and diabetes. Diabetologia (2017.0) 60 1304-1313. DOI: 10.1007/s00125-017-4261-z 158. De Brito G., Lupinacci F.C., Beraldo F.H., Santos T.G., Roffé M., Lopes M.H., de Lima V.C., Martins V.R., Hajj G.N.. Loss of prion protein is associated with the development of insulin resistance and obesity. Biochem. J. (2017.0) 474 2981-2991. DOI: 10.1042/BCJ20170137 159. Dang Z., Avolio E., Thomas A.C., Faulkner A., Beltrami A.P., Cervellin C., Carrizzo A., Maciag A., Gu Y., Ciaglia E.. Transfer of a human gene variant associated with exceptional longevity improves cardiac function in obese type 2 diabetic mice through induction of the SDF-1/CXCR4 signalling pathway. Eur. J. Heart Fail. (2020.0) 22 1568-1581. DOI: 10.1002/ejhf.1840 160. Catalan V., Gomez-Ambrosi J., Rodríguez A., Silva C., Rotellar F., Gil M.J., Cienfuegos J., Salvador J., Frühbeck G.. Expression of caveolin-1 in human adipose tissue is upregulated in obesity and obesity-associated type 2 diabetes mellitus and related to inflammation. Clin. Endocrinol. (2008.0) 68 213-219. DOI: 10.1111/j.1365-2265.2007.03021.x 161. Elkhidir A.E., Eltaher H.B., Mohamed A.O.. Association of lipocalin-2 level, glycemic status and obesity in type 2 diabetes mellitus. BMC Res. Notes (2017.0) 10. DOI: 10.1186/s13104-017-2604-y 162. Catalán V., Gómez-Ambrosi J., Pastor C., Rotellar F., Silva C., Rodríguez A., Gil M.J., Cienfuegos J.A., Salvador J., Vendrell J.. Influence of Morbid Obesity and Insulin Resistance on Gene Expression Levels of AQP7 in Visceral Adipose Tissue and AQP9 in Liver. Obes. Surg. (2008.0) 18 695-701. DOI: 10.1007/s11695-008-9453-7 163. Ingallinella P., Peier A.M., Pocai A., Di Marco A., Desai K., Zytko K., Qian Y., Du X., Cellucci A., Monteagudo E.. PEGylation of Neuromedin U yields a promising candidate for the treatment of obesity and diabetes. Bioorganic Med. Chem. (2012.0) 20 4751-4759. DOI: 10.1016/j.bmc.2012.06.003 164. Wittrisch S., Klöting N., Mörl K., Chakaroun R., Blüher M., Beck-Sickinger A.G.. NPY1R-targeted peptide-mediated delivery of a dual PPARα/γ agonist to adipocytes enhances adipogenesis and prevents diabetes progression. Mol. Metab. (2020.0) 31 163-180. DOI: 10.1016/j.molmet.2019.11.009 165. Bódis K., Kahl S., Simon M.-C., Zhou Z., Sell H., Knebel B., Tura A., Strassburger K., Burkart V., Müssig K.. Reduced expression of stearoyl-CoA desaturase-1, but not free fatty acid receptor 2 or 4 in subcutaneous adipose tissue of patients with newly diagnosed type 2 diabetes mellitus. Nutr. Diabetes (2018.0) 8 49. DOI: 10.1038/s41387-018-0054-9 166. Sánchez-Infantes D., White U.A., Elks C.M., Morrison R.F., Gimble J.M., Considine R.V., Ferrante A.W., Ravussin E., Stephens J.M.. Oncostatin M Is Produced in Adipose Tissue and Is Regulated in Conditions of Obesity and Type 2 Diabetes. J. Clin. Endocrinol. Metab. (2014.0) 99 E217-E225. DOI: 10.1210/jc.2013-3555 167. Subramanian S., Pallati P.K., Sharma P., Agrawal D.K., Nandipati K.C.. Significant association of TREM-1 with HMGB1, TLRs and RAGE in the pathogenesis of insulin resistance in obese diabetic populations. Am. J. Transl. Res. (2017.0) 9 3224-3244. PMID: 28804542 168. Trégouet D.-A., Groop P.-H., McGinn S., Forsblom C., Hadjadj S., Marre M., Parving H.-H., Tarnow L., Telgmann R., Godefroy T.. G/T Substitution in Intron 1 of the. Diabetes (2008.0) 57 2843-2850. DOI: 10.2337/db08-0073 169. Nomoto H., Pei L., Montemurro C., Rosenberger M., Furterer A., Coppola G., Nadel B., Pellegrini M., Gurlo T., Butler P.C.. Activation of the HIF1α/PFKFB3 stress response pathway in beta cells in type 1 diabetes. Diabetologia (2020.0) 63 149-161. DOI: 10.1007/s00125-019-05030-5 170. Anjosa Z.P., Santos M.M.S., Rodrigues N.J., DE Lacerda G.A.N., Araujo J., Silva J.D.A., Tavares N.D.A.C., Guimarães R.L., Crovella S., Brandão L.A.C.. Polymorphism in ficolin-1 (FCN1) gene is associated with an earlier onset of type 1 diabetes mellitus in children and adolescents from northeast Brazil. J. Genet. (2016.0) 95 1031-1034. DOI: 10.1007/s12041-016-0719-x 171. Paccagnini D., Sieswerda L., Rosu V., Masala S., Pacifico A., Gazouli M., Ikonomopoulos J., Ahmed N., Zanetti S., Sechi L.A.. Linking Chronic Infection and Autoimmune Diseases: Mycobacterium avium Subspecies paratuberculosis, SLC11A1 Polymorphisms and Type-1 Diabetes Mellitus. PLoS ONE (2009.0) 4. DOI: 10.1371/journal.pone.0007109 172. Cinaglia P., Cannataro M.. Network alignment and motif discovery in dynamic networks. Netw. Model. Anal. Health Inform. Bioinform. (2022.0) 11 38. DOI: 10.1007/s13721-022-00383-1 173. Cinaglia P., Guzzi P.H., Veltri P.. INTEGRO: An algorithm for data-integration and disease-gene association. Proceedings of the 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) 2076-2081. DOI: 10.1109/bibm.2018.8621193 174. Freedman B.I., Hicks P.J., Bostrom M.A., Comeau M.E., Divers J., Bleyer A.J., Kopp J.B., Winkler C.A., Nelson G.W., Langefeld C.D.. Non-muscle myosin heavy chain 9 gene MYH9 associations in African Americans with clinically diagnosed type 2 diabetes mellitus-associated ESRD. Nephrol. Dial. Transplant. (2009.0) 24 3366-3371. DOI: 10.1093/ndt/gfp316 175. Zhao H., Ma L., Yan M., Wang Y., Zhao T., Zhang H., Liu P., Liu Y., Li P.. Association between MYH9 and APOL1 Gene Polymorphisms and the Risk of Diabetic Kidney Disease in Patients with Type 2 Diabetes in a Chinese Han Population. J. Diabetes Res. (2018.0) 2018 5068578. DOI: 10.1155/2018/5068578 176. Ling C., Cai C., Chang B., Shi W., Wei F., Yu P., Chen L., Li W.. MYH9 gene polymorphisms may be associated with cerebrovascular blood flow in patients with type 2 diabetes. Genet. Mol. Res. (2015.0) 14 1008-1016. DOI: 10.4238/2015.February.6.4 177. Huang Y., Han X., Chang T., Li F.-F., Chen X., She Y.-Q.. Serum ErbB2 concentration positively correlated to the glycemic variations in newly diagnosed Type 2 diabetic patients. Sci. Rep. (2022.0) 12 4940. DOI: 10.1038/s41598-022-07549-x 178. De Kay J.T., Carver J., Shevenell B., Kosta A.M., Tsibulnikov S., Certo E., Sawyer D.B., Ryzhov S., Robich M.P.. Decreased expression of ErbB2 on left ventricular epicardial cells in patients with diabetes mellitus. Cell. Signal. (2022.0) 96 110360. DOI: 10.1016/j.cellsig.2022.110360 179. Akhtar S., Yousif M.H.M., Chandrasekhar B., Benter I.F.. Activation of EGFR/ERBB2 via Pathways Involving ERK1/2, P38 MAPK, AKT and FOXO Enhances Recovery of Diabetic Hearts from Ischemia-Reperfusion Injury. PLoS ONE (2012.0) 7. DOI: 10.1371/journal.pone.0039066 180. Akhtar S., Yousif M., Dhaunsi G.S., Sarkhouh F., Chandrasekhar B., Attur S., Benter I.F.. Activation of ErbB2 and Downstream Signalling via Rho Kinases and ERK1/2 Contributes to Diabetes-Induced Vascular Dysfunction. PLoS ONE (2013.0) 8. DOI: 10.1371/journal.pone.0067813 181. Wei H., Qu H., Wang H., Ji B., Ding Y., Liu D., Duan Y., Liang H., Peng C., Xiao X.. 1,25-Dihydroxyvitamin-D3 prevents the development of diabetic cardiomyopathy in type 1 diabetic rats by enhancing autophagy via inhibiting the β-catenin/TCF4/GSK-3β/mTOR pathway. J. Steroid Biochem. Mol. Biol. (2017.0) 168 71-90. DOI: 10.1016/j.jsbmb.2017.02.007 182. Kim S., Kim I., Cho W., Oh G.T., Park Y.M.. Vimentin Deficiency Prevents High-Fat Diet-Induced Obesity and Insulin Resistance in Mice. Diabetes Metab. J. (2020.0) 45 97-108. DOI: 10.4093/dmj.2019.0198 183. Roefs M.M., Carlotti F., Jones K., Wills H., Hamilton A., Verschoor M., Durkin J.M.W., Perez L.G., Brereton M.F., McCulloch L.. Increased vimentin in human α- and β-cells in type 2 diabetes. J. Endocrinol. (2017.0) 233 217-227. DOI: 10.1530/JOE-16-0588 184. Yang S., Xia C., Li S., Du L., Zhang L., Hu Y.. Mitochondrial dysfunction driven by the LRRK2-mediated pathway is associated with loss of Purkinje cells and motor coordination deficits in diabetic rat model. Cell Death Dis. (2014.0) 5 e1217. DOI: 10.1038/cddis.2014.184 185. Huang M.-H., Liu Y.-F., Nfor O.N., Hsu S.-Y., Lin W.-Y., Chang Y.-S., Liaw Y.-P.. Interactive Association Between Intronic Polymorphism (rs10506151) of the LRRK2 Gene and Type 2 Diabetes on Neurodegenerative Diseases. Pharm. Pers. Med. (2021.0) 14 839-847. DOI: 10.2147/PGPM.S316158 186. Reinbothe T., Alkayyali S., Ahlqvist E., Tuomi T., Isomaa B., Lyssenko V., Renström E.. The human L-type calcium channel Cav1.3 regulates insulin release and polymorphisms in CACNA1D associate with type 2 diabetes. Diabetologia (2013.0) 56 340-349. DOI: 10.1007/s00125-012-2758-z 187. Bonds J.A., Shetti A., Bheri A., Chen Z., Disouky A., Tai L., Mao M., Head B.P., Bonini M.G., Haus J.M.. Depletion of Caveolin-1 in Type 2 Diabetes Model Induces Alzheimer’s Disease Pathology Precursors. J. Neurosci. (2019.0) 39 8576-8583. DOI: 10.1523/JNEUROSCI.0730-19.2019 188. Luo M., Xu C., Luo Y., Wang G., Wu J., Wan Q.. Circulating miR-103 family as potential biomarkers for type 2 diabetes through targeting CAV-1 and SFRP4. Acta Diabetol. (2020.0) 57 309-322. DOI: 10.1007/s00592-019-01430-6 189. Oh Y.S., Lee T.S., Cheon G.J., Jang I.-S., Jun H.-S., Park S.C.. Modulation of Insulin Sensitivity and Caveolin-1 Expression by Orchidectomy in a Nonobese Type 2 Diabetes Animal Model. Mol. Med. (2011.0) 17 4-11. DOI: 10.2119/molmed.2009.00105 190. Oh Y.S., Khil L.-Y., Cho K.A., Ryu S.J., Ha M.K., Cheon G.J., Lee T.S., Yoon J.-W., Jun H.-S., Park S.C.. A potential role for skeletal muscle caveolin-1 as an insulin sensitivity modulator in ageing-dependent non-obese type 2 diabetes: Studies in a new mouse model. Diabetologia (2008.0) 51 1025-1034. DOI: 10.1007/s00125-008-0993-0 191. Elçioğlu K.H., Kabasakal L., Çetinel S., Conturk G., Sezen S.F., Ayanoğlu-Dülger G.. Changes in caveolin-1 expression and vasoreactivity in the aorta and corpus cavernosum of fructose and streptozotocin-induced diabetic rats. Eur. J. Pharmacol. (2010.0) 642 113-120. DOI: 10.1016/j.ejphar.2010.05.049 192. Carrillo-Sepulveda M.A., Matsumoto T.. Phenotypic Modulation of Mesenteric Vascular Smooth Muscle Cells from Type 2 Diabetic Rats is Associated with Decreased Caveolin-1 Expression. Cell. Physiol. Biochem. (2014.0) 34 1497-1506. DOI: 10.1159/000366354 193. Stadion M., Schwerbel K., Graja A., Baumeier C., Rödiger M., Jonas W., Wolfrum C., Staiger H., Fritsche A., Häring H.-U.. Increased Ifi202b/IFI16 expression stimulates adipogenesis in mice and humans. Diabetologia (2018.0) 61 1167-1179. DOI: 10.1007/s00125-018-4571-9 194. Mousa U., Onur H., Utkan G.. Is obesity always a risk factor for all breast cancer patients? c-erbB2 expression is significantly lower in obese patients with early stage breast cancer. Clin. Transl. Oncol. (2012.0) 14 923-930. DOI: 10.1007/s12094-012-0878-z 195. Roh E., Yoo H.J.. The Role of Adipose Tissue Lipolysis in Diet-Induced Obesity: Focus on Vimentin. Diabetes Metab. J. (2021.0) 45 43-45. DOI: 10.4093/dmj.2020.0293 196. Abaj F., Saeedy S.A.G., Mirzaei K.. Mediation role of body fat distribution (FD) on the relationship between CAV1 rs3807992 polymorphism and metabolic syndrome in overweight and obese women. BMC Med. Genom. (2021.0) 14. DOI: 10.1186/s12920-021-01050-6 197. Razani B., Combs T.P., Wang X.B., Frank P.G., Park D.S., Russell R.G., Li M., Tang B., Jelicks L.A., Scherer P.E.. Caveolin-1-deficient Mice Are Lean, Resistant to Diet-induced Obesity, and Show Hypertriglyceridemia with Adipocyte Abnormalities. J. Biol. Chem. (2002.0) 277 8635-8647. DOI: 10.1074/jbc.M110970200 198. Pandey V., Vijayakumar M.V., Ajay A.K., Malvi P., Bhat M.K.. Diet-induced obesity increases melanoma progression: Involvement of Cav-1 and FASN. Int. J. Cancer (2012.0) 130 497-508. DOI: 10.1002/ijc.26048 199. Czikora I., Feher A., Lucas R., Fulton D.J.R., Bagi Z.. Caveolin-1 prevents sustained angiotensin II-induced resistance artery constriction and obesity-induced high blood pressure. Am. J. Physiol. Circ. Physiol. (2015.0) 308 H376-H385. DOI: 10.1152/ajpheart.00649.2014 200. Pecci A., Ma X., Savoia A., Adelstein R.S.. MYH9: Structure, functions and role of non-muscle myosin IIA in human disease. Gene (2018.0) 664 152-167. DOI: 10.1016/j.gene.2018.04.048 201. Holbro T., Beerli R.R., Maurer F., Koziczak M., Barbas C.F., Hynes N.E.. The ErbB2/ErbB3 heterodimer functions as an oncogenic unit: ErbB2 requires ErbB3 to drive breast tumor cell proliferation. Proc. Natl. Acad. Sci. USA (2003.0) 100 8933-8938. DOI: 10.1073/pnas.1537685100 202. Forrest M.P., Hill M.J., Quantock A.J., Martin-Rendon E., Blake D.J.. The emerging roles of TCF4 in disease and development. Trends Mol. Med. (2014.0) 20 322-331. DOI: 10.1016/j.molmed.2014.01.010 203. Battaglia R., Delic S., Herrmann H., Snider N.T.. Vimentin on the move: New developments in cell migration. F1000Research (2018.0) 7 1796. DOI: 10.12688/f1000research.15967.1 204. Berwick D.C., Heaton G.R., Azeggagh S., Harvey K.. LRRK2 Biology from structure to dysfunction: Research progresses, but the themes remain the same. Mol. Neurodegener. (2019.0) 14 49. DOI: 10.1186/s13024-019-0344-2 205. Cao H., Alston L., Ruschman J., Hegele R.A.. Heterozygous. Lipids Health Dis. (2008.0) 7 3. DOI: 10.1186/1476-511X-7-3 206. Jønsson K.L., Laustsen A., Krapp C., Skipper K.A., Thavachelvam K., Hotter D., Egedal J.H., Kjolby M., Mohammadi P., Prabakaran T.. IFI16 is required for DNA sensing in human macrophages by promoting production and function of cGAMP. Nat. Commun. (2017.0) 8 14391. DOI: 10.1038/ncomms14391 207. Yan Y., Shi R., Yu X., Sun C., Zang W., Tian H.. Identification of atrial fibrillation-associated microRNAs in left and right atria of rheumatic mitral valve disease patients. Genes Genet. Syst. (2019.0) 94 23-34. DOI: 10.1266/ggs.17-00043 208. Wang S., Min J., Yu Y., Yin L., Wang Q., Shen H., Yang J., Zhang P., Xiao J., Wang Z.. Differentially expressed miRNAs in circulating exosomes between atrial fibrillation and sinus rhythm. J. Thorac. Dis. (2019.0) 11 4337-4348. DOI: 10.21037/jtd.2019.09.50 209. Yan Y., Song D., Zhang X., Hui G., Wang J.. GEO Data Sets Analysis Identifies COX-2 and Its Related Micro RNAs as Biomarkers for Non-Ischemic Heart Failure. Front. Pharmacol. (2020.0) 11 1155. DOI: 10.3389/fphar.2020.01155 210. Guo X., Wang X., Wang Y., Zhang C., Quan X., Zhang Y., Jia S., Ma W., Fan Y., Wang C.. Variants in the SMARCA4 gene was associated with coronary heart disease susceptibility in Chinese han population. Oncotarget (2017.0) 8 7350-7356. DOI: 10.18632/oncotarget.14387 211. Matsha T.E., Kengne A.P., Hector S., Mbu D.L., Yako Y.Y., Erasmus R.T.. MicroRNA profiling and their pathways in South African individuals with prediabetes and newly diagnosed type 2 diabetes mellitus. Oncotarget (2018.0) 9 30485-30498. DOI: 10.18632/oncotarget.25271 212. Tavano F., Fontana A., Mazza T., Gioffreda D., Biagini T., Palumbo O., Carella M., Andriulli A.. Early-Onset Diabetes as Risk Factor for Pancreatic Cancer: miRNA Expression Profiling in Plasma Uncovers a Role for miR-20b-5p, miR-29a, and miR-18a-5p in Diabetes of Recent Diagnosis. Front. Oncol. (2020.0) 10 1567. DOI: 10.3389/fonc.2020.01567 213. Ding L., Ai D., Wu R., Zhang T., Jing L., Lu J., Zhong L.. Identification of the differential expression of serum microRNA in type 2 diabetes. Biosci. Biotechnol. Biochem. (2016.0) 80 461-465. DOI: 10.1080/09168451.2015.1107460 214. Erlich H.A., Valdes A.M., Julier C., Mirel D., Noble J.A.. Evidence for association of the TCF7 locus with type I diabetes. Genes Immun. (2009.0) 10 S54-S59. DOI: 10.1038/gene.2009.92 215. Binia A., Van Stiphout N., Liang L., Michel S., Bhavsar P.K., Chung K.F., Brightling C.E., Barnes P.J., Kabesch M., Bush A.. A Polymorphism Affecting MYB Binding within the Promoter of the. Hum. Mutat. (2013.0) 34 1131-1139. DOI: 10.1002/humu.22340 216. Goldfine A.B., Crunkhorn S., Costello M., Gami H., Landaker E.J., Niinobe M., Yoshikawa K., Lo D., Warren A., Jimenez-Chillaron J.. Necdin and E2F4 Are Modulated by Rosiglitazone Therapy in Diabetic Human Adipose and Muscle Tissue. Diabetes (2006.0) 55 640-650. DOI: 10.2337/diabetes.55.03.06.db05-1015 217. Rakshit K., Matveyenko A.V.. Induction of Core Circadian Clock Transcription Factor Bmal1 Enhances β-Cell Function and Protects Against Obesity-Induced Glucose Intolerance. Diabetes (2020.0) 70 143-154. DOI: 10.2337/db20-0192 218. Stratigopoulos G., LeDuc C.A., Cremona M.L., Chung W.K., Leibel R.L.. Cut-like Homeobox 1 (CUX1) Regulates Expression of the Fat Mass and Obesity-associated and Retinitis Pigmentosa GTPase Regulator-interacting Protein-1-like (RPGRIP1L) Genes and Coordinates Leptin Receptor Signaling. J. Biol. Chem. (2011.0) 286 2155-2170. DOI: 10.1074/jbc.M110.188482 219. Sedaghat F., Cheraghpour M., Hosseini S.A., Pourvali K., Teimoori-Toolabi L., Mehrtash A., Talaei R., Zand H.. Hypomethylation of. Br. J. Nutr. (2019.0) 122 499-508. DOI: 10.1017/S000711451800212X 220. Patankar J.V., Chandak P.G., Obrowsky S., Pfeifer T., Diwoky C., Uellen A., Sattler W., Stollberger R., Hoefler G., Heinemann A.. Loss of intestinal GATA4 prevents diet-induced obesity and promotes insulin sensitivity in mice. Am. J. Physiol. Endocrinol. Metab. (2011.0) 300 E478-E488. DOI: 10.1152/ajpendo.00457.2010 221. Drareni K., Ballaire R., Barilla S., Mathew M.J., Toubal A., Fan R., Liang N., Chollet C., Huang Z., Kondili M.. GPS2 Deficiency Triggers Maladaptive White Adipose Tissue Expansion in Obesity via HIF1A Activation. Cell Rep. (2018.0) 24 2957-2971.e6. DOI: 10.1016/j.celrep.2018.08.032
--- title: Grape-Seed Proanthocyanidins Modulate Adipose Tissue Adaptations to Obesity in a Photoperiod-Dependent Manner in Fischer 344 Rats authors: - Èlia Navarro-Masip - Marina Colom-Pellicer - Francesca Manocchio - Anna Arola-Arnal - Francisca Isabel Bravo - Begoña Muguerza - Gerard Aragonès journal: Nutrients year: 2023 pmcid: PMC9967183 doi: 10.3390/nu15041037 license: CC BY 4.0 --- # Grape-Seed Proanthocyanidins Modulate Adipose Tissue Adaptations to Obesity in a Photoperiod-Dependent Manner in Fischer 344 Rats ## Abstract Seasonal rhythms drive metabolic adaptations that influence body weight and adiposity. Adipose tissue is a key regulator of energy homeostasis in the organism, and its healthiness is needed to prevent the major consequences of overweight and obesity. In this context, supplementation with proanthocyanidins has been postulated as a potential strategy to prevent the alterations caused by obesity. Moreover, the effects of these (poly)phenols on metabolism are photoperiod dependent. In order to describe the impact of grape-seed proanthocyanidins extract (GSPE) on important markers of adipose tissue functionality under an obesogenic environment, we exposed Fischer 344 rats to three different photoperiods and fed them a cafeteria diet for five weeks. Afterwards, we supplemented them with 25 mg GSPE/kg/day for four weeks. Our results revealed that GSPE supplementation prevented excessive body weight gain under a long photoperiod, which could be explained by increased lipolysis in the adipose tissue. Moreover, cholesterol and non-esterified fatty acids (NEFAs) serum concentrations were restored by GSPE under standard photoperiod. GSPE consumption slightly helped combat the obesity-induced hypertrophy in adipocytes, and adiponectin mRNA levels were upregulated under all photoperiods. Overall, the administration of GSPE helped reduce the impact of obesity in the adipose tissue, depending on the photoperiod at which GSPE was consumed and on the type of adipose depots. ## 1. Introduction Obesity is a public health issue worldwide that has nearly tripled since 1975, and in 2016, $39\%$ of adults were overweight and $13\%$ were obese [1]. The principal cause of obesity is an energy imbalance between food intake and energy expenditure [2,3,4]. The World Health Organization (WHO) and other entities report that obesity is preventable by consuming healthier foods, such as vegetables, legumes, whole grains, and nuts, and increasing physical activity by up to 150 min per week for adults. Moreover, seasonal changes affect the metabolic features that control energy balance, such as food intake, adiposity, or energy metabolism [5,6]. In this sense, humans have a higher biological susceptibility to gaining weight in the summer than in the winter, meaning that adopting unhealthy habits in the summer can contribute to weight gain and obesity. Indeed, children gained more weight in the summer holidays rather than in the Christmas holidays [7,8]. Despite the recommendations, the incidence of obesity is still increasing every year; therefore, the need for the development of new strategies to ameliorate obesity and obesity-related disorders becomes greater. A healthy functionality of the white adipose tissue (WAT) is crucial to prevent metabolic disorders since alterations in WAT metabolism can lead to an inflammatory response and systemic complications [9]. In an obesogenic environment, the healthiness of WAT is seriously threatened, and the general consequences that can arise from this situation depend on WAT location. In fact, the WAT is largely distributed throughout the organism, being classified into subcutaneous WAT and visceral WAT [10]. The subcutaneous WAT (sWAT) has a better adaptation to an obesogenic environment, yet it is capable of better modulating the inflammatory response and preventing lipotoxicity. However, the visceral WAT (vWAT) is less tolerant to a calory-excess situation: its development under obesity has a stronger impact on systemic inflammation, as it produces a considerable amount of pro-inflammatory cytokines and promotes an exacerbated immune response, followed by lipotoxicity [11]. The brown adipose tissue (BAT) appears as an important thermogenic tissue in mammals that strongly contributes to energy metabolism. Its role and characteristics are different from WAT: briefly, brown adipocytes contain high amounts of mitochondria and low amounts of small lipid droplets [12]. Moreover, they count on the presence of Uncoupling Protein 1 (UCP1) in the membrane of their mitochondria, which is a protein capable of dissipating the electron respiratory chain and using the energy to produce heat [12]. Consequently, its adaptation to distinct environments such as obesity or seasonality is different from the WAT and can help the organism face metabolic complications [6]. The consumption of specific bioactive compounds, such as (poly)phenols, is highly associated with health benefits [13]. These molecules have been related to a reduced incidence of metabolic disorders such as obesity, glucose intolerance, or cardiovascular disease [14,15]. In addition, phenolic compounds have shown synchronizing characteristics that can ease the metabolic adaptations of the organism to its environment [16,17]. In fact, an influence of photoperiod on the metabolic effects of grape-seed proanthocyanidin extract (GSPE) was recently demonstrated in healthy and obese animals. More specifically, GSPE modulated hepatic glucose and lipid metabolism in a photoperiod-dependent manner [18]. Moreover, GSPE effects on microbiota were dependent on photoperiod in obese animals [19]. GSPE contributed to the adaptation to new photoperiods by regulating body weight and energy expenditure [20]. However, the seasonal effects of GSPE consumption on adipose tissue adaptations to obesity have not been reported to date. Therefore, the aim of this study was to explore the effects of GSPE seasonal consumption on the functionality of different adipose tissue depots (BAT, visceral, and subcutaneous WAT) in animals fed a cafeteria (CAF) diet. ## 2.1. Grape-Seed Proanthocyanidin Extract GSPE was kindly provided by Les Dérivés Résiniques et Terpéniques (Dax, France). According to the manufacturer, the GSPE composition used in this study contained monomers ($21.3\%$), dimers ($17.4\%$), trimers ($16.3\%$), tetramers ($13.3\%$), and oligomers (5–13 units; $31.7\%$) of proanthocyanidins. The exact phenolic composition of GSPE was determined by HPLC-MS/MS (TOF 6210, Agilent) [21], according to what was described by Quiñones et al. [ 22], and can be found in Supplementary Table S1. ## 2.2. Animal Experimental Procedure Eight-week-old male Fischer 344 rats ($$n = 72$$) were purchased from Charles River Laboratories (Barcelona, Spain) and pair-housed in animal quarters at 22 °C with a light/dark period of 12 h. Animals had an adaptation period of 1 week fed with standard diets (STD) (Panlab, Barcelona, Spain) and tap water ad libitum. After the adaptation period, animals were submitted to three different light schedules for 9 weeks to mimic seasonal day lengths: L12 photoperiod (12 h light–12 h darkness); L18 photoperiod (18 h light–6 h darkness); and L6 photoperiod (6 h light–18 h darkness). Animals in each photoperiod group were further divided into three more groups ($$n = 8$$); one of them was fed an STD diet ad libitum, and the other two were fed cafeteria (CAF) diet ad libitum. The CAF diet consisted of biscuits with pâté, biscuits with cheese, ensaïmada (sweetened pastry), bacon, carrots, and sweetened milk ($20\%$ sucrose w/v) in addition to the standard diet (STD Panlab A04, Panlab, Barcelona, Spain). The CAF diet is a highly palatable diet that is able to induce voluntary hyperphagia. Its composition was $10\%$ protein, $31.9\%$ fat, and $58.1\%$ carbohydrates. Each component of the CAF diet was freshly provided to the animals daily, and they could choose and eat ad libitum. At week 6, the 4-week treatment period started while the animals kept eating the same type of diet. The three STD diet groups received the vehicle (VH) treatment, consisting of an oral dose of water and sweetened milk. Three of the CAF groups (one from each photoperiod) received VH, and the other three received 25 mg/kg/day of GSPE diluted in water and sweetened milk. At the end of the experiment, animals were fasted for 3 h and then sacrificed by live decapitation. Total blood was collected from the neck and then centrifuged (1500× g, 15 min, 4 ºC) to obtain serum. All adipose tissue depots were excised, weighted and immediately frozen into liquid nitrogen, one piece of epididymal white adipose tissue (eWAT) and inguinal white adipose tissue (iWAT) were also kept in formol for histological analysis. Both serum and tissues were stored at −80 °C until further use. A schematic representation of the experimental design can be found in Supplementary Figure S1. The Ethics Review Committee for Animal Experimentation of the University Rovira i Virgili (Tarragona, Spain) and the Generalitat de Catalunya approved all the procedures of the investigation (reference number 9495), which was carried out in accordance with the ethical standards and the Declaration of Helsinki. ## 2.3. Serum Analysis Serum glucose, total cholesterol, and triglycerides (TG) were measured with enzymatic colorimetric kits (QCA, Barcelona, Spain). Serum non-esterified fatty acids (NEFAs) were analyzed with the enzymatic colorimetric HR NEFA series kit (Wako, CA, USA). ## 2.4. Histology of Adipose Tissues Frozen iWAT and eWAT samples were thawed and fixed in $4\%$ formaldehyde. Tissues underwent successive dehydration and paraffin infiltration immersion (Citadel 2000, HistoStar, Thermo Scientific, Madrid, Spain) and the paraffin blocks were cut into 2-μm-thick sections using a microtome (Microm HM 355S, ThermoScientific). The sections were subjected to automated hematoxylin–eosin staining (Varistain Gemini, Shandom, Thermo Scientific). Sections were observed and acquired at ×10 magnification using AxioVision ZeissImaging software (Carl Zeiss Iberia, S.L., Madrid, Spain). The area of adipocytes was measured using the Adiposoft open-source software (CIMA, University of Navarra, Pamplona, Spain). Four fields per sample were measured, and six samples per group were analyzed. The adipocyte area was calculated from the average value of the area of cells in all measured fields for each sample. The total adipocyte volume was calculated using the formula π6×3σ2×d¯+d¯3, where d¯ is the mean diameter and σ is the standard deviation of the diameter. Afterwards, this value was converted to the adipocyte weight using the adipocyte density (0.92 g/mL). In order to obtain the total adipocyte number in each depot, the weights of iWAT and eWAT depots were divided by the adipocyte weight. The frequencies of adipocytes were calculated by dividing all counted cells per sample into two groups according to their area, <3000 μm2 or >3000 μm2; then, the total number of counted adipocytes to calculate the percentage of adipocytes in both categories. ## 2.5. Gene Expression Analysis Total RNA from iWAT, eWAT and BAT was extracted using TRIzol reagent (Thermo Fisher Scientific, Barcelona, Spain) following the manufacturer’s protocol. RNA was quantified in NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE, USA). The integrity of the RNA was evaluated by the RNA integrity number (RIN) trough 2100 Bioanalyzer Instrument (Agilent Technologies). A RIN higher than six was accepted for total RNA samples. cDNA was synthesized using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Barcelona, Spain) in a Multigene ThermalCycler (Labnet, Madrid, Spain). The cDNA was subjected to a quantitative reverse transcriptase polymerase chain reaction amplification using iTaq™Universal SYBR Green Supermix (Bio-Rad, Madrid, Spain) in the 7900HT Fast Real-Time PCR System (Applied Biosystems). The primers used for the different genes are described in Supplementary Table S2 and were obtained from Biomers.net (Ulm, Germany). The relative expression of each gene was calculated according to cyclophilin peptidylprolyl isomerase A (Ppia) mRNA levels and normalized to the levels measured in the corresponding control group. The ∆∆Ct method was used and corrected for primer efficiency [23]. ## 2.6. Statistical Analysis The effects of both CAF diet administration and seasonal GSPE supplementation on biometric, biochemical, and serum variables as well as gene expression were evaluated by a two-way ANOVA with Tukey’s post hoc test for multiple comparisons. In addition, for biometric, biochemical, and serum variables, a one-way ANOVA followed by Tukey’s post hoc test was conducted to detect significant differences between groups in the same photoperiod. GraphPad Prism 9 (GraphPad Software, La Jolla, CA, USA) was used for all statistical analysis. The values are expressed as means ± SEM. $p \leq 0.05$ was considered significant. ## 3.1. GSPE Consumption Restored Cholesterol and NEFAs Serum Concentrations in a Photoperiod-Dependent Manner in Obese Animals After nine weeks of study, animals fed the CAF diet significantly increased their body weight, body weight gain, and food intake (Table 1). However, the four-week GSPE supplementation was able to significantly reduce body weight gain and food intake in all groups. When we conducted one-way ANOVA tests to assess the individual effects of GSPE on each photoperiod, we detected that the greater effects on reducing body weight gain were in animals exposed to L18. Furthermore, the CAF diet significantly increased total fat mass and the mass of individual adipose tissue depots (iWAT, eWAT, and BAT) in all animals. In this case, GSPE supplementation did not significantly affect these values. A diet effect was also observed in all circulating metabolic parameters, where CAF consumption significantly increased serum concentrations of glucose, cholesterol, TG, and NEFAs. However, GSPE supplementation was able to restore cholesterol and NEFAs levels in a photoperiod-dependent manner. ## 3.2. GSPE Consumption Restored Adipocyte Morphology in iWAT in a Photoperiod-Independent Manner The histological analysis of iWAT showed a significant effect of the CAF diet on both area and volume of adipocytes in iWAT, which were increased in obese animals (Figure 1A,B, Supplementary Figure S2). Consequently, the adipocyte area distribution also changed, and the number of larger adipocytes increased with respect to the smaller ones (Figure 1D). However, no differences were detected in adipocyte number (Figure 1C), and, when we applied one-way ANOVA analyses to assess the significant effects within each photoperiod, we only observed statistical significance on L12 animals. Noteworthy, as indicated by two-way ANOVA analyses, GSPE consumption reduced the CAF diet impact because the cellular profile of CAF-fed animals supplemented with GSPE was more similar to the STD diet fed groups, showing reduced adipocyte area and volume. No interaction between GSPE and photoperiod was detected, but the GSPE effect only reached significance in L12 groups after applying one-way ANOVA tests. ## 3.3. Gene Expression of iWAT Was Affected by GSPE in a Photoperiod-Dependent Manner However, after exploring the expression of the key metabolic genes in iWAT, we detected an interaction effect between GSPE and photoperiod in the expression levels of the adipogenic genes Cebpα and Pparγ (Figure 2A). The effect was only observed in animals exposed to the L6 photoperiod, in which the expression of these genes was significantly restored in response to GSPE consumption. A similar effect was detected in the lipid transport-related genes Cd36 and Fabp4 (Figure 2B), the lipolysis-related genes Hsl and Atgl (Figure 2C), and the adipokine genes Lep and Adipoq (Figure 2D). Differently, no significant effects in response to GSPE consumption were detected in the gene expression of the lipogenic genes Acacα and Fasn (Supplementary Figure S3). ## 3.4. GSPE Consumption Restored Adipocyte Morphology in eWAT in a Photoperiod-Independent Manner Similar to iWAT, a significant effect of both the CAF diet and GSPE consumption was detected in the histological analysis of eWAT. Again, the CAF diet increased adipocyte area and volume, whereas GSPE consumption attenuated this significant increase regardless of the photoperiod (Figure 3A,B, Supplementary Figure S4). However, these changes in response to proanthocyanidins were not fully reflected by adipocyte number or adipocyte area frequency (Figure 3C,D), and one-way ANOVA tests did not show significancy between CAF-VH and CAF-GSPE groups on area and volume values under any photoperiod. ## 3.5. GSPE Consumption Influenced the Expression of Adipogenic Genes in eWAT in a Photoperiod-Dependent Manner As observed in iWAT, when we explored the gene expression profile of metabolic genes in eWAT, an interaction between GSPE and photoperiod was significantly detected in adipogenic genes (Figure 4A). Moreover, GSPE consumption significantly upregulated the gene expression of Fabp4 (in L18 and L6 animals) and of Cd36 (Figure 4B). Similarly, an interaction effect between GSPE consumption and photoperiod was observed in the lipolytic gene Hsl (Figure 4C). As expected, the CAF diet significantly downregulated the expression of the lipogenic genes Acacα and Fasn in all animal groups (Supplementary Figure S5). Interestingly, an interaction effect between GSPE and photoperiod was also detected in the expression of Lep, which was reduced in response to GSPE consumption in the L12 and L6 groups but significantly upregulated in the L18 animals (Figure 4D). In addition, Adipoq gene expression was reduced by the CAF diet in all groups, but a tendency for GSPE to reverse this decrease was observed in animals subjected to the L12 and L18 photoperiods. Finally, a significant effect of the CAF diet on the expression of the inflammatory genes Il6 and Tnfα was also observed (Supplementary Figure S5). ## 3.6. GSPE Consumption Reverted the Obesity-Induced Downregulation of Pgc1α in BAT in a Photoperiod-Dependent Manner Being BAT such an important tissue in energy homeostasis, especially in rodents, we analyzed the most important components that drive the main metabolic pathways of brown adipocytes’ activity. Despite the fact that no significant effects of CAF diet or GSPE administration on the gene expression of Ucp1 were detected (Supplementary Figure S6), an interaction effect between GSPE and photoperiod was observed in the gene expression of Pgc1α, Cpt1β, and Dio2 (Figure 5A–C). Particularly, GSPE supplementation subtly upregulated the expression of these genes under L12 and L6 photoperiods, while it downregulated its expression under L18. Moreover, the CAF diet downregulated the expression of Pgc1α in all photoperiods and the expression of Dio2 in the L18 and L6 groups. Finally, mRNA levels of Cpt1β were influenced by the CAF diet in a photoperiod-dependent manner. Moreover, Cd36, Lpl, Pparγ, and Prdm16 gene expression was significantly downregulated by the CAF diet in all animal groups (Supplementary Figure S6). However, an interaction effect between GSPE and photoperiod was observed in the gene expression of Pparα, which tended to be upregulated in the L12 and L6 groups but not in the L18 animals (Figure 5D). ## 4. Discussion The beneficial effects of (poly)phenols on health and their natural presence in vegetable foods make these molecules of high interest in research. Additionally, a (poly)phenol-rich diet or the development of natural-based extracts as an alternative for preventing or treating metabolic disorders is more likely to be accepted by society than synthetic options [24]. In this context, (poly)phenol consumption can help reduce the negative effects of a disrupted metabolism in the adipose tissue and, consequently, in the whole organism [25]. Moreover, changes in the atmosphere can drive metabolic adaptations in mammals [26], meaning that seasonal rhythms can impact the effects of (poly)phenol consumption in adipose tissue. However, these effects have not been reported yet, even if evidence exists about the impact of seasonal consumption of (poly)phenols in other metabolic tissues such as the liver or the gut microbiota [18,19]. Our study aimed to elucidate the effects of proanthocyanidin seasonal consumption on the adaptations of the adipose tissue to an obesogenic environment. For this purpose, Fischer 344 rats were exposed to different photoperiod schedules in order to mimic the seasons of the year. Then, animals were fed a CAF diet and supplemented with either VH or 25 mg GSPE/ kg/day for four weeks. The dose of GSPE was chosen based on previous studies that showed metabolic effects with the same dose [27]. After nine weeks of study, body weight gain was reduced in response to proanthocyanidin consumption in CAF-fed animals. These results were particularly marked when GSPE was consumed at L18 than when it was at L12 and L6 photoperiods. According to the seasonal adaptations to photoperiod, increased weight gain in long photoperiods appears as a natural metabolic modulation in mammals, which, contrarily, lose more weight in short photoperiods [28]. Nevertheless, GSPE consumption resulted in a significant reduction in food intake only in animals subjected to L12 and L6 photoperiods but not in L18 animals. The effects of proanthocyanidins on reducing food intake were previously reported in Wistar rats [29] and were attributed to the astringency of these (poly)phenols [30]. Other mechanisms that could explain the GSPE-derived inhibition of food intake involve the central nervous system. On the one hand, enhanced leptin signaling was detected in obese rats supplemented with GSPE [31]. Furthermore, proanthocyanidins were shown to induce glucagon-like peptide (GLP-1) secretion, which is capable of increasing satiety and inhibiting feeding [32]. Nevertheless, considering the strong seasonal impact of long photoperiods on feeding behavior, which remains unaltered even with increased leptin levels, reflecting reduced leptin sensitivity under summer-like conditions, it would be possible that the GSPE lowering effects on food intake were not adopted by animals under the L18 photoperiod [33,34,35]. Similarly, even if previous studies observed increased energy expenditure in animals supplemented with GSPE under standard photoperiods [29,36], a previous study from our group showed that the energy expenditure of CAF-fed rats exposed to L18 and supplemented with GSPE for one week tended to be reduced [20], which suggests that the observed decrease in body weight gain in L18-CAF animals supplemented with proanthocyanidins in our study should be explained by other mechanisms. For instance, it was recently reported that obese rats exposed to L18 showed greater changes in the gut microbiota composition in response to proanthocyanidin consumption compared to animals exposed to L12 and L6 photoperiods. Particularly, GSPE consumption under L18 resulted in increased levels of Bifidobacterium and Coprobacillus, while decreased Klebsiella to levels associated with reduced body weight gain and obesity [19]. Moreover, other studies indicated that proanthocyanidins can modulate nutrient absorption in the gastrointestinal tract, through inhibiting digestive enzymes such as α-Amylase and α-Glucosidase, resulting in a $20\%$ decrease in the amount of absorbed energy [29,37]. In future studies, it would be interesting to calculate energy/food efficiency in these animals in order to understand the proanthocyanidin-induced mechanism driving the reduction in body weight gain under the L18 photoperiod. Given that WAT functionality responds to changes in photoperiod and that obesity drives major changes to the structure and activity of this tissue, histological analyses of WAT depots were performed. As expected, we observed a strong impact of the CAF diet on the area and volume of adipocytes, which were larger compared to animals fed a chow diet. These changes are common in obesity and contribute to the altered functionality of the WAT. In fact, under an obesogenic environment, the WAT needs to store the excess calories, which can be done through two processes: hyperplasia or hypertrophy [38]. The hyperplasia expansion is considered the healthy expansion, where the number of adipocytes increases via adipogenesis and they maintain their functionality and insulin sensitivity [9,38]. However, when hypertrophy occurs, existing adipocytes need to incorporate the excess of fat, increasing their volume and disrupting their functionality [9]. These dysfunctional adipocytes are less glucose tolerant and become pro-inflammatory, producing cytokines such as IL-6 or TNFα, which can induce macrophage infiltration in the adipose tissue and provoke a systemic immune response [39]. Adipose tissue expansion partly depends on the adipose depot; indeed, under healthy conditions, hyperplasia is more commonly developed in subcutaneous depots, while visceral depots are more susceptible to hypertrophy [40]. Additionally, it has been studied that hypertrophic visceral adipose tissue has a stronger negative impact on the whole metabolism of the organism, being highly associated with the metabolic syndrome [39]. Therefore, it is important to maintain the integrity and healthiness of this adipose depot. Furthermore, in our study, proanthocyanidin consumption reversed the effect induced by the CAF diet in iWAT regardless of the photoperiod, as adipocyte areas and volumes were at similar levels to their healthy counterparts in all photoperiod conditions. Similar results have been reported with GSPE supplementation in iWAT [25]. Further, the administration of blueberry polyphenol extract resulted in a reduced adipocyte area in the eWAT [41]. Nonetheless, in our study, this effect was less clear in eWAT, where a significant effect of GSPE was observed but the individual differences within groups in each photoperiod were not significant. In addition, when we studied the gene expression of inflammatory markers in eWAT (Il6 and Tnfα), no effects were observed in response to GSPE consumption. One of the main roles of WAT is to regulate energy balance in the organism. Therefore, lipid turnover is constantly active, involving distinct metabolic pathways that control fat storage and release from the WAT. The process that drives the formation of new adipocytes is named adipogenesis and is mainly driven by the master adipogenic molecules PPARγ and C/EBPs [42]. In our study, gene expression of Pparγ and Cebpα in both iWAT and eWAT showed an interaction effect between GSPE administration and photoperiod. In iWAT, we observed that GSPE increased its gene expression only in the L12 and L18 groups. Similar effects were previously observed in response to cyanidin-3-glucoside administration [43] and could explain the differences observed in the histological analyses, where the presence of larger adipocytes provoked by the CAF diet was partly reversed in response to GSPE consumption in L12 and L18 animals. Hence, increased adipogenesis would favor hyperplasia expansion rather than hypertrophy. Furthermore, the lipid transport-related genes (Fabp4 and Cd36) were also affected by GSPE consumption in iWAT (L12 and L18 groups) and in eWAT (all groups), which could suggest a beneficial role for proanthocyanidins in helping release circulating TG. In fact, animals consuming GSPE at L12 and L18 photoperiods showed reduced serum TG concentrations compared to VH groups. However, another interaction effect between GSPE and photoperiod on the expression of lipolysis-related genes was detected in both iWAT and eWAT, which were upregulated in response to GSPE consumption only in the L12 and L18 groups. This activation of lipolysis could be involved in reducing the size of existing adipocytes, contributing to β-oxidation and global energy homeostasis. Moreover, the reduction in body weight gain observed in L18 animals could be partly explained by the increased lipolysis detected in these animals. In fact, supplementation with blueberry (poly)phenols suggested increased lipolysis in the WAT [41], and the lipolytic effects of other types of (poly)phenols such as epigallocatechin-gallate (EGCG) or resveratrol are already established [44]. Following these interaction effects between GSPE consumption and photoperiod, we observed that when proanthocyanidins were consumed at L6 photoperiod, the obesity-driven changes on the gene expression of iWAT were reverted. Thus, while the CAF diet induced the mRNA expression of genes related to adipogenesis (Pparγ, Cebpα), lipid transport (Fabp4, Cd36), lipolysis (Atgl, Hsl), and adipokines (Lep, Adipoq), GSPE was capable of returning the expression of these genes to a healthy situation. In this context, (poly)phenol-induced reduction of lipid turnover has been reported previously in human and animal studies [45,46,47], which partly agrees with our results in the L6 group. However, a significant effect was also observed only under L18, where GSPE consumption upregulated the expression of Adipoq. High adiponectin serum concentrations are associated with anti-inflammatory effects that help prevent the metabolic consequences of an obesogenic environment [48]. In this context, a previous study reported that GSPE administration lowered inflammatory levels and increased adiponectin mRNA levels in the mesenteric adipose tissue of obese rats, which agrees with our findings [49]. Importantly, in our study, we could not observe the same effects in animals exposed to L12 and L6, which reinforce the photoperiod-dependent effects of proanthocyanidins on the modulation of gene expression in WAT. Importantly, we also detected a browning effect in iWAT only in animals consuming proanthocyanidins under short photoperiods. In fact, GSPE was associated with increased thermogenesis in the adipose tissue of obese mice [36], but these results were obtained after supplementing with noticeable higher doses of GSPE. Therefore, in our study, it is possible that the dose of GSPE was too low to fully induce browning in the iWAT. In fact, proanthocyanidins did not seem to promote enhanced BAT activity in our study, as Ucp1 gene expression was unchanged after both the CAF diet and (poly)phenols consumption. However, other molecules involved in BAT thermogenesis were affected by GSPE in a photoperiod-dependent manner. Thus, proanthocyanidins upregulated the expression of Pgc1α, Cpt1β and Pparα in animals exposed to L12 and L6, while these genes were downregulated in L18 animals. Our results agree with previous findings [27], where GSPE reversed the obesity-induced mitochondrial dysfunction in BAT by increasing Pgc1α expression. However, in our study, we demonstrated that these effects are limited to L6 and L12 photoperiods. In conclusion, our results indicate that the beneficial effects associated with GSPE consumption on diet-induced obese animals are strongly influenced by the photoperiod of exposure in a tissue-specific manner. Particularly, our findings suggest that L18 photoperiod is highly sensitive to GSPE with respect to body weight adaptation and modulation of metabolic genes, especially adiponectin, in the iWAT. Meanwhile, animals under the L12 and L6 photoperiods showed reduced food intake in response to GSPE and enhanced BAT activity. Therefore, our study reinforces the relevance of considering seasonal rhythms when investigating the metabolic effects of (poly)phenols in the organism in order to properly potentiate the metabolic response of the adipose tissue. ## References 1. Abarca-Gómez L., Abdeen Z.A., Hamid Z.A., Abu-Rmeileh N.M., Acosta-Cazares B., Acuin C., Adams R.J., Aekplakorn W., Afsana K., Aguilar-Salinas C.A.. Worldwide Trends in Body-Mass Index, Underweight, Overweight, and Obesity from 1975 to 2016: A Pooled Analysis of 2416 Population-Based Measurement Studies in 128.9 Million Children, Adolescents, and Adults. Lancet (2017) 390 2627-2642. DOI: 10.1016/S0140-6736(17)32129-3 2. Global Health Observatory Data. Overweight and Obesity (2016) 3. Nicolaidis S.. Environment and Obesity. Metabolism (2019) 100 153942. DOI: 10.1016/j.metabol.2019.07.006 4. Rohde K., Keller M., la Cour Poulsen L., Blüher M., Kovacs P., Böttcher Y.. Genetics and Epigenetics in Obesity. Metabolism (2019) 92 37-50. DOI: 10.1016/j.metabol.2018.10.007 5. Yoshimura E., Tajiri E., Hatamoto Y., Tanaka S.. Changes in Season Affect Body Weight, Physical Activity, Food Intake, and Sleep in Female College Students: A Preliminary Study. Int. J. Environ. Res. Public Health (2020) 17. DOI: 10.3390/ijerph17238713 6. Au-Yong I.T.H., Thorn N., Ganatra R., Perkins A.C., Symonds M.E.. Brown Adipose Tissue and Seasonal Variation in Humans. Diabetes (2009) 58 2583-2587. DOI: 10.2337/db09-0833 7. Turicchi J., O’Driscoll R., Horgan G., Duarte C., Palmeira A.L., Larsen S.C., Heitmann B.L., Stubbs J.. Weekly, Seasonal and Holiday Body Weight Fluctuation Patterns among Individuals Engaged in a European Multi-Centre Behavioural Weight Loss Maintenance Intervention. PLoS ONE (2020) 15. DOI: 10.1371/journal.pone.0232152 8. Yokoya M., Higuchi Y.. Day Length May Make Geographical Difference in Body Size and Proportions: An Ecological Analysis of Japanese Children and Adolescents. PLoS ONE (2019) 14. DOI: 10.1371/journal.pone.0210265 9. Longo M., Zatterale F., Naderi J., Parrillo L., Formisano P., Raciti G.A., Beguinot F., Miele C.. Adipose Tissue Dysfunction as Determinant of Obesity-Associated Metabolic Complications. Int. J. Mol. Sci. (2019) 20. DOI: 10.3390/ijms20092358 10. Frigolet M.E., Gutiérrez-Aguilar R.. The Colors of Adipose Tissue. Gac. Med. Mex. (2020) 156 142-149. DOI: 10.24875/GMM.M20000356 11. Pellegrinelli V., Carobbio S., Vidal-Puig A.. Adipose Tissue Plasticity: How Fat Depots Respond Differently to Pathophysiological Cues. Diabetologia (2016) 59 1075-1088. DOI: 10.1007/s00125-016-3933-4 12. Bargut T.C.L., Aguila M.B., Mandarim-de-Lacerda C.A.. Brown Adipose Tissue: Updates in Cellular and Molecular Biology. Tissue Cell (2016) 48 452-460. DOI: 10.1016/j.tice.2016.08.001 13. Cardona F., Andrés-Lacueva C., Tulipani S., Tinahones F.J., Queipo-Ortuño M.I.. Benefits of Polyphenols on Gut Microbiota and Implications in Human Health. J. Nutr. Biochem. (2013) 24 1415-1422. DOI: 10.1016/j.jnutbio.2013.05.001 14. Mezhibovsky E., Knowles K.A., He Q., Sui K., Tveter K.M., Duran R.M., Roopchand D.E.. Grape Polyphenols Attenuate Diet-Induced Obesity and Hepatic Steatosis in Mice in Association With Reduced Butyrate and Increased Markers of Intestinal Carbohydrate Oxidation. Front. Nutr. (2021) 8 675267. DOI: 10.3389/fnut.2021.675267 15. Tresserra-Rimbau A., Medina-Remón A., Pérez-Jiménez J., Martínez-González M.A., Covas M.I., Corella D., Salas-Salvadó J., Gómez-Gracia E., Lapetra J., Arós F.. Dietary Intake and Major Food Sources of Polyphenols in a Spanish Population at High Cardiovascular Risk: The PREDIMED Study. Nutr. Metab. Cardiovasc. Dis. (2013) 23 953-959. DOI: 10.1016/j.numecd.2012.10.008 16. Ávila-Román J., Soliz-Rueda J.R., Bravo F.I., Aragonès G., Suárez M., Arola-Arnal A., Mulero M., Salvadó M.J., Arola L., Torres-Fuentes C.. Phenolic Compounds and Biological Rhythms: Who Takes the Lead?. Trends Food Sci. Technol. (2021) 113 77-85. DOI: 10.1016/j.tifs.2021.04.050 17. Arola-Arnal A., Cruz-Carrión Á., Torres-Fuentes C., Ávila-Román J., Aragonès G., Mulero M., Bravo F.I., Muguerza B., Arola L., Suárez M.. Chrononutrition and Polyphenols: Roles and Diseases. Nutrients (2019) 11. DOI: 10.3390/nu11112602 18. Rodríguez R.M., Colom-Pellicer M., Blanco J., Calvo E., Aragonès G., Mulero M.. Grape-Seed Procyanidin Extract (GSPE) Seasonal-Dependent Modulation of Glucose and Lipid Metabolism in the Liver of Healthy F344 Rats. Biomolecules (2022) 12. DOI: 10.3390/biom12060839 19. Arreaza-Gil V., Escobar-Martínez I., Muguerza B., Aragonès G., Suárez M., Torres-Fuentes C., Arola-Arnal A.. The Effects of Grape Seed Proanthocyanidins in Cafeteria Diet-Induced Obese Fischer 344 Rats Are Influenced by Faecal Microbiota in a Photoperiod Dependent Manner. Food Funct. (2022) 13 8363-8374. DOI: 10.1039/D2FO01206E 20. Soliz-Rueda J.R., López-Fernández-sobrino R., Bravo F.I., Aragonès G., Suarez M., Muguerza B.. Grape Seed Proanthocyanidins Mitigate the Disturbances Caused by an Abrupt Photoperiod Change in Healthy and Obese Rats. Nutrients (2022) 14. DOI: 10.3390/nu14091834 21. Serra A., MacI A., Romero M.P., Valls J., Bladé C., Arola L., Motilva M.J.. Bioavailability of Procyanidin Dimers and Trimers and Matrix Food Effects in in Vitro and in Vivo Models. Br. J. Nutr. (2010) 103 944-952. DOI: 10.1017/S0007114509992741 22. Quiñones M., Guerrero L., Suarez M., Pons Z., Aleixandre A., Arola L., Muguerza B.. Low-Molecular Procyanidin Rich Grape Seed Extract Exerts Antihypertensive Effect in Males Spontaneously Hypertensive Rats. Food Res. Int. (2013) 51 587-595. DOI: 10.1016/j.foodres.2013.01.023 23. Pfaffl M.W., Dorak R.. Relative Quantification. Real-Time PCR (2004) 64-82 24. Sachdeva V., Roy A., Bharadvaja N.. Current Prospects of Nutraceuticals: A Review. Curr. Pharm. Biotechnol. (2020) 21 884-896. DOI: 10.2174/1389201021666200130113441 25. Colom-Pellicer M., Rodríguez R.M., Navarro-Masip È., Bravo F.I., Mulero M., Arola L., Aragonès G.. Time-of-Day Dependent Effect of Proanthocyanidins on Adipose Tissue Metabolism in Rats with Diet-Induced Obesity. Int. J. Obes. (2022) 46 1394-1402. DOI: 10.1038/s41366-022-01132-0 26. Helfer G., Barrett P., Morgan P.J.. A Unifying Hypothesis for Control of Body Weight and Reproduction in Seasonally Breeding Mammals. J. Neuroendocrinol. (2019) 31 e12680. DOI: 10.1111/jne.12680 27. Pajuelo D., Quesada H., Díaz S., Fernández-Iglesias A., Arola-Arnal A., Bladé C., Salvadó J., Arola L.. Chronic Dietary Supplementation of Proanthocyanidins Corrects the Mitochondrial Dysfunction of Brown Adipose Tissue Caused by Diet-Induced Obesity in Wistar Rats. Br. J. Nutr. (2012) 107 170-178. DOI: 10.1017/S0007114511002728 28. Ebling F.J.P.. On the Value of Seasonal Mammals for Identifying Mechanisms Underlying the Control of Food Intake and Body Weight. Horm. Behav. (2014) 66 56-65. DOI: 10.1016/j.yhbeh.2014.03.009 29. Serrano J., Casanova-Martí À., Gual A., Pérez-Vendrell A.M., Blay M.T., Terra X., Ardévol A., Pinent M.. A Specific Dose of Grape Seed-Derived Proanthocyanidins to Inhibit Body Weight Gain Limits Food Intake and Increases Energy Expenditure in Rats. Eur. J. Nutr. (2017) 56 1629-1636. DOI: 10.1007/s00394-016-1209-x 30. Soares S., Brandão E., Mateus N., de Freitas V.. Sensorial Properties of Red Wine Polyphenols: Astringency and Bitterness. Crit. Rev. Food Sci. Nutr. (2017) 57 937-948. DOI: 10.1080/10408398.2014.946468 31. Ibars M., Ardid-Ruiz A., Suárez M., Muguerza B., Bladé C., Aragonès G.. Proanthocyanidins Potentiate Hypothalamic Leptin/STAT3 Signalling and Pomc Gene Expression in Rats with Diet-Induced Obesity. Int. J. Obes. (2017) 41 129-136. DOI: 10.1038/ijo.2016.169 32. Serrano J., Casanova-Martí À., Blay M., Terra X., Ardévol A., Pinent M.. Defining Conditions for Optimal Inhibition of Food Intake in Rats by a Grape-Seed Derived Proanthocyanidin Extract. Nutrients (2016) 8. DOI: 10.3390/nu8100652 33. Togo Y., Otsuka T., Goto M., Furuse M., Yasuo S.. Photoperiod Regulates Dietary Preferences and Energy Metabolism in Young Developing Fischer 344 Rats but Not in Same-Age Wistar Rats. Am. J. Physiol. Endocrinol. Metab. (2012) 303 777-786. DOI: 10.1152/ajpendo.00209.2012 34. Ibars M., Aragonès G., Ardid-Ruiz A., Gibert-Ramos A., Arola-Arnal A., Suárez M., Bladé C.. Seasonal Consumption of Polyphenol-Rich Fruits Affects the Hypothalamic Leptin Signaling System in a Photoperiod-Dependent Mode. Sci. Rep. (2018) 8 13572. DOI: 10.1038/s41598-018-31855-y 35. Mariné-Casadó R., Domenech-Coca C., del Bas J.M., Bladé C., Arola L., Caimari A.. The Exposure to Different Photoperiods Strongly Modulates the Glucose and Lipid Metabolisms of Normoweight Fischer 344 Rats. Front. Physiol. (2018) 9 416. DOI: 10.3389/fphys.2018.00416 36. Du H., Wang Q., Li T., Ren D., Yang X.. Grape Seed Proanthocyanidins Reduced the Overweight of C57BL/6J Mice through Modulating Adipose Thermogenesis and Gut Microbiota. Food Funct. (2021) 12 8467-8477. DOI: 10.1039/D1FO01361K 37. Yilmazer-Musa M., Griffith A.M., Michels A.J., Schneider E., Frei B.. Grape Seed and Tea Extracts and Catechin 3-Gallates Are Potent Inhibitors of α-Amylase and α-Glucosidase Activity. J. Agric. Food Chem. (2012) 60 8924-8929. DOI: 10.1021/jf301147n 38. Ursula W., Eric R.. Dynamics of Adipose Tissue Turnover in Human Metabolic Health and Disease HHS Public Access. Physiol. Behav. (2019) 176 139-148. DOI: 10.1007/s00125-018-4732-x.Dynamics 39. Despres J.P., Lemieux I.. Abdominal Obesity and Metabolic Syndrome. Nature (2006) 444 881-887. DOI: 10.1038/nature05488 40. Drolet R., Richard C., Sniderman A.D., Mailloux J., Fortier M., Huot C., Rhéaume C., Tchernof A.. Hypertrophy and Hyperplasia of Abdominal Adipose Tissues in Women. Int. J. Obes. (2008) 32 283-291. DOI: 10.1038/sj.ijo.0803708 41. Jiao X., Wang Y., Lin Y., Lang Y., Li E., Zhang X., Zhang Q., Feng Y., Meng X., Li B.. Blueberry Polyphenols Extract as a Potential Prebiotic with Anti-Obesity Effects on C57BL/6 J Mice by Modulating the Gut Microbiota. J. Nutr. Biochem. (2019) 64 88-100. DOI: 10.1016/j.jnutbio.2018.07.008 42. Ali A.T., Hochfeld W.E., Myburgh R., Pepper M.S.. Adipocyte and Adipogenesis. Eur. J. Cell Biol. (2013) 92 229-236. DOI: 10.1016/j.ejcb.2013.06.001 43. Matsukawa T., Inaguma T., Han J., Villareal M.O., Isoda H.. Cyanidin-3-Glucoside Derived from Black Soybeans Ameliorate Type 2 Diabetes through the Induction of Differentiation of Preadipocytes into Smaller and Insulin-Sensitive Adipocytes. J. Nutr. Biochem. (2015) 26 860-867. DOI: 10.1016/j.jnutbio.2015.03.006 44. Wang S., Moustaid-Moussa N., Chen L., Mo H., Shastri A., Su R., Bapat P., Kwun I.S., Shen C.L.. Novel Insights of Dietary Polyphenols and Obesity. J. Nutr. Biochem. (2014) 25 1-18. DOI: 10.1016/j.jnutbio.2013.09.001 45. Khalilpourfarshbafi M., Gholami K., Murugan D.D., Abdul Sattar M.Z., Abdullah N.A.. Differential Effects of Dietary Flavonoids on Adipogenesis. Eur. J. Nutr. (2019) 58 5-25. DOI: 10.1007/s00394-018-1663-8 46. Most J., Warnke I., Boekschoten M.v., Jocken J.W.E., de Groot P., Friedel A., Bendik I., Goossens G.H., Blaak E.E.. The Effects of Polyphenol Supplementation on Adipose Tissue Morphology and Gene Expression in Overweight and Obese Humans. Adipocyte (2018) 7 190-196. DOI: 10.1080/21623945.2018.1469942 47. de Lima L.P., de Paula Barbosa A.. A Review of the Lipolytic Effects and the Reduction of Abdominal Fat from Bioactive Compounds and Moro Orange Extracts. Heliyon (2021) 7 e07695. DOI: 10.1016/j.heliyon.2021.e07695 48. Fang H., Judd R.L.. Adiponectin Regulation and Function. Compr. Physiol. (2018) 8 1031-1063. DOI: 10.1002/cphy.c170046 49. Terra X., Montagut G., Bustos M., Llopiz N., Ardèvol A., Bladé C., Fernández-Larrea J., Pujadas G., Salvadó J., Arola L.. Grape-Seed Procyanidins Prevent Low-Grade Inflammation by Modulating Cytokine Expression in Rats Fed a High-Fat Diet. J. Nutr. Biochem. (2009) 20 210-218. DOI: 10.1016/j.jnutbio.2008.02.005
--- title: Tooth Loss and Caries Experience of Elderly Chileans in the Context of the COVID-19 Pandemic in Five Regions of Chile authors: - Víctor Beltrán - Marco Flores - Cristina Sanzana - Fernanda Muñoz-Sepúlveda - Eloy Alvarado - Bernardo Venegas - Juan Carlos Molina - Sandra Rueda-Velásquez - Alfredo von Marttens journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9967189 doi: 10.3390/ijerph20043001 license: CC BY 4.0 --- # Tooth Loss and Caries Experience of Elderly Chileans in the Context of the COVID-19 Pandemic in Five Regions of Chile ## Abstract Risk factors associated with tooth loss have been studied; however, the current status of the epidemiological profiles and the impact of the pandemic on the oral health of the elderly is still unknown. This study aims to determine the experience of caries and tooth loss among elderly Chilean citizens in five regions and to identify the risk factors associated with tooth loss. The sample includes 135 participants over 60 years old assessed during COVID-19 lockdown. Sociodemographic variables such as education and RSH (Social Registry of Households) were obtained through a teledentistry platform called TEGO. The history of chronic diseases such as diabetes, obesity, depression and dental caries reported by DMFT index scores were incorporated. The statistical analysis included Adjusted Odds Ratios (ORs) to assess risk factors associated with the lack of functional dentition. Multivariate hypothesis testing was used to compare the mean equality of DMFT and its components between regions (p-value < 0.05). Individuals with RSH ≤ $40\%$ were at higher risk of having no functional dentition with OR 4.56 ($95\%$ CI: 1.71, 12.17). The only mean difference between regions was the filled tooth component. Tooth loss was associated with multidimensional lower income, where the elderly belonging to the $40\%$ most vulnerable population had a higher prevalence of non-functional dentition. This study highlights the importance of implementing a National Oral Health Policy that focuses on oral health promotion and minimally invasive dentistry for the most vulnerable population. ## 1. Introduction The world population is experiencing an ageing process. The World Health Organization (WHO) declared that between 2015 and 2050, the percentage of the planet’s inhabitants over 60 years of age will almost double from $12\%$ to $22\%$ [1]. Chile will no longer be the exception. According to the projections of the Chilean Ministry of Health (MINSAL), by 2025, Chile will be the oldest country in the region [2]. This increase in the population’s life expectancy constitutes a great challenge for the health profession, given the high burden of oral disease and its relation to general health. Tooth loss reflects the endpoint of a lifetime of dental disease and the individual’s history or absence of dental treatment [3]. To prevent the consequences of edentulism the WHO, the World Dental Federation (FDI), and the International Association for Dental Research (IADR) urged countries on the importance of reducing the number of elderly edentulous people in ages ranging from 65 to 74 years. This effort has permitted an increase in the number of natural teeth present in the oral cavity and the number of individuals with functional dentitions [4], which has meant a significant decline in the prevalence and incidence of total tooth loss worldwide [5]. In Chile, more than half of the elderly population has 11 or less teeth, and the number of senior citizens over 65 years with no functional dentition remains elevated; in fact, $81.7\%$ have less than 20 natural teeth [6]. One way of reflecting the success of the public policies taken to diminish the number of elderly people experiencing tooth loss and caries is assessing the DMFT index [7], which summarizes the number of decayed teeth (DT), missing teeth(MT) and filled teeth (FT) to reflect the actual disease experience (past and present), and, if evaluated separately, indicates the management of the disease [8]. Maintaining optimal oral health by preserving a natural, healthy and functional dentition contributes to the survival of older adults [9], since missing unreplaced teeth has been associated with an increase in the risk of malnutrition, frailty and cardiovascular mortality [10]. Likewise, an association with the decline in cognitive function, dementia, obesity, diabetes and hypertension has been reported [11,12,13,14,15]. In this sense, the WHO’s Oral Health Program has encouraged national planners to strengthen the implementation of systematic programs aimed at improving healthy aging, oral health and a better quality of life for the elderly [16]. However, to this day, the neglect of oral health constitutes a failure of global health policies in delivering the basic human rights of older people [17]. In the context of the pandemic, due to the restrictions to prevent contagion by COVID-19 and the serious consequences this disease can bring to the elderly population in particular, measures such as social distancing and limitations on the number of face-to-face visits to access routine medical and dental check-ups were taken [18,19,20], so it is expected that the deficient general and oral health in these people would have become more severe than before the pandemic [21]. To provide a response to dental emergencies and priority treatments in the context of the COVID-19 pandemic, a technology based on teledentistry concepts called Tele-platform of Geriatric and Dental Specialties (TEGO) was developed. The pilot test of this web platform showed that a semi-presential teledentistry workflow can help elderly people who are impeded to look for traditional dental assistance during a pandemic, and made it possible to obtain a database with relevant information from the medical-dental examination carried out on the elderly population in five regions of Chile [22]. Former studies and National surveys related age, educational level, low income, depression and oral health inequalities as the main risk factors of dental caries and tooth loss in Chilean adults [23,24,25]. These studies are crucial due to the continuous need to provide adequate information for decision making. The objective of this study is to determine the experience of caries and tooth loss among the Chilean population over 60 years of age in five regions of Chile and identify the risk factors associated with tooth loss, aiming to update local epidemiological information. Our hypothesis assumed that despite the current pandemic context, we could expect a better state of oral health in the population studied. ## 2.1. Sampling Methodology In Chile, the elderly segment of the population is defined as people of 60 years and over [26]. The target population of this study were males and females aged 60 years or older, living in the regions of Antofagasta, Metropolitana, Maule, Bio Bio and La Araucanía. The national service for the elderly (SENAMA) promotes healthy ageing and equal rights for the elderly by articulating intersectoral networks in which groups and clubs of the elderly participate [27]. We recruited participants through databases of SENAMA regional coordination and community organizations between 2020 and 2021. We chose these regions due to the feasibility of implementing the teledentistry platform in these geographical areas, since most of them were near the universities that contributed to the development of this research. Within each region, we followed a stratified sampling design to select the specific districts that were to be included in the study. In this context, the allocation for each region was set proportionally to the resident population according to the projections made by the National Statistics Institute of Chile (INE). A final count of 135 participants were included in this study. This study was approved by the Scientific Ethics Committee of the Universidad de La Frontera, Temuco, Chile (Folio Number $\frac{090}{20}$). ## 2.2. Selection of Study Participants A new teledentistry workflow protocol proposed by the research team of this article [28] was executed, for which the recruitment and admission of subjects was carried out by a social worker, who synchronously collected basic socio-demographic data via phone calls, performed a COVID-19 triage, and selected the participants for this study using the inclusion criteria. The inclusion and exclusion criteria in the present study were as follows: Inclusion criteria:Older adult (over 60 years);Dental urgency or requirement of priority care: °Severe dental pain that does not yield to analgesics.°Recent trauma. Direct blow that involves teeth or mouth, accompanied by severe pain.°Oral bleeding.°Significant swelling of the mouth, face or neck.°Stains or wounds in any part of the mouth that do not disappear in a month.°Loss or fracture of restorations or dental prostheses.°Injuries to the mucosa, due to dental prosthesis mismatch.°Dental treatment required prior to urgent critical medical procedures. Exclusion criteria:Anticoagulant therapy;Chronic diseases without treatment;Cancer treatment;Dialysis. ## 2.3. Data Collection All subjects agreed to participate in the present study by signing an informed consent. Information was recorded on the platform named Geriatric Dental Specialties Teleplatform (TEGO by its acronym in Spanish: “Teleplataforma de Especialidades Geriátrico Odontológicas”) [29]. This web-based platform integrates anamnesis modules, a novel 3D standardized model for indexing relevant information for each case and also allows teleconsultations with specialists. Once the participants had been selected, full sociodemographic data were collected, including the Social Registry of Households in terms of percentile ranges (RSH) [30] and educational level (classified as complete basic education or less and incomplete secondary education or higher). The Barthel index [31] and an OHIP-14SP quality of life questionnaire [32] were applied. After the participants had been scheduled on the TEGO platform, the face-to-face care of the recruited subject was carried out by a general dentist in a mobile dental unit. The face-to-face care was carried out by three different general dentists with experience in primary health care, who carried out a complete medical-dental-geriatric anamnesis, which included the following history of chronic diseases: diabetes (obtained from the participants’’ reports), obesity (the subjects were measured and weighed inside the mobile clinic), and depression (the shortened Yesavage geriatric depression scale was applied [33]). Then, the dentist performed a complete general physical and dental examination and the information was recorded in a complete electronic clinical record on the TEGO platform. The dental examination included the application of the DMFT index scores that were recorded following WHO recommendations, where the M component comprises missing teeth due to caries or for any other reason [34]. The data were collected between February and June of 2021. ## 3. Results Using age categorization [34], frequently used in studies of this nature, Table 1 shows the absolute and relative frequencies of the 135 participants included in the research, differentiated by sex. A greater participation of women was shown in the study ($64.4\%$). Table 2 shows the DMFT index together with its $95\%$ confidence interval in parentheses. Additionally, t-student hypothesis tests were performed for the mean DMFT between both genders. No significant differences were found (p-value 0.6687). In Table 3, we present the Odds Ratios (ORs) with their respective confidence intervals and p-values of the univariate and multivariate logistic regression, including educational level (8 years of studies or more), RSH* according to variables mainly related to the economic income of the household or those variables that seek to reflect the income of a household [30]. For this study, subjects were classified as follows: RSH that corresponded to the up to $40\%$ more vulnerable population and the population over this segment (greater than $40\%$), and the presence of diabetes, depression, or obesity. Following a similar methodology to Urzúa et al. [ 23], our analysis established RSH* as the only statistically significant variable to predict the presence of tooth loss. Of the elderly participants, $8.1\%$ were edentulous. Table 4 shows the mean of the DMFT index by region and its respective breakdown into Decayed (D), Missing (M) and Filled (F) Teeth. The average for each region of Chile is shown in the age column. In order to determine if there is a significant statistical difference in the mean DMFT and their component between the different regions of Chile, multivariate hypothesis testing was performed, namely Wilks Lambda, Pillai’s Trace, Lawley–Hotelling Trace and Roy’s Largest Root showing there is not enough statistical evidence to infer that the mean DMFT and their components of Decayed (D) and Missing (M) teeth are different between regions. Nonetheless, the opposite conclusion was obtained for the Filled Teeth (F) component of the DMFT index (p-value 0.0268). ## 4. Discussion DMFT studies provide relevant information about emerging preventive dental practices in each country’s society [7,35,36,37,38,39]. The findings concerning the oral health of the elderly population (60 years and older) treated as part of our study in five regions of the country were similar to that observed in the Chilean adult population. The DMFT index (Table 2 and Table 4) was likewise similar to studies carried out in our country [23,40,41], demonstrating that the high caries experience of this group remains. Older adults with an age range of 65–74 years had DMFT scores (21.9) similar to those reported by Urzúa et al. ( 21.57) [23] and Mariño et al. ( 21.6) [40], and lower than those observed by Quinteros et al. ( 25.68) [39,41]. If we compare these data with other international studies under the same age segmentation, particularly from Spain [42], we observe that the rate of caries experienced in this European country is lower than that obtained in our study (16.38). In contrast with studies carried out in other Latin-American countries such as Brazil [43] and Uruguay [44], we found a higher rate than in Chile (29.24 and 24.1, respectively). Similarly, when compared with a study carried out in Mexico [45] under the same age segmentation, we found that the index of our study is similar to that obtained in that country (20.1). At a global level, an increase in dental retention has been reported [5] and according to the WHO, public health policies in the elderly population should lead to the improvement of the tooth loss index, which, in our sample, was 14.74, measured by mean MT (Table 2), being similar to the study by Urzúa et al. ( 17.46) [23] and Mariño et al. ( 17.9) [40]; much lower than Quinteros et al. ( 22.36) [41]. Since tooth loss has been associated with sociodemographic characteristics, as well as with individual factors such as lifestyle and general health [46,47,48,49,50], the difference between these studies could be explained due to the different sample sizes used in each study, in addition to the differences in the target population, among other factors. In relation to the rate of edentulism, a lower prevalence is observed ($8.1\%$) when compared with previous national studies that varied between $11.4\%$ and $25\%$ [6,23,40,51]. The prevalence detected may be lower than it is currently due to the objective of this project, which is not sought by edentulous patients. In relation to tooth loss, Urzúa et al. [ 23] studied the factors of educational level, personal and family income, and presence of depression and obesity and related it to functional tooth loss, finding family income as the only statistically significant variable within the context of logistic regression analysis. The same factors were incorporated in our study, except for the variables related to economic income, in which we used the Socioeconomic Classification (CSE) of the Household Social Registry (RSH), which attempts to measure socioeconomic vulnerability in different dimensions. This official classification is used by the Chilean government to support the selection processes of beneficiaries within a wide range of subsidies and social programs [30]. The Socioeconomic Classification (CSE) places each household in in one of seven segments related to their income: 0–$40\%$; 41–$50\%$; 51–$60\%$; 61–$70\%$; 71–$80\%$, 81–$90\%$ and 91–$100\%$. Each section reflects the level of vulnerability in the population, with the first section being the most affected. A model based only on family monetary income would place households in the same section of the CSE, not considering the number and the characteristics of the members within the household; for this reason, the RSH classification is more complete to determine the degree of vulnerability. Therefore, the variable included in the regression analysis refers to people that belong in Section 1 of this categorization. In our study, belonging to the most vulnerable group (up to $40\%$) was found to be the only statistically significant variable to describe the presence of tooth loss (p-value 0.002). To evaluate if there is a different burden of caries experience among the regions of Chile, hypothesis tests of mean equality of the DMFT and its components were performed. Our study found that there is not enough statistical evidence to assert that the means of DMFT, Decayed (DT) and Missing (MT), are different. However, this was not the case for the Filling component (FT). This difference could be explained by the great inequality that exists in the restorative treatment of dental caries (FT) in comparison with caries experience (DMFT) [8]. In Chile, as well as in other countries [52], the health status of people with a lower income is reported as worse and with more physical limitations [53]. Furthermore, they are faced with inequities in the access to dental health provided by the public sector [54], being even more serious regarding older adults, where 9 out of 10 are part of the public national health insurance system (FONASA) [55]. The relationship between general and oral health has already been established [56,57] through the convergence of risk factors such as sociodemographic, access to health, lifestyle, and others. The prevention and control of non-communicable diseases (NCDs), such as oral diseases, are integral parts of the COVID-19 response due to its association with the severity and fatality rates of COVID-19 [58]. Regarding this matter, one study compared the number of admissions for urgent dental care between the different periods of the pandemic, reporting a decrease in the use of services during the lockdown and the second wave [59]. To our knowledge, in our country, there is a lack of studies that evaluate the differences in the use of dental services prior to and during the pandemic; however, a survey was conducted before our study, reporting that $44\%$ of adults have had a dental problem during the pandemic; of these, $62\%$ indicated that their problems have not been resolved due to fear of contagion ($33\%$) and lack of clinical hours ($18\%$) [60]. This evidence and the fact that this study was carried out during the second wave, which occurred between the months of April and June 2021 [61], supports the idea that this modality of teledentistry was an important tool to implement during the pandemic. To this end, this project sought to provide care to a group with limited dental access during the COVID-19 lockdown and performed a comprehensive analysis of older people through the evaluation of sociodemographic factors and a complete medical-dental-geriatric anamnesis. Thus, it should be noted that although our DMFT scores are in line with what has been observed in several other countries, they are still very similar to those reported almost a decade ago in Chile [23,40,41], remaining very high in a global context and indicating a high burden of disease throughout the individual’s lifetime, since the main component of the DMFT corresponded to the MT [38].As expressed by Quinteros et al. [ 41], this could be explained by not being beneficiaries of the oral health promotion and prevention activities that are currently available, such as the project in which this study is contextualized. Even though our study did not find other associations with a lack of functional dentition through the logistical regression of our data, systematic and longitudinal studies related tooth loss with obesity [13,62,63], depression [64] and diabetes [14]. It is noteworthy that the prevalence of socioeconomic vulnerability, whether measured in net terms (personal or family economic income) or multidimensional (RSH), continues to be a constant in studies that measure the severity of tooth loss. It has also been said that in most countries, the health policies adopted seem inadequate to reduce the inequalities in oral health between socially vulnerable people [65]. Along with this, considering that Chile showed the lowest number of remaining teeth and the highest association with social inequity when comparing the prevalence of tooth loss with developed countries [66], it becomes necessary to adapt the former policies to new ones that focus on promotion, prevention and minimally invasive treatments. These activities must be accompanied by a strong focus on the most vulnerable strata in this age group, and should also incorporate the use of new technologies and mobile dental services in order to facilitate their access [22,67,68,69,70,71,72]. It is worth promoting further studies if the current health status of this age group is maintained due to the limited targeting of preventive public policies or in case it has worsened due to the current context of the pandemic; it has been hypothesized that loneliness and a lack of dental care could worsen the oral health of vulnerable groups [21]. Notwithstanding the relevance of finding the link described above, our study had some limitations that must be considered. Due to the confinement context, an exploratory cross-sectional study was applied, so caution must be exercised in the interpretation of the results, considering the relationship of these factors as a statistical association and not a causality. In addition, the subjects included come from regions selected by their convenience and cannot be extrapolated to the entire nation. These limitations emphasize the need to conduct a longitudinal representative survey including a larger number of regions from Chile. To our knowledge, this is the first study that relates the Social Registry of Households in terms of percentile ranges with oral health and that also compares the management of elderly individuals’ dental caries across regions. Another strength of this study is the evaluation of the oral health status of a highly vulnerable group through DMFT in a context in which a greater progression of oral diseases was projected. ## References 1. Aging and Health. Key Facts 2. Molina H., Sarmiento L., Aranco N., Jara P.. Envejecimiento y Atención a La Dependencia En Chile (2020.0) Volume 53. DOI: 10.18235/0002678 3. Peres M.A., Macpherson L.M.D., Weyant R.J., Daly B., Venturelli R., Mathur M.R., Listl S., Celeste R.K., Guarnizo-Herreño C.C., Kearns C.. Oral Diseases: A Global Public Health Challenge. Lancet (2019.0) 394 249-260. DOI: 10.1016/S0140-6736(19)31146-8 4. Hobdell M., Petersen P.E., Clarkson J., Johnson N.. Global Goals for Oral Health 2020. Int. Dent. J. (2003.0) 53 285-288. DOI: 10.1111/j.1875-595X.2003.tb00761.x 5. Kassebaum N.J., Bernabé E., Dahiya M., Bhandari B., Murray C.J.L., Marcenes W.. Global Burden of Severe Tooth Loss: A Systematic Review and Meta-Analysis. J. Dent. Res. (2014.0) 93 20S-28S. DOI: 10.1177/0022034514537828 6. 6. Ministerio de Salud Informe Encuesta Nacional de Salud 2016–2017: Salud BucalMinisterio de Salud Gobierno de ChileSantiago, Chile2019. Informe Encuesta Nacional de Salud 2016–2017: Salud Bucal (2019.0) 7. Do L.G., Roberts-Thomson K.F.. Dental Caries Experience in the Australian Adult Population. Aust. Dent. J. (2007.0) 52 249-251. DOI: 10.1111/j.1834-7819.2007.tb00496.x 8. Qin Y., Chen L., Li J., Wu Y., Huang S.. Greater Inequalities in Dental Caries Treatment than in Caries Experience: A Concentration Index Decomposition Approach. BMC Oral Health (2021.0) 21. DOI: 10.1186/s12903-021-01935-z 9. FDI Policy Statement on Oral Health for Healthy Ageing: Adopted by the FDI General Assembly: 24 September 2015, Bangkok, Thailand. Int. Dent. J. (2016.0) 66 7-8. DOI: 10.1111/idj.12231 10. Schwahn C., Polzer I., Haring R., Dörr M., Wallaschofski H., Kocher T., Mundt T., Holtfreter B., Samietz S., Völzke H.. Missing, Unreplaced Teeth and Risk of All-Cause and Cardiovascular Mortality. Int. J. Cardiol. (2013.0) 167 1430-1437. DOI: 10.1016/j.ijcard.2012.04.061 11. Galindo-Moreno P., Lopez-Chaichio L., Padial-Molina M., Avila-Ortiz G., O’Valle F., Ravida A., Catena A.. The Impact of Tooth Loss on Cognitive Function. Clin. Oral Investig. (2022.0) 26 3493-3500. DOI: 10.1007/s00784-021-04318-4 12. Thomson W.M., Barak Y.. Tooth Loss and Dementia: A Critical Examination. J. Dent. Res. (2021.0) 100 226-231. DOI: 10.1177/0022034520957233 13. Nascimento G.G., Leite F.R.M., Conceição D.A., Ferrúa C.P., Singh A., Demarco F.F.. Is There a Relationship between Obesity and Tooth Loss and Edentulism? A Systematic Review and Meta-Analysis. Obes. Rev. (2016.0) 17 587-598. DOI: 10.1111/obr.12418 14. Raju K., Taylor G.W., Tahir P., Hyde S.. Association of Tooth Loss with Morbidity and Mortality by Diabetes Status in Older Adults: A Systematic Review. BMC Endocr. Disord. (2021.0) 21. DOI: 10.1186/s12902-021-00830-6 15. Mendes J.J., Viana J., Cruz F., Pereira D., Ferreira S., Pereira P., Proença L., Machado V., Botelho J., Rua J.. Blood Pressure and Tooth Loss: A Large Cross-Sectional Study with Age Mediation Analysis. Int. J. Environ. Res. Public Health (2021.0) 18. DOI: 10.3390/ijerph18010285 16. Petersen P.E., Ogawa H.. Promoting Oral Health and Quality of Life of Older People—The Need for Public Health Action. Oral Health Prev. Dent. (2018.0) 16 113-124. DOI: 10.3290/j.ohpd.a40309 17. Patel J., Wallace J., Doshi M., Gadanya M., Ben Yahya I., Roseman J., Srisilapanan P.. Oral Health for Healthy Ageing. Lancet Healthy Longev. (2021.0) 2 e521-e527. DOI: 10.1016/S2666-7568(21)00142-2 18. Cuarentena Preventiva Obligatoria en Personas Mayores de 75 años 19. 19. COVID-19 Dental Services Evidence Review (CoDER) Working Group Recommendations for the Re-Opening of Dental Services: A Rapid Review of International SourcesCochrane Oral HealthManchester, UK2020. Recommendations for the Re-Opening of Dental Services: A Rapid Review of International Sources (2020.0) 20. 20. Ministerio de Salud COVID-19 En Chile. Pandemia 2020–2022Ministerio de Salud Gobierno de ChileSantiago, Chile2022. COVID-19 En Chile. Pandemia 2020–2022 (2022.0) 21. León S., Giacaman R.A.. Desigualdades En Salud Bucal Para Personas Mayores En Tiempos Del COVID-19. La Teleodontología y La Odontología de Mínima Intervención Como Caminos de Solución. Int. J. Interdiscip. Dent. (2020.0) 13 147-150. DOI: 10.4067/S2452-55882020000300147 22. Beltrán V., Acuña-Mardones P., Díaz J., Alvarado E., von Marttens A.. TEGO: A New Concept of Teledentistry for the Elderly through a Web Platform and Mobile App in the Context of the COVID-19 Pandemic. J. Oral Res. (2022.0) 1-8. DOI: 10.17126/joralres.2022.023 23. Urzua I., Mendoza C., Arteaga O., Rodríguez G., Cabello R., Faleiros S., Carvajal P., Muñoz A., Espinoza I., Aranda W.. Dental Caries Prevalence and Tooth Loss in Chilean Adult Population: First National Dental Examination Survey. Int. J. Dent. (2012.0) 2012 810170. DOI: 10.1155/2012/810170 24. 24. Ministerio de Salud Documento Segunda Entrega de Resultados. Tercera Encuesta Nacional de Salud (ENS) 2016–2017Ministerio de Salud Gobierno de ChileSantiago, Chile2018. Documento Segunda Entrega de Resultados. Tercera Encuesta Nacional de Salud (ENS) 2016–2017 (2018.0) 25. Borgeat Meza M., Espinoza I., Carvajal P., Cuevas R.. Changes in Oral Health Inequalities in Adults in Chile. Community Dent. Oral Epidemiol. (2021.0) 50 506-512. DOI: 10.1111/cdoe.12701 26. Glosario Gerontológico 27. Misión 28. Sanzana-luengo C., Díaz L., Abarza L., Alvarado E., Beltrán V.. Teledentistry Protocol for the Elderly in the Context of the COVID-19 Pandemic through a Web Platform/Mobile APP: Approach from the General Dentist. J. Oral Res. (2022.0) 1 1-8. DOI: 10.17126/joralres.2022.026 29. Beltrán V., von Marttens A., Acuña-Mardones P., Sanzana-Luengo C., Rueda-Velásquez S.J., Alvarado E., Flores M., Cerda A., Venegas B.. Implementation of a Teledentistry Platform for Dental Emergencies for the Elderly in the Context of the COVID-19 Pandemic in Chile. Biomed. Res. Int. (2022.0) 2022 6889285. DOI: 10.1155/2022/6889285 30. 30. Ministerio de Desarrollo Social Gobierno de Chile Orientaciones al RSH N° 8 Cálculo de La Calificación SocioeconómicaMinisterio de Desarrollo Social Gobierno de ChileSantiago, Chile2019. Orientaciones al RSH N° 8 Cálculo de La Calificación Socioeconómica (2019.0) 31. Bernaola-Sagardui I.. Validation of the Barthel Index in the Spanish Population. Enferm. Clin. (Engl. Ed.) (2018.0) 28 210-211. DOI: 10.1016/j.enfcli.2017.12.001 32. León S., Bravo-Cavicchioli D., Correa-Beltrán G., Giacaman R.A.. Validation of the Spanish Version of the Oral Health Impact Profile (OHIP-14Sp) in Elderly Chileans. BMC Oral Health (2014.0) 14. DOI: 10.1186/1472-6831-14-95 33. Yesavage J.A., Sheikh J.I.. Geriatric Depression Scale (GDS). Clin. Gerontol. (1986.0) 5 165-173. DOI: 10.1300/J018v05n01_09 34. 34. World Health Organization (WHO) Basic MethodsWHOGeneva, Switzerland2011Volume 510.1007/978-3-642-15352-5_3. Basic Methods (2011.0) Volume 5. DOI: 10.1007/978-3-642-15352-5_3 35. Bernabé E., Sheiham A.. Age, Period and Cohort Trends in Caries of Permanent Teeth in Four Developed Countries. Am. J. Public Health (2014.0) 104 e115-e121. DOI: 10.2105/AJPH.2014.301869 36. Yeung C.A., Chong L.Y., Glenny A.-M.. Fluoridated Milk for Preventing Dental Caries. Cochrane Database Syst. Rev. (2015.0) 2015 CD003876. DOI: 10.1002/14651858.CD003876.pub4 37. Kim H.-N., Kong W.-S., Lee J.-H., Kim J.-B.. Reduction of Dental Caries Among Children and Adolescents From a 15-Year Community Water Fluoridation Program in a Township Area, Korea. Int. J. Environ. Res. Public Health (2019.0) 16. DOI: 10.3390/ijerph16071306 38. Carvalho J.C., Schiffner U.. Dental Caries in European Adults and Senior Citizens 1996–2016: ORCA Saturday Afternoon Symposium in Greifswald, Germany—Part II. Caries Res. (2019.0) 53 242-252. DOI: 10.1159/000492676 39. Almerich-Torres T., Montiel-Company J.M., Bellot-Arcís C., Iranzo-Cortés J.E., Ortolá-Siscar J.C., Almerich-Silla J.M.. Caries Prevalence Evolution and Risk Factors among Schoolchildren and Adolescents from Valencia (Spain): Trends 1998–2018. Int. J. Environ. Res. Public Health (2020.0) 17. DOI: 10.3390/ijerph17186561 40. Mariño R.J., Cueto A., Badenier O., Acevedo R., Moya R.. Oral Health Status and Inequalities among Ambulant Older Adults Living in Central Chile. Community Dent. Health (2011.0) 28 143-148. PMID: 21780353 41. Quinteros M.E., Cáceres D.D., Soto A., Mariño R.J., Giacaman R.A.. Caries Experience and Use of Dental Services in Rural and Urban Adults and Older Adults from Central Chile. Int. Dent. J. (2014.0) 64 260-268. DOI: 10.1111/idj.12118 42. Eustaquio M.-V., Montiel J.-M., Almerich J.-M.. Oral Health Survey of the Adult Population of the Valencia Region (Spain). Med. Oral Patol. Oral Cir. Bucal. (2010.0) 15 e538-e544. DOI: 10.4317/medoral.15.e538 43. Rebelo M.A.B., Cardoso E.M., Robinson P.G., Vettore M.V.. Demographics, Social Position, Dental Status and Oral Health-Related Quality of Life in Community-Dwelling Older Adults. Qual. Life Res. (2016.0) 25 1735-1742. DOI: 10.1007/s11136-015-1209-y 44. Álvarez L., Liberman J., Abreu S., Mangarelli C., Correa M.B., Demarco F.F., Lorenzo S., Nascimento G.G.. Dental Caries in Uruguayan Adults and Elders: Findings from the First Uruguayan National Oral Health Survey. Cad. Saúde Pública (2015.0) 31 1663-1672. DOI: 10.1590/0102-311X00132214 45. Hernández-Palacios R.D., Ramírez-Amador V., Jarillo-Soto E.C., Irigoyen-Camacho M.E., Mendoza-Núñez V.M.. Relationship between Gender, Income and Education and Self-Perceived Oral Health among Elderly Mexicans. An Exploratory Study. Cienc. Saude Coletiva (2015.0) 20 997-1004. DOI: 10.1590/1413-81232015204.00702014 46. Gabiec K., Bagińska J., Łaguna W., Rodakowska E., Kamińska I., Stachurska Z., Dubatówka M., Kondraciuk M., Kamiński K.A.. Factors Associated with Tooth Loss in General Population of Bialystok, Poland. Int. J. Environ. Res. Public Health (2022.0) 19. DOI: 10.3390/ijerph19042369 47. Rodakowska E., Jamiolkowski J., Baginska J., Kaminska I., Gabiec K., Stachurska Z., Kondraciuk M., Dubatowka M., Kaminski K.A.. Oral Health–Related Quality of Life and Missing Teeth in an Adult Population: A Cross-Sectional Study from Poland. Int. J. Environ. Res. Public Health (2022.0) 19. DOI: 10.3390/ijerph19031626 48. Pitchika V., Jordan R.A., Norderyd O., Rolander B., Welk A., Völzke H., Holtfreter B., Kocher T.. Factors Influencing Tooth Loss in European Populations. J. Clin. Periodontol. (2022.0) 49 642-653. DOI: 10.1111/jcpe.13642 49. Bomfim R.A., Schneider I.J.C., de Andrade F.B., Lima-Costa M.F., Corrêa V.P., Frazão P., Watt R.G., Bastos J.L., de Oliveira C.. Racial Inequities in Tooth Loss among Older Brazilian Adults: A Decomposition Analysis. Community Dent. Oral Epidemiol. (2021.0) 49 119-127. DOI: 10.1111/cdoe.12583 50. Cunha-Cruz J., Hujoel P.P., Nadanovsky P.. Secular Trends in Socio-Economic Disparities in Edentulism: USA, 1972–2001. J. Dent. Res. (2007.0) 86 131-136. DOI: 10.1177/154405910708600205 51. Mariño R.J., Fu C.S., Giacaman R.A.. Prevalence of Root Caries among Ambulant Older Adults Living in Central Chile. Gerodontology (2015.0) 32 107-114. DOI: 10.1111/ger.12060 52. McMaughan D.J., Oloruntoba O., Smith M.L.. Socioeconomic Status and Access to Healthcare: Interrelated Drivers for Healthy Aging. Front. Public Health (2020.0) 8 231. DOI: 10.3389/fpubh.2020.00231 53. Vásquez F., Paraje G., Estay M.. Income-Related Inequality in Health and Health Care Utilization in Chile, 2000–2009. Pan Am. J. Public Health (2013.0) 33 98-106. DOI: 10.1590/S1020-49892013000200004 54. Gallego F., Larroulet C., Palomer L., Repetto A., Verdugo D.. Socioeconomic Inequalities in Self-Perceived Oral Health among Adults in Chile. Int. J. Equity Health (2017.0) 16 23. DOI: 10.1186/s12939-017-0519-9 55. 55. FONASA Cuenta Pública ParticipativaFondo Nacional de Salud (FONASA)Santiago, Chile2019. Cuenta Pública Participativa (2019.0) 56. Dörfer C., Benz C., Aida J., Campard G.. The Relationship of Oral Health with General Health and NCDs: A Brief Review. Int. Dent. J. (2017.0) 67 14-18. DOI: 10.1111/idj.12360 57. Wolf T.G., Cagetti M.G., Fisher J.-M., Seeberger G.K., Campus G.. Non-Communicable Diseases and Oral Health: An Overview. Front. Oral Health (2021.0) 2 725460. DOI: 10.3389/froh.2021.725460 58. Riad A., Boccuzzi M., Pold A., Krsek M.. The Alarming Burden of Non-Communicable Diseases in COVID-19 New Normal: Implications on Oral Health. Oral Dis. (2021.0) 27 791-792. DOI: 10.1111/odi.13491 59. Cagetti M.G., Balian A., Camoni N., Campus G.. Influence of the Covid-19 Pandemic on Dental Emergency Admissions in an Urgent Dental Care Service in North Italy. Int. J. Environ. Res. Public Health (2021.0) 18 1-10. DOI: 10.3390/ijerph18041812 60. 60. Centro de Epidemiología y Vigilancia de las Enfermedades Orales (CEVEO) Facultad de Odontología de la Universidad de Chile Colegio de Cirujanos Dentista de Chile Odontología En Pandemia COVID-19Centro de Epidemiología y Vigilancia de las Enfermedades Orales (CEVEO)Santiago, Chile2021. Odontología En Pandemia COVID-19 (2021.0) 61. Solis P., Adrián L.. Análisis Socioterritorial de la Distribución y Evolución Espacial del COVID-19 en las Comunas del Área Metropolitana de Santiago Periodo de Primera y Segunda Ola de Casos (2021.0) 62. Vallim A.C., Gaio E.J., Oppermann R.V., Rösing C.K., Albandar J.M., Susin C., Haas A.N.. Obesity as a Risk Factor for Tooth Loss over 5 Years: A Population-Based Cohort Study. J. Clin. Periodontol. (2021.0) 48 14-23. DOI: 10.1111/jcpe.13378 63. do Tôrres L.H.N., De Marchi R.J., Hilgert J.B., Hugo F.N., Ismail A.I., Antunes J.L.F., da Luz Rosário de Sousa M.. Oral Health and Obesity in Brazilian Elders: A Longitudinal Study. Community Dent. Oral Epidemiol. (2020.0) 48 540-548. DOI: 10.1111/cdoe.12566 64. Kunrath I., Silva A.E.R.. Oral Health and Depressive Symptoms among Older Adults: Longitudinal Study. Aging Ment. Health (2021.0) 25 2265-2271. DOI: 10.1080/13607863.2020.1855104 65. Cianetti S., Valenti C., Orso M., Lomurno G., Nardone M., Lomurno A.P., Pagano S., Lombardo G.. Systematic Review of the Literature on Dental Caries and Periodontal Disease in Socio-Economically Disadvantaged Individuals. Int. J. Environ. Res. Public Health (2021.0) 18. DOI: 10.3390/ijerph182312360 66. Elani H.W., Harper S., Thomson W.M., Espinoza I.L., Mejia G.C., Ju X., Jamieson L.M., Kawachi I., Kaufman J.S.. Social Inequalities in Tooth Loss: A Multinational Comparison. Community Dent. Oral Epidemiol. (2017.0) 45 266-274. DOI: 10.1111/cdoe.12285 67. Mariño R., Tonmukayakul U., Manton D., Stranieri A., Clarke K.. Cost-Analysis of Teledentistry in Residential Aged Care Facilities. J. Telemed. Telecare (2016.0) 22 326-332. DOI: 10.1177/1357633X15608991 68. Petcu R., Kimble C., Ologeanu-Taddei R., Bourdon I., Giraudeau N.. Assessing Patient’s Perception of Oral Teleconsultation. Int. J. Technol. Assess. Health Care (2017.0) 33 147-154. DOI: 10.1017/S0266462317000319 69. Aquilanti L., Santarelli A., Mascitti M., Procaccini M., Rappelli G.. Dental Care Access and the Elderly: What Is the Role of Teledentistry? A Systematic Review. Int. J. Environ. Res. Public Health (2020.0) 17. DOI: 10.3390/ijerph17239053 70. Oishi M.M., Childs C.A., Gluch J.I., Marchini L.. Delivery and Financing of Oral Health Care in Long-Term Services and Supports: A Scoping Review. J. Am. Dent. Assoc. (2021.0) 152 215-223.e2. DOI: 10.1016/j.adaj.2020.12.004 71. de Muniz I.A.F., Campos D.E.S., Shinkai R.S.A., da Trindade T.G., Cosme-Trindade D.C.. Case Report of Oral Mucosa Garlic Burn during COVID-19 Pandemic Outbreak and Role of Teledentistry to Manage Oral Health in an Older Adult Woman. Spec. Care Dent. (2021.0) 41 639-643. DOI: 10.1111/scd.12605 72. Ben-Omran M.O., Livinski A.A., Kopycka-Kedzierawski D.T., Boroumand S., Williams D., Weatherspoon D.J., Iafolla T.J., Fontelo P., Dye B.A.. The Use of Teledentistry in Facilitating Oral Health for Older Adults: A Scoping Review. J. Am. Dent. Assoc. (2021.0) 152 998-1011.e17. DOI: 10.1016/j.adaj.2021.06.005
--- title: 'Dietary Administration of Black Raspberries and Arsenic Exposure: Changes in the Gut Microbiota and Its Functional Metabolites' authors: - Pengcheng Tu - Qiong Tang - Zhe Mo - Huixia Niu - Yang Hu - Lizhi Wu - Zhijian Chen - Xiaofeng Wang - Bei Gao journal: Metabolites year: 2023 pmcid: PMC9967196 doi: 10.3390/metabo13020207 license: CC BY 4.0 --- # Dietary Administration of Black Raspberries and Arsenic Exposure: Changes in the Gut Microbiota and Its Functional Metabolites ## Abstract Mounting evidence has linked berries to a variety of health benefits. We previously reported that administration of a diet rich in black raspberries (BRBs) impacted arsenic (As) biotransformation and reduced As-induced oxidative stress. To further characterize the role of the gut microbiota in BRB-mediated As toxicity, we utilized the dietary intervention of BRBs combined with a mouse model to demonstrate microbial changes by examining associated alterations in the gut microbiota, especially its functional metabolites. Results showed that BRB consumption changed As-induced gut microbial alterations through restoring and modifying the gut microbiome, including its composition, functions and metabolites. A number of functional metabolites in addition to bacterial genera were significantly altered, which may be linked to the effects of BRBs on arsenic exposure. Results of the present study suggest functional interactions between dietary administration of black raspberries and As exposure through the lens of the gut microbiota, and modulation of the gut microbiota and its functional metabolites could contribute to effects of administration of BRBs on As toxicity. ## 1. Introduction Arsenic (As) exposure affects over 200 million members of the human population worldwide with geologically sourced contamination of drinking water being the major exposure route [1]. As exposure has been associated with a variety of human diseases including cardiovascular disease, diabetes, and bladder, lung, liver, and skin cancers [2]. It was previously established that perturbations of the gut microbiota and its metabolites could be a potential new mechanism by which As exposure leads to or exacerbates human diseases [3,4,5,6]. Therefore, modulation of the gut microbiota and its metabolic profile may affect As toxicity via impacting As-induced microbial perturbations. Mounting evidence has indicated the essential role of gut microbiota in human health and disease [7]. The gut microbiome is involved in immune cell development, energy production, and epithelial homeostasis [8,9,10,11,12]. More importantly, metabolic activities of gut bacteria have been linked to xenobiotic metabolism and toxicity of environmental chemicals. For example, previous studies have indicated that gut bacteria are involved in metabolism and biotransformation of As [13], polycyclic aromatic hydrocarbons [14], and polychlorinated biphenyls [15]. Also, changes in the gut microbiota have been mechanistically associated with toxic effects of environmental agents such as heavy metals including As, artificial sweeteners, and pesticides [16]. We previously reported that administration of a BRB-rich diet substantially changes the mouse gut microbiome at both compositional and functional levels [17,18,19], suggesting the potential of black raspberries to modulate the gut microbiome. In addition, we previously reported that the BRB-rich diet impacted As biotransformation and reduced levels of oxidative stress in As-treated mice [20]. As shown in Figure S1, dietary administration of BRBs increased As methylation thereby elevating urinary total As in As-treated mice. On the other hand, by measurement of levels of 8-oxo-2′-deoxyguanosine, one of the most commonly-used biomarkers of oxidative DNA damage [21], results showed that BRBs attenuated As-induced oxidative stress in mice. These previous results together supported the involvement of BRB consumption in As metabolism and toxicity. Given the key role of the gut microbiome in As toxicity coupled with the effects of BRBs on As biotransformation/toxicity, it is of significance to elucidate the changes in the gut microbiome upon interactions of As exposure and dietary administration of BRBs. In this follow-up study, we combined 16s rRNA gene sequencing and mass spectrometry-based metabolomics to extensively probe the changes in the gut microbiome of mice upon As exposure and dietary administration of black raspberries. Two complementary omic approaches were employed to achieve a comprehensive understanding of microbial changes. The 16S rRNA gene sequencing technique was used to identify bacterial changes in terms of abundance, which has been used as a mainstay of sequence-based bacterial analysis for decades. The other technique, activity-based metabolomics, allows the identification of metabolites that are differently abundant between groups with statistical significance. An untargeted approach enables the comprehensive comparison of metabolomes under different conditions, which is critical in understanding drivers of physiological activities related to gut microbiome. The results revealed that administration of BRBs reprogrammed the As-type gut microbiome, including alterations of various bacterial genera and key metabolites between groups with statistical significance, which could contribute to effects of BRBs on As biotransformation/toxicity. This follow-up study further elucidated changes of the gut microbiome in mice with dietary administration of black raspberries and As exposure, providing a connection with respect to diet, environmental agents, and the gut microbiota. ## 2.1. Preparation of Diets A BRB-rich diet was prepared as previously described [17,18,19,20]. Briefly, whole ripe BRBs (Rubus occidentalis) of the Jewel variety were picked mechanically, washed with water, and frozen at −20 °C on a single farm within 2 to 3 h of picking. The harvested berries were then shipped frozen to Van Drunen Farms in Momence, Illinois, where they were freeze-dried under anoxic conditions to protect the integrity of berry components. Next, seeds were removed by forcing the freeze-dried berries through a small sieve, and the dried pulp was ground into powder. The berry powder was shipped to Ohio State University, where it was stored at −20 °C until further use. For standardization purposes, each batch of powder underwent a quantitative chemical analysis of 26 randomly selected nutrients and nonnutrient components [22,23]. The levels of the 26 components remain relatively stable compared to the initial analyses for at least 2 years in powder stored at −20 °C [23]. The BRB powder was stored at −20 °C until being incorporated into custom purified American Institute of Nutrition (AIN)-76A animal diet (Dyets, Inc., Bethlehem, PA, USA) by $10\%$ w/w concentration at the expense of corn starch. AIN-76A diet was used as the control diet. Both diets were stored at 4 °C until being fed to mice. ## 2.2. Workflow to Investigate Functional Alterations of the Gut Microbiome by Dietary Administration of BRBs upon As Exposure As reported in our previous study [20] (Figure S1), dietary administration of BRBs successfully increased urinary excretion of As as well as modulated As biotransformation via facilitating As methylation. Moreover, BRB consumption reduced levels of oxidative stress in mice induced by As exposure. In this follow-up study, we aimed to further investigate functional alterations of the gut microbiome in mice upon As exposure and dietary administration of BRBs. The experimental design is shown in Figure 1A; briefly, 40 mice were randomly assigned into 4 groups: 76, 76+ As, BRB, BRB+ As. Of these, the 76 and 76+ As groups were fed AIN-76A diet (control diet), while BRB and BRB+ As groups were fed BRB diet. After 2 weeks of dietary administration, 76+ As and BRB+ As groups were switched to be exposed to As via drinking water (10 ppm). After another 4 weeks of As treatment, fecal samples were collected for taxonomic characterization and metabolite profiling. The experimental workflow combined high-through 16S rRNA gene sequencing and mass-spectrometry-based metabolomics for the examination of changes in the gut microbiome resulting from BRB-mediated As toxicity. ## 2.3. Animals The animal protocol was approved by the University of Georgia Institutional Animal Care and Use Committee (protocol No. A2013 06-033-Y3-A3). A total of 40 specific-pathogen-free (SPF) C57BL/6 mice (~8 weeks old) were purchased from Jackson Laboratories. The mice were housed in the animal facility of the University of Georgia. After 1 week of acclimation, mice were randomly assigned to 4 groups (76, 76+ As, BRB, BRB+ As). Of these, 76 and 76+ As groups were fed AIN-76A diet (control diet), while BRB and BRB+ As groups were fed BRB diet. Environmental conditions of 22 °C temperature, 40–$70\%$ humidity, with a $\frac{12}{12}$ h light/dark cycle were applied. 76+ As and BRB+ As groups were exposed to As via drinking water (10 ppm) after 2 weeks. Mice of 30 g have an average daily water intake of 2 mL [24]. After another 4 weeks of As treatment, fecal samples were collected individually, and stored at −80 °C for further experiments. ## 2.4. 16S rRNA Gene Sequencing Experiments of 16S rRNA gene sequencing were conducted as previously described [17]. Briefly, microbial DNA was extracted from mouse fecal samples (20–25 mg) using PowerSoil DNA isolation kit as per manufacturer’s instructions. For 16S rRNA gene sequencing, DNA was amplified using 515F and 806R primers to target the V4 regions of 16S rRNA gene. The DNA was then amplified, followed by normalization, barcoding, and the DNA was pooled, and quantified by Qubit 2.0 Fluorometer to construct the sequencing library. The resultant DNA was then paired-end sequenced using an Illumina MiSeq platform (Illumina, 500 cycles v2 kit, San Diego, CA, USA) in the Georgia Genomics Facility of University of Georgia. Paired-reads were assembled by the software Geneious (Biomatters, Auckland, New Zealand), followed by initial quality filtering with error probability of 0.01. The operational taxonomic unit (OTU) picking and diversity analysis was performed with a threshold of $97\%$ sequence similarity by the software of Quantitative Insights into Microbial Ecology (QIIME). A representative sequence from each OTU was selected for taxonomic assignment according to Greengenes database (version 13_5; http://greengenes.lbl.gov/; accessed on 1 May 2019). By default, QIIME uses uclust consensus taxonomy classifier to assign taxonomy. ## 2.5. Untargeted Metabolomic Analysis Experiments of untargeted metabolomic analysis were conducted as previously described [18]. Briefly, 20 mg of fecal samples and 50 mg of glass beads (Sigma-Aldrich, St. Louis, MO, USA) were mixed with 400 μL of cooled methanol solution (methanol/water 1:1). The mix was homogenized, and then centrifuged for 10 min at 12,000 rpm. The supernatant (~300 μL) was collected, dried in a SpeedVac (Savant SC110A; Thermo Electron, Waltham, MA, USA), and then resuspended in 30 μL 98:2 water:acetonitrile for MS analysis injection. LC-MS analysis was performed on a quadrupole time-of-flight (Q-TOF) 6530 (Agilent Technologies, Santa Clara, CA, USA) with an electrospray ionization source interfaced with an Agilent 1290 Infinity II UPLC system. The Q-TOF was calibrated daily using the standard tuning solution from Agilent Technologies. Metabolic features were analyzed in the positive ion mode using a C18 T3 reverse-phased column (Waters Corporation, Milford, MA, USA). The typical mass accuracy of the Q-TOF was <10 ppm. XCMS online server was applied for peak picking, alignment, integration, as well as extraction of peak intensities. MS/MS data were generated on Q-TOF for further identification of differently abundant features. Software packages MS-DIAL (version 2.9) and MS-FINDER (version 2.4) were applied for identification of metabolic features based on MS/MS spectrum [25,26]. ## 2.6. Statistical Analysis Alpha rarefaction and principal coordinate analysis (PCoA) were performed to assess alpha and beta diversities in the gut microbial communities, respectively. Alpha rarefaction analysis was performed using indices of observed OTUs, PD whole tree, and Chao1. PCoA was performed based on the unweighted UniFrac distance metric. Permutational multivariate analysis of variance (PERMANOVA) was applied to assess the difference between different cultivars. Also, principal component analysis (PCA) and hierarchical clustering algorithm were used for visualization of metabolite profiles. Differences in gut microbial abundances were assessed by a nonparametric test via Metastats. Two-tailed Welch’s t-test was used to analyze metabolites that were differently abundant between groups. False discovery rate (FDR) was used to correct multiple comparisons. ## 3.1. Gut Microbial Changes at Compositional Level Figure 1B,C shows the identified gut bacteria at order and family levels assigned from 16S rRNA sequencing reads with each color representing an individual bacterial order or family, respectively (legend at Figure S2). As exposure induced changes of consistent trends in several bacterial families regardless of different diets. For instance, in both 76 and BRB diet groups, Bifidobacteriaceae and Bacteroidales_f_S24-7 increased upon arsenic exposure. However, dietary difference contributes more significantly to microbiota changes compared to As exposure. Notably, Verrucomicrobiaceae was less than $0.1\%$ in 76 diet groups, while in BRB groups, the proportions of Verrucomicrobiaceae were $51.5\%$ and $30.2\%$ in BRB and BRB+ As groups, respectively. Moreover, As also induced contradicted changes in different diet groups. For example, in 76 diet groups, Clostridiaceae decreased upon As exposure with a 1.5-fold change; however, in BRB diet groups, Clostridiaceae increased four-fold. Taken together, although diet contributes more significantly to gut microbial changes at compositional level, As exposure-induced gut microbiota changes differed if mice were fed different diets. In addition, Table 1 shows the fold changes of significantly-altered bacteria in mice upon arsenic exposure with different diets: there were nine bacterial genera that were significantly altered, with three increased and six decreased genera. Notably, compared with 76+ As group, Akkermansia increased with a fold change of approximate 5000 in BRB+ As group, which is consistent with effects of BRB on the gut microbiota from our previous report [17]. ## 3.2. Principal Coordinate Analysis (PCoA) and Alpha Rarefaction Analysis To further assess the differences of the gut bacterial community, alpha rarefaction and PCoA analyses were performed on mouse fecal samples to assess alpha and beta diversities in the gut microbial communities, respectively. Figure 2A shows the 3D PCoA plot of gut microbial communities. PCoA analysis based on the UniFrac distance metric reflects beta diversity between groups. The four sample groups were separated majorly driven by diet as indicated by separation between groups on different diets ($p \leq 0.05$). As exposure also impacts the gut microbial communities according to PCoA analysis, $17.28\%$, $6.84\%$, and $4.15\%$ variation were explained by principal component (PC) 1, PC2, and PC3, respectively. In addition, alpha rarefaction analysis using indices of observed OTUs, PD whole tree, and Chao1 was shown in Figure 2B–D, respectively. Of these indices, observed OTUs and Chao1 reflect species richness in the community, and PD whole tree is a diversity calculated based on phylogenetic tree. Although there is no statistically significant difference, alpha diversities of the gut microbial community fluctuate upon arsenic exposure and dietary administration of BRBs. ## 3.3. Comparative Analysis of Metabolite Profiles Figure 3A shows the principal component analysis (PCA) plot. The PCA was calculated using the feature intensities from all samples with colors assigned based on sample groups. Consistent with the PCoA result of gut microbial communities, diet plays a more significant role in separating different cultivars. Moreover, Figure 3B,C shows the cloud plots, constructed by altered features with green bubbles representing up-regulated features and red bubbles representing down-regulated features (p-value is represented by how dark or light the color is; fold change is represented by the radius of each feature; retention time is represented by position on the x-axis; mass-to-charge ratio is represented by position on y-axis). Compared with 76 group, BRB group had 1180 significantly-regulated features (Figure 3B). BRB+ As group had 958 features that were significantly changed compared to 76+ As group (Figure 3C). Metabolite profiling of the gut microbiome in mice with dietary administration of BRBs was reported in one of our previous studies [18]. In addition, the hierarchical clustering heat map constructed using intensities of shared features showed consistent patterns within individual groups (Figure 4). Not only As exposure induced microbial alterations in mice, but also these metabolic perturbations induced by As treatment were partly reversed by BRB dietary administration, indicating that administration of BRBs modulates and potentially restores As-treated gut microbiome and functional metabolites. ## 3.4. Key Metabolites Associated with Dietary Administration of BRBs upon As Exposure To explore the role of BRB consumption in the gut microbiota of mice exposed to As, differently abundant metabolites between 76+ As and BRB+ As groups were profiled and identified. Table 2 lists the identified fecal metabolites. Between BRB+ As and 76+ As groups, there were a total of 18 identified metabolites that were significantly altered, including vitamin derivatives, bile acids, indoles, polyunsaturated fatty acids, bilirubins, and so forth. Of these metabolites, many of them are bioactive molecules that are involved in a number of metabolic processes and cellular functions. For example, flavins and tocopherols are vitamin derivatives related to the metabolic activities of gut bacteria and intestinal homeostasis. Riboflavin could be produced by the gut microbiota [27], and levels of tocopherols are associated with gut barrier functions [28]. Likewise, bile acids are a class of key metabolites for gut bacteria and diverse signaling pathways [29]. Taken together, key metabolites in the gut microbiota that were associated with dietary administration of BRBs upon As exposure could contribute to effects of BRB consumption on arsenic biotransformation/toxicity via host-gut microbiota axis. ## 4. Discussion With the increasingly recognized role of the gut microbiome in As toxicity, the knowledge of how gut microbiome alterations affect health outcomes in As exposure is critical to development of therapeutic approaches via modulation of the gut microbiome. This knowledge is of importance, because it may be applied in the future to develop gut microbiome-targeted therapeutic approaches via dietary intervention. We previously reported that BRB consumption effectively affects As biotransformation and possibly impact As toxicity [20]. The objective of this follow-up study was to determine the role of the gut microbiota in BRB-mediated As biotransformation/toxicity. Given the essential role of the gut microbiome in As toxicity coupled with the effects of BRBs on As biotransformation/toxicity, illustration of changes in the gut microbiota of mice upon interactions of As exposure and dietary administration of black raspberries is of significance and represents an important step toward understanding how diet affects environmental exposure through the lens of the gut microbiota. The gut microbiota not only directly impact intestinal homeostasis locally through microbial metabolic products, but also trigger systemic effects on remote tissues/organs such as the liver, adipose, or brain by producing metabolites that can act as signaling molecules [30,31]. Moreover, the role of the gut microbiota in transformation and metabolism of xenobiotics including As has been well recognized [3], indicated by previous reports on interactions of the gut microbiome with environmental toxic chemials such as As [3,13], diazinon [32], polycylic aromatic hydrocarbons [14], and polychlorinated biphenyls [15]. In addition, to perturb the gut microbiome and its functional metabolites is suggested to be a new mechanism of As toxicity [3,4,5,6]. Therefore, metabolic changes, especially perturbations in gut microbiota-related metabolites, play an essential role in As exposure and toxicity. In the present study, differently abundant metabolites in the gut microbiome of mice with or without administration of BRBs upon As exposure were profiled and identified. Significantly-altered metabolite fingerprints in fecal samples of mice were observed. Specifically, 12a-Hydroxy-3-oxocholadienic acid, belonging to the class of monohydroxy bile acids, alcohols and derivatives, increased with a fold change of two in fecal samples of mice fed BRBs compared to mice on control diet, upon As exposure. Bile acids are cholesterol derivatives synthesized in liver, which would undergo extensive enterohepatic recycling as well as modification by some gut bacteria. It is established that bile acids not only participate in digestion and absorption, but also serve as signaling molecules impacting a number of pathways by acting on diverse nuclear receptors [33]. Likewise, alpha-linolenic acid belongs to the class of lineolic acids and derivatives, which increased with a fold change of 1.8 in fecal samples of mice from BRB+ As group compared to mice from 76+ As group. Lineolic acids are polyunsaturated fatty acids with many beneficial effects associated with intestinal immunity and the gut microbiota [34]. Also, 13′-Carboxy-alpha-tocopherol is a derivative of tocopherol, which increased with a fold change of 1.5 in fecal samples of mice upon As exposure if fed BRBs compared to control diet. Tocopherols are known to confer protective effects on oxidative stress and inflammation [35,36,37]. More importantly, tocopherols may exhibit anti-cancer effect via several different cellular and molecular mechanisms [38], which could counter increased risks of bladder cancer and skin cancer associated with As exposure as well as possibly accounting for anti-cancer effects of BRBs, although the evidence remains controversial [39]. Thus, metabolic changes in the gut microbiota could contribute to effects of BRBs on mice upon As exposure. The human gut microbiome contributes to human health and disease in a significant way, including key functions involved in immune cell development, energy production, and epithelial homeostasis [8,9,10,11,12]. Moreover, its role in transformation and metabolism of xenobiotics has been well recognized [3,16]. The gut microbiota continues to be an attractive therapeutic target. Currently, our knowledge of targeted and predictable modulation of the gut microbiome is in its infancy [40]. It is of significance to provide a diet-based approach for gut microbiome modulation, to elucidate how the modulated microbiome differently reacts to environmental chemicals, and to understand the role of microbiome-derived metabolites in these interactions [16,40]. Diet emerges as an essential determinant of gut microbial structure and functions [41]. Dietary modulation of the gut microbiome received considerable attention due to the advantages of low toxicity profiles and high patient compliance. Dietary recommendations to tackle gut microbiota-associated diseases such as inflammatory bowel disease (IBD) are usually based on inconclusive or controversial evidence [42], resulting from the complexity and variability of IBD disease pathogenesis including disease phenotype, gut microbiome, host genetic susceptibility, and environmental factors [43]. Thus, a therapeutic approach to treat these diseases through gut microbiome modulation is still highly desirable. On the other hand, previous studies showed that As exposure perturbed the gut microbiome and its metabolic functions [3], which may contribute to its toxicity. Modulation of the gut microbiome, in particular the microbial metabolites, could potentially alter As toxicity. Using a standardized BRB-rich diet and a mouse model, we have previously reported that BRB consumption effectively modulated the gut microbiota, including its composition, functions and metabolites [17,18,19]. Moreover, biotransformation and toxic effects of As were altered in mice that were fed BRBs [20]. Results of this follow-up study further supported the role of BRBs in mediating As toxicity by elucidating gut microbiota changes upon the interactions of dietary administration of black raspberries and arsenic exposure. Taken together, the potential of BRBs in modulating the gut microbiome, and specifically in intervening in the toxicity of environmental chemicals including As warrants future studies. Admittedly this study was based on observation of effects of dietary administration of BRBs on mice upon As exposure and the gut microbiota, with no specific mechanism clearly illustrated. Further studies are needed to clarify the mechanism of interrelationships among those factors including the gut microbiome, its functional metabolites and brought effects for As-related adverse impact. Nevertheless, we profiled and identified metabolic changes in the gut microbiota of mice, and identified key metabolites that could contribute to effects of BRBs on mice upon As exposure. Although further investigation on mechanisms needs to be pursued, the present study is of significance for depicting the involvement of the gut microbiota in BRB-mediated As toxicity. ## 5. Conclusions In the present study, we further analyzed changes in the gut microbiota and its functional metabolites upon interactions of As exposure and BRB administration to potentially identify microbial alterations that were mechanistically associated with BRB-mediated As toxicity. 16S rRNA gene sequencing and metabolomic profiling techniques were used to probe alterations in the gut microbiota and its metabolic profiles. The results clearly show that BRB significantly changed As-type gut microbiota at both compositional and functional levels. Microbial alterations induced by As exposure were restored or modified. In addition, alterations in a variety of gut microbiota-related metabolic products, including vitamin derivatives, bile acids, indoles, polyunsaturated fatty acids, and bilirubins, may be associated with effects on As toxicity by BRBs. Taken together, these findings may provide insights regarding the connection among diet, environmental exposure, and the gut microbiota, as well as offer evidence for future development of approaches for gut microbiome modulation. ## References 1. Naujokas M.F., Anderson B., Ahsan H., Aposhian H.V., Graziano J.H., Thompson C., Suk W.A.. The broad scope of health effects from chronic arsenic exposure: Update on a worldwide public health problem. Environ. Health Perspect. (2013) 121 295-302. DOI: 10.1289/ehp.1205875 2. Hughes M.F., Beck B.D., Chen Y., Lewis A.S., Thomas D.J.. Arsenic exposure and toxicology: A historical perspective. Toxicol. Sci. Off. J. Soc. Toxicol. (2011) 123 305-332. DOI: 10.1093/toxsci/kfr184 3. Lu K., Abo R.P., Schlieper K.A., Graffam M.E., Levine S.S., Wishnok J.S., Swenberg J.A., Tannenbaum S.R., Fox J.G.. Arsenic exposure perturbs the gut microbiome and its metabolic profile in mice: An integrated metagenomics and metabolomics analysis. Environ. Health Perspect. (2014) 122 284-291. DOI: 10.1289/ehp.1307429 4. Chi L., Bian X., Gao B., Ru H., Tu P., Lu K.. Sex-Specific Effects of Arsenic Exposure on the Trajectory and Function of the Gut Microbiome. Chem. Res. Toxicol. (2016) 29 949-951. DOI: 10.1021/acs.chemrestox.6b00066 5. Chi L., Bian X., Gao B., Tu P., Ru H., Lu K.. The Effects of an Environmentally Relevant Level of Arsenic on the Gut Microbiome and Its Functional Metagenome. Toxicol. Sci. Off. J. Soc. Toxicol. (2017) 160 193-204. DOI: 10.1093/toxsci/kfx174 6. Chi L., Lai Y., Tu P., Liu C.-W., Xue J., Ru H., Lu K.. Lipid and Cholesterol Homeostasis after Arsenic Exposure and Antibiotic Treatment in Mice: Potential Role of the Microbiota. Environ. Health Perspect. (2019) 127 97002. DOI: 10.1289/EHP4415 7. Nicholson J.K., Holmes E., Kinross J., Burcelin R., Gibson G., Jia W., Pettersson S.. Host-gut microbiota metabolic interactions. Science (2012) 336 1262-1267. DOI: 10.1126/science.1223813 8. Pickard J.M., Zeng M.Y., Caruso R., Núñez G.. Gut microbiota: Role in pathogen colonization, immune responses, and inflammatory disease. Immunol. Rev. (2017) 279 70-89. DOI: 10.1111/imr.12567 9. Round J.L., Mazmanian S.K.. The gut microbiota shapes intestinal immune responses during health and disease. Nat. Rev. Immunol. (2009) 9 313-323. DOI: 10.1038/nri2515 10. Tremaroli V., Bäckhed F.. Functional interactions between the gut microbiota and host metabolism. Nature (2012) 489 242-249. DOI: 10.1038/nature11552 11. Turnbaugh P.J., Ley R.E., Mahowald M.A., Magrini V., Mardis E.R., Gordon J.I.. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature (2006) 444 1027-1031. DOI: 10.1038/nature05414 12. Shi N., Li N., Duan X., Niu H.. Interaction between the gut microbiome and mucosal immune system. Mil. Med. Res. (2017) 4 14. DOI: 10.1186/s40779-017-0122-9 13. Van de Wiele T., Gallawa C.M., Kubachk K.M., Creed J.T., Basta N., Dayton E.A., Whitacre S., Laing G.D., Bradham K.. Arsenic metabolism by human gut microbiota upon in vitro digestion of contaminated soils. Environ. Health Perspect. (2010) 118 1004-1009. DOI: 10.1289/ehp.0901794 14. Van de Wiele T., Vanhaecke L., Boeckaert C., Peru K., Headley J., Verstraete W., Siciliano S.. Human colon microbiota transform polycyclic aromatic hydrocarbons to estrogenic metabolites. Environ. Health Perspect. (2005) 113 6-10. DOI: 10.1289/ehp.7259 15. Choi J.J., Eum S.Y., Rampersaud E., Daunert S., Abreu M.T., Toborek M.. Exercise attenuates PCB-induced changes in the mouse gut microbiome. Environ. Health Perspect. (2013) 121 725-730. DOI: 10.1289/ehp.1306534 16. Tu P., Chi L., Bodnar W., Zhang Z., Gao B., Bian X., Stewart J., Fry R., Lu K.. Gut Microbiome Toxicity: Connecting the Environment and Gut Microbiome-Associated Diseases. Toxics (2020) 8. DOI: 10.3390/toxics8010019 17. Tu P., Bian X., Chi L., Gao B., Ru H., Knobloch T.J., Weghorst C.M., Lu K.. Characterization of the Functional Changes in Mouse Gut Microbiome Associated with Increased Population Modulated by Dietary Black Raspberries. ACS Omega (2018) 3 10927-10937. DOI: 10.1021/acsomega.8b00064 18. Tu P., Bian X., Chi L., Xue J., Gao B., Lai Y., Ru H., Lu K.. Metabolite Profiling of the Gut Microbiome in Mice with Dietary Administration of Black Raspberries. ACS Omega (2020) 5 1318-1325. DOI: 10.1021/acsomega.9b00237 19. Tu P., Chi L., Bian X., Gao B., Ru H., Lu K.. A Black Raspberry-Rich Diet Protects From Dextran Sulfate Sodium-Induced Intestinal Inflammation and Host Metabolic Perturbation in Association With Increased Aryl Hydrocarbon Receptor Ligands in the Gut Microbiota of Mice. Front. Nutr. (2022) 9 842298. DOI: 10.3389/fnut.2022.842298 20. Tu P., Xue J., Bian X., Chi L., Gao B., Leng J., Ru H., Knobloch T.J., Weghorst C.M., Lu K.. Dietary administration of black raspberries modulates arsenic biotransformation and reduces urinary 8-oxo-2′-deoxyguanosine in mice. Toxicol. Appl. Pharmacol. (2019) 377 114633. DOI: 10.1016/j.taap.2019.114633 21. Valavanidis A., Vlachogianni T., Fiotakis C.. 8-hydroxy-2′-deoxyguanosine (8-OHdG): A critical biomarker of oxidative stress and carcinogenesis. J. Environ. Sci. Health. Part C Environ. Carcinog. Ecotoxicol. Rev. (2009) 27 120-139. DOI: 10.1080/10590500902885684 22. Kresty L.A., Morse M.A., Morgan C., Carlton P.S., Lu J., Gupta A., Blackwood M., Stoner G.D.. Chemoprevention of esophageal tumorigenesis by dietary administration of lyophilized black raspberries. Cancer Res. (2001) 61 6112-6119. PMID: 11507061 23. Stoner G.D., Wang L.S., Sardo C., Zikri N., Hecht S.S., Mallery S.R.. Cancer prevention with berries: Role of anthocyanins, in Bioactive compounds and cancer. Bioactive Compounds and Cancer (2010) 703-723 24. Suez J., Korem T., Zeevi D., Zilberman-Schapira G., Thaiss C.A., Maza O., Israeli D., Zmora N., Gilad S., Weinberger A.. Artificial sweeteners induce glucose intolerance by altering the gut microbiota. Nature (2014) 514 181-186. DOI: 10.1038/nature13793 25. Tsugawa H., Cajka T., Kind T., Ma Y., Higgins B., Ikeda K., Kanazawa M., VanderGheynst J., Fiehn O., Arita M.. MS-DIAL: Data-independent MS/MS deconvolution for comprehensive metabolome analysis. Nat. Methods (2015) 12 523-526. DOI: 10.1038/nmeth.3393 26. Tsugawa H., Kind T., Nakabayashi R., Yukihira D., Tanaka W., Cajka T., Saito K., Fiehn O., Arita M.. Hydrogen Rearrangement Rules: Computational MS/MS Fragmentation and Structure Elucidation Using MS-FINDER Software. Anal. Chem. (2016) 88 7946-7958. DOI: 10.1021/acs.analchem.6b00770 27. LeBlanc J.G., Milani C., de Giori G.S., Sesma F., van Sinderen D., Ventura M.. Bacteria as vitamin suppliers to their host: A gut microbiota perspective. Curr. Opin. Biotechnol. (2013) 24 160-168. DOI: 10.1016/j.copbio.2012.08.005 28. Liu K.Y., Nakatsu C.H., Jones-Hall Y., Kozik A., Jiang Q.. Vitamin E alpha- and gamma-tocopherol mitigate colitis, protect intestinal barrier function and modulate the gut microbiota in mice. Free. Radic. Biol. Med. (2021) 163 180-189. DOI: 10.1016/j.freeradbiomed.2020.12.017 29. Cai J., Sun L., Gonzalez F.J.. Gut microbiota-derived bile acids in intestinal immunity, inflammation, and tumorigenesis. Cell Host Microbe (2022) 30 289-300. DOI: 10.1016/j.chom.2022.02.004 30. Claus S.P., Ellero S.L., Berger B., Krause L., Bruttin A., Molina J., Paris A., Want E.J., de Waziers I., Cloarec O.. Colonization-induced host-gut microbial metabolic interaction. MBio (2011) 2 e00271-10. DOI: 10.1128/mBio.00271-10 31. Heijtz R.D., Wang S., Anuar F., Qian Y., Björkholm B., Samuelsson A., Hibberd M.L., Forssberg H., Pettersson S.. Normal gut microbiota modulates brain development and behavior. Proc. Natl. Acad. Sci. USA. (2011) 108 3047-3052. DOI: 10.1073/pnas.1010529108 32. Gao B., Bian X., Mahbub R., Lu K.. Sex-Specific Effects of Organophosphate Diazinon on the Gut Microbiome and Its Metabolic Functions. Environ. Health Perspect. (2017) 125 198-206. DOI: 10.1289/EHP202 33. Nguyen A., Bouscarel B.. Bile acids and signal transduction: Role in glucose homeostasis. Cell. Signal. (2008) 20 2180-2197. DOI: 10.1016/j.cellsig.2008.06.014 34. Fu Y., Wang Y., Gao H., Li D., Jiang R., Ge L., Tong C., Xu K.. Associations among Dietary Omega-3 Polyunsaturated Fatty Acids, the Gut Microbiota, and Intestinal Immunity. Mediat. Inflamm. (2021) 2021 8879227. DOI: 10.1155/2021/8879227 35. Wallert M., Ziegler M., Wang X., Maluenda A., Xu X., Yap M.L., Witt R., Giles C., Kluge S., Hortmann M.. α-Tocopherol preserves cardiac function by reducing oxidative stress and inflammation in ischemia/reperfusion injury. Redox Biol. (2019) 26 101292. DOI: 10.1016/j.redox.2019.101292 36. Demirel-Yalciner T., Sozen E., Ozaltin E., Sahin A., Ozer N.K.. Alpha-Tocopherol supplementation reduces inflammation and apoptosis in high cholesterol mediated nonalcoholic steatohepatitis. BioFactors (2021) 47 403-413. DOI: 10.1002/biof.1700 37. Lewis E.D., Meydani S.N., Wu D.. Regulatory role of vitamin E in the immune system and inflammation. IUBMB Life (2019) 71 487-494. DOI: 10.1002/iub.1976 38. Abraham A., Kattoor A.J., Saldeen T., Mehta J.L.. Vitamin E and its anticancer effects. Crit. Rev. Food Sci. Nutr. (2019) 59 2831-2838. DOI: 10.1080/10408398.2018.1474169 39. Yang C.S., Luo P., Zeng Z., Wang H., Malafa M., Suh N.. Vitamin E and cancer prevention: Studies with different forms of tocopherols and tocotrienols. Mol. Carcinog. (2020) 59 365-389. DOI: 10.1002/mc.23160 40. Schmidt T.S.B., Raes J., Bork P.. The Human Gut Microbiome: From Association to Modulation. Cell (2018) 172 1198-1215. DOI: 10.1016/j.cell.2018.02.044 41. Wolter M., Grant E.T., Boudaud M., Steimle A., Pereira G.V., Martens E.C., Desai M.S.. Leveraging diet to engineer the gut microbiome. Nat. Rev. Gastroenterol. Hepatol. (2021) 18 885-902. DOI: 10.1038/s41575-021-00512-7 42. Lee M., Chang E.B.. Inflammatory Bowel Diseases (IBD) and the Microbiome-Searching the Crime Scene for Clues. Gastroenterology (2021) 160 524-537. DOI: 10.1053/j.gastro.2020.09.056 43. Knox N.C., Forbes J.D., Van Domselaar G., Bernstein C.N.. The Gut Microbiome as a Target for IBD Treatment: Are We There Yet?. Curr. Treat. Options Gastroenterol. (2019) 17 115-126. DOI: 10.1007/s11938-019-00221-w
--- title: Ethosomal Gel for Topical Administration of Dimethyl Fumarate in the Treatment of HSV-1 Infections authors: - Mariaconcetta Sicurella - Walter Pula - Karolina Musiał - Katarzyna Cieślik-Boczula - Maddalena Sguizzato - Agnese Bondi - Markus Drechsler - Leda Montesi - Elisabetta Esposito - Peggy Marconi journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9967198 doi: 10.3390/ijms24044133 license: CC BY 4.0 --- # Ethosomal Gel for Topical Administration of Dimethyl Fumarate in the Treatment of HSV-1 Infections ## Abstract The infections caused by the HSV-1 virus induce lesions on the lips, mouth, face, and eye. In this study, an ethosome gel loaded with dimethyl fumarate was investigated as a possible approach to treat HSV-1 infections. A formulative study was conducted, evaluating the effect of drug concentration on size distribution and dimensional stability of ethosomes by photon correlation spectroscopy. Ethosome morphology was investigated by cryogenic transmission electron microscopy, while the interaction between dimethyl fumarate and vesicles, and the drug entrapment capacity were respectively evaluated by FTIR and HPLC. To favor the topical application of ethosomes on mucosa and skin, different semisolid forms, based on xanthan gum or poloxamer 407, were designed and compared for spreadability and leakage. Dimethyl fumarate release and diffusion kinetics were evaluated in vitro by Franz cells. The antiviral activity against HSV-1 was tested by plaque reduction assay in Vero and HRPE monolayer cells, while skin irritation effect was evaluated by patch test on 20 healthy volunteers. The lower drug concentration was selected, resulting in smaller and longer stable vesicles, mainly characterized by a multilamellar organization. Dimethyl fumarate entrapment in ethosome was $91\%$ w/w, suggesting an almost total recovery of the drug in the lipid phase. Xanthan gum $0.5\%$, selected to thicken the ethosome dispersion, allowed to control drug release and diffusion. The antiviral effect of dimethyl fumarate loaded in ethosome gel was demonstrated by a reduction in viral growth both 1 h and 4 h post-infection. Moreover, the patch test demonstrated the safety of the ethosomal gel applied on the skin. ## 1. Introduction The human pathogen herpes simplex virus type 1 (HSV-1) can induce frequent recurrent infections especially in the orofacial zone, resulting in the formation of epidermal lesions inside and around the mouth, in nose, as well as in other body districts, such as fingers [1,2,3]. In addition, HSV-1 affects nearly every ocular tissue, including the corneal epithelium, stroma, or endothelium, leading to herpes stromal keratitis and herpes endotheliitis, which can eventually result in loss of vision due to corneal scar formation and neovascularization, while in the case the retina infection, HSV-1 can lead to acute retinal necrosis [4,5,6,7]. Regrettably, HSV-1 is never completely eradicated from the host since it can establish latency, maintaining a relatively quiescent state during which the viral genome is retained without producing virus particles, while it can reactivate, causing recurrent diseases in response to certain stressor, evading host antiviral innate immune responses [8]. The current standard of care in the treatment of HSV-1 infections is based on topical antivirals, among which acyclovir is the first-line drug, beyond topical corticosteroids, in the case of stromal keratitis [9]. However, despite their efficacy, the long-term use of antivirals can induce drug resistant virus strains, while corticosteroids can cause serious side effects. In this respect the treatment of HSV-1 infections represents an unmet need, requiring alternate efficacious drugs [9]. Notably, it has been demonstrated that the course of herpetic keratitis can be improved with fumaric acid ester treatment [10,11]. Dimethyl fumarate (DMF) is a fumaric acid esters derivate that can be considered as a pleiotropic drug, possessing immuno-modulatory, anti-inflammatory, and antioxidant properties that make it efficacious in many conditions, including inflammatory, degenerative, neoplastic, and cardiovascular diseases [12]. Indeed, DMF is used in the systemic treatment of psoriasis with a safe profile for long-term therapy [13,14], it has been approved by the Food and Drug Administration in the USA to treat relapsing-remitting multiple sclerosis [15], has potential applications to limiting HIV disease progression [16], a strong potential in eye pathologies, suggesting its use in several ophthalmological context [17,18], while it is also able to improve wound healing under diabetic conditions [19,20]. Despite the effectiveness, good tolerability and bioavailability of orally administered DMF, some adverse gastrointestinal effects (i.e., diarrhea, vomiting and nausea) have been described [12]. To treat local pathologies, such as HSV-1 infections, affecting orofacial or eye regions, a topical administration should be preferable with respect to the systemic route, due to many pharmacokinetic and pharmacodynamic aspects, including the possibility to use a lower drug dosage, and to deliver the drug directly on the affected tissue, thus reducing systemic side effects [9]. On the other hand, DMF topical use has not been thoroughly investigated, probably due its possible side effects and to a scarce knowledge about its safety profile. Thus, in this respect, cytotoxicity studies, as well as patch tests, are necessary to explore DMF suitability in the treatment of cutaneous, oromucosal or ophthalmic pathologies [12]. In addition, semisolid formulation suitable to deliver DMF directly in the affected body district (e.g., lips or eye) should be specifically designed. A recent ex vivo and in vivo evaluation has demonstrated the possibility to load DMF in a nano-vesicular phosphatidylcholine (PC) based gel suitable for cutaneous administration. The selection of DMF safe dosage enabled to obtain a formulation potentially effective in the treatment of wounds caused by diabetes mellitus or peripheral vascular diseases [20]. The dispersion of PC in water under specific conditions can generate several lyotropic liquid crystalline phases, possessing interesting features as drug delivery systems, such as liposomes or ethosomes (ETHO). ETHO can be described as colloidal dispersions in which the disperse phase is constituted of PC organized as multilamellar vesicles, while the dispersing phase is an ethanol/water mixture (ethanol 20–$45\%$) [21,22]. These nano-vesicular systems can entrap hydrophilic and lipophilic drugs, controlling their release and promoting their transdermal delivery [23]. The presence of PC and ethanol confers softness and thermodynamical stability to the vesicles, while the penetration enhancer properties of ETHO components promote their passage through the biological barriers, allowing the intracellular delivery of the entrapped drugs [24,25,26]. In this respect, in the present study an ETHO based formulation is proposed as a DMF delivery system to treat HSV-1 infections. To elucidate the influence of DMF loading on ETHO, their physico-chemical features were investigated, evaluating size, morphology, entrapment capacity, distribution of functional groups and chemical structure evolution. Moreover, an ETHO hydrogel was specifically designed to obtain a safe biomaterial suitable for topical application. Indeed, due to their consistency and high-water content, hydrogels can be comfortably applied on biological surfaces, such as the skin, eye and mucosae [27]. Notably, some authors demonstrated that various topical infectious diseases can be treated using nanovesicle hydrogels that can longer sustain drug release with respect to the corresponding plain nanovesicles, prolonging the contact time with the biological surface [28,29,30]. Particularly, the treatment of ocular infectious diseases requires frequent eye drop administrations, possibly resulting in drug resistance and also in decrease of patient compliance [31]. Therefore, in order to minimize precorneal drainage and to increase drug bioavailability, thickening agent can be added to eye drop forms [32]. To this aim, both xanthan gum (x-gum) and poloxamer 407 (p-407) have been evaluated, being able to produce biocompatible and biodegradable hydrophilic gels, suitable for administration on skin, lips, and eye. X-gum is a natural heteropolysaccharide, constituted of d-glucose, d-mannose, d-glucuronic acid, acetal-linked pyruvic acid, and O-acetyl repeating units, employed in many fields, including food, cosmetics, and pharmaceutical applications, being able to produce transparent and stable gels upon dispersion in water [33]. P-407 is a non-ionic poly(oxyethylene)poly(oxypropylene) (PEO-PPO) block copolymer, mainly employed in pharmaceutics, due to its thermo-reversible character under dispersion in water [34]. Indeed, p-407 micellar solution is suitable for cutaneous and mucosal administration, due to its liquid state, while in contact with body temperature it assumes a semisolid consistency, allowing to control drug release [35]. The release and permeability profiles of DMF loaded in the selected gel were evaluated in vitro by Franz cells. Furthermore, an in vivo irritation test was conducted to verify ETHO gel safeness, while the antiviral activity of DMF loaded in ETHO or in ETHO gel was studied, evaluating their inhibitory capacity on plaque formation of HSV-1 in Vero (African green monkey kidney) and HRPE (Human Retinal Pigment Epithelial Cells)monolayer cells. ## 2.1. Preparation of Ethosomes ETHO were designed as lipid colloidal systems suitable for DMF delivery through the skin and mucosae. ETHO preparation was simply performed by adding water under stirring to PC ethanol solutions [26]. To load DMF in ETHO, the drug (0.5 or 1 mg/mL) was solubilized in PC ethanol solutions before the addition of water (Table 1). Both loaded and unloaded ETHO appear as whitish homogeneous dispersions, free from separation phenomena. A preformulative study was conducted to evaluate the effect of DMF on vesicle size distribution and stability. ## 2.2.1. Size Distribution Size distribution parameters of ETHO measured by Photon Correlation Spectroscopy (PCS) are reported in Table 2. Mean diameters of the vesicles ranged between 212 and 231 nm, the presence of DMF-induced a slight size increase, particularly DMF 1 mg/mL led to the formation of a low represented population of big vesicles with diameter larger than 4 μm. Dispersity indexes were anyway lower than 0.25. In order to detect the vesicle stability, PCS measurements were performed monthly for 3 months. As depicted in Figure 1, vesicle mean diameters underwent a modest increase, reaching 268 nm in the case of ETHO-DMF1.0. SIR values, calculated to determine the effect of DMF content on size stability (Table 2), suggest a longer stability in the case of ETHO-DMF0.5, followed by ETHO and ETHO-DMF1.0. For this reason, ETHO-DMF0.5 was selected for further experiments. ## 2.2.2. Morphology The morphology of ETHO-DMF0.5 was investigated by cryogenic transmission electron microscopy (cryo-TEM). The micrograph reported in Figure 2 shows the fingerprint like structure typical of ETHO, reflecting the PC supramolecular organization in double layered multilamellar spherical vesicles, as well as unilamellar, and multi vesicular vesicles [21]. ## 2.2.3. Fourier-Transform Infrared Spectroscopy (FTIR) Studies FTIR studies were conducted to structurally characterize ETHO. Indeed, this technique provides significant information on molecular structure, specifically on the chemical functional groups of organic compounds by identifying the vibrational signatures related to specific types of chemical bonds [36]. Particularly, FTIR studies were conducted on ETHO and ETHO-DMF0.5 prepared using D2O instead of H2O. The use of D2O did not affect the macroscopic aspect of samples, nor the size distribution of the vesicles. Indeed, PCS analyses revealed Z Average values of 229.2 ± 15.1 nm and 208.3 ± 7.2 for ETHO and ETHO-DMF0.5, respectively, while dispersity values were 0.26 ± 0.06 and 0.19 ± 0.01, for ETHO and ETHO-DMF0.5 respectively. ## Solvent-Removing Experiments The effect of the presence of DMF molecules on H-bond (hydrogen-bond) formation between PC and solvent (water/ethanol) molecules is shown in Figure 3. In the solvent-removing experiments, the solvent molecules were slowly evaporated from both ETHO and ETHO-DMF0.5 by incubation at 40 °C. The changes in the amount of solvent were monitored by changes in the integrated peak area of νOH (stretching vibrations of OH groups of solvent molecules) band. The alterations in the number of H-bonds formed between C=O and PO2− lipid groups and OH groups of solvent molecules were monitored by changes in the maximum position of νC=O and νasPO2− bands, respectively. A brief interpretation of these bands according to [37,38,39] is the following: the band at 1736 cm−1 is the contribution of stretching vibrations of lipid C=O groups; the band at 1255 cm−1 is attributed to asymmetric stretching vibrations of lipid PO2− groups. The maximum of both bands shifts to the low-wavenumber range with an increase in the number of H-bonds formed between the above-mentioned lipid groups and the OH groups of solvent molecules. As Figure 3A shows, the blue line, which represents the relationship between the maximum position of νC=O band and the integrated peak area of νOH band in ETHO-DMF0.5, is shifted to lower wavenumbers in comparison to the gray line which is attributed to ETHO. This indicates that the incorporation of DMF molecules into ETHO induces an increase in the number of H-bonds formed between lipid C=O groups, located in the interfacial region of PC membranes, and the OH groups of solvent molecules. A similar situation was observed for the polar headgroup region of lipid membranes. In Figure 3B the low-wavenumber shift of the blue line compared to the gray line indicates that the presence of DMF molecules induces a rise in the number of H-bonds formed between lipid PO2− groups, located in the headgroup region of PC membranes, and the OH groups of solvent molecules. A similar effect of DMF molecules on the headgroup region of lipid membranes was observed for egg PC liposomes [40]. In this study, a DMF-induced increase in hydration level of membrane headgroup region was manifested by the low-wavenumber shift of νPO2− band, because of an increase in the number of H-bonds formed between lipid PO2− groups and OH groups of water molecules. ## Temperature-Dependent Studies The structural changes in pure ETHO and ETHO-DMF0.5 triggered by an increase in temperature were studied using FTIR spectroscopy supported by Principal Component Analysis (PCA). PCA calculations were conducted to improve the structural information derived from the spectra of measured samples modulated by temperature. As reported in Figure S1A (Supplementary Materials), a positive loading peak assigned to νOH vibrations, with a maximum centered at lower-wavenumber range, indicates the formation a larger alcohol clusters in ETHO and ETHO-DMF0.5 under lower temperatures. These clusters reorganize into smaller ones under the influence of temperature increase, resulting in a shift of νOH band to higher wavenumbers [41]. ## 2.3. DMF Entrapment Capacity (EC) The EC of DMF in ETHO-DMF0.5 was evaluated separating the lipid vesicular phase from the aqueous one by ultrafiltration. The EC value, obtained by HPLC after disaggregation of the lipid vesicles, was 91.42 ± $2.5\%$, suggesting an almost total association of the drug within the PC vesicles, in agreement with FTIR results. ## 2.4. Preparation and Characterization of ETHO Gel ETHO were thickened by x-gum (0.5, $1\%$ w/w) or p-407 (15, $20\%$ w/w) to obtain semisolid forms suitable for topical application (Table 3). The resulting ETHO gels appeared whitish and homogeneous. The spreadability and leakage of ETHO gels were investigated in vitro to select the more suitable gel for mucosal, ocular, or cutaneous administration. Indeed, the spreadability affects the covering of mucosa (such as the oral one) and lips, or skin area, as well as the extrudability from the container. Moreover, since the spreadability can also influence the gel dosage transfer, it indirectly impacts the therapeutic efficacy [42]. On the other hand, to achieve a sustained effect, the ETHO gels should remain as long as possible on the application site, with minimal leakage. At this regard, the formulation running distance over the vertical plane reflects the leakage. Table 3 and Figure 4 compare spreadability and leakage parameters of ETHO gels and ETHO, taken as control. The highest spreadability and leakage values were found in the case of ETHO and ETHO p-40715, suggesting that p-407 $15\%$ w/w scarcely affected the ETHO liquid consistency. Conversely, p-407 $20\%$ w/w strongly reduced the ETHO spreadability and leakage values. The spreadability and leakage values found in the case of ETHO x-gum1.0 were almost superposable to those obtained by ETHO p-40720. The halving of x-gum to $0.5\%$ w/w scarcely affected leakage, while it increased spreadability value. Therefore, ETHO x-gum0.5 (hereafter named EG) was selected, being characterized by an intermediate spreadability with respect to the other formulations, and a lower leakage, maintaining its position on the slide also 1 h after placement. The dispersion of x-gum $0.5\%$ w/w into ETHO-DMF0.5 resulted in the formation of an ETHO gel (EG-DMF0.5) with the same technological characteristics of the corresponding unloaded one. Table 4 reports acronyms and compositions of the gel formulations employed for further studies. ## 2.5. In Vitro Release Test (IVRT) Franz cells associated to synthetic membranes constituted of PTFE were employed to compare the DMF release kinetics from ETHO-DMF0.5, EG-DMF0.5, G-DMF0.5, and SOL-DMF0.5, a 0.5 mg/mL DMF solution in ethanol:water 30:70, v/v. The PTFE porous synthetic membrane was assembled between the upper and lower compartment of the Franz cell to act as a physical support, to prevent the mixing of donor and receptor phases. As shown in Figure 5, DMF release kinetics followed the order SOL-DMF0.5 > G-DMF0.5 > ETHO-DMF0.5 > EG-DMF0.5. The release rates of DMF, reported in Table 5, were 1.5-fold slower in the case of DMF loaded in ETHO with respect to the drug in solution. As expected, EG-DMF0.5 enabled to control drug release, indeed the release rate of DMF was 2.8- and 1.82-fold slower with respect to SOL-DMF0.5, and ETHO-DMF0.5 respectively. All the differences between RDMF values were statistically significant ($p \leq 0.005$), apart from the difference between ETHO-DMF0.5 and G-DMF0.5. Several plots (zero order plot, first order plot, Higuchi plot and Peppas plots) were drawn in order to understand the DMF release mechanism from ETHO-DMF0.5, EG-DMF0.5, and G-DMF0.5. Equations are reported in S2 in Supplementary Materials, data are shown in Figure S2 and Table 6. From the results, considering the R2 values, in all cases the drug release followed Higuchi order kinetics. The fitting into Korsmeyer–Peppas equation revealed a Fickian diffusion in the case of G-DMF0.5 ($$n = 0$.5$) and a non-Fickian diffusion mechanism in the case of ETHO-DMF0.5 and EG-DMF0.5, (0.5 < n < 1) [43]. ## 2.6. In Vitro Permeation Test (IVPT) The permeability of DMF loaded in EG-DMF0.5 was evaluated and compared to ETHO-DMF0.5, and SOL-DMF0.5, using Franz cell associated to Strat-M®, a synthetic polymeric multimembrane system able to mimic the skin. Strat-M® is made of two polyether sulfone layers overlapped to one polyolefin bottom layer, conferring to the membrane system a skin-like tortuous porous structure [44]. The impregnation with synthetic lipids further imparts to this membrane a skin affinity, recreating hydrophilic and lipophilic compartments, that lend barrier properties. As reported in Figure 6, during the first hour, the DMF profile through ETHO-DMF0.5, G-DMF0.5, and EG-DMF0.5 were superposable, afterwards the fastest kinetic was found in the case of ETHO-DMF0.5, followed by SOL-DMF0.5, G-DMF0.5, and EG-DMF0.5. The DMF permeability profile in the case of SOL-DMF was characterized by a Tlag, absent in the case of ETHO-DMF0.5, G-DMF0.5, and EG-DMF0.5. DMF permeation was typically faster within the first 8 h, afterwards it got slower, reaching a plateau at 24 h, as previously found in a study evaluating the ketoprofen permeation from a semisolid dosage form by Franz cells associated to Strat-M membrane [45]. Jss values were calculated from the linear part of the diffusion profiles (1–5 h). ETHO-DMF0.5 displayed the highest Kp, as reported in Table 7. All the differences between Kp values were statistically significant ($p \leq 0.005$), apart from the difference between ETHO-DMF0.5 and SOL-DMF0.5. ## 2.7. Citotoxicity Evaluation DMF concentration and safety of ETHO-DMF0.5 and EG-DMF0.5 were determined on two different cell lines (Vero and HRPE cells). The cell viability was tested after a 24 h incubation period with the neutral red assay. The neutral red assay determines the accumulation of the neutral red dye in the lysosomes, viable cells can release the incorporated dye in under acidified extracted conditions. In Figure 7 the data show the viability cells at different concentration of SOL-DMF0.5, EG-DMF0.5, and ETHO-DMF0.5. The obtained results demonstrated that (i) the entrapment of DMF in ETHO-DMF0.5 and in EG-DMF0.5 enabled to reduce its toxicity, (ii) HRPE cells were more susceptible than Vero cells. The concentration of DMF 35 μg/mL was selected for further antiviral activity study, being suitable for both cell lines ($70\%$ cell viability). ## 2.8. In Vitro Antiviral Activity The DMF antiviral activity against HSV-1 was tested by plaque reduction assay in Vero and HRPE monolayer cells. Particularly the antiviral activity was evaluated adding SOL-DMF0.5, ETHO-DMF0.5, G-DMF0.5, and EG-DMF0.5 (DMF 35 μg/mL) both simultaneously with the virus at the time of viral absorption, to test the direct action of the formulations on virus, and on cells after viral entry, during the infection. The data in Figure 8 show a significant reduction of plaques when the Vero cells and virus were treated simultaneously with ETHO-DMF0.5 and EG-DMF0.5, with respect to control infected cells (KOS) (Figure 8a). On the other hand, a significant reduction on viral particles was observed under HRPE cell infection with the virus and simultaneous treatment with SOL-DMF0.5, ETHO-DMF0.5, and EG-DMF0.5, with respect to control infected cells (KOS) (Figure 8b). Figure 8: Evaluation of antiviral activity of the treatment during HSV-1 KOS infection on Vero (a) or HRPE cells (b). Data are expressed as mean ± s.d. of three different experiments: p values < 0.001, ** p values < 0.0001.To test the antiviral activity of the substance after viral entry and during the replication, SOL-DMF0.5, ETHO-DMF0.5, G-DMF0.5, and EG-DMF0.5 (DMF 35 μg/mL) were added 1 h or 4 h post-infection (Figure 9). A strong and significant reduction in viral growth was observed in the case of Vero cells treated with ETHO-DMF0.5 and EG-DMF0.5, with respect to untreated control (Figure 9).* Figure 9: Evaluation of antiviral activity of the treatment post HSV-1 KOS infection on Vero (a) or HRPE cells (b). Data are expressed as mean ± s.d. of three different experiments: p values < 0.001, ** p values < 0.0001.* In the case of HRPE cells, the highest antiviral activity was exerted by EG-DMF0.5, both 1 h and 4 h post-infection, with respect to untreated control (Figure 10). ## 2.9. Patch Test A patch test was performed to evaluate the safeness of EG-DMF0.5 for cutaneous application. The gel applied for 48 h under occlusive condition on the healthy skin of 20 volunteers can be classified as “not irritating”, since it resulted in a 0.15 average irritation index, therefore well below the threshold of 0.5. ## 3. Discussion The results of this study suggested the possibility to employ DMF loaded in an ETHO gel in the topical treatment of HSV-1 infections. The purposes of developing an ETHO formulation with respect to a conventional form were (i) to enhance DMF transdermal delivery and ocular absorption, (ii) to prolong DMF antiviral action by improving its interaction with the skin, lips or eye and (iii) to increase DMF bioavailability, reducing the therapeutic dosage, while minimizing toxic side effects. The ETHO excipients were chosen based on previous formulative studies investigating the influence of vesicular nanosystem composition on their size distribution, morphology and stability [46,47]. Both DMF 0.5 and 1.0 mg/mL concentrations were considered on the basis of previous investigations about the nanoencapsulation of DMF in solid lipid nanoparticles and in transethosomes [20,48]. Notably, the loading of DMF 0.5 mg/mL in transethosomes, resulted in stable vesicles, suitable for cutaneous administration. Accordingly, in the present formulative study the entrapment of DMF 0.5 mg/mL in ETHO resulted in stable vesicles, with mean diameter compatible for topical administration on skin, lips and eye, and a homogeneous size distribution. ETHO-DMF0.5 visualized by cryo-TEM revealed the presence of multilamellar bi-layered vesicles due to the PC self-organization in ethanol/water mixture. The ETHO-DMF0.5 structural characterization by FTIR further demonstrated that the presence of DMF changes the occupation of both the interface and the headgroup lipid membrane regions with solvent molecules, possibly stabilizing the vesicle ultrastructure. The multilamellar organization of ETHO-DMF0.5 enabled to sustain DMF release with respect to SOL-DMF0.5, as demonstrated by Franz cell experiments performed with PTFE membrane. On the other hand, the use of STRAT-M®, employed to mimic the biological epithelia, revealed that ETHO-DMF0.5 improved DMF permeability with respect to SOL-DMF0.5, suggesting the capability of ETHO vesicles to enhance DMF diffusion through the membrane. In order to thicken the ETHO liquid dispersion, both p-407 and x-gum were considered. At 15–$25\%$ w/w concentrations, p-407 in water leads to thermo-reversible gels, passing from a low viscosity solution, to a transparent viscous gel above the transition temperature, suitable for administration on skin and mucosae. Different studies report the use of p-407 as thickeners of nanocarrier systems [49,50,51]. Particularly, in a previous study, in order to thicken an ETHO dispersion for cutaneous administration of caffeic acid, we directly added p-407 $15\%$, w/w to the dispersion, resulting in an ETHO gel with semisolid consistency, maintaining the typical supramolecular organization of PC [51]. X-gum is widely employed for pharmaceutical applications, typically 0.5–$1\%$ w/w of x-gum can be employed to confer to a disperse system the suitable viscosity for topical administration, as previously demonstrated by our group [20,52,53]. The formulative study here described enabled to select x-gum $0.5\%$, w/w, as thickener to confer to ETHO-DMF0.5 the appropriate technological features (i.e., spreadability and leakage) for topical administration, prolonging the contact time with the biological surfaces. As expected, the obtained nanovesicle scaffold EG-DMF0.5 enabled to control DMF release with respect to SOL-DMF and ETHO-DMF0.5, as demonstrated by Franz cell associated with PTFE membrane. With regard to the mechanism of DMF release, all formulations followed the Higuchi’s square root model, as previously found in the case of vesicular gels [54]. Indeed, the drug release from nano-vesicular gels was reported to depend on the PC bilayer organization, resulting in a core-shell controlled-release delivery system. Considering the diffusional exponent “n”, which is indicative of the transport mechanism described by Korsmeyer–Peppas, the plain G-DMF0.5. displayed a Fickian release, suggesting that the DMF release could be governed by diffusion through the gel matrix, as previously observed [45]. On the other hand, the presence of vesicles in ETHO-DMF0.5 and EG-DMF0.5 resulted in an anomalous non Fickian release, suggesting a superposition of diffusion and relaxation phenomena [43]. Furthermore, the Kp value of DMF through EG-DMF0.5, evaluated by STRAT-M®, was 1.55-fold lower with respect to ETHO-DMF0.5, while the total amount of DMF diffused after 24 h in the case of EG-DMF0.5 was 1.72- and 1.28-fold lower with respect to ETHO-DMF0.5 and SOL-DMF0.5, respectively. This behavior suggests that both the multilamellar PC organization of the vesicles and the presence of x-gum network in the dispersing phase contribute to sustain the diffusion of DMF through EG-DMF0.5. STRAT-M®, was efficaciously employed as a characterization tool for pre-development stage, nonetheless the transepithelial effect, including skin and mucosa retention studies, will be further evaluated using mucosa and epidermal sheets of murine skin. The plaque reduction assay, conducted after selecting the safe drug concentration, demonstrated the anti-HSV-1 activity of DMF on Vero and HRPE cells. Particularly, ETHO-DMF0.5 and EG-DMF0.5 were able to inhibit the viral infection or growth, both at the time of viral absorption, or 1 h and 4 h after the cell infection. Notably, in the case of Vero cells both ETHO-DMF0.5 and EG-DMF0.5 induced a strong and significant reduction in viral infection and viral growth, while, in the case of HRPE cells the reduction was observed only by EG-DMF0.5. Conversely, G-DMF0.5 antiviral activity was less effective with respect to the ETHO gel, especially 4 h after the cell infection, suggesting that the entrapment of DMF in ETHO vesicle within the x-gum network can sustain its activity, in agreement with release and diffusion data. It is noteworthy that in the case of HRPE cells the ETHO-DMF0.5 dispersion exerted a scarce anti HSV-1 action, while the virucidal activity of its thickened form was almost double (Figure 10), possibly indicating that the presence of x-gum was crucial to promote the DMF contact with the infected cells. The results agree with many studies demonstrating the ability of nanovesicle forms to enhance dermal and ocular drug delivery [24,25,51,55,56]. For instance, a pilot clinical study demonstrated the improved clinical efficacy of acyclovir-loaded ETHO preparation with respect to a commercial acyclovir cream in the treatment of recurrent herpes labialis [57], while very recently a valacyclovir-loaded liposomal formulation has been developed to treat HSV-1 infections [58]. In the case of ophthalmic applications, nano-vesicular systems offer advantages over other delivery systems in promoting intimate contact with corneal and conjunctival surfaces, thus improving the ocular drug absorption [31,56,59,60,61]. For instance ganciclovir-loaded liposomes proposed in the treatment of ocular infections demonstrated a high drug transcorneal permeability, due to an interaction between liposomes and the corneal epithelial surface [56]. Moreover ketoconazole-loaded transethosomes were able to enhance the drug ocular permeation, promoting the antifungal activity, penetrating deeply into the posterior eye segment, without any toxic effects [60]. Beyond the antiviral efficacy of the EG-DMF0.5 formulation developed in the present study, its safeness, demonstrated by patch test assay, suggests its suitability in the treatment of HSV-1 infection, nonetheless further study will be required to confirm the formulation usefulness for ophthalmic administration. ## 4.1. Materials Dimethyl fumarate (dimethyl (E)-but-2-enedioate, DMF), poloxamer 407 (p-407), xanthan gum (x-gum), blue Coomassie, and deuterated water (D2O) were purchased from Merck Life Science S.r.l. ( Milan, Italy). The soybean lecithin (PC) ($90\%$ phosphatidylcholine) was Epikuron 200 from Lucas Meyer (Hamburg, Germany). Polytetrafluoroethylene (PTFE, Whatman®) (pore size 200 nm), and STRAT-M® membranes were purchased from Merck Life Science S.r.l. ( Milan, Italy). Solvents were of HPLC grade, and all other chemicals were of analytical grade. ## 4.2. Ethosome Preparation ETHO preparation was obtained by the cold method [55]. Briefly, PC was solubilized in ethanol (30 mg/mL) under stirring at 750 rpm (IKA RCT basic, IKA®-Werke GmbH & Co. KG, Staufen, Germany), after complete solubilization, bi-distilled water was dropwise added to the PC solution up to a final 70:30 (v/v) water/ethanol ratio. The magnetic stirring was maintained for 30 min. To load DMF, the drug (0.5 or 1 mg/mL) was solubilized in the PC ethanol solution before adding water. ETHO for FTIR experiments were prepared by the same protocol, employing D2O instead of H2O. ## 4.3. Photon Correlation Spectroscopy Vesicle size distribution was measured using a Zetasizer Nano-S90 (Malvern Instr., Malvern, UK) with a 5 mW helium neon laser and a wavelength output of 633 nm. Measurements were performed at 25 °C at a 90° angle and a run time of at least 180 s. Samples were diluted with bi-distilled water in a 1:10 v/v ratio. Data were analyzed using the “CONTIN” method [62]. Measurements were performed thrice for 3 months after ETHO production, on ETHO stored at 22 °C, calculating the mean ± standard deviation (s.d.). The SIR of vesicles was expressed calculating the difference between Z Average mean diameter of ETHO stored for 3 months and Z Average mean diameter of ETHO measured the day after preparation, as follows:[1]SIR=Z Averageday 90−Z Averageday 1Z Averageday 1×100 The statistical differences were evaluated by t student test, GraphPad Prism 9 software (GraphPad Software Inc., San Diego, CA, USA), considering values of $p \leq 0.05$ as statistically significant. ## 4.4. Cryo-Transmission Electron Microscopy For cryo-TEM analyses, samples were vitrified following a method previously reported [51]. Namely, a 2 μL aliquot of sample was put for few seconds on a lacey carbon filmed copper grid (Science Services, München, Germany). After removing most of the liquid by a blotting paper, a thin film stretched over the lace holes was obtained. Vitrification was achieved by rapid immersion of specimen into liquid ethane cooled to approximately 90 K (−180 °C) by liquid nitrogen in a temperature-controlled freezing unit (Leica EMGP, Leica, Germany). The sample preparation procedure was conducted at controlled constant temperature in the Leica EMGP chamber. The vitrified specimen was transferred to a Zeiss/Leo EM922 Omega EFTEM (Zeiss Microscopy GmbH, Jena, Germany) transmission electron microscope using a cryoholder (CT3500, Gatan, Munich, Germany). During the microscopy observations, sample temperature was kept below 100 K. Specimens were examined with reduced doses ≈1000–2000 e/nm2 at 200 kV. Zero-loss filtered images (∆$E = 0$ eV) were recorded by a CCD digital camera (Ultrascan 1000, Gatan, Munich, Germany) and analyzed by a GMS 1.9 software (Gatan, Munich, Germany). ## 4.5. Structural Characterization of ETHO by FTIR In temperature-dependent studies for ETHO and ETHO-DMF0.5 samples, FTIR spectra were collected on a Nicolet Avatar FTIR spectrometer (GMI, Ramsey, MN, USA). The 32 scans were collected for each spectrum at a resolution of 2 cm−1. During a heating cycle, CaF2 windows and a 56-µm spacer to ensure a constant sample thickness were used. Spectra were measured in temperature range from 10 to 80 °C, with intervals of 5 °C. The samples were allowed to equilibrate for 5 min prior to the acquisition of each spectrum. High Stability Automatic Temperature Controller P/N 20120 series (Specac Ltd., Orpington, UK) as an external heating system was used for FTIR measurements. For solvent-removing experiments samples was incubated at 40 °C in order to induce a slow evaporation of solvent molecules from ethosomes dispersions. ATR accessory (PIKE, Madison, WI, USA) with a ZnSe crystal with 10 reflections and a face angle of 45° was equipped additionally to the Nicolet Avatar FTIR spectrometer. An F25 Julabo water bath (Julabo, Labrotechnic, GMbH) was used as an external heating system. Prior to performing PCA calculations, the pretreatment steps of FTIR spectra of ethosomes were as follows: [1] the subtraction of the solvent spectrum from the spectrum of the sample under study; [2] noise reduction using the Savitzky–Golay fuction with the smoothing filter was with 17-point window and a polynomial of order 2; [3] a baseline correction process with a linear function; and [4] application of standard normal variate (SNV) normalization and mean center processes to the analytical data. These spectral pretreatments and PCA calculations were accomplished using version 8.0 of the GRAMS/32 AI software (Galactic Industries Corporation, Thermo Scientific, Warsaw, Poland) and version 6.1 of the PLS Toolbox (Eigenvector Research, ICN, Wenatchee, WA, USA) for the Matlab R2009a software (The MathWorks Inc., Natick, MA, USA). ## 4.6. Evaluation of DMF EC in Ethosome The EC of DMF in ETHO-DMF0.5 was determined by ultrafiltration, 24 h after preparation using a centrifugal filter device (Microcon centrifugal filter unit YM-10 membrane, NMWCO 10 kDa, Sigma-Aldrich, St. Louis, MO, USA) and HPLC analysis as below reported. Namely, 500 μL of DMF-loaded ETHO were poured in the sample reservoir part of the device and subjected to ultrafiltration (Spectrafuge™ 24D Digital Microcentrifuge, Woodbridge, NJ, USA) at 4000 rpm for 15 min. Afterwards, both retentate and filtrate fractions were withdrawn respectively from the sample reservoir part or the vial, and diluted with ethanol (1:10, v/v). Before HPLC analysis, the diluted retentate was stirred for 30 min and filtered by nylon syringe membranes (0.22 μm pore diameter), while the filtrate fraction was analyzed as such. The EC was determined as follows:EC = DMF/TDMF × 100[2] where DMF is the amount of drug measured by HPLC and TDMF is the total amount of DMF employed for ETHO preparation. ## 4.7. Preparation and Characterization of Ethosomal Gels To prepare a viscous gel, alternatively x-gum or p-407 were employed. X-gum (0.5 or $1\%$, w/w) was added to ETHO under magnetic stirring for 30 min, up to complete dispersion. P-407 (15 or $20\%$, w/w) was gradually added to ETHO at 4 °C in an ice bath DMF under magnetic stirring, up to complete dispersion. The resulting ETHO gels (ETHO x-gum0.5, ETHO x-gum1.0, ETHO p-40715 and ETHO p-40720) were tested for spreadability and leakage. ## 4.7.1. Spreadability Studies The spreading capacity of gels (ETHO x-gum0.5, ETHO x-gum1.0, ETHO p-40715 and ETHO p-40720) was evaluated at ambient temperature (25 °C), 24 h after gel preparation [42]. Precisely, 150 mg of gel were placed in the center of a Petri dish (3 cm diameter) and then subjected to pressure by a glass dish carrying a 50 g mass. The diameter of the area occupied by the formulation in a predetermined time (10 s) was measured. The spreadability test was performed three times and the mean values ± standard deviations were calculated using the following equation:S = m × l/t[3] where S is the spreadability of the gel formulation, m is the weight (g) tied on the upper plate, t is the time (10 s), and l is the diameter (cm) of the area occupied by the gel in 10 s under pressure [42]. ## 4.7.2. Leakage Test To test leakage and adhesion of gels (ETHO x-gum0.5, ETHO x-gum1.0, ETHO p-40715 and ETHO p-40720), phosphate buffer pH 7.4 was prepared, afterwards agar ($1.5\%$ w/w) was added and stirred at 95 °C until solubilization. The gels obtained after cooling were then cut to obtain rectangular agar slides. The gels were colored for the leakage test by dissolving blue coomassie ($0.05\%$ w/w), afterwards, 50 mg of colored formulations were placed onto one end of agar slide placed in a Petri plate. The Petri plate was vertically put at an angle of 90° on one of the inner walls of a transparent box, maintained at 37 °C ± 1 °C. The running distance along the slide was measured 10 s after the gel placement. Gel leakage, expressed as percentage of the difference between the total length of the agar slide and the running distance of the gel, was measured three times, and the mean values ± standard deviations were calculated. ## 4.8. Franz Cell Diffusion Experiments Franz cells (orifice diameter 0.9 cm; PermeGear Inc., Hellertown, PA, USA) were employed for IVRT and for IVPT. Notably PTFE membranes were used for IVRT, while STRAT-M® membranes were employed for IVPT [20]. Both for IVRT and IVPT, samples of dried membranes were rehydrated by immersion in ethanol/water 50:50, v/v for 1 h, before assembling in Franz-type diffusion cells. The receptor compartment of the cell contained 5 mL of ethanol:water 50:50, v/v in order to assure sink conditions [63], stirred at 500 rpm by a magnetic bar and thermostated at 32 ± 1 °C during the experiments [64]. Five hundred microliters of DMF-loaded ETHO (ETHO-DMF0.5), DMF ethanolic solution (ethanol:water 30:70, v/v) (DMF 0.5 mg/mL) (SOL-DMF0.5), or ETHO gel (EG-DMF0.5), were placed on the membrane surface in the donor compartment that was afterwards sealed to avoid evaporation. At predetermined time intervals comprised between 1 and 24 h, samples (0.5 mL) of receptor phase were withdrawn and analyzed by HPLC to evaluate the DMF content. Each removed sample was replaced with an equal volume of simple receptor phase. The DMF concentrations were determined six times in independent experiments, the mean values ± s.d. were calculated. ## 4.8.1. In Vitro Release Test (IVRT) For data analysis, in the case of IVRT, DMF amount (μg/cm2) was plotted as a function of the square root of time [63]. To compare the DMF release kinetics from the different dispersions, the following parameters were evaluated: “RDMF” the slope of the cumulative amount of DMF released versus the square root of time; lag-time “Tlag” extrapolated from the intercept of the release profile with x-axis; and “ADMF” the cumulative amount of DMF released at the last sampling time (8 h). The release kinetics parameters were evaluated by mathematical models (KinetDS, Aleksander Mendyk) [43], as reported in S2. Both R2 and n values were taken as reference, considering 0.5 as index of Fickian diffusion model, and 0.5 < n < 1.0 of non Fickian one. ## 4.8.2. In Vitro Skin Permeation Test (IVPT) In the case of IVPT, to analyze data, Fick’s law was considered since it describes the steady-state permeation through the skin, assuming that, under sink conditions, drug concentration in the receptor compartment is negligible with respect to that in the donor compartment [39]. The steady-state flux of drug per unit area corresponds to the slope of the linear part of the curves “Jss” (µg/cm2/h) [65]. DMF permeability coefficients “Kp” values were calculated considering the steady-state portion of DMF cumulative penetration profiles versus time. Kp was calculated according to Equation [4]:Kp = Jss/Cd[4] where Cd is the drug concentration in the donor compartment. ## 4.9. HPLC Analysis HPLC analyses were performed using Perkin Elmer, Series 200 HPLC Systems equipped with a micro-pump, an auto sampler, and an UV-detector operating at 216 nm. A stainless-steel C-18 reverse-phase column (15 × 0.46 cm) packed with 5 μm particles (Hypersil BDS C18 Thermo Fisher Scientific S.p. A., Milan, Italy) was eluted at a flow rate of 1 mL/min with a mobile phase containing acetonitrile/water 40:60 v/v. ## 4.10.1. Cell Culture Vero and HRPE were cultivated in Eagle’s minimum essential medium (DMEM), or DMEM-F12 supplemented with $10\%$ fetal bovine serum (FBS), 100 mg/mL penicillin and 100 mg/mL streptomycin, incubated at 37 °C under $5\%$ CO2 in an incubator. Cells were seed at 5 × 105 per well in a six- well plate, 24 h prior to plaque assay [66]. ## 4.10.2. Cytotoxicity Cell viability was determined by neutral red uptake assay. The cells were seeded in triplicate in a 96-well plate at a density of 20 × 103 in 100 mL of DMEM high glucose and $10\%$ FCS medium for Vero cell line and DMEM-F12 with $10\%$ FBS for HRPE cells. The next day the cells were treated with different concentrations (12.5, 20, 25, 30, 35, 40, 45, 50 μg/μL) of DMF solution or ETHO-DMF0.5 or EG-DMF0.5. After 24 h the medium was removed from the 96-well plates, the cells were gently rinsed with phosphate buffered saline (PBS), 250 μL neutral red (NR) dye medium was added to the wells (25 μg/mL NR concentration), and the plates were incubated (37 ± 1 °C, 90 ± $5\%$ humidity, and 5.0 ± $1\%$ CO2/air) for three hours. After incubation, the NR medium was removed, the cells were rinsed with PBS, and 150 μL of a desorbed solution ($50\%$ bi-distilled water, $49\%$ ethanol, $1\%$ acetic acid) was applied. The plates were shaken on a microtiter plate shaker for 45 min to extract NR from the cells and form a homogeneous solution. The absorption (i.e., OD measurement) of the resulting-colored solution was measured (within 60 min of adding the desorb solution) at 540 nm in a spectrophotometric microtiter plate reader. ## 4.10.3. Herpes Virus Stock Generation Vero cells 2 × 107 in 10–20 mL of cell culture medium were seeded into a 175 cm2 tissue culture flask and incubate overnight at 37 °C in a humidified $95\%$ air-$5\%$ CO2 incubator. Vero cells were infected with herpes simplex virus strain KOS, at a multiplicity of infection (M.O.I) of 0.01. The cells were incubated for 1 h at 35 °C to allow adsorption of the virus to the cells. The flasks were rocked every 15 min to evenly distribute the inoculum. The virus inoculum was aspirated, and cell culture medium was added up to a final volume of 10–20 mL per flask. Infected cells were incubated at 35 °C in a humidified $95\%$ air-$5\%$ CO2 incubator for 36–48 h, until complete cytopathic effect (CPE) was reached. Cells and supernatant were collected, cells were removed by low-speed centrifugation, and the supernatant was centrifugated a 20,000 min−1 for 30 min. The obtained pellet was resuspended in 1 mL of medium, aliquoted, and kept at −80 °C until use [66]. ## 4.10.4. Titration of Virus by Plaque Assay The viral preparation was titrated on Vero cells. One day prior to titration, 6-well tissue culture plates with 0.5 × 106 Vero cells per well were prepared. The virus was thawed on ice and sonicated for a few seconds prior to infection, to separate virus particles. A series of ten-fold dilutions (10−2–10−10) of the virus stock in 1 mL cell culture medium without serum in Eppendorf tubes were prepared and added to each well containing cells. The cells were incubated for 1 h at 35 °C, to allow adsorption of the virus to the cells. After 1 h of infection, the viral inoculum was removed, and the monolayer was overlayed with 3 mL of $1\%$ methylcellulose (Sigma-Aldrich). The plates were incubated for 3–4 days until well-defined plaques were visible. The methylcellulose medium was removed from the wells and stained for 10–20 min with 2 mL of crystal violet staining solution to fix the cells and the virus. The number of plaques was counted, the average for each dilution ($$n = 3$$) was determined and multiplied by 10 to the power of the dilution to obtain the number of plaque forming units per mL (PFU/mL) [66]. ## 4.10.5. Antiviral Activity Assay The inhibition of virus replication was measured by plaque assay. HSV-1 strain KOS 1 × 105 pfu/mL and cells were incubated with SOL-DMF0.5, ETHO-DMF0.5, G-DMF0.5, and EG-DMF0.5 at the time of infection. Alternatively, the same forms were added 1 h and 4 h after cell infection. The positive control were cells infected with the virus and medium. Twenty-four h post-infection the medium and cells of each sample were collected and titrated. After 1 h at 35 °C to allow viral adsorption, the plates were washed and the medium replaced with 3 mL of $1\%$ methylcellulose, to prevent the formation of secondary plaques, and incubated for 3–4 days at 35 °C until the appearance of lysis plaques. Afterwards cells were fixed and stained with a $1\%$ solution of crystal violet to determine the number of plaques [66]. The antiviral activity was evaluated as plaque reduction with respect to control cells. ## 4.11. Statistical Analysis All experiments were repeated 3–6 times and statistical values were expressed as the mean ± standard deviation (SD). For all data analysis, GraphPad Prism 9 software (GraphPad Software Inc., San Diego, CA, USA) was used. Values of $p \leq 0.05$ were considered statistically significant. ## 4.12. Patch Test An in vivo irritation test was performed to evaluate the effect of EG-DMF0.5 applied in single dose on the intact human skin. The occlusive patch test was conducted at the Cosmetology Center of the University of Ferrara, following the basic criteria of the protocols for the skin compatibility testing of potentially cutaneous irritant cosmetic ingredients on human volunteers (SCCNFP/$\frac{0245}{99}$). The protocol was approved by the Ethics Committee of the University of Ferrara, Italy (study number: 170583). The test was run on 20 healthy volunteers of both sexes, which gave a written consent to the experimentation. Subjects affected by dermatitis; with history of allergic skin reaction or under anti-inflammatory drug therapy (either steroidal or non-steroidal) were excluded. Ten milligrams of EG-DMF0.5, were posed into aluminum Finn chambers (Bracco, Milan, Italy), and applied onto the skin of the forearm or the back protected by self–sticking tape. Particularly samples were directly applied into the Finn chamber by an insulin syringe, left in contact with the skin surface for 48 h. Skin irritative reactions (erythematous and/or edematous) were evaluated 15 min and 24 h after removing the patch and cleaning the skin from sample residual. Erythematous reactions have been sorted out into three groups, according to the reaction degree: light, clearly visible and moderate/serious erythema. The average irritation index was calculated as the sum of erythema and edema scores and expressed according to a scale considering 0.5 as the threshold above which the product is to be classified as slightly irritating, 2.5–5 as moderately irritating and 5–8 as highly irritating. ## References 1. Everett R.D.. HSV-1 Biology and Life Cycle. Methods Mol. Biol. (2014.0) 1144 1-17. DOI: 10.1007/978-1-4939-0428-0_1 2. Amin I., Vajeeha A., Younas S., Afzal S., Shahid M., Nawaz R., Khan M.U., Idrees M.. HSV-1 Infection: Role of Viral Proteins and Cellular Receptors. Crit. Rev. Eukaryot. Gene Expr. (2019.0) 29 461-469. DOI: 10.1615/CritRevEukaryotGeneExpr.2019025561 3. Arduino P.G., Porter S.. Herpes Simplex Virus Type 1 infection: Overview on relevant clinico-pathological features: HSV-1 Literature Review. J. Oral Pathol. Med. (2007.0) 37 107-121. DOI: 10.1111/j.1600-0714.2007.00586.x 4. Li L., Li Y., Li X., Xia Y., Wang E., Gong D., Chen G., Yang L., Zhang K., Zhao Z.. HSV-1 infection and pathogenesis in the tree shrew eye following corneal inoculation. J. NeuroVirol. (2020.0) 26 391-403. DOI: 10.1007/s13365-020-00837-0 5. Knickelbein J.E., Hendricks R.L., Charukamnoetkanok P.. Management of Herpes Simplex Virus Stromal Keratitis: An Evidence-based Review. Surv. Ophthalmol. (2009.0) 54 226-234. DOI: 10.1016/j.survophthal.2008.12.004 6. Tornerup N.R., Fomsgaard A., Nielsen N.V.. HSV-1–induced acute retinal necrosis syndrome presenting with severe inflammatory orbitopathy, proptosis, and optic nerve involvement. Ophthalmology (2000.0) 107 397-401. DOI: 10.1016/S0161-6420(99)00053-6 7. Cruz G.P., Fonseca C., Oliveira J., Da Cunha J.S.. Acute retinal necrosis by herpes simplex virus type 1: An unusual presentation of a primary infection. BMJ Case Rep. (2019.0) 12 e232566. DOI: 10.1136/bcr-2019-232566 8. Zhu H., Zheng C.. The Race between Host Antiviral Innate Immunity and the Immune Evasion Strategies of Herpes Simplex Virus 1. Microbiol. Mol. Biol. Rev. (2020.0) 84 e00099-20. DOI: 10.1128/MMBR.00099-20 9. Arduino P.G., Porter S.. Oral and perioral herpes simplex virus type 1 (HSV-1) infection: Review of its management. Oral Dis. (2006.0) 12 254-270. DOI: 10.1111/j.1601-0825.2006.01202.x 10. Heiligenhaus A., Li H., Schmitz A., Wasmuth S., Bauer D.. Improvement of herpetic stromal keratitis with fumaric acid derivate is associated with systemic induction of T helper 2 cytokines. Clin. Exp. Immunol. (2005.0) 142 180-187. DOI: 10.1111/j.1365-2249.2005.02896.x 11. Heiligenhaus A., Li H., Wasmuth S., Bauer D.. Influence of dimethylfumarate on experimental HSV-1 necrotizing keratitis. Graefe’s Arch. Clin. Exp. Ophthalmol. (2004.0) 242 870-877. DOI: 10.1007/s00417-004-0933-8 12. Matteo P., Federico D., Emanuela M., Giulia R., Tommaso B., Alfredo G., Anna C., Annamaria O.. New and Old Horizons for an Ancient Drug: Pharmacokinetics, Pharmacodynamics, and Clinical Perspectives of Dimethyl Fumarate. Pharmaceutics (2022.0) 14. DOI: 10.3390/pharmaceutics14122732 13. Reszke R., Szepietowski J.C.. A safety evaluation of dimethyl fumarate in moderate-to-severe psoriasis. Expert Opin. Drug Saf. (2020.0) 19 373-380. DOI: 10.1080/14740338.2020.1736553 14. Mrowietz U., Barker J., Boehncke W.-H., Iversen L., Kirby B., Naldi L., Reich K., Tanew A., Van De Kerkhof P., Warren R.. Clinical use of dimethyl fumarate in moderate-to-severe plaque-type psoriasis: A European expert consensus. J. Eur. Acad. Dermatol. Venereol. (2018.0) 32 3-14. DOI: 10.1111/jdv.15218 15. Blair H.A.. Dimethyl Fumarate: A Review in Relapsing-Remitting MS. Drugs (2019.0) 79 1965-1976. DOI: 10.1007/s40265-019-01229-3 16. Gill A.J., Kolson D.L.. Dimethyl fumarate modulation of immune and antioxidant responses: Application to HIV therapy. Crit. Rev. Immunol. (2013.0) 33 307-359. DOI: 10.1615/CritRevImmunol.2013007247 17. Yu J., Li Y., Li Z., Li H., Chen Y., Chen X., Su W., Liang D.. Subconjunctival injections of dimethyl fumarate inhibit lymphangiogenesis and allograft rejection in the rat cornea. Int. Immunopharmacol. (2021.0) 96 107580. DOI: 10.1016/j.intimp.2021.107580 18. Manai F., Govoni S., Amadio M.. The Challenge of Dimethyl Fumarate Repurposing in Eye Pathologies. Cells (2022.0) 11. DOI: 10.3390/cells11244061 19. Li Y., Ma F., Li H., Song Y., Zhang H., Jiang Z., Wu H.. Dimethyl fumarate accelerates wound healing under diabetic conditions. J. Mol. Endocrinol. (2018.0) 61 163-172. DOI: 10.1530/JME-18-0102 20. Ferrara F., Benedusi M., Cervellati F., Sguizzato M., Montesi L., Bondi A., Drechsler M., Pula W., Valacchi G., Esposito E.. Dimethyl Fumarate-Loaded Transethosomes: A Formulative Study and Preliminary Ex Vivo and In Vivo Evaluation. Int. J. Mol. Sci. (2022.0) 23. DOI: 10.3390/ijms23158756 21. Natsheh H., Vettorato E., Touitou E.. Ethosomes for Dermal Administration of Natural Active Molecules. Curr. Pharm. Des. (2019.0) 25 2338-2348. DOI: 10.2174/1381612825666190716095826 22. Touitou E., Natsheh H.. Topical Administration of Drugs Incorporated in Carriers Containing Phospholipid Soft Vesicles for the Treatment of Skin Medical Conditions. Pharmaceutics (2021.0) 13. DOI: 10.3390/pharmaceutics13122129 23. Esposito E., Calderan L., Galvan A., Cappellozza E., Drechsler M., Mariani P., Pepe A., Sguizzato M., Vigato E., Pozza E.D.. Ex Vivo Evaluation of Ethosomes and Transethosomes Applied on Human Skin: A Comparative Study. Int. J. Mol. Sci. (2022.0) 23. DOI: 10.3390/ijms232315112 24. Shen L.-N., Zhang Y.-T., Wang Q., Xu L., Feng N.. Enhanced in vitro and in vivo skin deposition of apigenin delivered using ethosomes. Int. J. Pharm. (2014.0) 460 280-288. DOI: 10.1016/j.ijpharm.2013.11.017 25. Jain S., Tiwary A.K., Sapra B., Jain N.K.. Formulation and evaluation of ethosomes for transdermal delivery of lamivudine. AAPS PharmSciTech (2007.0) 8 249-257. DOI: 10.1208/pt0804111 26. Ferrara F., Benedusi M., Sguizzato M., Cortesi R., Baldisserotto A., Buzzi R., Valacchi G., Esposito E.. Ethosomes and Transethosomes as Cutaneous Delivery Systems for Quercetin: A Preliminary Study on Melanoma Cells. Pharmaceutics (2022.0) 14. DOI: 10.3390/pharmaceutics14051038 27. Rehman W.U., Asim M., Hussain S., Khan S.A., Khan S.B.. Hydrogel: A Promising Material in Pharmaceutics. Curr. Pharm. Des. (2020.0) 26 5892-5908. DOI: 10.2174/1381612826666201118095523 28. Ballell-Hosa L., González-Mira E., Santana H., Morla-Folch J., Moreno-Masip M., Martínez-Prieto Y., Revuelta A., Di Mauro P.P., Veciana J., Sala S.. DELOS Nanovesicles-Based Hydrogels: An Advanced Formulation for Topical Use. Pharmaceutics (2022.0) 14. DOI: 10.3390/pharmaceutics14010199 29. Thakur K., Sharma G., Singh B., Chhibber S., Katare O.P.. Current State of Nanomedicines in the Treatment of Topical Infectious Disorders. Recent Pat. Anti-Infect. Drug Discov. (2018.0) 13 127-150. DOI: 10.2174/1574891X13666180529103804 30. Qasim M., Lim D.-J., Park H., Na D.. Nanotechnology for Diagnosis and Treatment of Infectious Diseases. J. Nanosci. Nanotechnol. (2014.0) 14 7374-7387. DOI: 10.1166/jnn.2014.9578 31. Reimondez-Troitiño S., Csaba N., Alonso M., de la Fuente M.. Nanotherapies for the treatment of ocular diseases. Eur. J. Pharm. Biopharm. (2015.0) 95 279-293. DOI: 10.1016/j.ejpb.2015.02.019 32. Nsengiyumva E.M., Alexandridis P.. Xanthan gum in aqueous solutions: Fundamentals and applications. Int. J. Biol. Macromol. (2022.0) 216 583-604. DOI: 10.1016/j.ijbiomac.2022.06.189 33. Mishra G.P., Bagui M., Tamboli V., Mitra A.K.. Recent Applications of Liposomes in Ophthalmic Drug Delivery. J. Drug Deliv. (2011.0) 2011 1-14. DOI: 10.1155/2011/863734 34. Dumortier G., Grossiord J.L., Agnely F., Chaumeil J.C.. A Review of Poloxamer 407 Pharmaceutical and Pharmacological Characteristics. Pharm. Res. (2006.0) 23 2709-2728. DOI: 10.1007/s11095-006-9104-4 35. Russo E., Villa C.. Poloxamer Hydrogels for Biomedical Applications. Pharmaceutics (2019.0) 11. DOI: 10.3390/pharmaceutics11120671 36. He X., Liu X., Nie B., Song D.. FTIR and Raman spectroscopy characterization of functional groups in various rank coals. Fuel (2017.0) 206 555-563. DOI: 10.1016/j.fuel.2017.05.101 37. Cieślik-Boczula K., Szwed J., Jaszczyszyn A., Gasiorowski K., Koll A.. Interactions of Dihydrochloride Fluphenazine with DPPC Liposomes: ATR-IR and. J. Phys. Chem. B (2009.0) 113 15495-15502. DOI: 10.1021/jp904805t 38. Cieślik-Boczula K., Koll A.. The effect of 3-pentadecylphenol on DPPC bilayers ATR-IR and 31P NMR studies. Biophys. Chem. (2009.0) 140 51-56. DOI: 10.1016/j.bpc.2008.11.009 39. Ciesik K., Koll A., Grdadolnik J.. Structural characterization of a phenolic lipid and its derivative using vibrational spectroscopy. Vib. Spectrosc. (2006.0) 41 14-20. DOI: 10.1016/j.vibspec.2005.12.004 40. Sydykov B., Oldenhof H., de Oliveira Barros L., Sieme H., Wolkers W.F.. Membrane permeabilization of phosphatidylcholine liposomes induced by cryopreservation and vitrification solutions. Biochim. Biophys. Acta (BBA) Biomembr. (2018.0) 1860 467-474. DOI: 10.1016/j.bbamem.2017.10.031 41. Golub P., Doroshenko I., Pogorelov V., Sablinskas V., Balevicius V., Ceponkus J.. Temperature Evolution of Cluster Structures in Ethanol. Dataset Pap. Phys. (2013.0) 2013 473294. DOI: 10.1155/2013/473294 42. Sicurella M., Sguizzato M., Cortesi R., Huang N., Simelière F., Montesi L., Marconi P., Esposito E.. Mangiferin-Loaded Smart Gels for HSV-1 Treatment. Pharmaceutics (2021.0) 13. DOI: 10.3390/pharmaceutics13091323 43. Siepmann J., Siepmann F.. Modeling of diffusion controlled drug delivery. J. Control Release (2012.0) 161 351-362. DOI: 10.1016/j.jconrel.2011.10.006 44. Haq A., Goodyear B., Ameen D., Joshi V., Michniak-Kohn B.. Strat-M. Int. J. Pharm. (2018.0) 547 432-437. DOI: 10.1016/j.ijpharm.2018.06.012 45. Salamanca C.H., Barrera-Ocampo A., Lasso J.C., Camacho N., Yarce C.J.. Franz Diffusion Cell Approach for Pre-Formulation Characterisation of Ketoprofen Semi-Solid Dosage Forms. Pharmaceutics (2018.0) 10. DOI: 10.3390/pharmaceutics10030148 46. Sguizzato M., Ferrara F., Mariani P., Pepe A., Cortesi R., Huang N., Simelière F., Boldrini P., Baldisserotto A., Valacchi G.. “Plurethosome” as Vesicular System for Cutaneous Administration of Mangiferin: Formulative Study and 3D Skin Tissue Evaluation. Pharmaceutics (2021.0) 13. DOI: 10.3390/pharmaceutics13081124 47. Sguizzato M., Ferrara F., Hallan S., Baldisserotto A., Drechsler M., Malatesta M., Costanzo M., Cortesi R., Puglia C., Valacchi G.. Ethosomes and Transethosomes for Mangiferin Transdermal Delivery. Antioxidants (2021.0) 10. DOI: 10.3390/antiox10050768 48. Esposito E., Cortesi R., Drechsler M., Fan J., Fu B.M., Calderan L., Mannucci S., Boschi F., Nastruzzi C.. Nanoformulations for dimethyl fumarate: Physicochemical characterization and in vitro / in vivo behavior. Eur. J. Pharm. Biopharm. (2017.0) 115 285-296. DOI: 10.1016/j.ejpb.2017.04.011 49. Giuliano E., Paolino D., Fresta M., Cosco D.. Drug-Loaded Biocompatible Nanocarriers Embedded in Poloxamer 407 Hydrogels as Therapeutic Formulations. Medicines (2018.0) 6. DOI: 10.3390/medicines6010007 50. Cristiano M., Mancuso A., Giuliano E., Cosco D., Paolino D., Fresta M.. EtoGel for Intra-Articular Drug Delivery: A New Challenge for Joint Diseases Treatment. J. Funct. Biomater. (2021.0) 12. DOI: 10.3390/jfb12020034 51. Hallan S., Sguizzato M., Mariani P., Cortesi R., Huang N., Simelière F., Marchetti N., Drechsler M., Ruzgas T., Esposito E.. Design and Characterization of Ethosomes for Transdermal Delivery of Caffeic Acid. Pharmaceutics (2020.0) 12. DOI: 10.3390/pharmaceutics12080740 52. Esposito E., Sguizzato M., Bories C., Nastruzzi C., Cortesi R.. Production and Characterization of a Clotrimazole Liposphere Gel for Candidiasis Treatment. Polymers (2018.0) 10. DOI: 10.3390/polym10020160 53. Esposito E., Ravani L., Mariani P., Contado C., Drechsler M., Puglia C., Cortesi R.. Curcumin containing monoolein aqueous dispersions: A preformulative study. Mater. Sci. Eng. C (2013.0) 33 4923-4934. DOI: 10.1016/j.msec.2013.08.017 54. Jain A., Jain S.K.. In vitro release kinetics model fitting of liposomes: An insight. Chem. Phys. Lipids (2016.0) 201 28-40. DOI: 10.1016/j.chemphyslip.2016.10.005 55. Sguizzato M., Mariani P., Spinozzi F., Benedusi M., Cervellati F., Cortesi R., Drechsler M., Prieux R., Valacchi G., Esposito E.. Ethosomes for Coenzyme Q10 Cutaneous Administration: From Design to 3D Skin Tissue Evaluation. Antioxidants (2020.0) 9. DOI: 10.3390/antiox9060485 56. Shen Y., Tu J.. Preparation and ocular pharmacokinetics of ganciclovir liposomes. AAPS J. (2007.0) 9 E371-E377. DOI: 10.1208/aapsj0903044 57. Horwitz E., Pisanty S., Czerninski R., Helser M., Eliav E., Touitou E.. A clinical evaluation of a novel liposomal carrier for acyclovir in the topical treatment of recurrent herpes labialis. Oral Surg. Oral Med. Oral Pathol. Oral Radiol. Endodontol. (1999.0) 87 700-705. DOI: 10.1016/S1079-2104(99)70164-2 58. Mallick A., Sahu R., Nandi G., Dua T.K., Shaw T.K., Dhar A., Kanu A., Paul P.. Development of Liposomal Formulation for Controlled Delivery of Valacyclovir: An In Vitro Study. J. Pharm. Innov. (2023.0). DOI: 10.1007/s12247-022-09706-1 59. Cortesi R., Argnani R., Esposito E., Dalpiaz A., Scatturin A., Bortolotti F., Lufino M., Guerrini R., Cavicchioni G., Incorvaia C.. Cationic liposomes as potential carriers for ocular administration of peptides with anti-herpetic activity. Int. J. Pharm. (2006.0) 317 90-100. DOI: 10.1016/j.ijpharm.2006.02.050 60. Ahmed T., Alzahrani M., Sirwi A., Alhakamy N.. Study the Antifungal and Ocular Permeation of Ketoconazole from Ophthalmic Formulations Containing Trans-Ethosomes Nanoparticles. Pharmaceutics (2021.0) 13. DOI: 10.3390/pharmaceutics13020151 61. Javaid A., Zahra D., Asim A., Javaid N., Ashfaq U.A.. Recent Updates on the Role of Nanoparticles in the Treatment of Viral Diseases. Crit. Rev. Ther. Drug Carr. Syst. (2021.0) 38 75-102. DOI: 10.1615/CritRevTherDrugCarrierSyst.2020034715 62. Pecora R.. Dynamic Light Scattering Measurement of Nanometer Particles in Liquids. J. Nanoparticle Res. (2000.0) 2 123-131. DOI: 10.1023/A:1010067107182 63. Miranda M., Pais A., Cardoso C., Vitorino C.. aQbD as a platform for IVRT method development—A regulatory oriented approach. Int. J. Pharm. (2019.0) 572 118695. DOI: 10.1016/j.ijpharm.2019.118695 64. EMA Quality and Equivalence of Topical Products—Scientific Guideline 65. Le Guyader G., Do B., Vieillard V., Andrieux K., Paul M.. Comparison of the In Vitro and Ex Vivo Permeation of Existing Topical Formulations Used in the Treatment of Facial Angiofibroma and Characterization of the Variations Observed. Pharmaceutics (2020.0) 12. DOI: 10.3390/pharmaceutics12111060 66. Sutter S.O., Marconi P., Meier A.F.. Herpes Simplex Virus Growth, Preparation, and Assay. Methods Mol. Biol. (2019.0) 2060 57-72. DOI: 10.1007/978-1-4939-9814-2_3
--- title: Neurovascular Coupling in Hypertension Is Impaired by IL-17A through Oxidative Stress authors: - Jessica Youwakim - Diane Vallerand - Helene Girouard journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9967204 doi: 10.3390/ijms24043959 license: CC BY 4.0 --- # Neurovascular Coupling in Hypertension Is Impaired by IL-17A through Oxidative Stress ## Abstract Hypertension, a multifactorial chronic inflammatory condition, is an important risk factor for neurovascular and neurodegenerative diseases, including stroke and Alzheimer’s disease. These diseases have been associated with higher concentrations of circulating interleukin (IL)-17A. However, the possible role that IL-17A plays in linking hypertension with neurodegenerative diseases remains to be established. Cerebral blood flow regulation may be the crossroads of these conditions because regulating mechanisms may be altered in hypertension, including neurovascular coupling (NVC), known to participate in the pathogenesis of stroke and Alzheimer’s disease. In the present study, the role of IL-17A on NVC impairment induced by angiotensin (Ang) II in the context of hypertension was examined. Neutralization of IL-17A or specific inhibition of its receptor prevents the NVC impairment ($p \leq 0.05$) and cerebral superoxide anion production ($p \leq 0.05$) induced by Ang II. Chronic administration of IL-17A impairs NVC ($p \leq 0.05$) and increases superoxide anion production. Both effects were prevented with Tempol and NADPH oxidase 2 gene deletion. These findings suggest that IL-17A, through superoxide anion production, is an important mediator of cerebrovascular dysregulation induced by Ang II. This pathway is thus a putative therapeutic target to restore cerebrovascular regulation in hypertension. ## 1. Introduction Hypertension is the most prevalent modifiable risk factor for neurovascular and neurodegenerative diseases, including stroke and Alzheimer’s disease [1]. Although hypertension treatments greatly reduce stroke incidence [2], their impact on cognitive dysfunction is less clear [3]. This underlies the importance of better understanding the mechanisms by which hypertension affects the brain. In hypertensive humans and experimental models of hypertension, important alterations of cerebral blood flow (CBF) regulation, including neurovascular coupling (NVC), have been observed [4,5,6,7,8]. NVC is the dynamic link between neuronal activity and local blood supply. Its importance is of significance since the brain has high energy needs and no energy reserve; thus, slight alterations of this mechanism can negatively impact cerebral protein synthesis and neuronal functions [9]. Understanding the mechanisms underlying NVC impairment in hypertension is crucial to develop preventive approaches to preserve the brain’s health. Angiotensin (Ang) II, a peptide known to be involved in the development of hypertension, impairs NVC independently of blood pressure [5,6,10]. Interestingly, hypertension is now recognized as being a subclinical inflammatory condition, and Ang II is a powerful modulator of the immune system. In an experimental model of hypertension induced by chronic Ang II perfusion in mice, we previously revealed the impact of inflammation on NVC through anti-inflammatory treatments with T regulatory (Treg) lymphocytes (CD4+/CD25+) or interleukin (IL)-10 [4]. These treatments prevent gliosis and reactive oxygen species (ROS) production suggesting that the inflammatory conditions play a role in maintaining high ROS levels. IL-17A was shown to be upregulated in many experimental models of hypertension and in hypertensive humans. In 2010, Madhur et al. reported that the pro-inflammatory cytokine IL-17A, through its IL-17A receptor, is required for the maintenance of hypertension induced by Ang II and that serum IL-17 levels are correlated with blood pressure in humans [11]. The IL-17 cytokine family is composed of six isoforms (going from IL-17A to IL-17F), where the biological function and regulation of IL-17A and IL-17F are best understood. Even though these two cytokines share the strongest sequence homology and both modulate pro-inflammatory responses, IL-17A, but not IL-17F, has been demonstrated to play a role in hypertension-associated end-organ damages for cardiac, vascular, and renal injuries [12,13,14]. In the brain, in a murine model of a high-salt diet, anti-IL-17A treatments prevent reduced resting CBF and impaired NVC [15]. However, whether IL-17A could impair NVC by itself or be involved in the NVC impairment induced by Ang II remains to be investigated. Ang II impairs NVC by activating its AT1 receptor and nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, specifically the NADPH oxidase (NOX) 2 subtype [5]. Interestingly, IL-17 induces vascular inflammation through NOX2-derived ROS production [16]. Thus, ROS production from NOX2 is a possible converging mechanism by which Ang II and IL-17A impair NVC. The hypothesis is that NVC impairment induced by chronic Ang II administration is mediated by IL-17A and the subsequent increase in NOX2-dependent superoxide production. To explore this question, we determined the impact of IL-17A on NVC, as CBF changes in response to whiskers stimulation with laser Doppler flowmetry and ROS production using an IL-17A neutralizing antibody (Ab) and an antagonist against its receptor. After establishing that recombinant (Rb) IL-17A can by itself impair NVC, chronic treatments with the ROS scavenger, Tempol, and NOX2 deletion were tested in mice chronically receiving IL-17A. ## 2.1. Neutralization of IL-17A or Inhibition of Its Receptor Prevents the Ang II-Induced NVC Impairment To examine the involvement of IL-17A on NVC impairment induced by Ang II, injections of a neutralizing IL-17A Ab were administered in mice intraperitoneally (i.p.) every four days concomitantly with Ang II. As previously observed [6], Ang II attenuated CBF increases to 14.2 ± $0.6\%$ in response to whiskers stimulations compared with 18.5 ± $0.8\%$ in sham-operated mice (Figure 1A; $p \leq 0.01$). Chronic administration of the neutralizing IL-17A Ab prevented the NVC impairment induced by Ang II (Figure 1A,D; $p \leq 0.01$) without altering relative resting CBF (Figure 1B). However, it slightly attenuated the increase in systolic blood pressure (SBP) induced by Ang II at days 7 and 14 by 9.7 ± 3.0 and 9.4 ± 2.1 mmHg, respectively (Supplemental Figure S1). IL-17A Ab, on its own, did not elicit changes in cerebrovascular responses to neuronal stimulations (Figure 1A,C) or SBP (Supplemental Figure S1) in control mice. To assess whether IL-17A impairs NVC through its receptor, mice received an IL-17A receptor antagonist (IL-17RA mAB) simultaneously with the chronic systemic administration of Ang II. In this experimental group, CBF increased in response to whiskers stimulations by $13.6\%$ ± 0.6 and 18.3 ± $0.6\%$ in the presence or the absence of Ang II, respectively (Figure 2A; $p \leq 0.01$). Inhibiting IL-17RA prevented the NVC attenuation induced by Ang II (Figure 2A,D; $p \leq 0.01$) without modifying the relative resting CBF (Figure 2B). Nevertheless, inhibition of the IL-17A receptor attenuated the increased SBP induced by Ang II by 8.2 ± 4.7 and 7.9 ± 2.6 mmHg on days 7 and 14, respectively (Supplemental Figure S2). In control mice, the IL-17RA mAB did not elicit changes in cerebrovascular responses (Figure 2A–C) or SBP (Supplemental Figure S2). ## 2.2. Neutralization of IL-17A or Inhibition of Its Receptor Prevents the Superoxide Anion Production Induced by Ang II To determine whether the increase in ROS production induced by Ang II is mediated by IL-17A, we investigated whether IL-17A Ab reduces superoxide anion production induced by Ang II. As shown in Figure 3, the increased production of superoxide anion by Ang II seen in the somatosensory cortex ($$p \leq 0.079$$) and in the hippocampus ($p \leq 0.0001$) was prevented following IL-17A Ab administration. Similarly, the higher production of superoxide anion in the somatosensory cortex ($p \leq 0.01$) and the hippocampus ($p \leq 0.05$) in mice receiving chronic administration of Ang II was prevented by IL-17RA mAB administration (Figure 4). In the control groups, the production of superoxide anion did not change after IL-17A Ab or IL-17RA mAB administration. ## 2.3. Chronic Administration of an IL-17A Recombinant Impairs NVC To demonstrate that IL-17A can, on its own, impair the cerebrovascular response, we evaluated whether the IL-17A Rb impairs NVC. Systemic IL-17A Rb administration reduces CBF increase in a dose-dependent manner (Supplemental Figure S3). At the selected dose of 50 pg/kg/h, IL-17A Rb administration reduced CBF increase in response to whiskers stimulations from 20.0 ± $1.1\%$ in the sham group to 14.1 ± $1.1\%$ (Supplemental Figure S3A; $p \leq 0.01$). As shown in Supplemental Table S1, IL-17A Rb administration led to a comparable plasmatic concentration to the one observed in Ang II hypertensive mice (7.66 ± 0.80 pg/mL, 11.12 ± 2.60 pg/mL, 13.43 ± 3.75 pg/mL in sham, IL-17A Rb, and Ang II, respectively). Interestingly, no change was observed in brain homogenates. ## 2.4. Tempol Treatment or NOX2 Deletion Prevents Superoxide Anion Production and NVC Dysfunction Induced by IL-17A Chronic Ang II administration impairs NVC through NOX2-dependent oxidative stress [5]. Thus, since IL-17A neutralization and IL-17A receptor inhibition prevent NVC impairment and oxidative stress induced by chronic systemic administration of Ang II, we tested whether superoxide anions mediate the NVC impairment induced by IL-17A. We first evaluated the efficiency of Tempol, an antioxidant superoxide scavenger and superoxide dismutase-mimetic, and NOX2 deletion to normalize the superoxide anion production induced by IL-17A. Figure 5 showed a significantly higher production of superoxide anion in the somatosensory cortex ($p \leq 0.0001$) and the hippocampus ($p \leq 0.01$) in mice receiving the IL-17A Rb. In those regions, Tempol prevented this increase without modulating the superoxide levels in the control group. Similarly, NOX2−/− mice that received IL-17A Rb presented a similar level of superoxide anion production in the somatosensory cortex ($p \leq 0.001$) and the hippocampus ($p \leq 0.0001$) compared to the control mice (Figure 6). These results are complementary to those observed in Figure 7, where the disruption of NVC by the IL-17A Rb was prevented by the Tempol treatment ($p \leq 0.05$). Tempol alone did not modulate CBF responses (Figure 7A,C). The relative resting CBF was similar in all groups (Figure 7B). In the same manner, NOX2 deletion prevented NVC impairment induced by IL-17A Rb administration (Figure 8A,C,D; $p \leq 0.05$) without any difference in the laser Doppler perfusion units between the four groups (Figure 8B). Results from Supplemental Figure S4 showed that the deletion of the NOX2 gene did not prevent the increase in SBP observed at day 7 in response to IL-17A Rb administration (151.4 mmHg in C57BL/6 WT vs. 152.0 mmHg in NOX2−/− mice). ## 3. Discussion We tested the hypothesis of whether IL-17A mediates the NVC impairment induced by Ang II through NOX2-derived ROS. The major new findings of this study are that neutralizing IL-17A or its receptor prevents Ang II-induced ROS production and NVC impairment. This was supported by showing ROS increase and NVC impairment following chronic IL-17A administration. A ROS scavenger and NOX2 deletion prevented these effects suggesting that NOX2-derived ROS production is responsible for the NVC impairment induced by this cytokine. Ang II through the Ang II type 1 receptor (AT1R) signaling pathway is an important pro-inflammatory stimulus, triggering the production of cerebral and systemic pro-inflammatory cytokines [4,17,18,19], chemokines [20] and ROS [5,11,21,22,23,24]. A putative role of inflammation in the effect of Ang II on NVC emerged in a study showing that the adaptive transfer of Treg lymphocytes (CD4+/CD25+) or of IL-10 prevents NVC impairment. In that study, Ang II administration increased the production of circulating pro-inflammatory cytokines such as IL-1α, IL-6, IL-17A, TNF-α, and LIF [4]. Inversely, a systemic inflammatory state in mice characterized by higher circulating IL-17A levels and induced by a high-salt diet contributed to NVC impairment [15]. In the present study, neutralizing IL-17A or inhibiting its receptor prevented the NVC impairment observed in an Ang II slow pressor hypertension model. Overall, these results suggest an important contribution of IL-17A in NVC impairment in models of hypertension. A lower increase in blood pressure induced by Ang II was observed in mice receiving IL-17A Ab or IL-17RA mAB [14] or in IL-17−/− mice [11]. Our results confirmed the lower rise in blood pressure in mice receiving these Ab treatments. However, these slight changes in blood pressure most probably do not explain the NVC impairment induced by Ang II since it was previously demonstrated that the impact of Ang II on NVC is independent of its hypertensive effect [5,6,8]. In the periphery, IL-17−/− mice presented preserved vascular functions, decreased superoxide production, and reduced T-cell infiltration in response to Ang II [11]. Therefore, we hypothesized that IL-17A could mediate the Ang II-induced NVC impairment by increasing oxidative stress. Our results confirm an increased superoxide anion production in the brain (somatosensory cortex and hippocampus) induced by Ang II. Interestingly, neutralizing IL-17A or inhibiting its receptor normalizes the superoxide anion production. These results are coherent with the mediating effects of IL-17A on peripheral blood vessels at the same regimen of Ang II administration as observed by Madhur et al. [ 11]. Overall, these findings suggest that IL-17A is involved in the superoxide anion production induced by Ang II, which may be a missing puzzle piece in the mechanism by which Ang II impairs NVC. To demonstrate that IL-17A can, by itself, impair the cerebrovascular response, we tested whether IL-17A Rb impairs NVC. We first showed that IL-17A Rb administration impaired NVC in a dose-dependent manner, where the 50 pg/kg/h dose showed a decrease in CBF in response to whiskers stimulations to a similar extent to the one seen in Ang II-induced hypertensive mice. The chosen concentration of IL-17A Rb seems to correspond to physiopathological levels observed in humans. In hypertensive patients, the concentrations vary between 1.3 pg/mL [25] and 14.5 pg/mL in refractory hypertensive patients [26]. In the present study, plasma IL-17A levels reached 13.4 and 11.1 in the Ang II and IL-17A groups, respectively. It is, nonetheless, worth mentioning that IL-17A levels in the plasma of hypertensive participants substantially vary with the duration of hypertension, the antihypertensive medication, and comorbidities [25]. Furthermore, no study has established the link between IL-17A levels and end-organ damages. Thus, further studies with large clinical cohorts will be necessary to establish the levels of IL-17A associated with cerebrovascular dysfunctions. In the present study, mice receiving Ang II or IL-17A Rb did not show an increased level of brain IL-17A compared to the control group. An absence of IL-17A increase in the brain was also observed in a high-salt diet model of hypertension despite an increase in the circulating IL-17A. Thus, circulating IL-17A may interfere with NVC by acting on cerebral endothelial cells or by communicating with the neurovascular unit from the meninges [15,27,28]; both mechanisms could ultimately lead to higher NOX2-derived superoxide production and cerebrovascular dysfunction [15,28]. In addition to ROS production, IL-17A may cause endothelial dysfunction via Rho-kinase activation [29]. Ang II impairs NVC through an increase in NOX2-derived superoxide anion [30]. Therefore, we investigated the role of oxidative stress in IL-17A-induced NVC dysfunction. In this study, mice were treated with the antioxidant Tempol due to its ability to cross membranes easily and its stronger therapeutic effect compared with other frequently used antioxidants [31]. Tempol prevented the increase in cerebral superoxide anion production in mice receiving IL-17A Rb. This effect was accompanied by a normalized CBF response to neuronal activation. These results suggest that increased superoxide anion production is a key mediator by which IL-17A impairs NVC. Superoxide anion can react with nitric oxide to form the highly reactive oxidant peroxynitrite. This mechanism is also implicated in the alteration of NVC by Ang II [10]. However, the role of peroxynitrite in the IL-17A-induced NVC alteration remains to be established. NADPH oxidase, specifically NOX2, the isoform expressed in cerebral endothelial cells, perivascular macrophages, microglia, and astrocytes, is likely the main source of increased cerebral superoxide anion in Ang II-induced hypertension in mice [5,10,24].Moreover, NOX2−/− mice are protected from the Ang II-induced NVC alteration [5], further confirming the role of oxidative stress on cerebrovascular dysfunctions. We thus investigated the importance of NOX2 on cerebrovascular dysfunction and oxidative stress in response to IL-17A. We showed that IL-17A increases superoxide anion production. This increase was prevented by treatment with Tempol or in NOX2−/− mice. Interestingly, NOX2 expression in the cortex and the hippocampus remained the same in mice receiving chronic Ang II or the IL-17A Rb administration compared to the control group (Supplemental Figure S5). These results suggest that IL-17A increases superoxide anion production by modulating NOX2 activity. NVC impairment caused by IL-17A is also mediated by NOX2 since similar CBF responses to whiskers stimulations were observed in NOX2−/− mice compared with their corresponding wildtypes. This coincides with results observed in mouse aortic vascular smooth cells where IL-17A induces superoxide anion formation through NOX2 activation [16]. Finally, even though peripheral cardiovascular protection is possible, there was no difference in blood pressure between NOX2−/− mice and their wildtypes. This supports previous results where NOX2−/− mice receiving similar doses of Ang II as in the present study (764 ng/kg/min) do not present a lower SBP [32]. In conclusion, we have demonstrated that IL-17A, through superoxide anion production, is an important modulator of NVC impairment induced by Ang II. Altogether, our findings suggest that modulating the immune system and targeting inflammation in hypertension could be a promising approach for reducing cerebrovascular dysfunctions (Figure 9). Given that hypertension and chronic inflammation are important risk factors for stroke, vascular cognitive impairment, and Alzheimer’s disease, the results of this study could open the door for future investigations to examine the influence of the immune system and inflammation on brain degeneration. ## 4.1. Animals The study was approved by the Committee on Ethics of Animal Experiments of the Université de Montréal and performed in accordance with the guidelines of the Canadian Council for Animal Care and by the ARRIVE (Animal Research: Reporting of In Vivo Experiments). Ten- to twelve-weeks-old C57BL/6 male mice from Charles River Laboratories (Saint-Constant, Qc, Canada) were individually housed in a temperature-controlled room with ad libitum access to water and a standard protein rodent diet (Tekland global $18\%$ protein rodent diet). Ten-weeks-old C57BL/6 male mice with a targeted genetic deletion of NOX2 (B6.129S-Cybbtm1Din/J; stock No: 002365) and their controls were obtained from Jackson Laboratory (Bar Harbord, ME, USA). Given that the female mice are protected from the deleterious effects of Ang II on cerebrovascular functions [23], only male mice were used in this study. Following acclimatization, animals were randomly assigned to experimental groups. ## 4.2. Drugs Administration Osmotic minipumps (model 1002; Alzet, Cupertino, CA, USA) containing Ang II (MilliporeSigma, Oakville, ON, Canada) were subcutaneously implanted under isoflurane anesthesia as previously described [33]. Briefly, mice received bupivacaine hydrochloride (Marcaine; CDMV, Canada, 2 mg/kg s.c.) at the site of the incision before the osmotic pump implantation. Each osmotic pump delivered 600 ng/kg/min of Ang II for 14 days while the control group was sham-operated. NVC impairment induced by Ang II between sham-operated mice and mice receiving saline through an osmotic pump was compared in a separate group of experiments. The mice were injected i.p. every four days with an IL-17A neutralizing antibody (0.5 µg/µL; eBioMM17F3; eBioscience – Thermo Fisher Scientific, Burlington, ON, Canada), a specific IL-17A receptor antagonist (0.5 µg/µL; PL-31280; Amgen, Thousand Oaks, CA, USA), or an immunoglobulin G (IgG) isotype control (0.5 µg/µL Invitrogen – Thermo Fisher Scientific, Burlington, ON, Canada) starting on the day of the implantation (Supplemental Figure S6A). This administration regimen was chosen based on prior studies on murine models of hypertension and atherosclerosis [14,34,35]. In another group of animals, systemic infusion of 50 pg/kg/h of mouse-recombinant IL-17A (IL-17A Rb; 421-ML/CF; R&D system, Minneapolis, MN, USA) for 7 days was achieved via an osmotic minipump (model 1007D; Alzet) (Supplemental Figure S6B). Since no study has previously shown the effect of systemic infusion of IL-17A on NVC, a dose-response curve of IL-17A Rb on cerebrovascular responses was assessed. IL-17A Rb administration has shown a dose-dependent effect on CBF in response to whiskers stimulations. The 50 pg/kg/h dose was chosen because it showed a decrease in cerebrovascular response to the level seen in Ang II-induced hypertensive mice (Supplemental Figure S3). A subgroup of C57BL/6 mice was simultaneously treated with Tempol (4-hydroxy-TEMPO; Millipore Sigma,Oakville, ON, Canada; 1 mmol/L) dissolved in drinking water or with its vehicle (regular drinking water). Treatment with Tempol started 2 days before the osmotic pump implantation and ended at the time of sacrifice (one week after surgery) (Supplemental Figure S6C). ## 4.3. Systolic Blood Pressure Monitoring SBP was monitored in awake mice using tail-cuff plethysmography (Kent Scientific Corp, Torrington, CT, USA). Mice were warmed on a heating pad preheated at 37 °C for ten minutes before and during blood pressure recordings. Animals were habituated to the procedure three days before blood pressure assessment. Right before the implantation of osmotic minipumps (day 0) and weekly until the NVC analysis, ten blood pressure assessments per mouse were measured and averaged. Blood pressure was monitored by the same person at the same period of the day. ## 4.4. Neurovascular Coupling Anesthesia was initiated with isoflurane (induction: $5\%$, maintenance: $2\%$) and maintained with 50 mg/kg of α-chloralose i.p. ( MilliporeSigma, Oakville, ON, Canada) and 750 mg/kg of urethane i.p. ( MilliporeSigma, Oakville, ON, Canada). The depth of anesthesia was checked by testing corneal reflexes and motor responses to tail pinch. Mean blood pressure and blood sample collection for gas assessment were monitored through the catheterization of the femoral artery. Mice were artificially ventilated with a nitrogen/oxygen/CO2 mixture through tracheal intubation. Body temperature was maintained at 37 °C throughout the experiment. CBF was monitored with a laser Doppler probe (AD Instruments, Colorado Springs, CO, USA) placed on the thinned skull above the whisker-barrel area of the somatosensory cortex. The flowmeter and blood pressure transducer were connected to a computerized data acquisition system (MacLab; Colorado Springs, CO, USA). Analysis of CBF responses began 30 min after the end of the surgery to allow blood gases to stabilize. Animals with mean arterial blood pressure under 60 mmHg and/or blood gases outside the normal range (pH: 7.35–7.40; pCO2: 33–45; and pO2: 120–140) were eliminated from the study. CBF responses to neuronal activity were evaluated during whiskers stimulations. Three whiskers stimulations sessions of one minute were performed on the contralateral side of the CBF measurement. Three minutes of resting periods were left between each stimulation. CBF values were acquired with the LabChart6 Pro software (v6.1.3, AD Instruments, Colorado Springs, CO, USA). The percentage increase in CBF represents the peak CBF response relative to the resting CBF peak values during the 20 s before stimulations. ## 4.5. Superoxide Anion Production Superoxide anion production was assessed by hydroethidine microfluorography as previously described [36]. Hydroethidine (dihydroethidium) is cell permeable and is oxidized to become the fluorophore ethidium bromide that intercalates in double-stranded DNA [37]. Mice were anesthetized with sodium pentobarbital (100 mg/kg body weight, CDMV, Saint-Hyacinthe, Qc, Canada) and transcardially perfused with Phosphate-buffered saline (PBS) 1X, pH 7.4. Brains were carefully isolated, frozen on dry ice, and stored at −80 °C until further analysis. Frozen brains were cut with a cryostat (20 µm), and brain sections were mounted on slides and stored at −20 °C overnight. The slides were air dried at room temperature for 15 min followed by 15 min on a slide warmer set at 45 °C. The slides were then immersed in a dihydroethidium (DHE) solution (2 µM, MilliporeSigma, Oakville, ON, Canada) dissolved in PBS 1X at 37 °C for 2 min. The slides were rinsed in PBS for 5 min and dried on a slide warmer for 20 min before they were coverslipped with a Fluoromount-G mounting medium (Southern Biotech, Birmingham, AL, USA). Images were acquired in the somatosensory cortex and the hippocampus (average analysis of the lacunosum moleculare (LMol), the dentate gyrus (DG), the cornu ammonis 1 (CA1), and the cornu ammonis 3 (CA3)) using an epifluorescence microscope Leica DM2000 with the same acquisition parameters for all groups. Analysis of relative fluorescence intensity was conducted using the ImageJ software (version 1.53; National Institutes of Health). Briefly, following subtracting the background for all micrographs, the mean fluorescence intensity was measured. A ratio of the mean intensity for each mouse relative to the mean intensity of the daily controls was calculated to avoid possible variability between days of the experiment. As such, DHE results are expressed as a ratio relative to the control group for each set of experiments. ## 4.6. NOX2 Expression Hippocampal and cortex tissue lysates were prepared using a lysis buffer (Tris 50 mM, NP-40 $1\%$, NaCl 137 mM, glycerol $10\%$, MgCl2 5 mmol/L, sodium fluoride 20 mM, sodium pyrophosphate 1 Mm, sodium orthovanadate 1 mM, pH 7.4) complemented with a protease inhibitor EDTA-free tablet (MilliporeSigma, Oakville, ON, Canada). Proteins (25 µg) were loaded and run on polyacrylamide gels ($10\%$) and then transferred onto nitrocellulose membranes (Biorad, Saint-Laurent, Qc, Canada). The transferred proteins were detected using the specific primary antibodies anti-NOX2 (ab129068, Abcam, Toronto, ON, Canada) and Pan-Actin as a loading control (4968S, Cell Signaling Technology, Danvers, MA, USA) at a concentration of 1:5000 in Tris-Buffered Saline-Tween (TBST; Tris 20 mM, NaCl 137 mM, Tween-20 $0.1\%$, pH 7.6) containing $5\%$ skim milk and 1:1000 in TBST containing $5\%$ BSA, respectively. The secondary antibody was an HRP-linked antibody (7074, Cell Signaling Technology, Danvers, MA, USA) used at a concentration of 1:5000 in TBST containing $5\%$ skim milk and 1:10,000 in TBST containing $5\%$ BSA, respectively. Chemiluminescence was used to detect protein expression, and membranes were digitalized using a GE LAS 4000 mini. Band intensities (integrated optical density) were quantified with the ImageJ software (version 1.53; National Institutes of Health). NOX2 results are expressed relative to the Pan-Actin loading control. ## 4.7. Brain Homogenate and Plasmatic IL-17A Levels Brains were homogenized in PBS 1X complemented with a protease inhibitor EDTA-free tablet (MilliporeSigma, Oakville, ON, Canada). Brain homogenates and plasma samples were sent to Eve Technologies Corporation, where a Mouse High Sensitivity T-Helper Cells Custom Assay (Eve Technologies Corporation, Calgary, AB, Canada) was used for the quantitative analysis of mouse IL-17A. ## 4.8. Statistical Analysis Data analysis was performed with GraphPad Prism software (version 7.0, La Jolla, USA), and results are presented as mean ± SEM. CBF responses to whiskers stimulations, resting CBF, superoxide anion production, and SBP analysis were evaluated with an ANOVA for factorial design with repeated measures followed by a Bonferroni post-test for multiple group comparisons. The dose-response effect on CBF increases in response to whiskers stimulations, relative resting CBF, and mean arterial pressure, as well as the NOX2 expression and IL-17A plasmatic levels, were analyzed using a one-way ANOVA followed by a Dunnet’s post-test comparing each group with the Sham group. Significance was set at $p \leq 0.05.$ Sample size per group is presented in the results section as well as in the figure legends. ## References 1. Qiu C., Winblad B., Fratiglioni L.. The age-dependent relation of blood pressure to cognitive function and dementia. Lancet Neurol. (2005) 4 487-499. DOI: 10.1016/S1474-4422(05)70141-1 2. Wang X., Carcel C., Woodward M., Schutte A.E.. Blood Pressure and Stroke: A Review of Sex- and Ethnic/Racial-Specific Attributes to the Epidemiology, Pathophysiology, and Management of Raised Blood Pressure. Stroke (2022) 53 1114-1133. DOI: 10.1161/STROKEAHA.121.035852 3. Girouard H.. Hypertension and the Brain as an End-Organ Target (2016) 1-221. DOI: 10.1007/978-3-319-25616-0 4. Iulita M.F., Duchemin S., Vallerand D., Barhoumi T., Alvarez F., Istomine R., Laurent C., Youwakim J., Paradis P., Arbour N.. CD4(+) Regulatory T Lymphocytes Prevent Impaired Cerebral Blood Flow in Angiotensin II-Induced Hypertension. J. Am. Heart Assoc. (2019) 8 e009372. DOI: 10.1161/JAHA.118.009372 5. Kazama K., Anrather J., Zhou P., Girouard H., Frys K., Milner T.A., Iadecola C.. Angiotensin II impairs neurovascular coupling in neocortex through NADPH oxidase-derived radicals. Circ. Res. (2004) 95 1019-1026. DOI: 10.1161/01.RES.0000148637.85595.c5 6. Capone C., Faraco G., Park L., Cao X., Davisson R.L., Iadecola C.. The cerebrovascular dysfunction induced by slow pressor doses of angiotensin II precedes the development of hypertension. Am. J. Physiol. Heart Circ. Physiol. (2011) 300 H397-H407. DOI: 10.1152/ajpheart.00679.2010 7. Jennings J.R., Muldoon M.F., Ryan C., Price J.C., Greer P., Sutton-Tyrrell K., van der Veen F.M., Meltzer C.C.. Reduced cerebral blood flow response and compensation among patients with untreated hypertension. Neurology (2005) 64 1358-1365. DOI: 10.1212/01.WNL.0000158283.28251.3C 8. Monteiro A., Castro P., Pereira G., Ferreira C., Sorond F., Milstead A., Higgins J.P., Polónia J., Azevedo E.. Neurovascular Coupling Is Impaired in Hypertensive and Diabetic Subjects Without Symptomatic Cerebrovascular Disease. Front. Aging Neurosci. (2021) 13 728007. DOI: 10.3389/fnagi.2021.728007 9. Xie Y., Mies G., Hossmann K.A.. Ischemic threshold of brain protein synthesis after unilateral carotid artery occlusion in gerbils. Stroke (1989) 20 620-626. DOI: 10.1161/01.STR.20.5.620 10. Girouard H., Park L., Anrather J., Zhou P., Iadecola C.. Cerebrovascular nitrosative stress mediates neurovascular and endothelial dysfunction induced by angiotensin II. Arterioscler. Thromb. Vasc. Biol. (2007) 27 303-309. DOI: 10.1161/01.ATV.0000253885.41509.25 11. Madhur M.S., Lob H.E., McCann L.A., Iwakura Y., Blinder Y., Guzik T.J., Harrison D.G.. Interleukin 17 promotes angiotensin II-induced hypertension and vascular dysfunction. Hypertension (2010) 55 500-507. DOI: 10.1161/HYPERTENSIONAHA.109.145094 12. Orejudo M., García-Redondo A.B., Rodrigues-Diez R.R., Rodrigues-Díez R., Santos-Sanchez L., Tejera-Muñoz A., Egido J., Selgas R., Salaices M., Briones A.M.. Interleukin-17A induces vascular remodeling of small arteries and blood pressure elevation. Clin. Sci. (Lond.) (2020) 134 513-527. DOI: 10.1042/CS20190682 13. Du Y.N., Tang X.F., Xu L., Chen W.D., Gao P.J., Han W.Q.. SGK1-FoxO1 Signaling Pathway Mediates Th17/Treg Imbalance and Target Organ Inflammation in Angiotensin II-Induced Hypertension. Front. Physiol. (2018) 9 1581. DOI: 10.3389/fphys.2018.01581 14. Saleh M.A., Norlander A.E., Madhur M.S.. Inhibition of Interleukin 17-A but not Interleukin-17F Signaling Lowers Blood Pressure and Reduces End-organ Inflammation in Angiotensin II-induced Hypertension. JACC Basic. Transl. Sci. (2016) 1 606-616. DOI: 10.1016/j.jacbts.2016.07.009 15. Faraco G., Brea D., Garcia-Bonilla L., Wang G., Racchumi G., Chang H., Buendia I., Santisteban M.M., Segarra S.G., Koizumi K.. Dietary salt promotes neurovascular and cognitive dysfunction through a gut-initiated TH17 response. Nat. Neurosci. (2018) 21 240-249. DOI: 10.1038/s41593-017-0059-z 16. Pietrowski E., Bender B., Huppert J., White R., Luhmann H.J., Kuhlmann C.R.. Pro-inflammatory effects of interleukin-17A on vascular smooth muscle cells involve NAD(P)H- oxidase derived reactive oxygen species. J. Vasc. Res. (2011) 48 52-58. DOI: 10.1159/000317400 17. Shi P., Diez-Freire C., Jun J.Y., Qi Y., Katovich M.J., Li Q., Sriramula S., Francis J., Sumners C., Raizada M.K.. Brain microglial cytokines in neurogenic hypertension. Hypertension (2010) 56 297-303. DOI: 10.1161/HYPERTENSIONAHA.110.150409 18. Shen X.Z., Li Y., Li L., Shah K.H., Bernstein K.E., Lyden P., Shi P.. Microglia participate in neurogenic regulation of hypertension. Hypertension (2015) 66 309-316. DOI: 10.1161/HYPERTENSIONAHA.115.05333 19. Iulita M.F., Vallerand D., Beauvillier M., Haupert N., Ulysse C.A., Gagné A., Vernoux N., Duchemin S., Boily M., Tremblay M.-È.. Differential effect of angiotensin II and blood pressure on hippocampal inflammation in mice. J. Neuroinflammation (2018) 15 62. DOI: 10.1186/s12974-018-1090-z 20. Vital S.A., Terao S., Nagai M., Granger D.N.. Mechanisms underlying the cerebral microvascular responses to angiotensin II-induced hypertension. Microcirculation (2010) 17 641-649. DOI: 10.1111/j.1549-8719.2010.00060.x 21. Silva T.M.D., Broughton B.R.S., Drummond G.R., Sobey C.G., Miller A.A.. Gender Influences Cerebral Vascular Responses to Angiotensin II Through Nox2-Derived Reactive Oxygen Species. Stroke (2009) 40 1091-1097. DOI: 10.1161/STROKEAHA.108.531707 22. Jackman K.A., Miller A.A., Drummond G.R., Sobey C.G.. Importance of NOX1 for angiotensin II-induced cerebrovascular superoxide production and cortical infarct volume following ischemic stroke. Brain Res. (2009) 1286 215-220. DOI: 10.1016/j.brainres.2009.06.056 23. Girouard H., Lessard A., Capone C., Milner T.A., Iadecola C.. The neurovascular dysfunction induced by angiotensin II in the mouse neocortex is sexually dimorphic. Am. J. Physiol. Heart Circ. Physiol. (2008) 294 H156-H163. DOI: 10.1152/ajpheart.01137.2007 24. Girouard H., Park L., Anrather J., Zhou P., Iadecola C.. Angiotensin II attenuates endothelium-dependent responses in the cerebral microcirculation through nox-2-derived radicals. Arterioscler. Thromb. Vasc. Biol. (2006) 26 826-832. DOI: 10.1161/01.ATV.0000205849.22807.6e 25. Simundic T., Jelakovic B., Dzumhur A., Turk T., Sahinovic I., Dobrosevic B., Takac B., Barbic J.. Interleukin 17A and Toll-like Receptor 4 in Patients with Arterial Hypertension. Kidney Blood Press. Res. (2017) 42 99-108. DOI: 10.1159/000471900 26. Wang L., Du Y., Wang K., Xu G., Luo S., He G.. Chronic cerebral hypoperfusion induces memory deficits and facilitates Aβ generation in C57BL/6J mice. Exp. Neurol. (2016) 283 353-364. DOI: 10.1016/j.expneurol.2016.07.006 27. Cipollini V., Anrather J., Orzi F., Iadecola C.. Th17 and Cognitive Impairment: Possible Mechanisms of Action. Front. Neuroanat. (2019) 13 95. DOI: 10.3389/fnana.2019.00095 28. Kierdorf K., Masuda T., Jordão M.J.C., Prinz M.. Macrophages at CNS interfaces: Ontogeny and function in health and disease. Nat. Rev. Neurosci. (2019) 20 547-562. DOI: 10.1038/s41583-019-0201-x 29. Nguyen H., Chiasson V.L., Chatterjee P., Kopriva S.E., Young K.J., Mitchell B.M.. Interleukin-17 causes Rho-kinase-mediated endothelial dysfunction and hypertension. Cardiovasc. Res. (2013) 97 696-704. DOI: 10.1093/cvr/cvs422 30. Huppert J., Closhen D., Croxford A., White R., Kulig P., Pietrowski E., Bechmann I., Becher B., Luhmann H.J., Waisman A.. Cellular mechanisms of IL-17-induced blood-brain barrier disruption. FASEB j. (2010) 24 1023-1034. DOI: 10.1096/fj.09-141978 31. Luo Z., Chen Y., Chen S., Welch W.J., Andresen B.T., Jose P.A., Wilcox C.S.. Comparison of inhibitors of superoxide generation in vascular smooth muscle cells. Br. J. Pharmacol. (2009) 157 935-943. DOI: 10.1111/j.1476-5381.2009.00259.x 32. Murdoch C.E., Alom-Ruiz S.P., Wang M., Zhang M., Walker S., Yu B., Brewer A., Shah A.M.. Role of endothelial Nox2 NADPH oxidase in angiotensin II-induced hypertension and vasomotor dysfunction. Basic. Res. Cardiol. (2011) 106 527-538. DOI: 10.1007/s00395-011-0179-7 33. Duchemin S., Belanger E., Wu R., Ferland G., Girouard H.. Chronic perfusion of angiotensin II causes cognitive dysfunctions and anxiety in mice. Physiol. Behav. (2013) 109 63-68. DOI: 10.1016/j.physbeh.2012.10.005 34. Madhur M.S., Funt S.A., Li L., Vinh A., Chen W., Lob H.E., Iwakura Y., Blinder Y., Rahman A., Quyyumi A.A.. Role of interleukin 17 in inflammation, atherosclerosis, and vascular function in apolipoprotein e-deficient mice. Arterioscler. Thromb. Vasc. Biol. (2011) 31 1565-1572. DOI: 10.1161/ATVBAHA.111.227629 35. Cheng X., Taleb S., Wang J., Tang T.T., Chen J., Gao X.L., Yao R., Xie J.J., Yu X., Xia N.. Inhibition of IL-17A in atherosclerosis. Atherosclerosis (2011) 215 471-474. DOI: 10.1016/j.atherosclerosis.2010.12.034 36. Sadekova N., Vallerand D., Guevara E., Lesage F., Girouard H.. Carotid calcification in mice: A new model to study the effects of arterial stiffness on the brain. J. Am. Heart Assoc. (2013) 2 e000224. DOI: 10.1161/JAHA.113.000224 37. Münzel T., Afanas’ev I.B., Kleschyov A.L., Harrison D.G.. Detection of superoxide in vascular tissue. Arterioscler. Thromb. Vasc. Biol. (2002) 22 1761-1768. DOI: 10.1161/01.ATV.0000034022.11764.EC
--- title: Relation of Aortic Waveforms with Gut Hormones following Continuous and Interval Exercise among Older Adults with Prediabetes authors: - Daniel J. Battillo - Steven K. Malin journal: Metabolites year: 2023 pmcid: PMC9967213 doi: 10.3390/metabo13020137 license: CC BY 4.0 --- # Relation of Aortic Waveforms with Gut Hormones following Continuous and Interval Exercise among Older Adults with Prediabetes ## Abstract Prediabetes raises cardiovascular disease risk, in part through elevated aortic waveforms. While insulin is a vasodilatory hormone, the gut hormone relation to aortic waveforms is less clear. We hypothesized that exercise, independent of intensity, would favor aortic waveforms in relation to gut hormones. Older adults (61.3 ± 1.5 yr; 33.2 ± 1.1 kg/m2) with prediabetes (ADA criteria) were randomized to undertake 60 min of work-matched continuous (CONT, $$n = 14$$) or interval (INT, $$n = 14$$) exercise for 2 wks. During a 180 min 75-g OGTT, a number of aortic waveforms (applanation tonometry) were assessed: the augmentation pressure (AP) and index (AIx75), brachial (bBP) and central blood pressure (cBP), pulse pressure (bPP and cPP), pulse pressure amplification (PPA), and forward (Pf) and backward pressure (Pb) waveforms. Acylated-ghrelin (AG), des-acylated ghrelin (dAG), GIP, and GLP-1active were measured, and correlations were co-varied for insulin. Independent of intensity, exercise increased VO2peak ($$p \leq 0.01$$) and PPA120min ($$p \leq 0.01$$) and reduced weight ($p \leq 0.01$), as well as AP120min ($$p \leq 0.02$$) and AIx75120min ($p \leq 0.01$). CONT lowered bSBP ($p \leq 0.02$) and bDBP ($p \leq 0.02$) tAUC180min more than INT. There were decreases dAG0min related to Pb120min ($r = 0.47$, $$p \leq 0.03$$), cPP120min ($r = 0.48$, $$p \leq 0.02$$), and AP120min ($r = 0.46$, $$p \leq 0.02$$). Declines in AG tAUC60min correlated with lower Pb120min ($r = 0.47$, $$p \leq 0.03$$) and cPP120min ($r = 0.49$, $$p \leq 0.02$$) were also found. GLP-1active 0min was reduced associated with lowered AP180min ($r = 0.49$, $$p \leq 0.02$$). Thus, while CONT exercise favored blood pressure, both intensities of exercise improved aortic waveforms in relation to gut hormones after controlling for insulin. ## 1. Introduction Aortic waveforms are clinically relevant as they can reflect the load on the heart and/or compliance of peripheral vessels [1,2,3,4,5]. While pulse-wave velocity is considered the gold standard for arterial stiffness, the augmentation index (AIx75) is a surrogate measure that enables pulse-wave reflection for the understanding of central and peripheral hemodynamics [6]. AIx75 is thus influenced by central hemodynamics coupled with the peripheral arterial tree. Pressure waves in the aorta can be separated into forward (Pf) and backward waves (Pb). Pf is generated mainly by left ventricular contraction and pulse-wave velocity, while backward pressure (Pb) is caused by the reflection of the Pf back toward the heart due to varying characteristics of the vascular walls [4,5]. This is clinically relevant towards understanding pulsatile components that influence central (aortic) and peripheral (brachial) blood pressure, leading to CVD risk among older adults [7]. To date, though, most research on aortic waveform components has focused on fasting measures and have not considered the post-prandial state. This is physiologically important as the post-prandial state is considered a stronger predictor of cardiovascular disease (CVD) than the fasting state alone [8]. Insulin is the prevailing post-prandial hormone secreted from pancreatic beta-cells in response to carbohydrate and protein. It plays a key role in maintaining blood glucose levels and promoting arterial compliance of blood vessels [9]. Arterial compliance is an important mechanical property that contributes to regulation of blood pressure, flow, and hemodynamic load on the heart [4,10]. Despite insulin acutely lowering AIx75 and pulse wave velocity in healthy individuals [10], there is reduced endothelial responsiveness in some [11,12], but not all [13,14,15], studies of adults with obesity. In turn, this has raised questions on the possibility that other post-prandial hormones could influence vascular function. Ghrelin is often recognized as an appetite-stimulating hormone secreted from oxyntic glands in the stomach. Interestingly, exogenous ghrelin exerts beneficial hemodynamic effects in healthy participants [16], as well as those with congestive heart failure, [17] through, in part, the inhibition of proinflammatory cytokines [18]. In fact, ghrelin administration increased vasodilation in response to acetylcholine via a nitric oxide specific mechanism [19] in people with metabolic syndrome. Glucagon-like peptide (GLP-1), an established incretin with known effects to promote beta-cell insulin secretion and delayed gastric emptying, also increases macro- and micro-vascular dilatory effects with and without insulin [20]. Additionally, glucose-dependent insulinotropic polypeptide (GIP), a small-intestinal K cell derived hormone known to also increase insulin secretion, is noted to raise adipose tissue blood flow during conditions of hyperglycemia and hyperinsulinemia, although these effects may be attenuated in people with obesity [21,22,23]. Aerobic exercise improves blood pressure (BP), lipid profiles, and inflammation, often in the absence of clinically meaningful weight loss [24,25,26]. Furthermore, acute exercise can decrease fasting AIx75 during the immediate post-exercise period [27], although some suggest the acute effect of aerobic exercise on AIx75 may last up to 24 h following the last bout [28]. Lower AIx75 following exercise may be partially attributed to the working muscles promoting reduced vascular resistance via enhancement of nitric oxide [29,30]. We also have shown that short-term interval (INT) exercise training reduces AIx75 during the post-prandial, but not fasting, state in people with obesity [31]. While our later work suggested that insulin may have, in part, contributed to these favorable reductions in post-prandial AIx75, independent of exercise intensity, it is unknown whether ghrelin and/or incretins play a role in aortic waveform changes following exercise, independent of insulin. Furthermore, we did not determine if changes in AIx75 were depicted by improved Pf or Pb waveforms, nor did we assess central compared with brachial blood pressures to discern pulse pressure amplification—a CVD mortality risk factor [32]. Thus, we tested herein the hypothesis that exercise, independent of intensity, would reduce post-prandial aortic waveforms in older adults with prediabetes. We further hypothesized this change would relate to changes in gut hormones implicated in regulating vascular function. ## 2.1. Participants Twenty-eight older adults with obesity (61.3 ± 1.5 yr; 33.2 ± 1.1 kg/m2, Table 1) were recruited via advertisements. Some of the AIx75 related outcomes, gut hormones, and cardiometabolic data were previously reported [31,33,34]. Participants were screened for prediabetes based on the American Diabetes Association criteria (75g OGTT) and had to have impaired fasting glucose (100–125 mg/dL), impaired glucose tolerance (2-hr plasma glucose 140–200 mg/dL), and/or elevated HbA1c (5.7–$6.4\%$). Participants were non-smoking, sedentary (exercise < 60 min/wk), and weight stable over the prior six months (≤2 kg variation). People were excluded if they had chronic disease (i.e., renal, hepatic, cardiovascular, etc.) or were on anti-diabetic or weight-inducing medications (e.g., GLP-1 agonists, sulfonylureas, biguanides, etc.). All participants underwent a physical exam and stress test with an electrocardiogram to ensure their health status. Individuals provided written and verbal informed consent before participation as approved by the University of Virginia Institutional Review Board (IRB-HSR #17822). ## 2.2. Aerobic Fitness and Body Mass Peak oxygen consumption (VO2peak) and heart rate (HRpeak) were determined using a continuous incremental cycle ergometer exercise test and indirect calorimetry (Carefusion, Vmax Encore, Yorba Linda, CA, USA) as described previously [31,33,34]. Body weight was measured to the nearest 0.01 kg on a digital scale while height was measured with a stadiometer to assess body mass index (BMI). ## 2.3. Metabolic Control Participants were instructed to refrain from alcohol, caffeine, medication, and strenuous physical activity for 24 h prior to each study visit. Participants were also instructed to consume a diet containing approximately 250 g of carbohydrates during the 24 h period prior to the pre-intervention testing to minimize influence on alterations in insulin secretion and gut hormones. This diet was recorded and replicated on the day before post-intervention testing. Participants were instructed to maintain non-exercise physical activity and habitual diets throughout the intervention. ## 2.4. OGTT Participants reported to the Clinical Research Unit (CRU) after an approximate 10 h overnight fast. An IV catheter was placed in the antecubital fossa for blood draws to determine glucose and hormonal responses during a 75 g oral glucose load. Blood was collected at 0, 30, and 60 min to capture acylated and des-ghrelin, GLP-1active, and GIP [32,33], while glucose and insulin were additionally recorded at 90, 120, and 180 min. Post-intervention assessments were obtained approximately 24 h after the last training session. ## 2.5. Pulse Waveform Analysis The SphygmoCor XCEL system (AtCor Medical, Itasca, IL, USA) was used to characterize hemodynamic and aortic waveform responses, as described before [24]. In short, this included peripheral systolic (bSBP), diastolic (bDBP) and pulse pressure (bPP), heart rate (HR), central systolic (cSBP), diastolic (cDBP) and pulse pressure (cPP), and the augmentation index (AIx), as well as wave deconvolution aspects of forward (Pf) and backward (Pb) pressure and reflection magnitude (RM). The augmentation index was corrected to a standard HR of 75 bpm (AIx75) using the manufacture’s software. Pulse pressure amplification (PPA) was calculated as a ratio (brachial PP/central PP). All measurements occurred while individuals were resting quietly in a semi-supine position in a temperature-controlled room. A blood pressure cuff was placed on upper arm and measurements were recorded in triplicate over a 10 min period and averaged. tAUC for aortic waveform measures was calculated from the values obtained at 0, 60, 120, and 180 min of the OGTT. ## 2.6. Exercise Training Participants were randomly assigned to either supervised CONT or INT training, utilizing a block design that was stratified by a prediabetes phenotype. Twelve work-matched bouts of cycle ergometry exercise were performed for 60 min/d over thirteen days. CONT exercise was performed at an intensity of $70\%$ HRpeak; whereas INT exercise involved alternating 3 min intervals at $90\%$ HRpeak followed by $50\%$ HRpeak for the 60 min duration. The first 2 exercise sessions, however, were performed at 30 and 45 min, respectively, at the desired intensity to ease participants into the intervention. Ad-libitum water, but no food, was provided to the subjects. Heart rate (Polar Electro, Inc. Woodbury, NY) and rating of perceived exertion (RPE) were monitored throughout exercise to ensure appropriate intensity. Energy expenditure during CONT and INT exercise was calculated using HR-VO2 regression analysis, as previously described [34]. ## 2.7. Biochemical Analysis Plasma glucose was measured immediately after collection using the glucose oxidase method (YSI 2300 STAT Plus, Yellow Springs, OH, USA). Blood samples were collected in chilled vacutainers that contained protease inhibitors. AG and dAG samples contained aprotinin, DPP-IV, and AEBSF (EMD Millipore, Billerica, MA, USA). GLP-1 contained aprotinin and DPP-IV, while insulin contained only aprotinin. Blood was centrifuged at 4 °C for 10 min at 3000 RPM. Following centrifugation, HCl was immediately added to acidify the ghrelin sample. All blood was frozen at −80°C until subsequent analysis. AG and dAG concentrations, as well as GLP-1active and insulin, were determined using an enzyme-linked immunosorbent assay (ELISA), as described before [32]. ## 2.8. Statistical Analysis Data were analyzed using GraphPad Prism version 9 (GraphPad Software, San Diego, CA, USA). Non-normally distributed data were log-transformed for analysis. Baseline differences were assessed using independent samples, two-tailed t-test, while repeated measures analysis of variances (ANOVA) was used to determine group x time differences. Pearson correlations were used to examine relationships, and insulin was used as a co-variate for gut hormones to isolate effects. Statistical significance was accepted as p ≤ 0.05 and data are presented as mean ± SEM. ## 3.1. Participant and Exercise Training Characteristics Independent of intensity, exercise raised VO2peak ($p \leq 0.01$) and decreased BMI ($p \leq 0.01$; Table 1). Exercise session adherence was excellent and similar between CONT and INT (96.2 ± $2.2\%$ vs. 95.6 ± $1.5\%$; $$p \leq 0.83$$). Despite INT having a higher heart rate during training compared with CONT (77.8 ± $1.0\%$ vs. 72.6 ± $1.3\%$; $p \leq 0.01$), there were no significant differences between CONT and INT in RPE (12.5 ± 0.3 vs. 12.0 ± 0.5 a.u.; $$p \leq 0.46$$) or exercise energy expenditure (393.2 ± 16.3 vs. 384.5 ± 18.9 kcal/session; $$p \leq 0.62$$). ## 3.2. Glucose Tolerance and Insulin Although exercise did not reduce fasting glucose, it reduced both 120 min glucose ($$p \leq 0.02$$) and glucose tAUC180min ($$p \leq 0.03$$), independent of intensity (Table 1). Furthermore, fasting insulin was not altered, but insulin tAUC180min was significantly reduced following both exercise intensities ($p \leq 0.01$; Table 2). ## 3.3. Hemodynamics We report that AIx75 tAUC180min was lowered after both INT and CONT exercise ($p \leq 0.01$; Figure 1), independent of heart rate changes in response to exercise ($$p \leq 0.66$$ and $$p \leq 0.94$$, respectively; Table 3). CONT training elicited greater improvements than INT in both bSBP tAUC180min ($$p \leq 0.02$$) and bDBP tAUC180min ($$p \leq 0.04$$; Table 3). However, CONT and INT comparably reduced AP120min ($$p \leq 0.02$$) and increased PPA120min ($$p \leq 0.01$$), although there was no influence on 120 min bSBP ($$p \leq 0.57$$) and cSBP ($$p \leq 0.96$$), or 120 min bDBP ($$p \leq 0.45$$) and cDBP ($$p \leq 0.33$$; Table 3). ## 3.4. Gut Hormones Fasting GIP increased with CONT but decreased after INT ($$p \leq 0.03$$). However, there were no exercise-induced changes to GIP tAUC60min ($$p \leq 0.93$$; Table 2). Furthermore, there was no significant effect of CONT or INT on fasting AG ($$p \leq 0.32$$) or dAG ($$p \leq 0.66$$), nor AG or dAG tAUC60min after OGTT administration ($$p \leq 0.20$$ and $$p \leq 0.72$$, respectively). Additionally, neither exercise intervention altered fasting or tAUC60min GLP-1active ($$p \leq 0.38$$ and $$p \leq 0.73$$, respectively; Table 2). ## 3.5. Correlations Exercise-induced reductions in fasting insulin correlated with lower Pf120min ($r = 0.54$, $$p \leq 0.01$$; Figure 2). Lower Pb120min correlated with declines in dAG0min ($r = 0.47$, $$p \leq 0.03$$) and AG tAUC60min ($r = 0.47$, $$p \leq 0.03$$; Figure 2). Prior to covarying for insulin, however, reductions in Pb120min correlated with neither dAG0min ($r = 0.37$, $$p \leq 0.08$$) nor AG tAUC60min ($r = 0.30$, $$p \leq 0.18$$). Furthermore, reduced GLP-1active 0min was associated with lowered AP180min after covarying for insulin ($r = 0.49$, $$p \leq 0.02$$; Figure 2) but not before ($r = 0.32$, $$p \leq 0.19$$), and increased GLP-1active tAUC60min was associated with decreased Pf180min (r = −0.51, $$p \leq 0.03$$; Figure 2) but not before (r = −0.18, $$p \leq 0.43$$). ## 4. Discussion The primary finding from the present study is exercise, independent of intensity, reduced post-prandial AIx and AP, as well as increased post-prandial PPA in older adults with prediabetes. However, CONT exercise yielded lower blood pressure responses during the OGTT than INT. This contrasts prior work suggesting INT may be better at reducing fasted blood pressure, particularly during the immediate post-exercise period (~1 h) [35]. The exact cause of improved post-prandial blood pressure following CONT and INT is beyond the scope of this work, but we [36] and others have reported that CONT exercise favors increased conduit artery blood flow during an OGTT [37] and/or brachial flow-mediated dilation [38]. As flow-mediated dilation is a non-invasive measure of nitric oxide bioavailability, it is possible that the rhythmic nature of muscle contraction during CONT exercise promoted endothelial function. In either case, few data are available examining aortic waveforms following short-term exercise [39,40,41]. Our current findings of no change in Pf or Pb contrast some prior work displaying reduced Pf and Pb following lower body exercise in the immediate post-exercise period for up to 2 h [41]. However, this latter study was conducted in participants who were young, healthy adults (26.0 ± 3.0 yr), and aortic waveform measures were taken every 20 min up to 2 h after exercise, thereby making comparisons difficult to our older participants (61.3 ± 1.5 yr). Taken together with this immediate post-exercise work, our data suggests these effects are short-lived, potentially since we observed no effect on fasting indices 24 h after the last training session. Another study looking at resistance exercise in healthy adults on aortic waveforms reported that AIx increased 1 h following the bout [40], and other work saw similar increases in AIx 10 min after a bout of resistance exercise, but no changes in central or brachial blood pressures [39]. The mechanisms mediating this contrary response to aerobic exercise are unclear but might relate to upper versus lower body exercise and stimulation of muscle mass. Further investigation is warranted in this area given recent work suggesting cardiac adaptations to aerobic and resistance exercise are unique [42]. In either case, our findings extend upon this exercise work by showing favorable effects in the post-prandial period in older adults with prediabetes. Post-prandial gut hormones have been purported to influence vascular function [43,44,45]. Contrary to our hypothesis, gut hormones alone did not associate with any aortic waveforms measured in this study. This is surprising, given AG and dAG have both been implicated in vasomotor tone and nitric oxide-mediated endothelial function [43]. Indeed, AG and dAG were also both shown to inhibit ET-1-mediated vasoconstriction when applied to artery segments [43]. Additionally, GLP-1 has been shown to enhance muscle microvascular perfusion in healthy humans, as well as increase brachial artery diameter and flow velocity through PKA (protein kinase A)-mediated eNOS activation [45]. Lastly, GIP has been demonstrated to increase blood flow and triglyceride clearance in abdominal adipose tissue of lean humans via the recruitment of capillaries promoting lipoprotein lipase activity on triacylglycerol particles [43]. While much of these data demonstrate favorable effects of gut hormones on the vasculature, direct infusion, rather than oral ingestion, was used. Hence, we looked to expand on this work by testing if exercise would influence aortic waveforms via modulation of gut hormones during an OGTT. In our study, neither fasting or post-prandial AG, dAG, or GLP-1active was altered compared with pre-intervention, and the change in these hormones did not correlate independently with aortic waveform responses. Why these hormones did not change more robustly is difficult to address, but total ghrelin often increases following weight loss greater than 3 kg, which is considerably more than the present study [46]. Another possibility is that gut hormone sensitivity changes may have occurred after exercise training, such that changes in gut hormones were not observed [47]. Indeed, recent work highlights improved GLP-1 sensitivity following endurance exercise in women with obesity. Given we did not measure gut hormone sensitivity, we cannot determine if a lack of change in hormones reflects a more sensitive system [48]. Regardless, our work suggests that two weeks of exercise is capable of improving aortic waveforms and blood pressure in older adults with prediabetes, and the gut hormones measured herein do not appear independently related. In human endothelial cells, insulin binds to the insulin receptor (IR) and tyrosine kinase phosphorylates IRS-1. This phosphorylation leads to the downstream binding and activation of PI3K and Akt. Thereafter, Akt phosphorylates and activates eNOS for nitric oxide production [49]. Nitric oxide promotes the relaxation of smooth muscle cells lining the vessel walls, which ultimately increases perfusion and delivery of glucose and insulin to target tissues [37]. Endogenous insulin is influenced significantly by each of the gut hormones measured in this study. For example, a primary function of GLP-1 and GIP is the promotion of pancreatic insulin secretion [50]. Additionally, ghrelin has been reported to blunt beta-cell insulin secretion [51,52]. Thus, it would be reasonable to expect that changes in ambient insulin concentration during the post-prandial state might influence the relationship between gut hormones and aortic waveforms. Interestingly, dAG, AG, and GLP-1active correlated with changes in aortic waveforms only after covarying for changes in insulin tAUC. Specifically, reductions in fasting dAG and AG tAUC were both associated with lowered Pb120min. This is clinically relevant as lower Pb suggests reduced impedances from the vascular walls producing a partial wave reflection back towards the heart [4,5]. Given, though, that ghrelin infusion has been shown to promote endothelial function, it is interesting that reductions in these hormones were associated with favorable Pb120min results. A possible reason for this relates to ghrelin inducing reduced beta-cell function and/or promoting insulin resistance [53]. In the present study, insulin levels were reduced after both CONT and INT, perhaps suggesting the vasculature became more insulin responsive with less ambient ghrelin in circulation. Furthermore, this may explain the lack of change in post-prandial AG and dAG seen in both CONT and INT, as insulin infusion has been shown to decrease circulating total ghrelin [54]. Alternatively, it remains possible that the interaction between ghrelin and insulin influenced cellular signals (e.g., Akt) to modulate vessel function [55]. Indeed, lower fasting insulin correlated with lower Pf120min. This would be consistent with the reduced left ventricular workload and higher PPA seen with our intervention. In fact, increases in PPA are favorable as they demonstrate central arterial compliance leading to lower cPP relative to bPP [56]. Interestingly, increases in GLP-1active tAUC60min were associated with reduced Pf180min. Consistent with prior work, infusion of GLP-1 into healthy participants improved endothelial function [46]. In turn, better peripheral blood flow may enable greater delivery of insulin to reduce load on the left ventricle [57], which mirrors our reduced post-prandial AIx and AP, independent of heart rate, following exercise. Collectively, insulin appears to be a central post-prandial hormone regulating vascular function following exercise training in older adults with prediabetes. This study has limitations that may impact our interpretations. The present study may be underpowered to detect statistical differences in some vascular outcomes. Based on tAUC data for Pf and Pb, the sample size required to detect an effect with 0.80 power at 0.05 significance was calculated for Pf (delta = 254, standard deviation (SD) = 687, $$n = 59$$) and Pb (delta = 117, SD = 446, $$n = 59$$) to inform future studies examining exercise and pressure waveforms. Interestingly, other work from our lab in women with obesity has similarly reported reductions in AIx75 following 2 weeks of exercise without concurrent changes in Pb or Pf [24]. Why we detect changes in AIx75, as well as blood pressure, is beyond the scope of this study but may be attributable to the software’s sensitivity to detect changes in the indirect measure of the waveforms during analysis and/or the use of an OGTT vs. direct hormone infusion, given insulin infusion has previously been shown to lower Pb in about 20 subjects [58]. A 75 g OGTT, rather than a mixed meal, was used to characterize post-prandial gut hormones. This may limit generalizability of the findings as macronutrients have been demonstrated to affect post-prandial ghrelin. For instance, ghrelin suppression occurs more from protein than carbohydrate or lipid-based meals [59]. Furthermore, food intake sequence has been shown to influence incretin responses. Incretin responses to carbohydrates in a meal are blunted when protein is consumed beforehand [60]. In either case, OGTT and mixed-meal tolerance tests show similar directional post-prandial gut hormone responses [61,62], with some alterations in magnitude of GLP-1 stimulation [63] and ghrelin suppression [64]. Another consideration is that gut hormones were measured at 0, 30, and 60 min of a 75 g OGTT. These limited timepoints may underestimate the effects of exercise on the hormones of interest. However, studies have demonstrated peak suppression of ghrelin, in addition to stimulation of GIP and GLP-1, occurs within the first 60 min of the OGTT [65], suggesting we are likely to depict initial gut hormone responses. However, it is possible that differences in gut hormone clearance rates could influence the vasculature. This study was also completed with an absence of healthy controls. Obesity status itself may blunt GLP-1 secretory responses to aerobic exercise [66], as well as mitigate AG increase following exercise [67]. Moreover, the present study features a modest sample size and primarily white women, highlighting additional attention to diverse groups of people is warranted. A non-exercise control was not included, so the independent effects of exercise may be over-/under-estimated. The study also consisted mostly of older white women, thereby limiting these findings to younger men and women from diverse backgrounds. Lastly, we used aortic waveforms to characterize vascular function. While some suggest AIx may be used as an indicator of arterial stiffness, it is worth nothing that pulse wave velocity (PWV) is considered the better non-invasive measure of arterial stiffness. Thus, we are not able to state whether gut hormones impact arterial stiffness, but instead focused on analysis of changes in aortic load and/or peripheral arterial compliance. In conclusion, two weeks of exercise improved post-prandial aortic waveforms in older adults with prediabetes, independent of intensity. Furthermore, CONT exercise favored reductions in post-prandial blood pressure when compared with INT exercise. While gut hormone changes after exercise training were not independently related to improvements in central hemodynamics, covarying for insulin revealed significant relationships. This observation suggests gut hormones may interact with insulin to influence aortic waveforms in older adults with prediabetes. Therefore, additional studies are necessary to elucidate the underlying pancreatic-gut “cross-talk” mechanism with the vasculature to optimize CVD risk reduction. ## References 1. Palatini P., Casiglia E., Gąsowski J., Głuszek J., Jankowski P., Narkiewicz K., Saladini F., Stolarz-Skrzypek K., Tikhonoff V., Bortel L.V.. Arterial Stiffness, Central Hemodynamics, and Cardiovascular Risk in Hypertension. Vasc. Health Risk Manag. (2011) 7 725. DOI: 10.2147/VHRM.S25270 2. Chester R.C., Gornbein J.A., Hundley W.G., Srikanthan P., Watson K.E., Horwich T.. Reflection Magnitude, a Measure of Arterial Stiffness, Predicts Incident Heart Failure in Men But Not Women: Multi-Ethnic Study of Atherosclerosis (MESA). J. Card. Fail. (2017) 23 353-362. DOI: 10.1016/j.cardfail.2017.01.002 3. Mitchell G.F., Hwang S.-J., Vasan R.S., Larson M.G., Pencina M.J., Hamburg N.M., Vita J.A., Levy D., Benjamin E.J.. Arterial Stiffness and Cardiovascular Events: The Framingham Heart Study. Circulation (2010) 121 505-511. DOI: 10.1161/CIRCULATIONAHA.109.886655 4. Townsend R.R., Wilkinson I.B., Schiffrin E.L., Avolio A.P., Chirinos J.A., Cockcroft J.R., Heffernan K.S., Lakatta E.G., McEniery C., Mitchell G.F.. Recommendations for Improving and Standardizing Vascular Research on Arterial Stiffness. Hypertension (2015) 66 698-722. DOI: 10.1161/HYP.0000000000000033 5. Stock J.M., Chouramanis N.V., Chirinos J.A., Edwards D.G.. Dynamic and Isometric Handgrip Exercise Increases Wave Reflection in Healthy Young Adults. J. Appl. Physiol. (1985) (2020) 129 709-717. DOI: 10.1152/japplphysiol.00281.2020 6. Laurent S.. Surrogate Measures of Arterial Stiffness. Hypertension (2006) 47 325-326. DOI: 10.1161/01.HYP.0000200701.43172.9a 7. Widmer R.J., Lerman A.. Endothelial Dysfunction and Cardiovascular Disease. Glob. Cardiol. Sci. Pract. (2014) 2014 291-308. DOI: 10.5339/gcsp.2014.43 8. Cavalot F., Petrelli A., Traversa M., Bonomo K., Fiora E., Conti M., Anfossi G., Costa G., Trovati M.. Post-prandial Blood Glucose Is a Stronger Predictor of Cardiovascular Events than Fasting Blood Glucose in Type 2 Diabetes Mellitus, Particularly in Women: Lessons from the San Luigi Gonzaga Diabetes Study. J. Clin. Endocrinol. Metab. (2006) 91 813-819. DOI: 10.1210/jc.2005-1005 9. Shirwany N.A., Zou M.. Arterial Stiffness: A Brief Review. Acta Pharmacol. Sin. (2010) 31 1267. DOI: 10.1038/aps.2010.123 10. Papaioannou T.G., Protogerou A.D., Stergiopulos N., Vardoulis O., Stefanadis C., Safar M., Blacher J.. Total Arterial Compliance Estimated by a Novel Method and All-Cause Mortality in the Elderly: The PROTEGER Study. Age (Dordr.) (2014) 36 9661. DOI: 10.1007/s11357-014-9661-0 11. Westerbacka J., Yki-Järvinen H.. Arterial Stiffness and Insulin Resistance. Semin. Vasc. Med. (2002) 2 157-164. DOI: 10.1055/s-2002-32039 12. Jahn L.A., Hartline L., Rao N., Logan B., Kim J.J., Aylor K., Gan L.-M., Westergren H.U., Barrett E.J.. Insulin Enhances Endothelial Function Throughout the Arterial Tree in Healthy But Not Metabolic Syndrome Subjects. J. Clin. Endocrinol. Metab. (2016) 101 1198-1206. DOI: 10.1210/jc.2015-3293 13. Westerbacka J., Vehkavaara S., Bergholm R., Wilkinson I., Cockcroft J., Yki-Järvinen H.. Marked Resistance of the Ability of Insulin to Decrease Arterial Stiffness Characterizes Human Obesity. Diabetes (1999) 48 821-827. DOI: 10.2337/diabetes.48.4.821 14. Jatic Z., Skopljak A., Hebibovic S., Sukalo A., Rustempasic E., Valjevac A.. Effects of Different Antihypertensive Drug Combinations on Blood Pressure and Arterial Stiffness. Med. Arch. (2019) 73 157-162. DOI: 10.5455/medarh.2019.73.157-162 15. Dotson B.L., Heiston E.M., Miller S.L., Malin S.K.. Insulin Stimulation Reduces Aortic Wave Reflection in Adults with Metabolic Syndrome. Am. J. Physiol. Heart Circ. Physiol. (2021) 320 H2305-H2312. DOI: 10.1152/ajpheart.00975.2020 16. Nagaya N., Kojima M., Uematsu M., Yamagishi M., Hosoda H., Oya H., Hayashi Y., Kangawa K.. Hemodynamic and Hormonal Effects of Human Ghrelin in Healthy Volunteers. Am. J. Physiol.-Regul. Integr. Comp. Physiol. (2001) 280 R1483-R1487. DOI: 10.1152/ajpregu.2001.280.5.R1483 17. Nagaya N., Moriya J., Yasumura Y., Uematsu M., Ono F., Shimizu W., Ueno K., Kitakaze M., Miyatake K., Kangawa K.. Effects of Ghrelin Administration on Left Ventricular Function, Exercise Capacity, and Muscle Wasting in Patients with Chronic Heart Failure. Circulation (2004) 110 3674-3679. DOI: 10.1161/01.CIR.0000149746.62908.BB 18. Dixit V.D., Schaffer E.M., Pyle R.S., Collins G.D., Sakthivel S.K., Palaniappan R., Lillard J.W., Taub D.D.. Ghrelin Inhibits Leptin- and Activation-Induced Proinflammatory Cytokine Expression by Human Monocytes and T Cells. J. Clin. Investig. (2004) 114 57-66. DOI: 10.1172/JCI200421134 19. Tesauro M., Schinzari F., Iantorno M., Rizza S., Melina D., Lauro D., Cardillo C.. Ghrelin Improves Endothelial Function in Patients with Metabolic Syndrome. Circulation (2005) 112 2986-2992. DOI: 10.1161/CIRCULATIONAHA.105.553883 20. Tan A.W.K., Subaran S.C., Sauder M.A., Chai W., Jahn L.A., Fowler D.E., Patrie J.T., Aylor K.W., Basu A., Liu Z.. GLP-1 and Insulin Recruit Muscle Microvasculature and Dilate Conduit Artery Individually But Not Additively in Healthy Humans. J. Endocr. Soc. (2018) 2 190-206. DOI: 10.1210/js.2017-00446 21. Asmar M., Asmar A., Simonsen L., Dela F., Holst J.J., Bülow J.. GIP-Induced Vasodilation in Human Adipose Tissue Involves Capillary Recruitment. Endocr. Connect. (2019) 8 806-813. DOI: 10.1530/EC-19-0144 22. Asmar M., Simonsen L., Madsbad S., Stallknecht B., Holst J.J., Bülow J.. Glucose-Dependent Insulinotropic Polypeptide May Enhance Fatty Acid Re-Esterification in Subcutaneous Abdominal Adipose Tissue in Lean Humans. Diabetes (2010) 59 2160-2163. DOI: 10.2337/db10-0098 23. Asmar M., Simonsen L., Arngrim N., Holst J.J., Dela F., Bülow J.. Glucose-Dependent Insulinotropic Polypeptide Has Impaired Effect on Abdominal, Subcutaneous Adipose Tissue Metabolism in Obese Subjects. Int. J. Obes. (Lond.) (2014) 38 259-265. DOI: 10.1038/ijo.2013.73 24. Heiston E.M., Gilbertson N.M., Eichner N.Z.M., Malin S.K.. A Low-Calorie Diet with or without Exercise Reduces Post-prandial Aortic Waveform in Females with Obesity. Med. Sci. Sports Exerc. (2021) 53 796-803. DOI: 10.1249/MSS.0000000000002515 25. Gilbertson N.M., Eichner N.Z.M., Heiston E.M., Gaitán J.M., Francois M.E., Mehaffey J.H., Hassinger T.E., Hallowell P.T., Weltman A., Malin S.K.. A Low-Calorie Diet with or without Interval Exercise Training Improves Adiposopathy in Obese Women. Appl. Physiol. Nutr. Metab. (2019) 44 1057-1064. DOI: 10.1139/apnm-2018-0717 26. Gaesser G.A., Angadi S.S., Sawyer B.J.. Exercise and Diet, Independent of Weight Loss, Improve Cardiometabolic Risk Profile in Overweight and Obese Individuals. Phys. Sportsmed. (2011) 39 87-97. DOI: 10.3810/psm.2011.05.1898 27. Stock J.M., Chirinos J.A., Edwards D.G.. Lower-Body Dynamic Exercise Reduces Wave Reflection in Healthy Young Adults. Exp. Physiol. (2021) 106 1720-1730. DOI: 10.1113/EP089581 28. Hanssen H., Nussbaumer M., Moor C., Cordes M., Schindler C., Schmidt-Trucksäss A.. Acute Effects of Interval versus Continuous Endurance Training on Pulse Wave Reflection in Healthy Young Men. Atherosclerosis (2015) 238 399-406. DOI: 10.1016/j.atherosclerosis.2014.12.038 29. Munir S., Jiang B., Guilcher A., Brett S., Redwood S., Marber M., Chowienczyk P.. Exercise Reduces Arterial Pressure Augmentation through Vasodilation of Muscular Arteries in Humans. Am. J. Physiol.-Heart Circ. Physiol. (2008) 294 H1645-H1650. DOI: 10.1152/ajpheart.01171.2007 30. Goto C., Nishioka K., Umemura T., Jitsuiki D., Sakagutchi A., Kawamura M., Chayama K., Yoshizumi M., Higashi Y.. Acute Moderate-Intensity Exercise Induces Vasodilation through an Increase in Nitric Oxide Bioavailiability in Humans. Am. J. Hypertens.* (2007) 20 825-830. DOI: 10.1016/j.amjhyper.2007.02.014 31. Eichner N.Z.M., Gaitán J.M., Gilbertson N.M., Khurshid M., Weltman A., Malin S.K.. Post-prandial Augmentation Index Is Reduced in Adults with Prediabetes Following Continuous and Interval Exercise Training. Exp. Physiol. (2019) 104 264-271. DOI: 10.1113/EP087305 32. Benetos A., Thomas F., Joly L., Blacher J., Pannier B., Labat C., Salvi P., Smulyan H., Safar M.E.. Pulse Pressure Amplification a Mechanical Biomarker of Cardiovascular Risk. J. Am. Coll. Cardiol. (2010) 55 1032-1037. DOI: 10.1016/j.jacc.2009.09.061 33. Heiston E.M., Eichner N.Z.M., Gilbertson N.M., Gaitán J.M., Kranz S., Weltman A., Malin S.K.. Two Weeks of Exercise Training Intensity on Appetite Regulation in Obese Adults with Prediabetes. J. Appl. Physiol. (1985) (2019) 126 746-754. DOI: 10.1152/japplphysiol.00655.2018 34. Malin S.K., Francois M.E., Eichner N.Z.M., Gilbertson N.M., Heiston E.M., Fabris C., Breton M.. Impact of Short-Term Exercise Training Intensity on β-Cell Function in Older Obese Adults with Prediabetes. J. Appl. Physiol. (1985) (2018) 125 1979-1986. DOI: 10.1152/japplphysiol.00680.2018 35. John A.T., Chowdhury M., Islam M.R., Mir I.A., Hasan M.Z., Chong C.Y., Humayra S., Higashi Y.. Effectiveness of High-Intensity Interval Training and Continuous Moderate-Intensity Training on Blood Pressure in Physically Inactive Pre-Hypertensive Young Adults. J. Cardiovasc. Dev. Dis. (2022) 9. DOI: 10.3390/jcdd9080246 36. Malin S.K., Gilbertson N.M., Eichner N.Z.M., Heiston E., Miller S., Weltman A.. Impact of Short-Term Continuous and Interval Exercise Training on Endothelial Function and Glucose Metabolism in Prediabetes. J. Diabetes Res. (2019) 2019 4912174. DOI: 10.1155/2019/4912174 37. Mikus C.R., Fairfax S.T., Libla J.L., Boyle L.J., Vianna L.C., Oberlin D.J., Uptergrove G.M., Deo S.H., Kim A., Kanaley J.A.. Seven Days of Aerobic Exercise Training Improves Conduit Artery Blood Flow Following Glucose Ingestion in Patients with Type 2 Diabetes. J. Appl. Physiol. (2011) 111 657-664. DOI: 10.1152/japplphysiol.00489.2011 38. Shenouda N., Gillen J.B., Gibala M.J., MacDonald M.J.. Changes in Brachial Artery Endothelial Function and Resting Diameter with Moderate-Intensity Continuous but Not Sprint Interval Training in Sedentary Men. J. Appl. Physiol. (2017) 123 773-780. DOI: 10.1152/japplphysiol.00058.2017 39. Tai Y.L., Gerhart H., Mayo X., Kingsley J.D.. Acute Resistance Exercise Using Free Weights on Aortic Wave Reflection Characteristics. Clin. Physiol. Funct. Imaging (2018) 38 145-150. DOI: 10.1111/cpf.12396 40. Marshall E.M., Parks J.C., Singer T.J., Tai Y.L., DeBord A.R., Humm S.M., Kingsley J.D.. Vascular Responses to High-Intensity Battling Rope Exercise between the Sexes. J. Sports Sci. Med. (2021) 20 349-356. DOI: 10.52082/jssm.2021.349 41. Patik J.C., Stock J.M., Shenouda N., Chouramanis N.V., Mehrer J.D., Chirinos J.A., Edwards D.G.. Pulsatile Load and Wasted Pressure Effort Are Reduced Following an Acute Bout of Aerobic Exercise. J. Appl. Physiol. (1985) (2021) 131 184-191. DOI: 10.1152/japplphysiol.00946.2020 42. Marsh C.E., Thomas H.J., Naylor L.H., Dembo L.G., Scurrah K.J., Green D.J.. Left Ventricular Adaptation to Exercise Training via Magnetic Resonance Imaging: Studies of Twin Responses to Understand Exercise THerapy. Med. Sci. Sports Exerc. (2022) 54 1095-1104. DOI: 10.1249/MSS.0000000000002899 43. Pearson J.T., Shirai M., Sukumaran V., Du C.-K., Tsuchimochi H., Sonobe T., Waddingham M.T., Katare R., Schwenke D.O.. Ghrelin and Vascular Protection. Vasc. Biol. (2019) 1 H97-H102. DOI: 10.1530/VB-19-0024 44. Broom D.R., Stensel D.J., Bishop N.C., Burns S.F., Miyashita M.. Exercise-Induced Suppression of Acylated Ghrelin in Humans. J. Appl. Physiol. (2007) 102 2165-2171. DOI: 10.1152/japplphysiol.00759.2006 45. Subaran S.C., Sauder M.A., Chai W., Jahn L.A., Fowler D.E., Aylor K.W., Basu A., Liu Z.. GLP-1 at Physiological Concentrations Recruits Skeletal and Cardiac Muscle Microvasculature in Healthy Humans. Clin. Sci. (2014) 127 163-170. DOI: 10.1042/CS20130708 46. Leidy H.J., Gardner J.K., Frye B.R., Snook M.L., Schuchert M.K., Richard E.L., Williams N.I.. Circulating Ghrelin Is Sensitive to Changes in Body Weight during a Diet and Exercise Program in Normal-Weight Young Women. J. Clin. Endocrinol. Metab. (2004) 89 2659-2664. DOI: 10.1210/jc.2003-031471 47. Mani B.K., Castorena C.M., Osborne-Lawrence S., Vijayaraghavan P., Metzger N.P., Elmquist J.K., Zigman J.M.. Ghrelin Mediates Exercise Endurance and the Feeding Response Post-Exercise. Mol. Metab. (2018) 9 114-130. DOI: 10.1016/j.molmet.2018.01.006 48. Åkerström T., Stolpe M.N., Widmer R., Dejgaard T.F., Højberg J.M., Møller K., Hansen J.S., Trinh B., Holst J.J., Thomsen C.. Endurance Training Improves GLP-1 Sensitivity and Glucose Tolerance in Overweight Women. J. Endocr. Soc. (2022) 6 bvac111. DOI: 10.1210/jendso/bvac111 49. Muniyappa R., Montagnani M., Koh K.K., Quon M.J.. Cardiovascular Actions of Insulin. Endocr. Rev. (2007) 28 463-491. DOI: 10.1210/er.2007-0006 50. Vilsbøll T., Holst J.J.. Incretins, Insulin Secretion and Type 2 Diabetes Mellitus. Diabetologia (2004) 47 357-366. DOI: 10.1007/s00125-004-1342-6 51. McLaughlin T., Abbasi F., Lamendola C., Frayo R.S., Cummings D.E.. Plasma Ghrelin Concentrations Are Decreased in Insulin-Resistant Obese Adults Relative to Equally Obese Insulin-Sensitive Controls. J. Clin. Endocrinol. Metab. (2004) 89 1630-1635. DOI: 10.1210/jc.2003-031572 52. Broglio F., Arvat E., Benso A., Gottero C., Muccioli G., Papotti M., van der Lely A.J., Deghenghi R., Ghigo E.. Ghrelin, a Natural GH Secretagogue Produced by the Stomach, Induces Hyperglycemia and Reduces Insulin Secretion in Humans. J. Clin. Endocrinol. Metab. (2001) 86 5083-5086. DOI: 10.1210/jcem.86.10.8098 53. Vestergaard E.T., Gormsen L.C., Jessen N., Lund S., Hansen T.K., Moller N., Jorgensen J.O.L.. Ghrelin Infusion in Humans Induces Acute Insulin Resistance and Lipolysis Independent of Growth Hormone Signaling. Diabetes (2008) 57 3205-3210. DOI: 10.2337/db08-0025 54. McCowen K.C., Maykel J.A., Bistrian B.R., Ling P.R.. Circulating Ghrelin Concentrations Are Lowered by Intravenous Glucose or Hyperinsulinemic Euglycemic Conditions in Rodents. J. Endocrinol. (2002) 175 R7-R11. DOI: 10.1677/joe.0.175r007 55. Baldanzi G., Filigheddu N., Cutrupi S., Catapano F., Bonissoni S., Fubini A., Malan D., Baj G., Granata R., Broglio F.. Ghrelin and Des-Acyl Ghrelin Inhibit Cell Death in Cardiomyocytes and Endothelial Cells through ERK1/2 and PI 3-Kinase/AKT. J. Cell Biol. (2002) 159 1029-1037. DOI: 10.1083/jcb.200207165 56. Nijdam M.-E., Plantinga Y., Hulsen H.T., Bos W.J.W., Grobbee D.E., van der Schouw Y.T., Bots M.L.. Pulse Pressure Amplification and Risk of Cardiovascular Disease. Am. J. Hypertens. (2008) 21 388-392. DOI: 10.1038/ajh.2007.89 57. Love K.M., Liu J., Regensteiner J.G., Reusch J.E.B., Liu Z.. GLP-1 and Insulin Regulation of Skeletal and Cardiac Muscle Microvascular Perfusion in Type 2 Diabetes. J. Diabetes (2020) 12 488-498. DOI: 10.1111/1753-0407.13045 58. Remchak M.-M.E., Heiston E.M., Ballantyne A., Dotson B.L., Malin S.K.. Aortic. Waveform Responses to Insulin in Late versus Early Chronotype with Metabolic Syndrome. Physiol. Rep. (2022) 10 e15473. DOI: 10.14814/phy2.15473 59. Foster-Schubert K.E., Overduin J., Prudom C.E., Liu J., Callahan H.S., Gaylinn B.D., Thorner M.O., Cummings D.E.. Acyl and Total Ghrelin Are Suppressed Strongly by Ingested Proteins, Weakly by Lipids, and Biphasically by Carbohydrates. J. Clin. Endocrinol. Metab. (2008) 93 1971-1979. DOI: 10.1210/jc.2007-2289 60. Sun L., Goh H.J., Govindharajulu P., Leow M.K.-S., Henry C.J.. Post-prandial Glucose, Insulin and Incretin Responses Differ by Test Meal Macronutrient Ingestion Sequence (PATTERN Study). Clin. Nutr. (2020) 39 950-957. DOI: 10.1016/j.clnu.2019.04.001 61. Brown M.A., Green B.P., James L.J., Stevenson E.J., Rumbold P.L.S.. The Effect of a Dairy-Based Recovery Beverage on Post-Exercise Appetite and Energy Intake in Active Females. Nutrients (2016) 8. DOI: 10.3390/nu8060355 62. Yau A.M.W., McLaughlin J., Gilmore W., Maughan R.J., Evans G.H.. The Acute Effects of Simple Sugar Ingestion on Appetite, Gut-Derived Hormone Response, and Metabolic Markers in Men. Nutrients (2017) 9. DOI: 10.3390/nu9020135 63. Koopman A.D.M., Rutters F., Rauh S.P., Nijpels G., Holst J.J., Beulens J.W., Alssema M., Dekker J.M.. Incretin Responses to Oral Glucose and Mixed Meal Tests and Changes in Fasting Glucose Levels during 7 Years of Follow-up: The Hoorn Meal Study. PLoS ONE (2018) 13. DOI: 10.1371/journal.pone.0191114 64. Codella R., Benedini S., Paini S., Caumo A., Adamo M., Terruzzi I., Ferrulli A., Macrì C., Andreoni L., Sterlicchio M.. Effect of Sugar versus Mixed Breakfast on Metabolic and Neurofunctional Responses in Healthy Individuals. J. Diabetes Res. (2017) 2017 e9634585. DOI: 10.1155/2017/9634585 65. Gjesing A.P., Ekstrøm C.T., Eiberg H., Urhammer S.A., Holst J.J., Pedersen O., Hansen T.. Fasting and Oral Glucose-Stimulated Levels of Glucose-Dependent Insulinotropic Polypeptide (GIP) and Glucagon-like Peptide-1 (GLP-1) Are Highly Familial Traits. Diabetologia (2012) 55 1338-1345. DOI: 10.1007/s00125-012-2484-6 66. Adam T.C.M., Westerterp-Plantenga M.S.. Activity-Induced GLP-1 Release in Lean and Obese Subjects. Physiol. Behav. (2004) 83 459-466. DOI: 10.1016/j.physbeh.2004.08.035 67. Mackelvie K.J., Meneilly G.S., Elahi D., Wong A.C.K., Barr S.I., Chanoine J.-P.. Regulation of Appetite in Lean and Obese Adolescents after Exercise: Role of Acylated and Desacyl Ghrelin. J. Clin. Endocrinol. Metab. (2007) 92 648-654. DOI: 10.1210/jc.2006-1028
--- title: 'Association between Dairy Consumption and Psychological Symptoms: Evidence from a Cross-Sectional Study of College Students in the Yangtze River Delta Region of China' authors: - Zhimin Zhao - Ruibao Cai - Yongxing Zhao - Yanyan Hu - Jingzhi Liu - Minghao Wu journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9967214 doi: 10.3390/ijerph20043261 license: CC BY 4.0 --- # Association between Dairy Consumption and Psychological Symptoms: Evidence from a Cross-Sectional Study of College Students in the Yangtze River Delta Region of China ## Abstract Background: Assessing the dairy consumption and psychological symptoms of Chinese college students as a reference for the mental health of Chinese college students. Methods: A three-stage stratified whole-group sampling method was used to investigate dairy consumption and psychological symptoms among 5904 (2554 male students, accounting for $43.3\%$ of the sample) college students in the Yangtze River Delta region. The mean age of the subjects was 20.13 ± 1.24 years. Psychological symptoms were surveyed using the Brief Questionnaire for the Assessment of Adolescent Mental Health. The detection rates of emotional problems, behavioral symptoms, social adaptation difficulties and psychological symptoms among college students with different dairy consumption habits were analyzed using chi-square tests. The association between dairy consumption and psychological symptoms was assessed using a logistic regression model. Results: College students from the “Yangtze River Delta” region of China participated in the study, of which 1022 ($17.31\%$) had psychological symptoms. The proportions of participants with dairy consumption of ≤2 times/week, 3–5 times/week, and ≥6 times/week were $25.68\%$, $42.09\%$, and $32.23\%$, respectively. Using dairy consumption ≥6 times/week as a reference, multifactor logistic regression analysis showed that college students with dairy consumption ≤2 times/week (OR = 1.42, $95\%$ CI: 1.18, 1.71) were at higher risk of psychological symptoms ($p \leq 0.001$). Conclusion: During the COVID-19 pandemic, Chinese college students with lower dairy consumption exhibited higher detection rates of psychological symptoms. Dairy consumption was negatively associated with the occurrence of psychological symptoms. Our study provides a basis for mental health education and increasing knowledge about nutrition among Chinese college students. ## 1. Introduction Milk, milk powder, cheese, and other dairy products are rich in nutrients, such as calcium, protein, potassium, and phosphorus [1]. Dairy products are an important part of the human diet [2], providing between 9.0 and $12.0\%$ of an individual’s energy needs [3]. For example, in Western countries, up to two-thirds of the population’s calcium intake comes from dairy products, demonstrating the importance of dairy products for bone health [4]. Recent studies suggest that the consumption of dairy products appears to have beneficial effects on building muscle, cardiovascular disease risk reduction, prevention of tooth decay, diabetes, cancer, and obesity [2,5]. Epidemiologically-based studies have also demonstrated a significant negative association between the intake of dairy products (especially the low-fat variety) and the prevalence of metabolic syndromes, such as hypertension, dyslipidemia, and abdominal obesity [6]. Dairy consumption has been associated with a $13\%$ reduction in the risk of all-cause mortality (RR: 0.87, $95\%$ CI: 0.77, 0.98) [7]. Given the importance of dairy products for health, dairy consumption has increased significantly in developed countries [8], and the average intake of dairy products by Chinese residents increased from 14.9 g/d in 1992 to 24.7 g/d in 2012 [9]. The World Health Organization describes mental health as the foundation of human health, and mental health disorders can seriously affect physical health, personal well-being, and daily life [10,11,12]. According to the World Health Organization, as many as 80,000 people worldwide die by suicide each year because of depression among young people aged from 15–29 [13]. However, psychological symptoms such as depression, anxiety, and stress are prevalent in the college student population [14], and during the coronavirus disease of 2019 (COVID-19) pandemic, approximately $40.0\%$ of Chinese college students experienced anxiety symptoms [15]. Some studies have shown that the rates of depression and anxiety symptoms among college students have increased each year over approximately the past decade, from $15.6\%$ [16] in 2005 to $31.0\%$ in 2018, representing an increase of $15.0\%$ [17]. A growing number of studies have suggested that healthy eating habits may reduce the risk of developing mental illness [18,19]. For example, the intake of dairy products (skim milk) in adults was negatively associated with depressive symptoms, whereas whole milk and low-calcium dairy products were positively associated with symptoms of depression and insomnia in adults [20,21]. Clinical trials conducted in various countries have reported that the consumption of probiotic-containing dairy products or dairy products in general by patients with depression is associated with a reduction in depressive symptoms [22,23,24]. However, a prospective analysis of Italian adults found no association between depressive symptoms and dairy intake [25]. This suggests that the relationship between dairy products and mental health warrants further investigation. The above findings are from studies of adults, and there is limited research on college student populations. A study from the Azores showed that the consumption of fermented dairy products had a positive effect on reducing anxiety in young college students [26]. In addition, probiotics have been reported to improve panic anxiety, neurophysiological anxiety, negative emotions, and worry, and to increase the regulation of negative emotions in college students [27]. To the best of our knowledge, no previous studies have examined the relationship between dairy product consumption and mental health among Chinese college students. College students are undertaking an important period of transition from adolescence to adulthood, and their dietary habits and psychological health are closely related to their later life development. To this end, this study investigated dairy consumption and psychological symptoms among 5904 college students (2554 males, $43.3\%$) in the Yangtze River Delta region of China. The aim was to assess the relationship between dairy consumption and psychological symptoms among Chinese college students, with the aim of providing a reference and basis for the healthy development of college students. ## 2.1. Participants To ensure the representativeness of the data, a stratified whole-group sampling method was used to select the sample size for this study. First, we selected Hefei, Shanghai, Suzhou, and Hangzhou in the Yangtze River Delta region as the survey cities. Second, we selected two colleges in each city as the surveyed schools. Third, electronic questionnaires were distributed in each college using online publicity and posters. Finally, a total of 5997 college students were surveyed from eight colleges. After excluding students with incomplete demographic information, a total of 5904 valid data sets were obtained. The effective recovery rate of the questionnaire was $98.45\%$. There were 2554 male students, accounting for $43.3\%$ of the sample. The mean age of the subjects was 20.13 ± 1.24 years. In the current study, the survey was conducted after college students gave written informed consent. The survey was implemented after obtaining approval from the Human Ethics Committee of Chizhou University [202104090]. ## 2.2. Data Collection Process Before the test, we introduced the test content of the project and the purpose and significance of the study to the school in detail. The participants were informed of the purpose and significance of the study before the survey. The participants filled out the questionnaire by scanning the QR code using their cell phones. To ensure the authenticity of the data, the study participants were given sufficient time to fill out the relevant contents independently and were given consultation and answers to questions in the process of filling out the questionnaire via WeChat or telephone. ## 2.3. Dairy Consumption Dairy consumption data were obtained by conducting a questionnaire survey on the subjects. The specific questions were as follows: How many times have you consumed dairy in the past week? For example, liquid milk (yogurt, etc.), dairy powder (whole milk powder, skim milk powder, etc.), milk fat, cheese, other dairy products, etc. For the convenience of data analysis, our study divided dairy consumption into three groups: ≤2 times/week, 3–5 times/week, and ≥6 times/week. ## 2.4. Psychological Symptoms We assessed college students’ psychological symptoms using the Brief Questionnaire for the Assessment of Adolescent Mental Health, which was divided into three dimensions of emotional problems (seven questions), behavioral symptoms (four questions), and social adaptation difficulties (four questions), with a total of fifteen questions, and the total score of all questions was the total score of psychological symptoms. The questionnaire has been widely adopted among Chinese university students and has good reliability and validity. Each question was filled in according to the participants’ actual situation, and each question had six options, in the order of [1] “lasted more than 3 months”, [2] “lasted more than 2 months”, [3] “lasted more than 1 month”, [4] “lasted more than 2 weeks”, [5] “lasted more than 1 week”, and [6] “none or less than 1 week”; selecting [1]–[3] scored 1 point, selecting [4]–[6] scored 0 points. When the score of emotional problems dimension was ≥4, the presence of emotional problems was judged; when the score of character problems dimension was ≥1, the presence of social use difficulties was judged; when the score of social adaptation difficulties was ≥2, the positive result of this dimension was indicated. The presence of psychological symptoms was determined when the total score of the three dimensions was ≥7. ## 2.5. Covariates The investigation of covariates in our study included age, sex, socioeconomic status (SES), screen time, body mass index (BMI), and moderate and vigorous physical activity (MVPA). Age was calculated using the participants’ last birthday, and SES was judged by the education, occupation, and income of college students’ parents, which were divided into low (<25th percentile), medium (25–75th percentile), and high (>75th percentile). Height (m) and weight (kg) were measured for college students according to the testing requirements of the National Physical Fitness Standards for Students in China, with an accuracy of 0.1 cm and 0.1 kg, respectively. BMI was calculated as weight (kg)/height(m)2. MVPA time per day for college students was calculated by investigating the average daily frequency and time of moderate and vigorous physical activity of the participants for the previous week using a questionnaire. ## 2.6. Statistical Analysis Categorical data were expressed by percentage and continuous data by mean ± standard deviation. The cardinality test was used to analyze the dairy consumption status of different groups. The continuous type variables in this study conformed to a normal distribution. Comparisons of continuous variables such as BMI and MVPA between different dairy consumption habits were performed by single factor variance. Logistic regression models were used to assess the association between dairy consumption and psychological symptoms. The Crude Model, Model 1 (controlling for age and SES), and Model 2 (controlling for screen time, sugar-sweetened beverages, sleep quality, BMI, and MVPA on the basis of Model 1) were used to analyze the association between college students’ dairy consumption and emotional symptoms, behavioral symptoms, social adaptation difficulties, and psychological symptoms, respectively, reporting the odds ratio and $95.0\%$ confidence interval. Analysis was performed using SPSS 25.0 (SPSS Inc., Chicago, IL, USA) software with a two-sided test level of α = 0.05. ## 3. Results This was a cross-sectional study in which we investigated the dairy consumption of 5904 (2554 males, $43.3\%$) college students during the COVID-19 pandemic. The results of the study showed that the proportions of college students that consumed dairy ≤2 times/week, 3–5 times/week, and ≥6 times/week were $25.68\%$, $42.09\%$, and $32.23\%$, respectively. Our study showed that, in terms of count variables, the differences in dairy consumption detection rates among college students significantly differed by sex, SES, screen time, SSB consumption, and sleep quality (χ2 = 43.115, 87.779, 39.678, 87.275, 54.136, $p \leq 0.001$). In terms of continuous variables, the differences in dairy consumption detection rates of MVPA among college students were significant ($F = 29.821$, $p \leq 0.001$) (Table 1). Our survey revealed that during the COVID-19 pandemic, the detection rate of psychological symptoms among college students in the Yangtze River Delta region was $17.31\%$ ($\frac{1022}{5904}$). The detection rate of psychological symptoms for male students was $16.05\%$ ($\frac{410}{2554}$) and for female students was $18.27\%$ ($\frac{612}{3350}$); the differences in the detection rate of psychological symptoms between males and females was significant (χ2 = 4.969, $p \leq 0.05$). Overall, the lowest detection rate of psychological symptoms was $13.8\%$ for dairy consumption ≥6 times/week, followed by $16.8\%$ for 3–5 times/week. The highest detection rate of psychological symptoms was $22.5\%$ for individuals with dairy consumption of ≤2 times/week, and the difference was significant (χ2 = 45.062, $p \leq 0.001$). In terms of different dimensions, the trend of the magnitude of detection rates of emotional symptoms, behavioral symptoms, and social adaptation difficulties among college students was consistent with that for psychological symptoms, with ≤2 times/week ($23.6\%$, $23.7\%$. $19.7\%$) exhibiting the highest percentage of college students, followed by 3–5 times/week and ≥6 times/week, which all showed significant differences (χ2 = 38.278, 40.611, 34.521, $p \leq 0.001$). In addition, the same trend was found between the males and females (Table 2). After adjusting for relevant covariates, the Model 2 analysis showed that college students with dairy consumption ≤2 times/week (OR = 1.42, $95\%$ CI: 1.18, 1.71), using dairy consumption ≥6 times/week as a reference, were at a higher risk of psychological symptoms ($p \leq 0.001$). Similarly, the risk of emotional symptoms (OR = 1.38, $95\%$ CI: 1.08, 1.77) was higher among female students with dairy consumption ≤2 times/week compared with college students with dairy consumption ≥6 times/week ($p \leq 0.001$). Our findings suggest that dairy consumption was negatively associated with the occurrence of psychological symptoms among Chinese college students during the COVID-19 pandemic (Table 3). ## 4. Discussion Dairy is closely related to human health, and the results of our study showed that the proportion of college students with dairy consumption ≥6 times/week was $32.23\%$, whereas the proportion of college students with dairy consumption ≤2 times/week was $25.68\%$. This result indicated that the proportion of college students with regular dairy consumption was generally low, which is consistent with data from the China Statistical Yearbook, 2021, published by the National Bureau of Statistics of China, but substantially different to the recommended amount of dietary nutrition guidelines for China in 2021 [9]. Although the living standards of Chinese residents have been improving, and the intake of dairy products has increased, the overall level is still low [9]. The proportion of dairy consumption among Chinese college students is higher than that of Chinese adults, and much lower than that of European and American countries. The results showed that $20.4\%$ of Chinese adults consume dairy products at least once a week, and $68.5\%$ of respondents reported they never or rarely consume dairy products [28]. Among college students in Azorean, $91.8\%$ of female students and $98.8\%$ of male students were reported to consume at least one serving of dairy products between two and four times per week [26]. The reason for this difference may be related to differences in the geographical dietary habits of the study population. Our study showed that the detection rate of psychological symptoms among Chinese college students was $17.31\%$, which was higher than that reported in previous studies. A study conducted during the COVID-19 pandemic reported that the detection rate of psychological symptoms among Chinese college students was $8.10\%$ [29]. A study prior to the COVID-19 pandemic reported that the detection rate of psychological problems among college students was $9.98\%$ [30]. The studies indicated that the detection rate of mental health problems among Chinese college students ranged from approximately $45.0\%$ to $47.8\%$ [31] from the beginning of the COVID-19 pandemic to the present. The psychological symptom detection rates among college students decreased as COVID-19 was controlled, and the rates of lethality and severe illness decreased. During the COVID-19 pandemic, many colleges conducted substantial public health education and mental health education to promote students’ mental health, resulting in a gradual decrease in the detection rate of psychological symptoms among college students. The current study also revealed that the detection rate of psychological symptoms was slightly higher in female students ($18.26\%$) compared with that in male students ($16.05\%$). This difference could be explained by the stress reactivity model to some extent, which attributes this difference to the differences between males and females in reactivity under stress [32]. Many previous studies have shown that dietary habits and diet types have an effect on psychological symptoms. Our study showed that college students with dairy consumption ≥6 times/week had the lowest detection rate of psychological symptoms compared with those with ≤2 times/week, and less dairy consumption was associated with a higher detection rate of psychological symptoms. This result is consistent with the findings of previous studies [33]. Some previous studies have reported an inverse relationship between dairy consumption and anxiety [34], with skim milk and moderate dairy-based dessert intake being negatively associated with depressive symptoms, whereas whole milk was positively associated with depressive symptoms in adults [35]. However, some studies have reported that dairy consumption had no positive effect on mental health [36]. This situation suggests that more attention should be paid to the mental health of college students and that more in-depth research on the effects of dairy consumption on mental health is required. The results of the current study revealed an association between college students’ dairy consumption and psychological symptoms, and college students with dairy consumption ≤2 times/week had the highest detection rate of psychological symptoms. Thus, dairy consumption was an important factor affecting college students’ mental health, possibly related to the combined effects of nutrients in dairy and intestinal microorganisms. Dairy products are rich in tryptophan, which is an important raw material for the production of 5-hydroxytryptamine (serotonin) in the body, and low serotonin levels are associated with higher anxiety [37,38]. The gut microbiome produced by the fermentation of dairy products in the intestine has been reported to play a positive role in depressive symptoms, anxiety, cognitive function, sleep, and brain function [39]. It has been reported that gut microbes can alleviate stress and depression-related symptoms by modulating brain function [40]. The intake of dairy products can increase blood tryptophan levels and improve sleep quality and mood, thus, promoting mental health among college students. The current research has several advantages. On the one hand, the sample size of our study was relatively large and representative. The sample was selected from several groups of college students with different majors in the Yangtze River Delta region, which has significant representativeness. In addition, our study effectively adjusted for additional factors that could potentially affect psychological symptoms, such as SES, screen time, sugar-sweetened beverage consumption, sleep quality, MVPA, and body mass index. On this basis, we further analyzed the association between dairy consumption and psychological symptoms among Chinese college students so that the results were more reliable. However, the current study has some limitations. First, this was a cross-sectional study of a college student population, and we only analyzed the association between the two factors rather than the causal relationship. The causal relationship between dairy consumption and psychological symptoms should be explored in depth in future research in conjunction with longitudinal studies. Second, the questionnaire in our study was used to recall dairy consumption, psychological symptoms, and other related factors in the past; thus, the results may have been biased because of the influence of individual recall ability. Third, we collected data on the frequency of weekly dairy intake, but not on the total weekly dairy intake and type of dairy products. In previous studies, a higher frequency of low-fat dairy consumption was associated with a lower prevalence of depressive symptoms [41], and full-fat dairy products may be associated with poorer mental health [42]. Different intakes of dairy products and different types of dairy products may have effects of different magnitudes on promoting mental health, and these gaps should be improved in future studies. Our study provides a reference for better dairy consumption by Chinese college students in the future to better promote healthy physical and mental development. ## 5. Conclusions The results revealed that Chinese college students with lower dairy consumption had a higher detection rate of psychological symptoms. Dairy consumption was negatively associated with the occurrence of psychological symptoms. These findings suggest that the dairy consumption of Chinese college students should be increased in the future. Additionally, dietary health education and psychological health education should be provided to ensure a healthy diet for Chinese college students, while simultaneously reducing their psychological symptoms and promoting healthy physical and mental development. ## References 1. Rozenberg S., Body J.J., Bruyere O., Bergmann P., Brandi M.L., Cooper C., Devogelaer J.P., Gielen E., Goemaere S., Kaufman J.M.. Effects of Dairy Products Consumption on Health: Benefits and Beliefs--A Commentary from the Belgian Bone Club and the European Society for Clinical and Economic Aspects of Osteoporosis, Osteoarthritis and Musculoskeletal Diseases. Calcified Tissue Int. (2016) 98 1-17. DOI: 10.1007/s00223-015-0062-x 2. Tunick M.H., Van Hekken D.L.. Dairy Products and Health: Recent Insights. J. Agric. Food Chem. (2015) 63 9381-9388. DOI: 10.1021/jf5042454 3. Bonjour J.P.. Calcium and phosphate: A duet of ions playing for bone health. J. Am. Coll. Nutr. (2011) 30 438S-448S. DOI: 10.1080/07315724.2011.10719988 4. Gueguen L., Pointillart A.. The bioavailability of dietary calcium. J. Am. Coll. Nutr. (2000) 19 119S-136S. DOI: 10.1080/07315724.2000.10718083 5. Lovegrove J.A., Hobbs D.A.. New perspectives on dairy and cardiovascular health. Proc. Nutr. Soc. (2016) 75 247-258. DOI: 10.1017/S002966511600001X 6. Soedamah-Muthu S.S., Verberne L.D., Ding E.L., Engberink M.F., Geleijnse J.M.. Dairy consumption and incidence of hypertension: A dose-response meta-analysis of prospective cohort studies. Hypertension (2012) 60 1131-1137. DOI: 10.1161/HYPERTENSIONAHA.112.195206 7. Elwood P.C., Pickering J.E., Givens D.I., Gallacher J.E.. The consumption of milk and dairy foods and the incidence of vascular disease and diabetes: An overview of the evidence. Lipids (2010) 45 925-939. DOI: 10.1007/s11745-010-3412-5 8. Rehm C.D., Penalvo J.L., Afshin A., Mozaffarian D.. Dietary Intake Among US Adults, 1999-2012. JAMA-J. Am. Med. Assoc. (2016) 315 2542-2553. DOI: 10.1001/jama.2016.7491 9. He Y., Yang X., Xia J., Zhao L., Yang Y.. Consumption of meat and dairy products in China: A review. Proc. Nutr. Soc. (2016) 75 385-391. DOI: 10.1017/S0029665116000641 10. Correll C.U., Solmi M., Veronese N., Bortolato B., Rosson S., Santonastaso P., Thapa-Chhetri N., Fornaro M., Gallicchio D., Collantoni E.. Prevalence, incidence and mortality from cardiovascular disease in patients with pooled and specific severe mental illness: A large-scale meta-analysis of 3,211,768 patients and 113,383,368 controls. World Psychiatry (2017) 16 163-180. DOI: 10.1002/wps.20420 11. Kessler R.C.. The costs of depression. Psychiatr. Clin. N. Am. (2012) 35 1-14. DOI: 10.1016/j.psc.2011.11.005 12. 12. World Health Organization Mental Health Action Plan 2013–2020World Health OrganizationGeneva, Switzerland2013. Mental Health Action Plan 2013–2020 (2013) 13. 13. World Health Organization DepressionWorld Health OrganizationGeneva, Switzerland2019. Depression (2019) 14. Alkhatatbeh M.J., Khwaileh H.N., Abdul-Razzak K.K.. High prevalence of low dairy calcium intake and association with insomnia, anxiety, depression and musculoskeletal pain in university students from Jordan. Public Health Nutr. (2021) 24 1778-1786. DOI: 10.1017/S1368980020002888 15. Fu W., Yan S., Zong Q., Anderson-Luxford D., Song X., Lv Z., Lv C.. Mental health of college students during the COVID-19 epidemic in China. J. Affect Disorders (2021) 280 7-10. DOI: 10.1016/j.jad.2020.11.032 16. Eisenberg D., Gollust S.E., Golberstein E., Hefner J.L.. Prevalence and correlates of depression, anxiety, and suicidality among university students. Am. J. Orthopsychiatry (2007) 77 534-542. DOI: 10.1037/0002-9432.77.4.534 17. Auerbach R.P., Mortier P., Bruffaerts R., Alonso J., Benjet C., Cuijpers P., Demyttenaere K., Ebert D.D., Green J.G., Hasking P.. WHO World Mental Health Surveys International College Student Project: Prevalence and distribution of mental disorders. J. Abnorm. Psychol. (2018) 127 623-638. DOI: 10.1037/abn0000362 18. Opie R.S., Itsiopoulos C., Parletta N., Sanchez-Villegas A., Akbaraly T.N., Ruusunen A., Jacka F.N.. Dietary recommendations for the prevention of depression. Nutr. Neurosci. (2017) 20 161-171. DOI: 10.1179/1476830515Y.0000000043 19. Lai J.S., Hiles S., Bisquera A., Hure A.J., McEvoy M., Attia J.. A systematic review and meta-analysis of dietary patterns and depression in community-dwelling adults. Am. J. Clin. Nutr. (2014) 99 181-197. DOI: 10.3945/ajcn.113.069880 20. Song Y., Chavarro J.E., Cao Y.. Whole milk intake is associated with prostate cancer-specific mortality among U.S. male physicians. J. Nutr. (2013) 143 189-196. DOI: 10.3945/jn.112.168484 21. Li Y., Zhang C., Li S., Zhang D.. Association between dietary protein intake and the risk of depressive symptoms in adults. Br. J. Nutr. (2020) 123 1290-1301. DOI: 10.1017/S0007114520000562 22. Huang J., Yuan C.M., Xu X.R., Wang Y., Hong W., Wang Z.W., Su Y.S., Hu Y.Y., Cao L., Wang Y.. The relationship between lifestyle factors and clinical symptoms of bipolar disorder patients in a Chinese population. Psychiatry Res. (2018) 266 97-102. DOI: 10.1016/j.psychres.2018.04.059 23. Kazemi A., Noorbala A.A., Azam K., Eskandari M.H., Djafarian K.. Effect of probiotic and prebiotic vs placebo on psychological outcomes in patients with major depressive disorder: A randomized clinical trial. Clin. Nutr. (2019) 38 522-528. DOI: 10.1016/j.clnu.2018.04.010 24. Benton D., Williams C., Brown A.. Impact of consuming a milk drink containing a probiotic on mood and cognition. Eur. J. Clin. Nutr. (2007) 61 355-361. DOI: 10.1038/sj.ejcn.1602546 25. Elstgeest L., Visser M., Penninx B., Colpo M., Bandinelli S., Brouwer I.A.. Bidirectional associations between food groups and depressive symptoms: Longitudinal findings from the Invecchiare in Chianti (InCHIANTI) study. Br. J. Nutr. (2019) 121 439-450. DOI: 10.1017/S0007114518003203 26. Sousa R., Baptista J., Silva C.. Consumption of fermented dairy products is associated with lower anxiety levels in Azorean university students. Front. Nutr. (2022) 9 930949. DOI: 10.3389/fnut.2022.930949 27. Tran N., Zhebrak M., Yacoub C., Pelletier J., Hawley D.. The gut-brain relationship: Investigating the effect of multispecies probiotics on anxiety in a randomized placebo-controlled trial of healthy young adults. J. Affect Disorders (2019) 252 271-277. DOI: 10.1016/j.jad.2019.04.043 28. Kakkoura M.G., Du H., Guo Y., Yu C., Yang L., Pei P., Chen Y., Sansome S., Chan W.C., Yang X.. Dairy consumption and risks of total and site-specific cancers in Chinese adults: An 11-year prospective study of 0.5 million people. BMC Med. (2022) 20. DOI: 10.1186/s12916-022-02330-3 29. Wang Y., Bi C., Liu H., Lin H., Cai R., Zhang J.. Association of sugar-sweetened beverage consumption with psychological symptoms among Chinese university students during the COVID-19 pandemic. Front. Psychol. (2022) 13 1024946. DOI: 10.3389/fpsyg.2022.1024946 30. Fan X., Zhou D.. Evaluation of Mental health Status of Normal College students in Qingdao University. Chin. Sch. Health (2006) 326-328 31. Bi C., Lin H., Zhang J., Zhao Z.. Association between Sleep Quality and Depression Symptoms in Chinese College Students during the COVID-19 Lockdown Period. Children (2022) 9. DOI: 10.3390/children9081237 32. Hankin B.L., Mermelstein R., Roesch L.. Sex differences in adolescent depression: Stress exposure and reactivity models. Child Dev. (2007) 78 279-295. DOI: 10.1111/j.1467-8624.2007.00997.x 33. Mahdavifar B., Hosseinzadeh M., Salehi-Abargouei A., Mirzaei M., Vafa M.. The association between dairy products and psychological disorders in a large sample of Iranian adults. Nutr. Neurosci. (2022) 25 2379-2389. DOI: 10.1080/1028415X.2021.1969065 34. Ramon-Arbues E., Martinez A.B., Granada L.J., Echaniz S.E., Pellicer G.B., Juarez V.R., Guerrero P.S., Saez G.M.. Eating behavior and relationships with stress, anxiety, depression and insomnia in university students. Nutr. Hosp. (2019) 36 1339-1345. DOI: 10.20960/nh.02641 35. Sun J., Wang W., Zhang D.. Associations of different types of dairy intakes with depressive symptoms in adults. J. Affect Disorders (2020) 274 326-333. DOI: 10.1016/j.jad.2020.05.095 36. Du C., Hsiao P.Y., Ludy M.J., Tucker R.M.. Relationships between Dairy and Calcium Intake and Mental Health Measures of Higher Education Students in the United States: Outcomes from Moderation Analyses. Nutrients (2022) 14. DOI: 10.3390/nu14040775 37. Nongonierma A.B., FitzGerald R.J.. Milk proteins as a source of tryptophan-containing bioactive peptides. Food Funct. (2015) 6 2115-2127. DOI: 10.1039/C5FO00407A 38. Reimold M., Batra A., Knobel A., Smolka M.N., Zimmer A., Mann K., Solbach C., Reischl G., Schwarzler F., Grunder G.. Anxiety is associated with reduced central serotonin transporter availability in unmedicated patients with unipolar major depression: A [11C]DASB PET study. Mol. Psychiatry (2008) 13 606-613. DOI: 10.1038/sj.mp.4002149 39. Tabouy L., Getselter D., Ziv O., Karpuj M., Tabouy T., Lukic I., Maayouf R., Werbner N., Ben-Amram H., Nuriel-Ohayon M.. Dysbiosis of microbiome and probiotic treatment in a genetic model of autism spectrum disorders. Brain Behav. Immun. (2018) 73 310-319. DOI: 10.1016/j.bbi.2018.05.015 40. Tyagi P., Tasleem M., Prakash S., Chouhan G.. Intermingling of gut microbiota with brain: Exploring the role of probiotics in battle against depressive disorders. Food Res. Int. (2020) 137 109489. DOI: 10.1016/j.foodres.2020.109489 41. Cui Y., Huang C., Momma H., Ren Z., Sugiyama S., Guan L., Niu K., Nagatomi R.. Consumption of low-fat dairy, but not whole-fat dairy, is inversely associated with depressive symptoms in Japanese adults. Soc. Psychiatry Psychiatr. Epidemiol. (2017) 52 847-853. DOI: 10.1007/s00127-016-1333-1 42. Crichton G.E., Murphy K.J., Bryan J.. Dairy intake and cognitive health in middle-aged South Australians. Asia Pac. J. Clin. Nutr. (2010) 19 161-171. PMID: 20460228
--- title: The Psychological Impact of the COVID-19 Pandemic on Alcohol Abuse and Drunkorexia Behaviors in Young Adults authors: - Daniele Di Tata - Dora Bianchi - Sara Pompili - Fiorenzo Laghi journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9967230 doi: 10.3390/ijerph20043466 license: CC BY 4.0 --- # The Psychological Impact of the COVID-19 Pandemic on Alcohol Abuse and Drunkorexia Behaviors in Young Adults ## Abstract The COVID-19 outbreak negatively affected young adults’ psychological well-being, increasing their stress levels and symptoms of anxiety and depression, and potentially triggering health-risk behaviors. The present study was aimed at investigating the psychological impact of the COVID-19 pandemic on alcohol abuse and drunkorexia behaviors among young adults living in Italy. Participants were 370 emerging adults ($63\%$ women, $37\%$ men; Mage = 21.00, SDage = 2.96, range: 18–30) who were recruited through an online survey between November 2021 and March 2022. Participants completed measures of alcohol abuse, drunkorexia behaviors, negative life experiences, and post-traumatic symptoms related to the COVID-19 outbreak. The results showed that the emotional impact and negative life experiences associated with the pandemic predicted both alcohol abuse and drunkorexia behaviors, albeit in different ways. Specifically, the number of negative life experiences during the pandemic and the tendency to avoid COVID-19–related negative thoughts positively predicted alcohol abuse; and the presence of intrusive thoughts associated with the pandemic significantly predicted the frequency of drunkorexia behaviors. Implications for research and clinical practice are discussed. ## 1. Introduction The COVID-19 pandemic and the emergency measures adopted to counteract it profoundly impacted lifestyles, with negative effects on psychological well-being [1]. In addition to generating health risks and concerns, the COVID-19 pandemic also led to feelings of instability and insecurity towards the future and imposed dramatic and sudden changes in social, academic, and professional realms. For example, the onset of the outbreak in 2020 quickly caused a global economic crisis, increasing unemployment rates worldwide and leaving millions of people suddenly furloughed or unemployed. Moreover, schools and universities closed for a long period, forcing the education system to transition to online learning. Additionally, the social restrictions adopted to combat the pandemic negatively impacted individuals’ ability to spend quality time with others, including friends and romantic partners. This situation may have been particularly stressful for young adults, who were already facing challenging and complex age-related tasks, such as achieving independence from the family, working towards career goals, and developing romantic relationships [2,3]. Thus, the forced isolation, home confinement, and other strict pandemic rules may have interfered with the developmental tasks of young adulthood, generating stress and insecurity in youths [4]. In line with this, research has shown that, although older adults were more at risk of contracting a severe form of the disease, emerging adults reported relatively increased concerns about the health of their loved ones [5] and greater vulnerability to pandemic-related stress [6]. Studies have found a higher prevalence of psychological problems, such as anxiety, depression, and mood disorders among emerging adults during the COVID-19 outbreak, compared with pre-pandemic levels [7,8,9,10]. These emotional issues may have had cascade effects on health risk behaviors, increasing the likelihood that young adults would develop alcohol-related problems and dysfunctional eating patterns, such as those associated with drunkorexia. The long-term psychological impact of the COVID-19 pandemic on drunkorexia behaviors has not yet been studied, despite recent evidence that maladaptive emotion regulation strategies significantly predicted drunkorexia behaviors during the lockdown [11]. Moreover, a trend of increased alcohol use during the pandemic [12,13] highlights the need to identify COVID-19–specific distress factors associated with alcohol abuse in order to better prevent and monitor these behaviors in the post-pandemic phase. ## 1.1. Alcohol Use and the COVID-19 Pandemic Studies have produced mixed results with regard to changes in alcohol consumption during the COVID-19 pandemic [12]. Some studies have reported a decrease in alcohol use due to the social restrictions imposed by governments to prevent contagion [14,15,16], while others have shown that drinking behaviors increased in individuals who used alcohol to cope with pandemic-related distress [17,18]. Both of these patterns are aligned with the motivational model of alcohol use [19,20,21], which suggests that drinking behaviors are driven by different needs and serve several functions. For instance, individuals may drink to achieve a positive outcome (e.g., social acceptance, peer approval) or to avoid a negative one. According to this model, the COVID-19 health emergency affected individuals’ motivation to drink in different ways: on the one hand, social drinking was reduced by the limits imposed on social gatherings, while on the other hand, emotional distress associated with the pandemic may have increased emotional drinking, as a method of coping. In addition, the interaction between different motivations to drink and the lifestyle changes imposed by the pandemic must be considered. During emerging adulthood, the need to be socially accepted and connected with others may facilitate alcohol consumption episodes, often during parties or small social gatherings. In this vein, the reduction of social activities due to the health emergency may have contributed to decreased alcohol use among youths [14]. Nevertheless, studies have shown that the COVID-19 pandemic negatively impacted young adults’ psychological well-being and mental health [6,7,8], and this may have led some youths to use substances (e.g., alcohol) to alleviate distress and negative feelings associated with the dramatic situation. For instance, Vera et al. [ 22] documented that young people with higher levels of depression were more resistant to decreasing their alcohol consumption during the pandemic relative to the general population. Research has not yet examined the possible long-term effects of pandemic-related emotional symptoms on alcohol abuse patterns in young people, although drinking for coping was highly documented in research conducted during the COVID-19 outbreak [17,18,23,24]. The self-medication hypothesis (SMH) [25] proposes that individuals with difficulty self-regulating may use alcohol to relieve or alter their subjective state of distress. Specifically, the anesthetic effect of alcohol could supplant the inability of some individuals to recognize and regulate intolerable emotions. Moreover, several studies have provided support for the SMH regarding trauma-related drinking to cope (see [26] for a review). In particular, research has shown significant associations between post-traumatic symptoms and a coping motivation to drink [27]. Relative to the COVID-19 pandemic, studies have found an association between higher levels of psychological distress and increased alcohol consumption during the health emergency [28,29]. However, evidence is lacking on the long-term effect of COVID-19 trauma-related symptoms on alcohol abuse. ## 1.2. Drunkorexia Behaviors and the COVID-19 Pandemic As Rodgers and colleagues [30] suggested, the COVID-19 pandemic also exacerbated problem eating behaviors, increasing risk conditions, and decreasing protective factors. During the pandemic, individuals experienced higher levels of emotional distress, anxiety, and depression symptoms [31], which are triggering factors for problematic eating behaviors [32]. Conversely, the protective effects of emotional and social support were reduced by the forced isolation. As a consequence, there was an increase in the incidence of eating disorders characterized by binge and restrictive behaviors (e.g., bulimia nervosa and anorexia, respectively) [33,34,35,36,37,38], and a worsening of symptoms in individuals who had already been experiencing these conditions [39,40]. Drunkorexia (also known as food and alcohol disturbance) refers to a pattern of restrictive and compensatory behaviors that are practiced when drinking is planned, in order to compensate for the calories consumed and/or to enhance the intoxicating effect of the alcohol [41,42,43]. These behaviors include self-imposed calorie restriction, fasting, purging, and excessive exercising [44]. Research has identified several commonalities between drunkorexia and eating disorders [45], suggesting the presence of similar dysfunctional patterns [46,47,48]. Drunkorexia behaviors have also been shown to be positively associated with affective variables, including emotion dysregulation, psychological distress, and post-traumatic symptoms [49,50,51,52,53]. Thus, drunkorexia can be considered an eating disorder combined with alcohol use, related to negative emotional experiences and maladaptive coping strategies. Research has shown that drunkorexia is very common among young adults and is mostly driven by a motivation to drink alcohol in order to cope [54,55]. Despite recent evidence of the pandemic’s role in contributing to an upsurge in eating disorders, drinking problems, and other negative psychological conditions [7,12,16,30,40], little is known about the impact of pandemic-related stress on drunkorexia behaviors [11]. However, recent studies have documented an association between drunkorexia, psychological distress, and emotion regulation difficulties [56,57,58,59]. For instance, Qi and colleagues [59] observed higher mean scores of affective problems (e.g., anxiety, depression) in young adults engaging in drunkorexia behaviors. Thus, it is worth exploring the possible impact of the COVID-19 pandemic and the emotional consequences of the health emergency on young adults’ drunkorexia behaviors. ## 1.3. The Current Study To fill this gap in the literature, the present study aimed to examine the psychological impact of the COVID-19 pandemic on alcohol abuse and drunkorexia behaviors in a sample of young adults living in Italy. Specifically, the effects of the number of COVID-19–related negative experiences in the prior year and their long-term emotional effects, as experienced during the prior week, were examined. The study aimed at investigating the predictive roles played by COVID-19-related experiences and two long-term emotional symptoms (i.e., intrusive thoughts about the pandemic, and avoidance of pandemic thoughts) on young adults’ alcohol abuse and drunkorexia behaviors, respectively. Given the trend of increased alcohol consumption during the pandemic [12], positive associations were expected between alcohol abuse and both COVID-19–related negative experiences (H1) and the investigated long-term emotional symptoms (H2). In line with this, previous studies [60,61] have shown a significant positive relationship between alcohol use and COVID-19–specific distress factors, as well as increased alcohol consumption in individuals with psychological problems [62,63,64]. Moreover, based on evidence of a relationship of drunkorexia with psychological distress [51,59] and post-traumatic stress symptoms [52], positive associations were expected between drunkorexia behaviors and both COVID-19–related negative life experiences (H3) and the investigated long-term emotional symptoms (H4). ## 2.1. Participants and Procedure In total, 370 emerging adults living in Italy participated in the study. In accordance with Arnett’s definition of emerging adulthood [2,3], participants included men and women aged 18–30 years. The inclusion criteria were: (a) aged 18–30 years, (b) currently living in Italy, and (c) at least occasionally drinking alcohol. Data were gathered from November 2021 to March 2022, and the recruitment was conducted online using a snowball sampling method. A link to the anonymous survey was disseminated via the university website, and each participant was asked to share the link with their acquaintances and friends. The first page of the survey explained the research procedures and guaranteed complete anonymity and voluntariness of participation. Participants indicated their informed consent before proceeding with the survey. The remainder of the online questionnaire took approximately 20 min to complete. Initially, 414 youths provided informed consent and completed the entire questionnaire. Of these, 40 were removed because they did not drink alcohol (i.e., exclusion criterion). Moreover, 4 participants were excluded for not correctly completing the survey. Therefore, only 370 young adults met all the criteria for inclusion in the research, representing a response rate of $89.4\%$. The study and its procedure were reviewed and approved by the Ethics Committee of the Department of Developmental and Social Psychology, Sapienza University of Rome. Power analyses were conducted using GPower software, version 3.1. Considering the conventional $80\%$ power and 0.05 alpha significance level [65], the a priori power analysis indicated a required sample size of 311 to detect small effect sizes (Cohen’s $d = 0.20$). ## 2.2.1. Individual Information Participants reported their biological sex (0 = man; 1 = woman), age, country of origin, area of residence, and education level. ## 2.2.2. Alcohol Abuse The Alcohol Use Disorders Identification Test (AUDIT) [66,67] was used to assess participants’ alcohol abuse. The 10-item scale was designed by the World Health Organization to evaluate hazardous and harmful drinking. Items investigate the amount and frequency of alcohol intake, the presence of alcohol dependence symptoms, and the presence of alcohol-related problems, with answers rated on a 5-point scale ranging from 0 (never) to 4 (frequently or daily). The present study used the AUDIT mean total score. Male participants with a score ≥ 8 and female participants with a score ≥ 6 were considered high-risk drinkers in accordance with the guidelines set by previous studies [68,69]. The AUDIT has been shown to have good psychometric properties in the Italian context [29,69,70]. The present study confirmed the high reliability of the scale (Cronbach’s alpha = 0.80). ## 2.2.3. Drunkorexia The frequency of drunkorexia was evaluated using the Drunkorexia Behaviors subscale (12 items; sample item: “On a* day I planned to drink, I controlled my eating by avoiding fatty foods”) of the Drunkorexia Motives and Behaviors Scale (DMBS) [71]. On this subscale, questions investigate the frequency with which compensatory behaviors are engaged around alcohol consumption episodes, with answers rated on a Likert-type scale ranging from 1 (never) to 5 (always). The DMBS has been shown to have good psychometric properties in the Italian context [11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53]. In the present study, the instrument showed high reliability (Cronbach’s alpha = 0.97). ## 2.2.4. COVID-19-Related Negative Experiences A 19-item checklist was used to assess the number of stressors experienced by participants during the prior year due to the COVID-19 pandemic (i.e., “Please answer the following questions about your experiences with COVID-19 over the past year”). The instrument was inspired by the 16-COVID Stressors Checklist [72], which was modified to accommodate the Italian context and to meet the needs of the present study. Three additional items were added from the SARS-related Stressors Checklist [73]. Items 1–12 assessed participants’ experiences and whether their relatives, friends, and/or acquaintances had been infected (or were suspected of having been infected) by the virus (e.g., “During the last year… did you contract COVID-19 and test positive for a swab?”). Questions 13–16 assessed the degree to which participants had experienced distress due to information they had received about COVID-19 (e.g., “Have you had a hard time understanding the authenticity of online information regarding the COVID-19 pandemic?”), and questions 17–19 assessed the impact of the pandemic on participants’ daily life (e.g., “Have you had conflicts and quarrels with your family due to the pandemic?”). Each item had a dichotomous response option to indicate whether the COVID-19–related negative event had been experienced (0 = no; 1 = yes). The total score indicated the number of stressors participants had experienced in the prior year due to the pandemic. ## 2.2.5. Psychological Impact of the COVID-19 Pandemic The Impact of Event Scale with Modifications for COVID-19 (IES-COVID-19) [74] was administered to assess the long-term psychological impact of the pandemic on participants over the prior seven days. The original version of the instrument comprises 15 items, measuring two dimensions: Avoidance, which investigates the tendency to avoid COVID-19–related negative thoughts (8 items; sample item: “Over the last seven days … I stayed away from things that made me think about the COVID-19 pandemic”); and Intrusion, which assesses the presence of negative intrusive thoughts regarding the COVID-19 pandemic (7 items; sample item: “Over the last seven days … I thought about the COVID-19 pandemic when I didn’t mean to”). Answers are rated on a 4-point scale ranging from 1 (never) to 4 (often). In the absence of an Italian version of the instrument, a blinded forward-backward translation procedure was conducted to adapt the original Dutch version to the Italian cultural context (guidelines by [75]). The few discrepancies that emerged were discussed by psychology researchers in a focus group. Thereafter, the Italian version of the IES-COVID-19 was included in the present study. A confirmatory factor analysis (CFA) was run on the data using the Jamovi software (version 2.3.16) to verify the adequacy of the two-factor model for Italian participants. Goodness-of-fit was estimated by the relative chi-square test statistic (χ2/df, values expected to range between 1–3 for good model fit) [76]; by the CFI, NFI, and NNFI indices (≥0.90 for acceptable fit) [77]; and by the RMSEA and SRMR indices (≤0.08 for acceptable fit) [78]. One item (i.e., item 15) was removed from the Italian version to improve the reliability of the two dimensions, and the final 14-item model obtained acceptable fit indices, confirming the presence of two correlated factors: χ2 [71] = 209.057, $p \leq 0.001$; χ2/df = 2.94; CFI = 0.94; NFI = 0.91; NNFI = 0.92; RMSEA = 0.07; $90\%$ CI [0.06, 0.08]; SRMR = 0.05. The two factors were positively correlated (Pearson’s $r = 0.51$, $p \leq 0.001$) and showed good reliability in the Italian sample (Cronbach’s alphas of 0.84 for Avoidance and 0.86 for Intrusion). ## 2.3. Data Analysis Data were analyzed using the statistical package SPSS, version 27. To preliminarily investigate the effects of gender and age differences on alcohol abuse and drunkorexia behaviors, two univariate analyses of covariance (ANCOVA) were performed, including gender as a between-subjects factor and age as a covariate. In the first ANCOVA, a continuous score of problematic alcohol use was entered as a dependent variable. In the second ANCOVA, drunkorexia was included as a dependent variable. Bivariate Pearson’s correlations were computed among the study variables, with both drunkorexia and alcohol abuse included as continuous scores. Subsequently, to verify the first and second research hypotheses, a hierarchical regression analysis was performed on the continuous score of alcohol abuse. Gender and age were entered as covariates in step 1, while COVID-19–related negative experiences, Avoidance, and Intrusion were added to the regression equation in step 2. Finally, to investigate the third and fourth research hypotheses, a negative binomial regression analysis was conducted to test the effects of COVID-19–related experiences and emotional symptoms on the frequency of drunkorexia behaviors. Gender, age, and alcohol abuse were included in the model to control for their effects, and COVID-19–related negative experiences, Intrusion, and Avoidance were added to test the hypotheses. ## 3. Results Participants were 370 young adults (Mage = 21.00, SDage = 2.96, age range: 18–30; $63\%$ women, $37\%$ men) living in Italy. In terms of geography, $67\%$ of participants resided in central Italy, $26\%$ in southern Italy, and $7\%$ in northern Italy. Most participants ($97\%$) had been born in Italy, while $3\%$ had an immigrant background. Regarding education level, $2\%$ had finished middle school, $82\%$ had graduated from high school, and the remaining $16\%$ held a bachelor’s degree; moreover, $98\%$ currently were university students. Regarding the alcohol abuse groups, 149 participants ($40\%$) were classified as “high-risk drinkers” and 221 as “low-risk drinkers,” based on the gender-adjusted cut-off scores for the AUDIT questionnaire. As regards drunkorexia, 139 participants ($38\%$) reported enacting compensatory behaviors around episodes of alcohol consumption. The results of the first ANCOVA on alcohol abuse indicated significant univariate effects of gender, F[1] = 9.45, $p \leq 0.01$, η2 = 03, with men reporting a higher mean score than women, and age, F[1] = 7.44, $p \leq 0.01$, η2 = 02. With respect to drunkorexia behaviors, the results of the second ANCOVA showed a significant univariate effect of age, F[1] = 6.88, $p \leq 0.01$, η2 = 02. Net of this effect, a significant difference between men and women was found, F[1] = 6.11, $p \leq 0.05$, η2 = 02. Specifically, women reported higher scores on drunkorexia relative to men (see Table 1). Table 2 summarizes the descriptive statistics and bivariate Pearson’s correlations. The analysis showed a significant and positive correlation between alcohol abuse and drunkorexia behaviors, with both behaviors positively and significantly associated with the number of COVID–related negative experiences during the health emergency, the presence of negative intrusive thoughts about the pandemic, and the tendency to avoid such negative thoughts. Regarding the hierarchical regression analysis on the alcohol abuse continuous score, the results in step 1 explained $3\%$ of the variance in alcohol abuse, detecting significant negative effects of gender (with men reporting higher scores than women) and age (with younger people reporting higher scores). Step 2 contributed a significant $6\%$ to the explained variance. Gender and age were still significant covariates, and controlling for their effects, COVID-19–related negative life experiences and Avoidance significantly and positively predicted alcohol abuse. Conversely, Intrusion was not significant. The final model explained $9\%$ of the variance in alcohol abuse (see Table 3 for statistics). The negative binomial regression model explained a significant $17\%$ of the variance, detecting a significant effect of gender (with women reporting higher scores) and age (with younger participants reporting higher rates of drunkorexia). Moreover, alcohol abuse showed a significant and positive effect on drunkorexia behaviors. Net of these effects, the presence of intrusive negative thoughts about the pandemic emerged as a significant predictor of drunkorexia behaviors, whilst the effects of COVID-19 experiences and Avoidance were null. Table 4 reports the full model statistics. ## 4. Discussion The present study investigated the psychological impact of the COVID-19 pandemic on alcohol abuse and drunkorexia behaviors in young adults. Specifically, the number of COVID-19–related negative experiences, the presence of negative intrusive thoughts regarding the pandemic, and the tendency to avoid these negative thoughts were explored to understand their roles in predicting these two health-risk behaviors. Recent studies have observed an increase in alcohol consumption and eating disorders during the COVID-19 pandemic [12,35]. Thus, it is reasonable to expect a long-term emotional impact of the pandemic on alcohol abuse and drunkorexia behaviors. To the best of our knowledge, the present study was the first to investigate the relationship between drunkorexia and COVID-19–related stressors and emotional symptoms. The results of the preliminary analyses showed that women reported more frequent drunkorexia behaviors compared to men. This finding contradicts the results of previous international studies finding no gender differences in drunkorexic tendencies [79]. Conversely, men showed higher levels of alcohol abuse than women, in line with recent research on young Italians during the COVID-19 pandemic [80]. Moreover, the preliminary analyses found that younger participants reported higher scores for both drunkorexia behaviors and alcohol abuse. The first regression analysis found a significant effect of age (with younger participants reporting higher rates) and gender (with men reporting higher rates than women) on alcohol abuse. These findings confirm the results of previous studies that have found men to be significantly more likely to use alcohol than women during the pandemic [14,81,82]. Regarding age, previous research has produced mixed results, with most studies indicating higher alcohol consumption in older people during the COVID-19 outbreak [12]. However, the relaxation of containment measures in the last year recreated social drinking opportunities for younger people, who consequently reported a greater tendency to engage in problematic alcohol use. Controlling for gender and age effects, alcohol abuse was significantly and positively predicted by COVID-19–related negative life experiences and the avoidance of thinking about these events. These findings suggest that pandemic-related stress may have contributed to increased alcohol consumption aimed at coping with negative emotions [18,83], in line with the SMH [25]. Specifically, both the higher number of negative stressors caused by the pandemic and the tendency to avoid COVID-19–related negative thoughts predicted higher alcohol abuse, suggesting that alcohol consumption served as a maladaptive emotion regulation strategy in response to negative experiences. However, the percentage of variance explained by the model ($9\%$) suggests that the results should be interpreted with caution. Other predictors not considered by the study (e.g., lack of social support and mental health problems) may also contribute to explaining the risk of alcohol abuse related to the pandemic. Thus, future studies should investigate the relationship between alcohol use and COVID-19–related psychological distress more deeply by including more control variables (e.g., mental health problems, perceived social support, and living conditions during the pandemic). Regarding the second regression analysis on drunkorexia behaviors, a significant association with gender was found, with women reporting a greater likelihood of engaging in compensatory behaviors around an episode of alcohol consumption relative to men. This finding aligns with the results of previous research [84]. A significant negative effect of age on drunkorexia behaviors was also found, with younger participants reporting a significantly greater likelihood of restricting calories when deciding to drink alcohol, suggesting that drunkorexia risk may decrease with age. Indeed, previous research has shown that risky behaviors are more frequent during adolescence and emerging adulthood and decrease with the transition into adulthood [85]. In line with previous studies [45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86], the present study found that alcohol abuse positively predicted drunkorexia, whereby participants with greater alcohol-related problems exhibited more frequent dysfunctional compensatory behaviors with respect to COVID-19–related variables, the long-term emotional impact of the pandemic also affected drunkorexia behaviors, albeit in a very different way, confirming the uniqueness of this behavior. Specifically, only the frequency of negative intrusive thoughts about the pandemic positively predicted drunkorexia behaviors, while avoidance of these thoughts and the number of COVID-19 negative experiences were not significant. Thus, drunkorexia was predicted by ruminative thinking about the pandemic, regardless of the actual negative events that were experienced. Previous studies have shown that ruminative thoughts predict different eating disorders in the short and long term (review by [87]). The present findings uncovered a similar pattern in drunkorexic tendencies. This result, together with the very different findings of the first regression on alcohol abuse, provides new evidence for the higher affinity of drunkorexia with other eating disorders rather than alcohol-related problems. Moreover, the present findings corroborate previous evidence showing that psychological distress is a risk factor for engaging in compensatory behaviors around episodes of alcohol consumption [46,88]. ## 5. Conclusions The present study represented one of the first attempts to explore the long-term emotional impact of the COVID-19 pandemic on alcohol abuse and drunkorexia behaviors. However, it also suffered from some limitations. First, data were obtained through self-report instruments, which may have been affected by social desirability bias. Consequently, some undesirable behaviors (e.g., alcohol misuse) might have been underreported. Second, the data were collected in Italy, and thus, the findings may not be generalizable to populations of countries with different experiences of the pandemic. Moreover, participants were not representative of the target population but constituted a convenience sample recruited through the dissemination of an anonymous online questionnaire. Third, due to the cross-sectional research design, only correlational associations between study variables, and not causal relationships, could be drawn. Future studies should attempt to confirm the present results with representative samples from different countries. In addition, future research should continue to monitor the psychological impact of the pandemic on health-risk behaviors among young people by implementing mixed research methodologies. Despite these study limitations, the present results provide useful information for the implementation of prevention programs to monitor alcohol abuse and drunkorexia behaviors in the post-pandemic phase, helping to identify at-risk individuals. It is recommended that public health interventions focus on reducing psychological distress and supporting individuals who suffered from the pandemic to prevent further problematic behavioral consequences. ## References 1. Talevi D., Socci V., Carai M., Carnaghi G., Faleri S., Trebbi E., di Bernardo A., Capelli F., Pacitti F.. Mental Health Outcomes of the CoVID-19 Pandemic. Riv. Psichiatr. (2020) 55 137-144. DOI: 10.1708/3382.33569 2. Arnett J.J.. Emerging Adulthood: The Winding Road from the Late Teens through the Twenties (2014). DOI: 10.1093/oxfordhb/9780199795574.013.9 3. Arnett J.J.. Emerging adulthood: A theory of development from the late teens through the twenties. Am. Psychol. (2000) 55 469-480. DOI: 10.1037/0003-066X.55.5.469 4. Ohannessian C.M.. Introduction to the Special Issue: The Impact of the COVID-19 Pandemic on the Lives of Emerging Adults. Emerg. Adulthood (2021) 9 431-432. DOI: 10.1177/21676968211030875 5. Germani A., Buratta L., Delvecchio E., Mazzeschi C.. Emerging Adults and COVID-19: The Role of Individualism-Collectivism on Perceived Risks and Psychological Maladjustment. Int. J. Environ. Res. Public Health (2020) 17. DOI: 10.3390/ijerph17103497 6. Varma P., Junge M., Meaklim H., Jackson M.L.. Younger people are more vulnerable to stress, anxiety and depression during COVID-19 pandemic: A global cross-sectional survey. Prog. Neuro-Psychopharmacol. Biol. Psychiatry (2021) 109 110236. DOI: 10.1016/j.pnpbp.2020.110236 7. Buizza C., Bazzoli L., Ghilardi A.. Changes in College Students Mental Health and Lifestyle During the COVID-19 Pandemic: A Systematic Review of Longitudinal Studies. Adolesc. Res. Rev. (2022) 7 537-550. DOI: 10.1007/s40894-022-00192-7 8. Deng J., Zhou F., Hou W., Silver Z., Wong C.Y., Chang O., Drakos A., Zuo Q.K., Huang E.. The prevalence of depressive symptoms, anxiety symptoms and sleep disturbance in higher education students during the COVID-19 pandemic: A systematic review and meta-analysis. Psychiatry Res. (2021) 301 113863. DOI: 10.1016/j.psychres.2021.113863 9. Bussone S., Pesca C., Tambelli R., Carola V.. Psychological Health Issues Subsequent to SARS-Cov 2 Restrictive Measures: The Role of Parental Bonding and Attachment Style. Front. Psychiatry (2020) 11 589444. DOI: 10.3389/fpsyt.2020.589444 10. Fruehwirth J.C., Biswas S., Perreira K.M.. The COVID-19 pandemic and mental health of first-year college students: Examining the effect of COVID-19 stressors using longitudinal data. PLoS ONE (2021) 16. DOI: 10.1371/journal.pone.0247999 11. Pompili S., Di Tata D., Bianchi D., Lonigro A., Zammuto M., Baiocco R., Longobardi E., Laghi F.. Food and alcohol disturbance among young adults during the COVID-19 lockdown in Italy: Risk and protective factors. Eat. Weight. Disord.-Stud. Anorex. Bulim. Obes. (2021) 27 769-780. DOI: 10.1007/s40519-021-01220-6 12. Roberts A., Rogers J., Mason R., Siriwardena A.N., Hogue T., Whitley G.A., Law G.R.. Alcohol and other substance use during the COVID-19 pandemic: A systematic review. Drug Alcohol Depend. (2021) 229 109150. DOI: 10.1016/j.drugalcdep.2021.109150 13. Charles N.E., Strong S.J., Burns L.C., Bullerjahn M.R., Serafine K.M.. Increased mood disorder symptoms, perceived stress, and alcohol use among college students during the COVID-19 pandemic. Psychiatry Res. (2021) 296 113706. DOI: 10.1016/j.psychres.2021.113706 14. Patin A., Ladner J., Tavolacci M.-P.. Change in University Student Health Behaviours after the Onset of the COVID-19 Pandemic. Int. J. Environ. Res. Public Health (2023) 20. DOI: 10.3390/ijerph20010539 15. Rehm J., Kilian C., Ferreira-Borges C., Jernigan D., Monteiro M., Parry C.D.H., Sanchez Z.M., Manthey J.. Alcohol use in times of the COVID 19: Implications for monitoring and policy. Drug Alcohol Rev. (2020) 39 301-304. DOI: 10.1111/dar.13074 16. Sohi I., Chrystoja B.R., Rehm J., Wells S., Monteiro M., Ali S., Shield K.D.. Changes in alcohol use during the COVID-19 pandemic and previous pandemics: A systematic review. Alcohol. Clin. Exp. Res. (2022) 46 498-513. DOI: 10.1111/acer.14792 17. Rodriguez L.M., Litt D.M., Stewart S.H.. Drinking to cope with the pandemic: The unique associations of COVID-19-related perceived threat and psychological distress to drinking behaviors in American men and women. Addict. Behav. (2020) 110 106532. DOI: 10.1016/j.addbeh.2020.106532 18. Wardell J.D., Kempe T., Rapinda K.K., Single A., Bilevicius E., Frohlich J.R., Hendershot C.S., Keough M.T.. Drinking to Cope During COVID-19 Pandemic: The Role of External and Internal Factors in Coping Motive Pathways to Alcohol Use, Solitary Drinking, and Alcohol Problems. Alcohol. Clin. Exp. Res. (2020) 44 2073-2083. DOI: 10.1111/acer.14425 19. Cox W.M., Klinger E.. A motivational model of alcohol use. J. Abnorm. Psychol. (1988) 97 168-180. DOI: 10.1037/0021-843X.97.2.168 20. Cooper M.L.. Motivations for alcohol use among adolescents: Development and validation of a four-factor model. Psychol. Assess. (1994) 6 117-128. DOI: 10.1037/1040-3590.6.2.117 21. Kuntsche E., Knibbe R., Gmel G., Engels R.. Why do young people drink? A review of drinking motives. Clin. Psychol. Rev. (2005) 25 841-861. DOI: 10.1016/j.cpr.2005.06.002 22. Vera B.D.V., Carmona-Márquez J., Lozano-Rojas M., Parrado-González A., Vidal-Giné C., Pautassi R.M., Fernández-Calderón F.. Changes in Alcohol Use during the COVID-19 Pandemic among Young Adults: The Prospective Effect of Anxiety and Depression. J. Clin. Med. (2021) 10. DOI: 10.3390/jcm10194468 23. Bollen Z., Pabst A., Creupelandt C., Fontesse S., Lannoy S., Pinon N., Maurage P.. Prior drinking motives predict alcohol consumption during the COVID-19 lockdown: A cross-sectional online survey among Belgian college students. Addict. Behav. (2021) 115 106772. DOI: 10.1016/j.addbeh.2020.106772 24. McPhee M.D., Keough M.T., Rundle S., Heath L.M., Wardell J.D., Hendershot C.S.. Depression, Environmental Reward, Coping Motives and Alcohol Consumption During the COVID-19 Pandemic. Front. Psychiatry (2020) 11 1128. DOI: 10.3389/fpsyt.2020.574676 25. Khantzian E.J.. The Self-Medication Hypothesis of Substance Use Disorders: A Reconsideration and Recent Applications. Harv. Rev. Psychiatry (1997) 4 231-244. DOI: 10.3109/10673229709030550 26. Hawn S.E., Cusack S.E., Amstadter A.B.. A Systematic Review of the Self-Medication Hypothesis in the Context of Posttraumatic Stress Disorder and Comorbid Problematic Alcohol Use. J. Trauma. Stress (2020) 33 699-708. DOI: 10.1002/jts.22521 27. Hawn S.E., Bountress K.E., Sheerin C.M., Dick D.M., Amstadter A.B.. Trauma-related drinking to cope: A novel approach to the self-medication model. Psychol. Addict. Behav. (2020) 34 465-476. DOI: 10.1037/adb0000552 28. Acuff S.F., Strickland J.C., Tucker J.A., Murphy J.G.. Changes in alcohol use during COVID-19 and associations with contextual and individual difference variables: A systematic review and meta-analysis. Psychol. Addict. Behav. (2022) 36 1-19. DOI: 10.1037/adb0000796 29. Laghi F., Di Tata D., Bianchi D., Lonigro A., Pompili S., Zammuto M., Baiocco R.. Problematic alcohol use in young adults during the COVID-19 lockdown in Italy. Psychol. Health Med. (2022) 27 139-149. DOI: 10.1080/13548506.2021.1975785 30. Rodgers R.F., Lombardo C., Cerolini S., Franko D.L., Omori M., Fuller-Tyszkiewicz M., Linardon J., Courtet P., Guillaume S.. The impact of the COVID -19 pandemic on eating disorder risk and symptoms. Int. J. Eat. Disord. (2020) 53 1166-1170. DOI: 10.1002/eat.23318 31. Lakhan R., Agrawal A., Sharma M.. Prevalence of Depression, Anxiety, and Stress during COVID-19 Pandemic. J. Neurosci. Rural. Pract. (2020) 11 519-525. DOI: 10.1055/s-0040-1716442 32. Leehr E.J., Krohmer K., Schag K., Dresler T., Zipfel S., Giel K.E.. Emotion regulation model in binge eating disorder and obesity—A systematic review. Neurosci. Biobehav. Rev. (2015) 49 125-134. DOI: 10.1016/j.neubiorev.2014.12.008 33. Castellini G., Cassioli E., Rossi E., Innocenti M., Gironi V., Sanfilippo G., Felciai F., Monteleone A.M., Ricca V.. The impact of COVID-19 epidemic on eating disorders: A longitudinal observation of pre versus post psychopathological features in a sample of patients with eating disorders and a group of healthy controls. Int. J. Eat. Disord. (2020) 53 1855-1862. DOI: 10.1002/eat.23368 34. Giel K.E., Schurr M., Zipfel S., Junne F., Schag K.. Eating behaviour and symptom trajectories in patients with a history of binge eating disorder during COVID-19 pandemic. Eur. Eat. Disord. Rev. (2021) 29 657-662. DOI: 10.1002/erv.2837 35. Linardon J., Messer M., Rodgers R.F., Fuller-Tyszkiewicz M.. A systematic scoping review of research on COVID-19 impacts on eating disorders: A critical appraisal of the evidence and recommendations for the field. Int. J. Eat. Disord. (2022) 55 3-38. DOI: 10.1002/eat.23640 36. Taquet M., Geddes J.R., Luciano S., Harrison P.J.. Incidence and outcomes of eating disorders during the COVID-19 pandemic. Br. J. Psychiatry (2021) 220 262-264. DOI: 10.1192/bjp.2021.105 37. Termorshuizen J.D., Watson H.J., Thornton L.M., Borg S., Flatt R.E., MacDermod C.M., Harper L.E., van Furth E.F., Peat C.M., Bulik C.M.. Early impact of COVID-19 on individuals with eating disorders: A survey of _1000 individuals in the United States and The Netherlands. Int. J. Eat. Disord. (2020) 53 1780-1790. DOI: 10.1002/eat.23353 38. Tavolacci M.-P., Ladner J., Déchelotte P.. Sharp Increase in Eating Disorders among University Students since the COVID-19 Pandemic. Nutrients (2021) 13. DOI: 10.3390/nu13103415 39. Bianchi D., Baiocco R., Pompili S., Lonigro A., Di Norcia A., Cannoni E., Longobardi E., Zammuto M., Di Tata D., Laghi F.. Binge Eating and Binge Drinking in Emerging Adults During COVID-19 Lockdown in Italy: An Examination of Protective and Risk Factors. Emerg. Adulthood (2021) 10 291-303. DOI: 10.1177/21676968211058501 40. Devoe D.J., Han A., Anderson A., Katzman D.K., Patten S.B., Soumbasis A., Flanagan J., Paslakis G., Vyver E., Marcoux G.. The impact of the COVID-19 pandemic on eating disorders: A systematic review. Int. J. Eat. Disord. (2022) 56 5-25. DOI: 10.1002/eat.23704 41. Burke S.C., Cremeens J., Vail-Smith K., Woolsey C.. Drunkorexia: Calorie restriction prior to alcohol consumption among college freshmen. J. Alcohol Drug Educ. (2010) 54 17-35 42. Ma E.M.C., Rancourt D., Thompson J.K.. From fad to FAD: A theoretical formulation and proposed name change for “drunkorexia” to food and alcohol disturbance (FAD). Int. J. Eat. Disord. (2018) 51 831-834. DOI: 10.1002/eat.22926 43. Thompson-Memmer C., Glassman T., Diehr A.. Drunkorexia: A new term and diagnostic criteria. J. Am. Coll. Health (2019) 67 620-626. DOI: 10.1080/07448481.2018.1500470 44. Rahal C.J., Bryant J.B., Darkes J., Menzel J.E., Thompson J.K.. Development and validation of the Compensatory Eating and Behaviors in Response to Alcohol Consumption Scale (CEBRACS). Eat. Behav. (2012) 13 83-87. DOI: 10.1016/j.eatbeh.2011.11.001 45. Hunt T.K., Forbush K.T.. Is “drunkorexia” an eating disorder, substance use disorder, or both?. Eat. Behav. (2016) 22 40-45. DOI: 10.1016/j.eatbeh.2016.03.034 46. Hill E.M., Martin J.D., Lego J.E.. College students’ engagement in drunkorexia: Examining the role of sociocultural attitudes toward appearance, narcissism, and Greek affiliation. Curr. Psychol. (2019) 40 4468-4478. DOI: 10.1007/s12144-019-00382-y 47. Hill E.M., Lego J.E.. Examining the role of body esteem and sensation seeking in drunkorexia behaviors. Eat. Weight. Disord.-Stud. Anorex. Bulim. Obes. (2020) 25 1507-1513. DOI: 10.1007/s40519-019-00784-8 48. Peralta R.L.. Alcohol use and the fear of weight gain in college: Reconciling two social norms. Gend. Issues (2002) 20 23-42. DOI: 10.1007/s12147-002-0021-5 49. Horvath S.A., Shorey R.C., Racine S.E.. Emotion dysregulation as a correlate of food and alcohol disturbance in undergraduate students. Eat. Behav. (2020) 38 101409. DOI: 10.1016/j.eatbeh.2020.101409 50. Laghi F., Pompili S., Bianchi D., Lonigro A., Baiocco R.. Drunkorexia: An Examination of the Role of Theory of Mind and Emotional Awareness among Adolescents. Dev. Neuropsychol. (2021) 46 70-81. DOI: 10.1080/87565641.2020.1869743 51. Laghi F., Pompili S., Bianchi D., Lonigro A., Baiocco R.. Exploring the association between psychological distress and drunkorexia behaviors in non-clinical adolescents: The moderating role of emotional dysregulation. Eat. Weight. Disord.-Stud. Anorex. Bulim. Obes. (2021) 26 797-806. DOI: 10.1007/s40519-020-00912-9 52. Michael M.L., Witte T.H.. Traumatic stress and alcohol-related disordered eating in a college sample. J. Am. Coll. Health (2020) 69 806-811. DOI: 10.1080/07448481.2019.1706534 53. Pompili S., Laghi F.. Drunkorexia among adolescents: The role of motivations and emotion regulation. Eat. Behav. (2018) 29 1-7. DOI: 10.1016/j.eatbeh.2018.01.001 54. Lupi M., Martinotti G., Di Giannantonio M.. Drunkorexia: An emerging trend in young adults. Eat. Weight. Disord.-Stud. Anorex. Bulim. Obes. (2017) 22 619-622. DOI: 10.1007/s40519-017-0429-2 55. Simons R.M., Hansen J.M., Simons J.S., Hovrud L., Hahn A.M.. Drunkorexia: Normative behavior or gateway to alcohol and eating pathology?. Addict. Behav. (2021) 112 106577. DOI: 10.1016/j.addbeh.2020.106577 56. Malaeb D., Bianchi D., Pompili S., Berro J., Laghi F., Azzi V., Akel M., Obeid S., Hallit S.. Drunkorexia behaviors and motives, eating attitudes and mental health in Lebanese alcohol drinkers: A path analysis model. Eat. Weight. Disord.-Stud. Anorex. Bulim. Obes. (2021) 27 1787-1797. DOI: 10.1007/s40519-021-01321-2 57. Azzi V., Bianchi D., Pompili S., Laghi F., Gerges S., Akel M., Malaeb D., Obeid S., Hallit S.. Emotion regulation and drunkorexia behaviors among Lebanese adults: The indirect effects of positive and negative metacognition. BMC Psychiatry (2022) 22. DOI: 10.1186/s12888-022-04030-x 58. Azzi V., Hallit S., Malaeb D., Obeid S., Brytek-Matera A.. Drunkorexia and Emotion Regulation and Emotion Regulation Difficulties: The Mediating Effect of Disordered Eating Attitudes. Int. J. Environ. Res. Public Health (2021) 18. DOI: 10.3390/ijerph18052690 59. Qi B., Humphrey A., Bulik C.M., Baker J.H., Munn-Chernoff M.A.. Food-restricted alcohol consumption: Relation to psychopathology in college students. J. Am. Coll. Health (2021) 69 1-10. DOI: 10.1080/07448481.2021.1891915 60. Gritsenko V., Skugarevsky O., Konstantinov V., Khamenka N., Marinova T., Reznik A., Isralowitz R.. COVID 19 Fear, Stress, Anxiety, and Substance Use Among Russian and Belarusian University Students. Int. J. Ment. Health Addict. (2021) 19 2362-2368. DOI: 10.1007/s11469-020-00330-z 61. Rahman M.A., Hoque N., Alif S.M., Salehin M., Islam S.M.S., Banik B., Sharif A., Nazim N.B., Sultana F., Cross W.. Factors associated with psychological distress, fear and coping strategies during the COVID-19 pandemic in Australia. Glob. Health (2020) 16 95. DOI: 10.1186/s12992-020-00624-w 62. Callinan S., Smit K., Mojica-Perez Y., D’Aquino S., Moore D., Kuntsche E.. Shifts in alcohol consumption during the COVID-19 pandemic: Early indications from Australia. Addiction (2021) 116 1381-1388. DOI: 10.1111/add.15275 63. Pérez-Carbonell L., Meurling I.J., Wassermann D., Gnoni V., Leschziner G., Weighall A., Ellis J., Durrant S., Hare A., Steier J.. Impact of the novel coronavirus (COVID-19) pandemic on sleep. J. Thorac. Dis. (2020) 12 S163-S175. DOI: 10.21037/jtd-cus-2020-015 64. Tran T.D., Hammarberg K., Kirkman M., Nguyen H.T.M., Fisher J.. Alcohol use and mental health status during the first months of COVID-19 pandemic in Australia. J. Affect. Disord. (2020) 277 810-813. DOI: 10.1016/j.jad.2020.09.012 65. Cohen J.. Statistical Power Analysis for the Behavioral Sciences (1988) 66. Babor T.F., de la Fuente J.R., Saunders J., Grant M.. AUDIT: The Alcohol Use Disorders Identification Test: Guidelines for Use in Primary Health Care (1992) 67. Saunders J.B., Aasland O.G., Babor T.F., De La Fuente J.R., Grant M.. Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO Collaborative Project on Early Detection of Persons with Harmful Alcohol Consumption-II. Addiction (1993) 88 791-804. DOI: 10.1111/j.1360-0443.1993.tb02093.x 68. García Carretero M.Á., Novalbos Ruiz J.P., Delgado J., Martínez M., González C.O.F.. Validation of the Alcohol Use Disorders Identification Test in university students: AUDIT and AUDIT-C. Adicciones (2016) 28 194-204. DOI: 10.20882/adicciones.775 69. Struzzo P., De Faccio S., Moscatelli E., Scafat E.. Identificazione precoce dei bevitori a rischio in assistenza primaria in Italia: Adattamento del questionario AUDIT e verifica dell’efficacia d’uso dello short-AUDIT test nel contesto nazionale. Boll. Farmacodipend. L’alcoolismo (2006) 29 20-25 70. Morelli M., Bianchi D., Baiocco R., Pezzuti L., Chirumbolo A.. Sexting behaviors and cyber pornography addiction among adolescents: The moderating role of alcohol consumption. Sex. Res. Soc. Policy (2013) 14 113-121. DOI: 10.1007/s13178-016-0234-0 71. Ward R.M., Galante M.. Development and initial validation of the Drunkorexia Motives and Behaviors scales. Eat. Behav. (2015) 18 66-70. DOI: 10.1016/j.eatbeh.2015.04.003 72. Li X., Wu H., Meng F., Li L., Wang Y., Zhou M.. Relations of COVID-19-Related Stressors and Social Support with Chinese College Students’ Psychological Response During the COVID-19 Pandemic. Front. Psychiatry (2020) 11 551315. DOI: 10.3389/fpsyt.2020.551315 73. Main A., Zhou Q., Ma Y., Luecken L.J., Liu X.. Relations of SARS-related stressors and coping to Chinese college students’ psychological adjustment during the 2003 Beijing SARS epidemic. J. Couns. Psychol. (2011) 58 410-423. DOI: 10.1037/a0023632 74. Vanaken L., Scheveneels S., Belmans E., Hermans D.. Validation of the Impact of Event Scale with Modifications for COVID-19 (IES-COVID19). Front. Psychiatry (2020) 11 738. DOI: 10.3389/fpsyt.2020.00738 75. Sousa V.D., Rojjanasrirat W.. Translation, adaptation and validation of instruments or scales for use in cross-cultural health care research: A clear and user-friendly guideline. J. Eval. Clin. Pract. (2011) 17 268-274. DOI: 10.1111/j.1365-2753.2010.01434.x 76. Schumacker R.E., Lomax R.G.. A Beginner’s Guide to Structural Equation Modeling (2004) 77. Hu L.T., Bentler P.M.. Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Struct. Equ. Model. Multidiscip. J. (1999) 6 1-55. DOI: 10.1080/10705519909540118 78. Schermelleh-Engel K., Moosbrugger H., Müller H.. Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures. Methods Psychol. Res. Online (2003) 8 23-74 79. Speed S., Ward R.M., Haus L., Branscum P., Barrios V., Budd K.M., Lemons K., Humenay E.. A Systematic Review of Common Drunkorexia Measures: Examining Gender Differences across Scales. Am. J. Health Educ. (2022) 53 342-359. DOI: 10.1080/19325037.2022.2100846 80. Mazzarella C., Federico A., Spina A., Dallio M., Romeo M., Masarone M., Persico M., Gravina A.G., Luzza F., Abenavoli L.. Alcoholic consumption of young Italians during the SARS-CoV-2 pandemic. Rev. Recent Clin. Trials (2022) 17 126-135. DOI: 10.2174/1574887117666220602100325 81. Panno A., Carbone G.A., Massullo C., Farina B., Imperatori C.. COVID-19 Related Distress Is Associated with Alcohol Problems, Social Media and Food Addiction Symptoms: Insights From the Italian Experience During the Lockdown. Front. Psychiatry (2020) 11 577135. DOI: 10.3389/fpsyt.2020.577135 82. Wang Y., Lu H., Hu M., Wu S., Chen J., Wang L., Luo T., Wu Z., Liu Y., Tang J.. Alcohol Consumption in China Before and During COVID-19: Preliminary Results From an Online Retrospective Survey. Front. Psychiatry (2020) 11 597826. DOI: 10.3389/fpsyt.2020.597826 83. Mojica-Perez Y., Livingston M., Pennay A., Callinan S.. Examining the relationship between alcohol consumption, psychological distress and COVID-19 related circumstances: An Australian longitudinal study in the first year of the pandemic. Addict. Behav. (2022) 135 107439. DOI: 10.1016/j.addbeh.2022.107439 84. Barry A., Whiteman S., Ms A.K.P.-G., Jensen A.C.. Gender Differences in the Associations Among Body Mass Index, Weight Loss, Exercise, and Drinking Among College Students. J. Am. Coll. Health (2013) 61 407-413. DOI: 10.1080/07448481.2013.823973 85. Arain M., Haque M., Johal L., Mathur P., Nel W., Rais A., Sandu R., Sharma S.. Maturation of the adolescent brain. Neuropsychiatr. Dis. Treat. (2013) 9 449. DOI: 10.2147/NDT.S39776 86. Pompili S., Laghi F.. Drunkorexia: Disordered eating behaviors and risky alcohol consumption among adolescents. J. Health Psychol. (2020) 25 2222-2232. DOI: 10.1177/1359105318791229 87. Smith K.E., Mason T.B., Lavender J.M.. Rumination and eating disorder psychopathology: A meta-analysis. Clin. Psychol. Rev. (2018) 61 9-23. DOI: 10.1016/j.cpr.2018.03.004 88. Roosen K.M., Mills J.S.. Exploring the motives and mental health correlates of intentional food restriction prior to alcohol use in university students. J. Health Psychol. (2015) 20 875-886. DOI: 10.1177/1359105315573436
--- title: 'Risk Factors for Anal Continence Impairment Following a Second Delivery after a First Traumatic Delivery: A Prospective Cohort Study' authors: - Gabriel Marcellier - Axelle Dupont - Agnes Bourgeois-Moine - Arnaud Le Tohic - Celine De Carne-Carnavalet - Olivier Poujade - Guillaume Girard - Amélie Benbara - Laurent Mandelbrot - Laurent Abramowitz journal: Journal of Clinical Medicine year: 2023 pmcid: PMC9967240 doi: 10.3390/jcm12041531 license: CC BY 4.0 --- # Risk Factors for Anal Continence Impairment Following a Second Delivery after a First Traumatic Delivery: A Prospective Cohort Study ## Abstract Postpartum anal incontinence is common. After a first delivery (D1) with perineal trauma, follow-up is advised to reduce the risk of anal incontinence. Endoanal sonography (EAS) may be considered to evaluate the sphincter and in case of sphincter lesions to discuss cesarean section for the second delivery (D2). Our objective was to study the risk factors for anal continence impairment following D2. Women with a history of traumatic D1 were followed before and 6 months after D2. Continence was measured using the Vaizey score. An increase ≥2 points after D2 defined a significant deterioration. A total of 312 women were followed and 67 ($21\%$) had worse anal continence after D2. The main risk factors for this deterioration were the presence of urinary incontinence and the combined use of instruments and episiotomy during D2 (OR 5.12, $95\%$ CI 1.22–21.5). After D1, 192 women ($61.5\%$) had a sphincter rupture revealed by EAS, whereas it was diagnosed clinically in only 48 ($15.7\%$). However, neither clinically undiagnosed ruptures nor severe ruptures were associated with an increased risk of continence deterioration after D2, and cesarean section did not protect against it. One woman out of five in this population had anal continence impairment after D2. The main risk factor was instrumental delivery. Caesarean section was not protective. Although EAS enabled the diagnosis of clinically-missed sphincter ruptures, these were not associated with continence impairment. Anal incontinence should be systematically screened in patients presenting urinary incontinence after D2 as they are frequently associated. ## 1. Introduction Anal incontinence (AI) is a taboo condition and a frequent cause of handicap, present in up to $20\%$ of the population [1,2]. Its detection requires a dedicated investigation and can be difficult [3]. Continence is a complex mechanism involving the anal sphincter, rectal compliance, anorectal angulation, pudendal nerve innervation and the nature of the stools. Incontinence occurs when one or more of these mechanisms is altered beyond compensation [4]. Aging leads to a decrease in muscle and perineal trophicity and is one of the main risk factors for AI [5]. Various events in perineal life will accelerate this aging. Childbirth is one of these disruptors of pelvic integrity. The damage it causes to the perineum (mostly anal sphincter injury and pudendal nerve stretching) is frequent, affecting 12 to $28\%$ of parturients [6,7] and can lead to AI in 4 to $40\%$ of women giving birth [6,7,8,9,10]. Repeated deliveries lead to a risk of cumulative damages [7,11,12,13] and obstetrical life events contribute to a higher risk of AI in women compared to men [5,14,15]. Assessment of obstetrical anal sphincter injuries (OASI) is carried out immediately after delivery by the obstetrician/gynecologist, who describes any extension of a perineal tear to the anal sphincter. However, ruptures identified during an endoanal sonography (EAS) are undiagnosed at delivery in up to $30\%$ of cases, and may be at risk of AI [6,9]. The extension of the sphincter damage is also estimated by EAS. Despite its potential advantages, the benefits of performing this procedure are debated because the diagnosis of a sphincter tear in EAS (EAS rupture) will not always have an impact on patient care [8]. Many studies in the literature focus on AI in primiparous women, but few address the risk factors of AI after a second delivery (D2), despite the fact that the fertility rate in France is close to two and that second child deliveries approach 250,000 annually [16,17]. An even smaller proportion of these studies have been conducted prospectively and none on large cohorts [18,19,20,21]. Moreover, there is a need to further investigate the risk factors of AI after D2 among women who have become vulnerable after a first traumatic delivery (D1), since a proctologist (surgeon or gastroenterologist) is then more likely to be consulted to advise about the risks and modalities of a new delivery. The objective of our study was to determine the risk factors of continence impairment after D2 among women who had a traumatic D1, in particular the impact of EAS ruptures. ## 2.1. Population Our work is an ancillary study of the prospective, randomized, multicenter “Prevention of Anal Incontinence by Caesarean Section” (EPIC) study, which compared the benefit of prophylactic caesarean section (CS) to vaginal delivery (VD) at D2 in women with a history of a traumatic D1 with sphincter rupture confirmed by EAS [8]. Women were recruited in six maternity units in the Paris area between April 2008 and December 2014. They were included by their gynecologist during the 3rd trimester of their 2nd pregnancy if they met the inclusion criteria, which were as follows: a single history of traumatic vaginal delivery (VD), defined as forceps extraction or with grade III perineal tear (reaching the sphincters), age above 18 years old, informed written consent and no AI at inclusion, based on a YES/NO answer to the question asked by the gynecologist. They were excluded if they had a history of grade IV perineum tear (which corresponds to the most severe grade of OASI with sphincter and anal mucosa damage) and if a CS was indicated for their future delivery for a non-proctological reason. After inclusion, systematic prospective follow-up was carried out, with a proctological examination during the 3rd trimester of their 2nd pregnancy (referred to as before D2 visit) and then 6 months after the second delivery (referred to as after D2 visit). This visit included questionnaires measuring the Vaizey score for anal continence and the measure of urinary handicap score (MHU) for urinary continence, as well as an EAS. In the EPIC study, women with EAS sphincter rupture were randomized to perform D2 by VD or CS. Some women included in EPIC were not randomized, either because EAS did not reveal a sphincter rupture or because they refused randomization. The mode of delivery was then discussed between the obstetrician and the patient. In this study, we included all women who were explored by EAS before D2 and for whom a Vaizey score was calculated before and after D2, regardless of their randomization status. ## 2.2. Objective and Thresholds The analysis of anal continence was based on the Vaizey score [22] (Appendix A). Data differs in the literature to assess which Vaizey value significantly defines incontinence [23]. In the EPIC study, based on this literature and on expert opinion, a score ≥5 defined AI [8,24]. Our population being inhomogeneous regarding continence before D2, we selected as the primary endpoint worsening of the Vaizey score after D2, defined as an increase ≥2 points in the score between the two proctological examinations. Comparable definitions were used in previous proctologic studies [25]. Because transient AI (lasting less than 2 months) is common in the immediate postpartum period [26,27], the assessment 6 months after D2 was used to measure persistent continence deterioration. EAS was performed by a single trained operator, using a rotating rectal probe (7–10 MHz, Brüel and Kjaer). Upper, middle and lower anal canal were studied. A sphincter lesion was identified as a loss of continuity visible by a change in echogenicity within the sphincter ring [28]. Severity was assessed based on the Starck score (Appendix B). A score ≥9 was used to define a severe sphincter rupture [29,30]. The clinical description of perineal lesions was based on the Royal College of Obstetricians and Gynecologists classification, where the anal sphincter is considered impaired in grades III and IV (Appendix C). We defined a “hidden sphincter rupture” as a tear undiagnosed in the delivery room (or under-diagnosed as a grade I or II) but observed by EAS. After D2, ruptures were considered “de novo” if no EAS defect was visible after D1. The analysis of urinary continence was based on the MHU score (Appendix D) [31], treated as a continuous variable ranging from 0 to 28 points. Macrosomia was defined by birthweight >4 kg [32]. Birthweight was not collected in D2 in the case of CS. Instrumental delivery referred to the use of all types of forceps or vacuum but the type of forceps was not specified. Details of the episiotomy were not collected. We defined “abnormal transit” as the presence of diarrhea, constipation or dyschesia. We asked the patients whether or not they had undergone perineal rehabilitation, but the modalities were not collected (number of sessions or technique used). ## 2.3. Statistical Analysis Categorical variables were described as numbers and percentages and quantitative variables were described as median and interquartile ranges. We compared median Vaizey scores at the two visits with Wilcoxon paired tests. To assess the association between the primary outcome and the characteristics of the women, univariate logistic regressions were performed to determine unadjusted odds ratios (OR) and their $95\%$ confidence intervals. Variables with a univariate p value < 0.20 were tested in multivariate models. Variable selection for the final multivariate model was performed using top-down selection with the Akaike information criterion. The linearity assumption was tested graphically and with the Wald test for the MHU score. An analysis on the subgroup of women giving birth by VD at D2 was conducted using the same methodology, to study the impact of a second vaginal delivery and its characteristics. p values < 0.05 were considered significant. The tests were two-sided. All analyses were performed using R software (v.3.4). ## 3.1. Participants A total of 549 parturients were included in the EPIC study, of which 312 had a Vaizey score completed before and after D2 and were included in our ancillary work (Figure 1). Characteristics of the population are described in Table 1. ## 3.2.1. First Delivery (D1) Among the 312 women included, 285 ($92.2\%$) delivered with forceps at D1 and 266 ($86.9\%$) had episiotomies. After D1, there were 27 ($8.9\%$) grade I, 15 ($4.9\%$) grade II and 48 ($15.7\%$) grade III perineal tears. An OASI was therefore only clinically visible in these 48 women. However, the EAS performed before D2 showed sphincter tears in 192 ($61.5\%$) of these women. After D1, 225 ($74.3\%$) women had perineal rehabilitation. ## 3.2.2. Second Delivery (D2) A total of 103 women ($34.8\%$) delivered by CS and 193 ($65.2\%$) by VD. For 16 women ($5.1\%$), the mode of delivery was not recorded. Among the 193 VD, there were 13 forceps ($6.7\%$), 12 vacuum ($6.2\%$), 49 episiotomies ($25.4\%$) and 87 grade I ($45.1\%$), 24 grade II ($12.4\%$) and 2 grade III ($1.0\%$) perineal tears. After D2, 201 ($64.4\%$) women underwent perineal rehabilitation. Characteristics of the second delivery among the subgroup of women delivering vaginally is described in Table 2. ## 3.3. Characteristics of Continence Before D2, 43 ($13.8\%$) women had a Vaizey score ≥ 5 and the median Vaizey score was 1 [0–3]. After D2, the median Vaizey score remained unchanged across the total population, with no significant worsening ($$p \leq 0.33$$). However, 67 women ($21.5\%$) significantly worsened their continence score (≥2 points) and increased their median Vaizey score from 1 [0–2] before D2 to 5 [3–7] afterwards. The median MHU score was $\frac{4}{28}$ [2–8] during the 3rd trimester of the 2nd pregnancy and $\frac{1}{28}$ [0–4] 6 months after D2. Details of the outcomes measured 6 months after D2 are presented in Table 3. ## 3.4.1. Among the 312 Included Patients Maternal weight, ethnicity, fetal presentation, macrosomia at D1, transient AI after D1, lack of perineal rehabilitation and the presence of a sphincter rupture diagnosed in EAS before D2 (whether occult or not and whether the EAS rupture was severe or not), were not associated with an increased Vaizey score after D2 (Table 4) in univariate analysis. In multivariate analysis, delivering by CS did not impact significantly the risk of worsening continence after D2 compared to VD (Table 5). Women who were already incontinent before D2 were less likely to deteriorate in their Vaizey score after D2 (OR 0.27–$95\%$ CI 0.08–0.86). Increase in MHU score after D2 was linearly associated with a risk of continence deterioration (OR increased by 1.24 per MHU point-$95\%$ CI 1.08–1.42). ## 3.4.2. Among the 193 Patients That Delivered Vaginally at D2 In this subgroup, in univariate analysis, macrosomia, the absence of perineal rehabilitation after D1 or D2 and the presence of an endosonographic or clinical sphincter rupture were not associated with an increased Vaizey score after D2 (Table 6). Instrumental delivery, with or without episiotomy, increased the risk of worsening anal continence after D2. We did not find a significant association with episiotomy itself. In multivariate analysis (Table 7), instrumental delivery or episiotomy were associated with a deterioration of anal continence only when performed together (OR 4.18–$95\%$ CI 1.05–16.59). Increase in MHU score after D2 was linearly associated with a risk of AI (OR1.25 per MHU point-$95\%$ CI 1.11–1.43). ## 3.5.1. Sphincter Tears before D2 192 ($61.5\%$) women had an EAS sphincter rupture before D2, whereas only 48 ($15.7\%$) had a grade III OASI clinically recognized by the obstetrician in the delivery room during D1. These hidden ruptures were not accompanied by an increased risk of continence impairment after D2 (Table 2). The description of sphincter ruptures is provided Appendix E. ## 3.5.2. Sphincter Tears after D2 A total of 143 ($71.9\%$) women had sphincter ruptures revealed by the EAS performed after D2 (out of 199 who underwent EAS). Only six had “de novo” sphincter ruptures. These were women who had delivered vaginally, including one with episiotomy. The subgroup was too small to evidence a significant association between these ruptures and continence impairment in univariate analysis. ## 4.1. Main Results To our knowledge, this study is the largest prospective cohort describing risk factors for continence impairment after a second delivery in at-risk parturients. The main risk factor was instrumental extraction coupled with an episiotomy during a second vaginal delivery. These results are consistent with the literature and the absence of an increased risk related to episiotomy when analyzed independently is reassuring regarding the controversial impact of this procedure [7,24]. The risk factors of postpartum AI are still debated. Forceps delivery and perineal tears are frequently associated with AI [7,11,13,33]. But the association is less clear for episiotomy, multiparity [7,11], macrosomia and nulliparity [13]. In our work, none of these factors were associated with an increased risk of continence impairment after D2, possibly because we studied at D2 the impact of some events occurring at D1 and because we relied on a large and prospective cohort, unlike previous studies. We found a strong association between the presence of urinary incontinence after D2 and the presence of AI. Patients and their physicians are often aware of the risk of urinary incontinence after delivery, while the diagnosis of AI is more challenging. The diagnosis of urinary incontinence could initiate dialogue between the obstetrician and the proctologist. At inclusion, several women denied having AI during non-specific questioning, while a more focalized interview during the proctological consultation enabled a proper continence evaluation. Surprisingly, the presence of AI before D2 was associated with a decreased risk of continence impairment after D2. We assume that some women had signs of AI related to the pregnancy itself, due to transit and pelvic static disorders, which improved after the delivery. CS was not associated with a decreased risk of continence impairment after D2. This is consistent with the EPIC study [8] and recent literature [34,35]. ## 4.2. The Place of EAS Women with traumatic deliveries are routinely offered EAS examination in order to detect a clinically undiagnosed sphincter defect. We indeed observed that among half of the patients, an anal sphincter rupture had been missed clinically but was evidenced by EAS, which is consistent with the literature [6,9]. The EPIC study demonstrated however, that it was not beneficial to perform prophylactic cesarean section to these patients to prevent the occurrence of AI after D2 [8]. Our study goes further, showing that neither the presence of these hidden sphincter ruptures diagnosed before a second delivery, nor even the most severe sphincter tears, were associated with a significant risk of continence impairment after D2. We therefore have no argument for advocating the use of EAS after a first traumatic delivery. However, in our practice, performing an EAS, as frequently requested by maternities, allows these women at-risk to have access to a proctologist, who can inform them about long-term risk factors for AI and avoidable cofactors. These results can be challenged by the relative scarcity of severe sphincter tears in our study, mostly because of the low percentage of internal sphincter injuries, which strongly impacts Starck’s score [29,30]. In Sultan’s cohort [9], $16\%$ of women had an internal sphincter defect versus only $5.8\%$ in our study. These results may be explained by the exclusion of patients with grade IV perineal tears or already reporting AI to their gynecologist after D1. There is therefore still room to debate the benefits of EAS among these patients. ## 4.3. Strengths and Limitations We collected data in medical centers representative of the population pools of our region and we relied on validated clinical criteria (Vaizey score, MHU score) and paraclinical criteria (Starck’s score [29,30]). EAS were performed by a single expert operator, which limited ranking bias by preventing inter-operator variability. Moreover, the inter-operator concordance of this practitioner had already been validated in a previous work [7]. There is no consensus on thresholds defining AI in literature. We chose to define a continence impairment as an increase in two points of the Vaizey score after D2. This is a sensitive threshold for continence deterioration but not a very specific one. Anal incontinence tends to decrease in the weeks following delivery and to reappear after menopause. One of the limitations of this work is the lack of follow-up beyond one year after D2. Although the follow-up was prospective, we conducted this work after the legal closing date of the original study, which prevented us from completing the missing data. Therefore, $\frac{122}{434}$ ($28\%$) of women who underwent EAS before D2 were not analyzed. However, we did not evidence any difference between the baseline characteristics of our population and this unanalyzed population (Appendix F). ## 4.4. What Can Be Offered to a Woman at Risk of Continence Impairment after D2 Potential perineal damage risk can be reduced by measures such as regulating transit, treating dyschesia (source of pudendal stretch or prolapse) and avoiding additional sphincter trauma as much as possible. It is thus advisable to refer women at risk or presenting postpartum AI to proctologists for multi-disciplinary management. Although the main risk factor for continence impairment is forceps delivery with episiotomy, these obstetrical interventions are sometimes necessary and cannot be completely avoided. In case of proven AI and after managing possible cofactors, pelvic rehabilitation with biofeedback or neuromodulation of the sacral roots may be beneficial [36,37,38]. ## 5. Conclusions After a first traumatic delivery, instrumental delivery with episiotomy is the main risk factor for anal continence impairment during a second delivery. Women who present symptoms of urinary incontinence after their second delivery also are at greater risk of developing symptoms of anal incontinence. Referral to a proctologist will improve the detection and management of this incontinence. Caesarian section does not prevent it. In a woman wishing to start a second pregnancy after a first traumatic delivery, the presence of a sphincter lesion on endoanal ultrasound will not change the course of treatment. ## References 1. Macmillan A.K., Merrie A., Marshall R., Parry B.R.. The Prevalence of Fecal Incontinence in Community-Dwelling Adults: A Systematic Review of the Literature. Dis. Colon Rectum (2004) 47 1341-1349. DOI: 10.1007/s10350-004-0593-0 2. Menees S.B., Almario C.V., Spiegel B.M., Chey W.D.. Prevalence of and Factors Associated with Fecal Incontinence: Results From a Population-Based Survey. Gastroenterology (2018) 154 1672-1681.e3. DOI: 10.1053/j.gastro.2018.01.062 3. Damon H., Guye O., Seigneurin A., Long F., Sonko A., Faucheron J.-L., Grandjean J.-P., Mellier G., Valancogne G., Fayard M.-O.. Prevalence of anal incontinence in adults and impact on quality-of-life. Gastroenterol. Clin. Biol. (2006) 30 37-43. DOI: 10.1016/S0399-8320(06)73076-7 4. Rao S.S.. Pathophysiology of adult fecal incontinence. Gastroenterology (2004) 126 S14-S22. DOI: 10.1053/j.gastro.2003.10.013 5. E Bharucha A., Dunivan G., Goode P.S., Lukacz E.S., Markland A., A Matthews C., Mott L., Rogers R.G., Zinsmeister A.R., E Whitehead W.. Epidemiology, Pathophysiology, and Classification of Fecal Incontinence: State of the Science Summary for the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) Workshop. Am. J. Gastroenterol. (2015) 110 127-136. DOI: 10.1038/ajg.2014.396 6. Faltin D.L., Boulvain M., Irion O., Bretones S., Stan C., Weil A.. Diagnosis of anal sphincter tears by postpartum endosonography to predict fecal incontinence. Obstet. Gynecol. (2000) 95 643-647. PMID: 10775721 7. Abramowitz L., Sobhani I., Ganansia R., Vuagnat A., Benifla J.L., Darai E., Madelenat P., Mignon M.. Are sphincter defects the cause of anal incontinence after vaginal delivery? Results of a prospective study. Dis. Colon. Rectum. (2000) 43 590-596. DOI: 10.1007/BF02235567 8. Abramowitz L., Mandelbrot L., Moine A.B., Le Tohic A., Carnavalet C.D.C., Poujade O., Roy C., Tubach F.. Caesarean section in the second delivery to prevent anal incontinence after asymptomatic obstetric anal sphincter injury: The EPIC multicentre randomised trial. BJOG Int. J. Obstet. Gynaecol. (2020) 128 685-693. DOI: 10.1111/1471-0528.16452 9. Sultan A.H., Kamm M.A., Hudson C.N., Thomas J.M., Bartram C.I.. Anal-Sphincter Disruption during Vaginal Delivery. N. Engl. J. Med. (1993) 329 1905-1911. DOI: 10.1056/NEJM199312233292601 10. Oberwalder M., Connor J., Wexner S.D.. Meta-analysis to determine the incidence of obstetric anal sphincter damage. Br. J. Surg. (2003) 90 1333-1337. DOI: 10.1002/bjs.4369 11. Peleg D., Kennedy C.M., Merrill D., Zlatnik F.J.. Risk of Repetition of a Severe Perineal Laceration. Obstet. Gynecol. (1999) 93 1021-1024. DOI: 10.1097/00006250-199906000-00025 12. Edozien L., Gurol-Urganci I., Cromwell D., Adams E., Richmond D., Mahmood T., van der Meulen J.. Impact of third- and fourth-degree perineal tears at first birth on subsequent pregnancy outcomes: A cohort study. BJOG Int. J. Obstet. Gynaecol. (2014) 121 1695-1703. DOI: 10.1111/1471-0528.12886 13. Farrar D., Tuffnell D.J., Ramage C.. Interventions for women in subsequent pregnancies following obstetric anal sphincter injury to reduce the risk of recurrent injury and associated harms. Cochrane Database Syst. Rev. (2014) 11 CD010374. DOI: 10.1002/14651858.CD010374.pub2 14. Pretlove S.J., Radley S., Toozs-Hobson P.M., Thompson P.J., Coomarasamy A., Khan K.S.. Prevalence of anal incontinence according to age and gender: A systematic review and meta-regression analysis. Int. Urogynecol. J. (2006) 17 407-417. DOI: 10.1007/s00192-005-0014-5 15. Nevler A.. The epidemiology of anal incontinence and symptom severity scoring. Gastroenterol. Rep. (2014) 2 79-84. DOI: 10.1093/gastro/gou005 16. Fradet-Menard C., Deparis J., Gachon B., Sichitiu J., Pierre F., Fritel X., Desseauve D.. Obstetrical anal sphincter injuries and symptoms after subsequent deliveries: A 60 patient study. Eur. J. Obstet. Gynecol. Reprod. Biol. (2018) 226 40-46. DOI: 10.1016/j.ejogrb.2018.05.007 17. Bellanger M., Lima S.M., Cowppli-Bony A., Molinié F., Terry M.B.. Effects of fertility on breast cancer incidence trends: Comparing France and US. Cancer Causes Control (2021) 32 903-910. DOI: 10.1007/s10552-021-01440-2 18. Fynes M., Donnelly V., Behan M., O’Connell P.R., O’Herlihy C.. Effect of second vaginal delivery on anorectal physiology and faecal continence: A prospective study. Lancet (1999) 354 983-986. DOI: 10.1016/S0140-6736(98)11205-9 19. Fitzpatrick M., Cassidy M., Barassaud M.L., Hehir M.P., Hanly A.M., O’Connell P.R., O’Herlihy C.. Does anal sphincter injury preclude subsequent vaginal delivery?. Eur. J. Obstet. Gynecol. Reprod. Biol. (2016) 198 30-34. DOI: 10.1016/j.ejogrb.2015.12.004 20. Jordan P.A., Naidu M., Thakar R., Sultan A.H.. Effect of subsequent vaginal delivery on bowel symptoms and anorectal function in women who sustained a previous obstetric anal sphincter injury. Int. Urogynecology J. (2018) 29 1579-1588. DOI: 10.1007/s00192-018-3601-y 21. Scheer I., Thakar R., Sultan A.H.. Mode of delivery after previous obstetric anal sphincter injuries (OASIS)—A reappraisal?. Int. Urogynecol. J. Pelvic. Floor Dysfunct. (2009) 20 1095-1101. DOI: 10.1007/s00192-009-0908-8 22. Vaizey C.J., Carapeti E., A Cahill J., A Kamm M.. Prospective comparison of faecal incontinence grading systems. Gut (1999) 44 77-80. DOI: 10.1136/gut.44.1.77 23. Perry S., Shaw C., McGrother C., Matthews R., Assassa R.P., Dallosso H., Williams K., Brittain K., Azam U., Clarke M.. Prevalence of faecal incontinence in adults aged 40 years or more living in the community. Gut (2002) 50 480-484. DOI: 10.1136/gut.50.4.480 24. Gommesen D., Nohr E.A., Qvist N., Rasch V.. Obstetric perineal ruptures—Risk of anal incontinence among primiparous women 12 months postpartum: A prospective cohort study. Am. J. Obstet. Gynecol. (2019) 222 165.e1-165.e11. DOI: 10.1016/j.ajog.2019.08.026 25. Bouchard D., Abramowitz L., Castinel A., Suduca J.M., Staumont G., Soudan D., Devulder F., Pigot F., Varastet M., Ganansia R.. One-year outcome of haemorrhoidectomy: A prospective multicentre French study. Color. Dis. (2013) 15 719-726. DOI: 10.1111/codi.12090 26. Villot A., Deffieux X., Demoulin G., Rivain A.L., Trichot C., Thubert T.. Management of postpartum anal incontinence: A systematic review. Prog. Urol. (2015) 25 1191-1203. DOI: 10.1016/j.purol.2015.06.002 27. Torrisi G., Minini G., Bernasconi F., Perrone A., Trezza G., Guardabasso V., Ettore G.. A prospective study of pelvic floor dysfunctions related to delivery. Eur. J. Obstet. Gynecol. Reprod. Biol. (2012) 160 110-115. DOI: 10.1016/j.ejogrb.2011.10.010 28. Law P.J., Bartram C.I.. Anal endosonography: Technique and normal anatomy. Abdom. Imaging (1989) 14 349-353. DOI: 10.1007/BF01889235 29. Starck M., Bohe M., Valentin L.. Results of endosonographic imaging of the anal sphincter 2-7 days after primary repair of third- or fourth-degree obstetric sphincter tears. Ultrasound Obstet. Gynecol. (2003) 22 609-615. DOI: 10.1002/uog.920 30. Norderval S., Markskog A., Røssaak K., Vonen B.. Correlation between anal sphincter defects and anal incontinence following obstetric sphincter tears: Assessment using scoring systems for sonographic classification of defects. Ultrasound Obstet. Gynecol. (2007) 31 78-84. DOI: 10.1002/uog.5155 31. Marquis P., Amarenco G., Sapède C., Josserand F., McCarthy C., Zerbib M., Richard F., Jacquetin B., Villet R., Leriche B.. Elaboration and validation of a specific quality of life questionnaire for urination urgency in women. Prog. Urol. (1997) 7 56-63. PMID: 9116740 32. Hage-Fransen M.A.H., Wiezer M., Otto A., Wieffer-Platvoet M.S., Slotman M.H., Nijhuis-van der Sanden M.W.G., Pool-Goudzwaard A.L.. Pregnancy- and obstetric-related risk factors for urinary incontinence, fecal incontinence, or pelvic organ prolapse later in life: A systematic review and meta-analysis. Acta Obstet. Gynecol. Scand. (2021) 100 373-382. DOI: 10.1111/aogs.14027 33. Boggs E.W., Berger H., Urquia M., McDermott C.D.. Recurrence of Obstetric Third-Degree and Fourth-Degree Anal Sphincter Injuries. Obstet. Gynecol. (2014) 124 1128-1134. DOI: 10.1097/AOG.0000000000000523 34. Jangö H., Langhoff-Roos J., Rosthøj S., Saske A.. Long-term anal incontinence after obstetric anal sphincter injury—Does grade of tear matter?. Am. J. Obstet. Gynecol. (2018) 218 232.e1-232.e10. DOI: 10.1016/j.ajog.2017.11.569 35. Nelson R.L., E Furner S., Westercamp M., Farquhar C.. Cesarean delivery for the prevention of anal incontinence. Cochrane Database Syst. Rev. (2010) 2010 CD006756. DOI: 10.1002/14651858.CD006756.pub2 36. Chin K.. Obstetrics and Fecal Incontinence. Clin. Colon Rectal Surg. (2014) 27 110-112. DOI: 10.1055/s-0034-1383903 37. Altomare D.F., De Fazio M., Giuliani R.T., Catalano G., Cuccia F.. Sphincteroplasty for fecal incontinence in the era of sacral nerve modulation. World J. Gastroenterol. (2010) 16 5267-5271. DOI: 10.3748/wjg.v16.i42.5267 38. Damon H., Siproudhis L., Faucheron J.-L., Piche T., Abramowitz L., Eléouet M., Etienney I., Godeberge P., Valancogne G., Denis A.. Perineal retraining improves conservative treatment for faecal incontinence: A multicentre randomized study. Dig. Liver Dis. (2014) 46 237-242. DOI: 10.1016/j.dld.2013.11.002
--- title: Impact of Body Composition and Sarcopenia on Mortality in Chronic Obstructive Pulmonary Disease Patients authors: - Manuel Gómez-Martínez - Wendy Rodríguez-García - Dulce González-Islas - Arturo Orea-Tejeda - Candace Keirns-Davis - Fernanda Salgado-Fernández - Samantha Hernández-López - Angelia Jiménez-Valentín - Alejandra Vanessa Ríos-Pereda - Juan Carlos Márquez-Cordero - Mariana Salvatierra-Escobar - Iris López-Vásquez journal: Journal of Clinical Medicine year: 2023 pmcid: PMC9967244 doi: 10.3390/jcm12041321 license: CC BY 4.0 --- # Impact of Body Composition and Sarcopenia on Mortality in Chronic Obstructive Pulmonary Disease Patients ## Abstract Background: Patients with chronic obstructive pulmonary disease (COPD) have alterations in body composition, such as low cell integrity, body cell mass, and disturbances in water distribution evidenced by higher impedance ratio (IR), low phase angle (PhA), as well as low strength, low muscle mass, and sarcopenia. Body composition alterations are associated with adverse outcomes. However, according to the European Working Group on Sarcopenia in Older People 2 (EWGSOP2), the impact of these alterations on mortality in COPD patients is not well-established. Our aims were to evaluate whether low strength, low muscle mass, and sarcopenia impacted mortality in COPD patients. Methods: A prospective cohort study performance was conducted with COPD patients. Patients with cancer, and asthma were excluded. Body composition was assessed by bioelectrical impedance analysis. Low strength and muscle mass, and sarcopenia were defined according to EWGSOP2. Results: 240 patients were evaluated, of whom $32\%$ had sarcopenia. The mean age was 72.32 ± 8.24 years. The factors associated with lower risk of mortality were handgrip strength (HR:0.91, CI $95\%$; 0.85 to 0.96, $$p \leq 0.002$$), PhA (HR:0.59, CI $95\%$; 0.37 to 0.94, $$p \leq 0.026$$) and exercise tolerance (HR:0.99, CI $95\%$; 0.992 to 0.999, $$p \leq 0.021$$), while PhA below the 50th percentile (HR:3.47, CI $95\%$; 1.45 to 8.29, $$p \leq 0.005$$), low muscle strength (HR:3.49, CI $95\%$; 1.41 to 8.64, $$p \leq 0.007$$) and sarcopenia (HR:2.10, CI $95\%$; 1.02 to 4.33, $$p \leq 0.022$$) were associated with a higher risk of mortality. Conclusion: Low PhA, low muscle strength, and sarcopenia are independently associated with poor prognosis in COPD patients. ## 1. Introduction Chronic Obstructive Pulmonary Disease (COPD) is a treatable and avoidable illness characterized by persistent respiratory symptoms and progressive and irreversible airflow limitation [1]. COPD is considered a global public health problem because it is the third leading cause of death [2]. Patients with COPD have characteristics such as low cell integrity, body cell mass, disturbances in water distribution between intracellular and extracellular compartments as evidenced by low phase angle (PhA) and higher impedance ratio (IR), as well as low muscle strength, low muscle mass, and sarcopenia [3,4,5,6,7]. Body composition compartments can be assessed by raw bioelectrical impedance analysis (BIA) variables, such as IR and PhA, which provide information on hydration status, cellular mass and quality, nutritional status, and quality of life [8,9,10]. In COPD, some studies have reported significant alterations in these parameters [6,10,11]. In addition, they are independent predictors of mortality [4,6]. According to the updated consensus of the European Working Group on Sarcopenia in Older People 2 (EWGSOP2) published in 2019, sarcopenia is a progressive and generalized skeletal muscle disorder defined as low muscle strength and diminished muscle mass associated with an increased likelihood of adverse outcomes in healthy people and in those with pathologies. In clinical practice for diagnostic sarcopenia, low muscle strength is the primary parameter that defines the role of low muscle mass [12,13,14]. Sarcopenia can occur as a secondary condition in the presence of some diseases such as COPD. In affected patients, multiple factors have been associated with sarcopenia: age, disease severity, disuse, functional performance, oxidative stress, hypoxia, malnutrition, a higher catabolic state, and glucocorticoid use and pro-inflammatory states such as tumor necrosis factor (TNF-α), interleukin (IL)-6, and IL-1β [3,15,16]. The prevalence of sarcopenia in COPD patients is $27.5\%$ ($95\%$ CI; 18.4 to $36.5\%$) across different population settings and other defining categories [17]. Regarding the clinical impact of sarcopenia in COPD patients, several studies have reported lower pulmonary function, functional capacity, quality of life, and higher levels of inflammatory markers [10,11,17,18]. However, the impact of low muscle strength, low muscle mass, and sarcopenia, according to EWGSOP2, on mortality in COPD patients is not well-established. In this study, our aims were to evaluate whether low muscle strength, low muscle mass, and sarcopenia, according to EWGSOP2, impacted mortality in COPD patients. ## 2. Materials and Methods A prospective cohort study was performed with COPD out-patients at the Instituto Nacional de Enfermedades Respiratorias “Ismael Cosío Villegas” in Mexico City, Mexico, from 30 July 2015 to 31 March 2022. According to the recommendations of the Global Initiative for Chronic Obstructive Lung Disease (GOLD) [19], patients with a confirmed diagnosis of COPD were included. The subjects were >40 years old, and spirometry with a forced expiratory volume over 1 s (FEV1)/forced vital capacity ratio (FVC) ratio < 0.70 post-bronchodilator. Patients with diagnoses of cancer, human immunodeficiency virus, and asthma were excluded. ## 2.1. Data Collection Body composition, anthropometry, clinical and demographic variables were evaluated; these are an integral part of the clinical management of patients who come to our Institute. With respect to clinical variables, exacerbations in the previous year were characterized by worsening respiratory symptoms for two or more consecutive days with medication changes and requiring hospital admission. Hospitalization during follow-up was due to acute exacerbation of COPD or acute heart failure. ## 2.2. Anthropometry Weight and height were measured according to the manual reference of anthropometric standardization [20]. All subjects wore light clothing and were barefoot. Body mass index was calculated by dividing the total body weight (kilograms) by the squared height (meters). ## 2.3. Bioelectrical Impedance Analysis (BIA) Total body composition and raw variables were measured with whole-body BIA using four-pole multifrequency equipment BodyStat QuadScan 4000 (BodyStat, Isle of Man, UK) by standard technique [21]. The BIA method is based on injecting an alternating electric current of a minimal intensity below the sensing thresholds. The impedance Z represents the opposition shown by biological materials to the passage of an alternating electric current. The electrical impedance Z comprises the resistance (R) and reactance (Xc). The current passage determines R through the intracellular and extracellular electrolyte solutions, and the Xc is the delay in current flow measured as a phase shift, reflecting the dielectric properties of the cell mass and integrity of the cell membranes [22]. The measurements were conducted by the same operator, in the morning, in a comfortable area, free of drafts, with portable electric heaters. The area was cleaned before the study. The subjects were fasting and should not have exercised eight hours before or consumed alcohol 12 h before the study. During the entire study, the person was in a supine position with the arms separated from the trunk by about 30° and the legs separated by about 45°. Electrodes were placed on the hand and ipsilateral foot. PhA was calculated according to the previous equation: phase angle (degrees) = arctan (Xc/R) · (180/π) [23]. In our population, PhA below the 50th percentile value was <5.1 in men and <4.8 in women. IR was calculated as follows: the ratio of high (200 kHz) to low frequency (5 kHz) of multifrequency BIA [10]. In our population, IR above the 50th percentile was >0.83 in men and >0.84 in women. ## 2.4. Muscle Mass Appendicular skeletal muscle mass index (ASMMI) was assessed according to Sergi’s formula [24]: ASMMI (kg/m2) = [−3.964 + (0.227 * (Height2 (cm)/Resistance) + (0.095 * Weight) + (1.384 * Sex) + (0.064 * Reactance)/Height (m2)]. Low muscle mass was defined according to EWGSOP2 [14] as ASMMI < 7.0 kg/m2, and in women, ASMMI < 6.0 kg/m2. ## 2.5. Muscle Strength Muscle strength was measured with a mechanical Smedley Hand Dynamometer (Stoelting, Wood Dale, UK) according to the technique described in Rodríguez et al. [ 25]. Low muscle strength was defined according to EWGSOP2 [14] as the presence of low handgrip strength (in men, handgrip strength < 27 kg, and in women, handgrip strength < 16 kg). ## 2.6. Sarcopenia Sarcopenia was defined according to EWGSOP2 [14] as the presence of low muscle strength (in men, handgrip strength < 27 kg, and in women, handgrip strength < 16 kg) and low muscle mass (in men, ASMMI < 7.0 kg/m2, and in women, ASMMI < 6.0 kg/m2). ## 2.7. Exercise Tolerance Exercise tolerance was assessed by a 6-min walk test, which was performed according to American Thoracic Society standards [26]. ## 2.8. Pulmonary Function Spirometry testing was conducted by an experienced pulmonary technician using a portable spirometer (EasyOnePC, Ndd Medical Technologies Inc., Zürich, Switzerland) according to the criteria of the American Thoracic Society/European Respiratory Society standards [27]. The reference values used for spirometry were obtained in Mexican-American individuals [28]. ## 2.9. Endpoint The endpoint was defined as mortality from all causes. ## 2.10. Statistical Analysis Analyses were performed using a commercially available STATA version 14 (Stata Corp., College Station, TX, USA). Categorical variables were presented as frequencies and percentages. The Shapiro–Wilk test was used to test the normality of continuous variables; continuous variables with normal distribution were presented as mean and standard deviation, and non-normal variables were presented as median and percentiles 25–75. A comparison among study groups was analyzed with a chi-square test or Fisher’s F test for categorical variables and unpaired Student’s t-test or Mann–Whitney U tests for continuous variables. Finally, bivariate Cox’s proportional hazards analysis was performed to evaluate the impact of body composition and sarcopenia on mortality. Subsequently, the multivariate Cox proportional hazards model was adjusted by clinical variables $p \leq 0.100$ in the bivariate model. Due to collinearity, body composition and BODE index variables were not included in the model. $p \leq 0.05$ was considered statistically significant. ## 3. Results Two hundred forty patients with COPD were evaluated. During 6.66 years of follow-up, 31 subjects died. The median survival time was 941 days (range 411–1343.5). The mean population age was 72.32 ± 8.24 years; $51.25\%$ were men; $25.83\%$ had diabetes, $49.17\%$ had systemic hypertension, $69.58\%$ were overweight or obese, $42.08\%$ were in heart failure, and $32\%$ had sarcopenia; $22.5\%$ had hospitalization during the follow-up, and $17.92\%$ had exacerbations in the previous year. The mean of the 6-min walk test distance was 319.04 m ± 133.06 (Table 1). Patients who succumbed were older and had a higher BODE index, more hospitalizations during the follow-up and more exacerbations in the previous year, as well as lower PhA. They had a higher prevalence of low muscle strength, sarcopenia, and less exercise tolerance than patients who survived (Table 1). In the bivariate model, age, BODE index, PhA below the 50th percentile, low muscle strength, and sarcopenia were associated with more mortality risk. In contrast, PhA was associated with less mortality risk (Table 2). Multivariate analysis showed that handgrip strength (HR: 0.91, CI $95\%$; 0.85 to 0.96, $$p \leq 0.002$$), PhA (HR: 0.59, CI $95\%$; 0.37 to 0.94, $$p \leq 0.026$$), and exercise tolerance (HR: 0.99, CI $95\%$; 0.992 to 0.999, $$p \leq 0.022$$) were associated with a lower risk of mortality, while PhA below the 50th percentile (HR: 3.47, CI $95\%$; 1.45 to 8.29, $$p \leq 0.005$$), low muscle strength (HR: 3.49, CI $95\%$; 1.41 to 8.64, $$p \leq 0.007$$) and sarcopenia (HR: 2.10, CI $95\%$; 1.02 to 4.33, $$p \leq 0.043$$) were associated with a higher risk of mortality (Table 3 and Figure 1). ## 4. Discussion The main finding was the impact of body composition components such as handgrip strength, PhA, low muscle strength, and sarcopenia on prognosis in COPD patients. COPD patients have significant body composition alterations, including sarcopenia. In our study, the prevalence was $32\%$, independently associated with a two-fold increase in the risk of death (HR: 2.10, CI $95\%$; 1.02 to 4.33, $$p \leq 0.043$$). Similar results were found by Schols et al., who showed that COPD patients with cachexia, evaluated by low fat-free mass index (FFMI) and low BMI, have a greater mortality risk than patients without impaired FFMI [29]. Benz et al. showed that sarcopenia is an independent risk factor for mortality in COPD patients [30]. Similar results have been found in diverse populations, such as geriatric [13] and dialysis patients [31]. In addition, sarcopenia is associated with functional disability, a higher rate of falls, a risk of hospitalization incidents, and lower pulmonary function [11,12]. Muscle strength is presently the most reliable measure of muscle function; low muscle strength is the primary parameter for diagnosing sarcopenia, according to EWGSOP2, and it is defined as probable sarcopenia [14]. In addition, decreased handgrip strength is significantly associated with moderate-to-very severe airflow limitation in the general population [32]. Our results showed that subjects who did not survive had lower handgrip strength than patients who survived. Subjects with low muscle strength had a threefold increase in the risk of death adjusted for confounding variables. Similar results have been observed in COPD and other populations [5,33]. In a healthy aging population, muscular strength declines 1–$2\%$ per year [32], while for COPD patients this decline is estimated at $4.3\%$ per year [34]. The pathological mechanisms involved in this muscle dysfunction in COPD subjects are corticosteroid use, hypoxemia, systemic inflammation, and oxidative stress. These factors play a fundamental role in affecting anabolism, and are associated with protein catabolism, mitochondrial dysfunction, lower plasma amino acids, and reduction of type 1 fibers in the peripheral muscles, among others [35]. Byun et al. showed a negative correlation between handgrip strength and skeletal muscle mass index with IL-6 and TNF-α but this was not found with skeletal muscle mass index. Higher TNF-α levels were also associated with sarcopenia in patients with stable COPD [36]. Skeletal muscle is the human body’s largest organ and accounts for 40–$50\%$ of total body weight under physiological conditions. In COPD patients with low FFMI, lower muscle strength has been observed compared with normal FFMI [4]. Regarding the impact of low muscle mass on prognosis, the evidence is not conclusive; some studies show that low muscle mass is a predictor of overall mortality in some populations, such as COPD patients and geriatric people [37,38]. In contrast, other studies have observed no association with exacerbations, days of hospitalization, quality of life, and mortality [6,18]. In our study, no association was observed between mortality and ASMMI or low muscle mass defined according to EWGSOP2 (in men, ASMMI < 7.0 kg/m2 and in women, ASMMI < 6.0 kg/m2) [5]. Although the impact of muscle mass is controversial, skeletal muscle is an endocrine organ with multiple metabolic functions such as energy homeostasis, heat regulation, insulin sensitivity, and amino acid metabolism [39]. Another of the determining factors in the prognosis of COPD patients is exercise tolerance. In our study, we observed that per meter increase in the 6-min walk test, there is a $1\%$ reduction in the probability of mortality (HR: 0.99, CI $95\%$; 0.992 to 0.999); similar results have been observed in different studies [6,40]. Low BMI is a significant risk factor for mortality in COPD [41,42] This implies that the subject has a cachexia diagnosis [43]. However, BMI has severe limitations: it is calculated as the ratio of body weight to height squared, which does not allow determining the distribution between the different components of body composition, such as muscle mass, fat mass, and hydration status, which have an independent impact on the prognosis of the subjects [6,29]. In our study, BMI was not associated with a worse prognosis, probably due to the high prevalence of overweight or obesity. A high prevalence of low muscle mass and muscle strength were observed. When determining the impact of these variables on the prognosis of patients with COPD, we observed that muscle strength is a stronger predictor for mortality than BMI or muscle mass. Similar results have been observed in other studies [6,29]. PhA and IR are raw BIA variables that provide information on water distribution between intracellular and extracellular compartments and cellular integrity [6,8,9]. Concerning PhA, in this study, non-surviving patients had lower PhA than patients who survived, and PhA below the 50th percentile was independently associated with mortality. Similar results were observed by Maddocks et al., who showed that COPD patients with low PhA (<4.5°) had lower quadriceps strength and quality of life, as well as more exacerbations and hospital admissions [4]. Moreover, low PhA is associated with an impaired pulmonary function [44]. Likewise, COPD patients have been observed to be malnourished and malnourished with systemic inflammation, and sarcopenic subjects have higher values of IR. Other studies on COPD patients showed that subjects with IR > 0.84 had decreased handgrip strength, skeletal muscle mass index, and diffusion capacity for carbon monoxide than subjects with IR < 0.84. In addition, IR > 0.84 was associated with lower FEV1 and FVC [44]. Blasio et al. showed that IR is an independent predictor of all-cause mortality in COPD [6]. However, in our study, IR above the 50th percentile was not associated with mortality in the bivariate and multivariate model. It is vital to evaluate and identify alterations in body composition, such as low PhA, low muscle strength and sarcopenia, since early attention to these alterations can positively impact clinical variables and the prognosis of patients with COPD. In a meta-analysis of randomized controlled clinical trials, Bernardes et al. evaluated the effect of energy and/or protein oral nutritional supplements or food fortification on body composition parameters. They found improvement in midarm muscle circumference, handgrip strength, and lean body mass [44]. In addition, pulmonary rehabilitation also showed improvement in exercise tolerance, quality of life, respiratory muscle strength, etc. [ 45,46]. A low FEV1 has been associated with an increased risk of hospitalizations and death in COPD patients [41,47]. However, in this study, no differences were observed between the survivor and non-survivor groups. These may be because other risk factors impact the prognosis of COPD subjects, such as short-distance walking, low body mass index, and a high degree of functional breathlessness. The BODE index is a multidimensional index of disease severity in COPD that incorporates these risk factors, which is a better predictor of prognosis in COPD patients [41,48]. In our study, the BODE index was associated with a worse prognosis. ## Strengths and Limitations of the Study Among the limitations of this study is the moderate sample size and a long recruitment period, which could impact the representativeness of the data. However, our study’s strength is that it is a prospective cohort, so we were able to assess the causality, and we also performed multivariate models, which allowed us to adjust for possible confounding variables. ## 5. Conclusions Low PhA, low muscle strength, and sarcopenia are independently associated with poor prognosis in COPD patients. ## References 1. Mirza S., Clay R.D., Koslow M.A., Scanlon P.D.. COPD Guidelines: A Review of the 2018 GOLD Report. Mayo Clin. Proc. (2018.0) 93 1488-1502. DOI: 10.1016/j.mayocp.2018.05.026 2. The Top 10 Causes of Death 3. Jones S.E., Maddocks M., Kon S.S., Canavan J.L., Nolan C.M., Clark A.L., Polkey M.I., Man W.D.. Sarcopenia in COPD: Prevalence, clinical correlates and response to pulmonary rehabilitation. Thorax (2015.0) 70 213-218. DOI: 10.1136/thoraxjnl-2014-206440 4. Maddocks M., Kon S.S., Jones S.E., Canavan J.L., Nolan C.M., Higginson I.J., Gao W., Polkey M.I., Man W.D.. Bioelectrical impedance phase angle relates to function, disease severity and prognosis in stable chronic obstructive pulmonary disease. Clin. Nutr. (2015.0) 34 1245-1250. DOI: 10.1016/j.clnu.2014.12.020 5. Burtin C., Ter Riet G., Puhan M.A., Waschki B., Garcia-Aymerich J., Pinto-Plata V., Celli B., Watz H., Spruit M.A.. Handgrip weakness and mortality risk in COPD: A multicentre analysis. Thorax (2016.0) 71 86-87. DOI: 10.1136/thoraxjnl-2015-207451 6. De Blasio F., Scalfi L., Di Gregorio A., Alicante P., Bianco A., Tantucci C., Bellofiore B., de Blasio F.. Raw bioelectrical impedance analysis variables are independent predictors of early all-cause mortality in patients with COPD. Chest (2019.0) 155 1148-1157. DOI: 10.1016/j.chest.2019.01.001 7. Orea-Tejeda A., Gómez-Martínez M., González-Islas D., Flores-Cisneros L., Keirns-Davis C., Sánchez-Santillán R., Pérez-García I., Martínez-Luna N., Robles-Hernández R., Sánchez-Moreno C.. The impact of hydration status and fluid distribution on pulmonary function in COPD patients. Sci. Rep. (2022.0) 12 1216. DOI: 10.1038/s41598-022-05192-0 8. Norman K., Stobäus N., Pirlich M., Bosy-Westphal A.. Bioelectrical phase angle and impedance vector analysis--clinical relevance and applicability of impedance parameters. Clin. Nutr. (2012.0) 31 854-861. DOI: 10.1016/j.clnu.2012.05.008 9. Norman K., Wirth R., Neubauer M., Eckardt R., Stobäus N.. The bioimpedance phase angle predicts low muscle strength, impaired quality of life, and increased mortality in old patients with cancer. J. Am. Med. Dir. Assoc. (2015.0) 16 173.E17-173.E22. DOI: 10.1016/j.jamda.2014.10.024 10. De Blasio F., Di Gregorio A., de Blasio F., Bianco A., Bellofiore B., Scalfi L.. Malnutrition and sarcopenia assessment in patients with chronic obstructive pulmonary disease according to international diagnostic criteria, and evaluation of raw BIA variables. Respir. Med. (2018.0) 134 1-5. DOI: 10.1016/j.rmed.2017.11.006 11. Martínez-Luna N., Orea-Tejeda A., González-Islas D., Flores-Cisneros L., Keirns-Davis C., Sánchez-Santillán R., Pérez-García I., Gastelum-Ayala Y., Martínez-Vázquez V., Martínez-Reyna Ó.. Association between body composition, sarcopenia and pulmonary function in chronic obstructive pulmonary disease. BMC Pulm. Med. (2022.0) 22. DOI: 10.1186/s12890-022-01907-1 12. Beaudart C., Zaaria M., Pasleau F., Reginster J.Y., Bruyère O.. Health Outcomes of Sarcopenia: A Systematic Review and Meta-Analysis. PLoS ONE (2017.0) 12. DOI: 10.1371/journal.pone.0169548 13. Zhang X., Wang C., Dou Q., Zhang W., Yang Y., Xie X.. Sarcopenia as a predictor of all-cause mortality among older nursing home residents: A systematic review and meta-analysis. BMJ Open (2018.0) 8 e021252. DOI: 10.1136/bmjopen-2017-021252 14. Cruz-Jentoft A.J., Bahat G., Bauer J., Boirie Y., Bruyère O., Cederholm T., Cooper C., Landi F., Rolland Y., Sayer A.A.. Sarcopenia: Revised European consensus on definition and diagnosis. Age Ageing (2019.0) 48 16-31. DOI: 10.1093/ageing/afy169 15. Eagan T.M., Gabazza E.C., D’Alessandro-Gabazza C., Gil-Bernabe P., Aoki S., Hardie J.A., Bakke P.S., Wagner P.D.. TNF-α is associated with loss of lean body mass only in already cachectic COPD patients. Respir. Res. (2012.0) 13 48. DOI: 10.1186/1465-9921-13-48 16. Langen R.C., Gosker H.R., Remels A.H., Schols A.M.. Triggers and mechanisms of skeletal muscle wasting in chronic obstructive pulmonary disease. Int. J. Biochem. Cell Biol. (2013.0) 45 2245-2256. DOI: 10.1016/j.biocel.2013.06.015 17. Sepúlveda-Loyola W., Osadnik C., Phu S., Morita A.A., Duque G., Probst V.S.. Diagnosis, prevalence, and clinical impact of sarcopenia in COPD: A systematic review and meta-analysis. J. Cachexia Sarcopenia Muscle (2020.0) 11 1164-1176. DOI: 10.1002/jcsm.12600 18. Lin B., Bai L., Wang S., Lin H.. The association of systemic interleukin 6 and interleukin 10 levels with sarcopenia in elderly patients with chronic obstructive pulmonary disease. Int. J. Gen. Med. (2021.0) 14 5893-5902. DOI: 10.2147/IJGM.S321229 19. Vogelmeier C.F., Criner G.J., Martinez F.J., Anzueto A., Barnes P.J., Bourbeau J., Celli B.R., Chen R., Decramer M., Fabbri L.M.. Global strategy for the diagnosis, management, and prevention of chronic obstructive lung disease 2017 report. GOLD executive summary. Am. J. Respir. Crit. Care Med. (2017.0) 195 557-582. DOI: 10.1164/rccm.201701-0218PP 20. Lohman T.G., Roche A.F., Martorell R.. Anthropometric Standardization Reference Manual (1988.0) 21. Lukaski H.C., Johnson P.E., Bolonchuk W.W., Lykken G.I.. Assessment of fat-free mass using bioelectrical impedance measurements of the human body. Am. J. Clin. Nutr. (1985.0) 41 810-817. DOI: 10.1093/ajcn/41.4.810 22. Piccoli A., Nescolarde L., Rosell J.. Análisis convencional y vectorial de bioimpedancia en la práctica clínica. Nefrología (2002.0) 22 228-238. PMID: 12123122 23. Kyle U.G., Genton L., Pichard C.. Low phase angle determined by bioelectrical impedance analysis is associated with malnutrition and nutritional risk at hospital admission. Clin. Nutr. (2013.0) 32 294-299. DOI: 10.1016/j.clnu.2012.08.001 24. Sergi G., De Rui M., Veronese N., Bolzetta F., Berton L., Carraro S., Bano G., Coin A., Manzato E., Perissinotto E.. Assessing appendicular skeletal muscle mass with bioelectrical impedance analysis in free-living Caucasian older adults. Clin. Nutr. (2015.0) 34 667-673. DOI: 10.1016/j.clnu.2014.07.010 25. Rodríguez-García W.D., García-Castañeda L., Orea-Tejeda A., Mendoza-Núñez V., González-Islas D.G., Santillán-Díaz C., Castillo-Martínez L.. Handgrip strength: Reference values and its relationship with bioimpedance and anthropometric variables. Clin. Nutr. ESPEN (2017.0) 19 54-58. DOI: 10.1016/j.clnesp.2017.01.010 26. ATS statement: Guidelines for the six-minute walk test. Am. J. Respir. Crit. Care Med. (2002.0) 166 111-117. DOI: 10.1164/ajrccm.166.1.at1102 27. Miller M.R., Hankinson J., Brusasco V., Burgos F., Casaburi R., Coates A., Crapo R., Enright P., van der Grinten C.P., Gustafsson P.. Standardisation of spirometry. Eur. Respir. J. (2005.0) 26 319-338. DOI: 10.1183/09031936.05.00034805 28. Hankinson J.L., Odencrantz J.R., Fedan K.B.. Spirometric reference values from a sample of the general U.S. population. Am. J. Respir. Crit. Care Med. (1999.0) 159 179-187. DOI: 10.1164/ajrccm.159.1.9712108 29. Schols A.M., Broekhuizen R., Weling-Scheepers C.A., Wouters E.F.. Body composition and mortality in chronic obstructive pulmonary disease. Am. J. Clin. Nutr. (2005.0) 82 53-59. DOI: 10.1093/ajcn/82.1.53 30. Benz E., Wijnant S.R.A., Trajanoska K., Arinze J.T., de Roos E.W., de Ridder M., Williams R., van Rooij F., Verhamme K.M.C., Ikram M.A.. Sarcopenia, systemic immune-inflammation index and all-cause mortality in middle-aged and older people with COPD and asthma: A population-based study. ERJ Open Res. (2022.0) 8 628-2021. DOI: 10.1183/23120541.00628-2021 31. Ribeiro H.S., Neri S.G.R., Oliveira J.S., Bennett P.N., Viana J.L., Lima R.M.. Association between sarcopenia and clinical outcomes in chronic kidney disease patients: A systematic review and meta-analysis. Clin. Nutr. (2022.0) 41 1131-1140. DOI: 10.1016/j.clnu.2022.03.025 32. Kim S., Yoon H.K., Rhee C.K., Jung H.W., Lee H., Jo Y.S.. hand grip strength and likelihood of moderate-to-severe airflow limitation in the general population. Int. J. Chronic Obstr. Pulm. Dis. (2022.0) 17 1237-1245. DOI: 10.2147/COPD.S364351 33. Celis-Morales C.A., Welsh P., Lyall D.M., Steell L., Petermann F., Anderson J., Iliodromiti S., Sillars A., Graham N., Mackay D.F.. Associations of grip strength with cardiovascular, respiratory, and cancer outcomes and all cause mortality: Prospective cohort study of half a million UK Biobank participants. BMJ (2018.0) 361 k1651. DOI: 10.1136/bmj.k1651 34. Hopkinson N.S., Tennant R.C., Dayer M.J., Swallow E.B., Hansel T.T., Moxham J., Polkey M.I.. A prospective study of decline in fat free mass and skeletal muscle strength in chronic obstructive pulmonary disease. Respir. Res. (2007.0) 8 25. DOI: 10.1186/1465-9921-8-25 35. Couillard A., Prefaut C.. From muscle disuse to myopathy in COPD: Potential contribution of oxidative stress. Eur. Respir. J. (2005.0) 26 703-719. DOI: 10.1183/09031936.05.00139904 36. Byun M.K., Cho E.N., Chang J., Ahn C.M., Kim H.J.. Sarcopenia correlates with systemic inflammation in COPD. Int. J. Chronic Obstr. Pulm. Dis. (2017.0) 12 669-675. DOI: 10.2147/COPD.S130790 37. Vestbo J., Prescott E., Almdal T., Dahl M., Nordestgaard B.G., Andersen T., Sørensen T.I., Lange P.. Body mass, fat-free body mass, and prognosis in patients with chronic obstructive pulmonary disease from a random population sample: Findings from the Copenhagen City Heart Study. Am. J. Respir. Crit. Care Med. (2006.0) 173 79-83. DOI: 10.1164/rccm.200506-969OC 38. Batsis J.A., Mackenzie T.A., Emeny R.T., Lopez-Jimenez F., Bartels S.J.. Low lean mass with and without obesity, and mortality: Results from the 1999-2004 national health and nutrition examination survey. J. Gerontol. Ser. A Biol. Sci. Med. Sci. (2017.0) 72 1445-1451. DOI: 10.1093/gerona/glx002 39. Schnyder S., Handschin C.. Skeletal muscle as an endocrine organ: PGC-1α, myokines and exercise. Bone (2015.0) 80 115-125. DOI: 10.1016/j.bone.2015.02.008 40. Höglund J., Boström C., Sundh J.. Six-minute walking test and 30 s chair-stand-test as predictors of mortality in COPD—A cohort study. Int. J. Chronic Obstr. Pulm. Dis. (2022.0) 17 2461-2469. DOI: 10.2147/COPD.S373272 41. Celli B.R., Cote C.G., Marin J.M., Casanova C., Montes de Oca M., Mendez R.A., Pinto Plata V., Cabral H.J.. The body-mass index, airflow obstruction, dyspnea, and exercise capacity index in chronic obstructive pulmonary disease. N. Engl. J. Med. (2004.0) 350 1005-1012. DOI: 10.1056/NEJMoa021322 42. Tsimogianni A.M., Papiris S.A., Stathopoulos G.T., Manali E.D., Roussos C., Kotanidou A.. Predictors of outcome after exacerbation of chronic obstructive pulmonary disease. J. Gen. Intern. Med. (2009.0) 24 1043-1048. DOI: 10.1007/s11606-009-1061-2 43. Evans W.J., Morley J.E., Argilés J., Bales C., Baracos V., Guttridge D., Jatoi A., Kalantar-Zadeh K., Lochs H., Mantovani G.. Cachexia: A new definition. Clin. Nutr. (2008.0) 27 793-799. DOI: 10.1016/j.clnu.2008.06.013 44. Bernardes S., Eckert I.C., Burgel C.F., Teixeira P.J.Z., Silva F.M.. Increased energy and/or protein intake improves anthropometry and muscle strength in COPD patients: A systematic review with meta-analysis on randomized controlled clinical trials. Br. J. Nutr. (2022.0) 1-55. DOI: 10.1017/S0007114522000976 45. Beaumont M., Forget P., Couturaud F., Reychler G.. Effects of inspiratory muscle training in COPD patients: A systematic review and meta-analysis. Clin. Respir. J. (2018.0) 12 2178-2188. DOI: 10.1111/crj.12905 46. Lee E.N., Kim M.J.. Meta-analysis of the effect of a pulmonary rehabilitation program on respiratory muscle strength in patients with chronic obstructive pulmonary disease. Asian Nurs. Res. (2019.0) 13 1-10. DOI: 10.1016/j.anr.2018.11.005 47. Siafakas N.M., Vermeire P., Pride N.B., Paoletti P., Gibson J., Howard P., Yernault J.C., Decramer M., Higenbottam T., Postma D.S.. Optimal assessment and management of chronic obstructive pulmonary disease (COPD). The European respiratory society task force. Eur. Respir. J. (1995.0) 8 1398-1420. DOI: 10.1183/09031936.95.08081398 48. Marin J.M., Carrizo S.J., Casanova C., Martinez-Camblor P., Soriano J.B., Agusti A.G., Celli B.R.. Prediction of risk of COPD exacerbations by the BODE index. Respir. Med. (2009.0) 103 373-378. DOI: 10.1016/j.rmed.2008.10.004
--- title: 'Wound Healing after Acellular Dermal Substitute Positioning in Dermato-Oncological Surgery: A Prospective Comparative Study' authors: - Alessia Paganelli - Andrea Giovanni Naselli - Laura Bertoni - Elena Rossi - Paola Azzoni - Alessandra Pisciotta - Anna Maria Cesinaro - Luisa Benassi - Shaniko Kaleci - Federico Garbarino - Barbara Ferrari - Chiara Fiorentini - Camilla Reggiani - Cristina Magnoni journal: Life year: 2023 pmcid: PMC9967245 doi: 10.3390/life13020463 license: CC BY 4.0 --- # Wound Healing after Acellular Dermal Substitute Positioning in Dermato-Oncological Surgery: A Prospective Comparative Study ## Abstract Background: MatriDerm and Integra are both widely used collagenic acellular dermal matrices (ADMs) in the surgical setting, with similar characteristics in terms of healing time and clinical indication. The aim of the present study is to compare the two ADMs in terms of clinical and histological results in the setting of dermato-oncological surgery. Methods: Ten consecutive patients with medical indications to undergo surgical excision of skin cancers were treated with a 2-step procedure at our Dermatologic Surgery Unit. Immediately after tumor removal, both ADMs were positioned on the wound bed, one adjacent to the other. Closure through split-thickness skin grafting was performed after approximately 3 weeks. Conventional histology, immunostaining and ELISA assay were performed on cutaneous samples at different timepoints. Results: No significant differences were detected in terms of either final clinical outcomes or in extracellular matrix content of the neoformed dermis. However, Matriderm was observed to induce scar retraction more frequently. In contrast, Integra was shown to carry higher infectious risk and to be more slowly reabsorbed into the wound bed. Sometimes foreign body-like granulomatous reactions were also observed, especially in Integra samples. Conclusions: Even in the presence of subtle differences between the ADMs, comparable global outcomes were demonstrated after dermato-oncological surgery. ## 1. Introduction Human skin acts as a barrier against external agents and pathogens [1], prevents water loss [2] and is crucial for vitamin D metabolism [3]. Therefore, loss of cutaneous integrity triggers an evolutionary-conserved sequential mechanism of wound healing aimed at restoring skin architecture [4,5]. Wound healing is composed of four different phases (hemostasis, inflammation, proliferation, and remodeling), and requires a complex orchestration of interactions among many different types of cells, including not only keratinocytes and fibroblasts, but also immune cells, endothelial cells, macrophages, and mesenchymal stromal cells (MSCs) [6,7,8]. Impairment in any phase of the wound healing process could lead to chronic ulcer formation and/or aberrant scarring (e.g., excessive retraction, hypertrophic/keloidal scars). The presence of specific risk factors, such as diabetes and peripheral vascular disease, has proven to be associated with impaired wound healing [9]. Such conditions do not only have a direct impact on patient quality of life, but also represent a burden in term of costs for the healthcare system [10]. For these reasons, dermatological research in the regenerative setting is aimed at finding innovative strategies for complete skin regeneration with restoral of physiological skin architecture [11]. However, at present, only up to about $80\%$ of the skin’s original tensile strength is regained in optimal healing conditions [12]. In the dermato-oncological setting, surgical excision of cutaneous neoplasms brings the implicit need for therapeutic interruption of skin integrity, which sometimes poses great challenges in terms of reconstructive and healing strategies. For small wounds, primary suture immediately after surgical excision represents the gold-standard treatment [13]. Larger skin defects often impose a need for using different techniques for skin reconstruction [14]. One possibility is using advanced dressings for inducing faster secondary-intention healing [15,16] Possible alternatives include covering the wound surface through in vitro expanded epidermal sheets [17], skin grafts or flaps [18,19], and even the use of nanotechnologies [20], and/or stem-cell based therapies [21]. Currently, skin grafts and flaps represent the most commonly used strategies in dermatologic surgery. The use of cutaneous flaps in dermato-oncology is sometimes limited by tumor and subsequent wound size, while the only fundamental requirement for successful grafting is represented by a well-vascularized wound bed. Moreover, grafts also represent the reconstructive strategy of choice when tumor margins are not clearly definable with non-invasive techniques (such as dermoscopy, confocal microscopy, and OCT) and/or Mohs surgery is not feasible. However, grafts alone sometimes provide unsatisfactory functional and aesthetic results, possibly due to both scar contractures and excessive wound depth, with subsequent apparent depression of the grafted area [22]. In recent decades, acellular dermal matrices (ADMs) have widely been employed in this setting with the aim of reducing scarring and replacing the excised dermal compartment [23]. Various bioengineered scaffolding materials have been developed in order to provide a provisional template for skin cell migration and proliferation, therefore promoting wound healing and reducing scar tissue formation [24]. Two collagenic ADMs, Matriderm (MedSkin Solution Dr. Suwelack AG, Billerbeck, Germany) and Integra (Integra LifeSciences, Plainsboro, NJ, USA) are currently commonly used in dermato-oncological reconstructive surgery, and represent the two available options in our center [25,26]. The primary goal of the present study was to evaluate and compare the clinical outcomes of the two different ADMs using the Vancouver Scar Scale (VSS) and assessing the occurrence of ADM-specific adverse events. We also aimed at assessing ADM-induced architectural changes in the extracellular matrix (ECM) at a histopathological level. Detection of specific cell populations (e.g., myo-fibroblasts, endothelial cells, MSCs) and quantification of ECM components were also considered as secondary objectives of our research. ## 2.1. Study Design An interventional, prospective, comparative clinical pilot study was performed at the Dermatological Surgery Unit of Modena University Hospital. The study was approved by our institutional review board (Protocol n. CE $\frac{4342}{20}$), and all study procedures were performed in accordance with Helsinki declaration principles. Written informed consent was obtained from all participants before undergoing any study procedures. Ten consecutive patients with medical indications to undergo demolitive dermatologic surgery for large skin cancers, not suitable for classical reconstruction with flaps, were included in the present study. Exclusion criteria included: age < 18 years old; inability to give informed consent or to complete the procedures required for study completion; pregnancy or breastfeeding; known allergy to any component of the dermal substitute; lesions located on palms, soles, genitalia, or in the face area; lesions involving bone and/or periosteum; immunodeficiency; heavy smoking (>10 pack/year); uncontrolled diabetes, osteomalacia, thyroid disorders; connective tissue diseases. ## 2.2. Investigational Study Devices Two ADMs were used, MatriDerm and Integra. Integra (Integra LifeSciences, Plainsboro, NJ, USA) is a device composed of two layers: a synthetic dermis made of a bovine collagen lattice covalently linked to chondroitin-6-sulfate, derived from shark cartilage, and covered with a silastic epidermis (silicone sheet) [27]. Matriderm (MedSkin Solution Dr. Suwelack AG, Billerbeck, Germany) is a highly porous membrane composed of three-dimensionally coupled collagen and elastin. Bovine dermis is used to obtain the collagenic part of the matrix, while elastin is obtained from the bovine nuchal ligament by hydrolysis [28]. Despite Matriderm being often applied underneath a split-thickness skin graft in the setting of one-stage surgery [29,30], we used both ADMs according to a two-step protocol. Trans-epidermal water loss (TEWL) in the Matriderm-treated area was minimized through external dressings with paraffin gauze and external bandages, due to a lack of silicone coverage of this device. In both cases, 2 mm-thick templates were used. ## 2.3. Study Protocol Patients underwent a two-step surgical procedure with initial dermal substitute positioning and subsequent skin grafting (for protocol flowchart, see Figure 1). Both the ADMs were positioned in the wound bed, with half of the treated area being covered with MatriDerm and the other half with Integra, in order to perform intra-patient comparison, therefore eliminating potential biases due to interindividual variability. After ADM positioning, all patients included in the study underwent standard wound-care visits at our center twice weekly. Iodopovidone and saline solution were used respectively on Integra and Matriderm and non-adherent dressings were then positioned directly on the ADMs. Microbiological swabs were performed in case of clinical signs of infection. Skin-graft reconstruction (with 0.5–0.6 mm grafts taken either from the thigh or the axillary region) was performed after 3 weeks from the first surgical intervention, after histological confirmation of complete excision with clear margins. External non adherent dressings were applied directly on the graft and changed twice weekly until suture stitch removal. Clinical pictures were collected immediately before and after the first surgical procedure (t0), at the first two post-surgical visits (t1 and t2), at the time of grafting (t3), after 3 months from surgery (t4) (See Figure 2). ADM samples were also collected at t1 and t2 on days 2 and 7 (±2), respectively, after ADM positioning. Skin biopsies were performed intraoperatively during the second surgical session with a 4-mm punch (t3). Long-term clinical outcomes were also evaluated 3 months later, during follow-up visits (t4); the Vancouver Scar Scale [31] was used for surgical scar evaluation at t4. ## 2.4. Conventional Histology ADM and/or skin samples obtained at t1, t2 and t3 were fixed in formalin, paraffin-embedded, and stained with conventional hematoxylin and eosin (HE). At t1 and t2, ADM colonization and the presence of inflammatory infiltrate were evaluated. At t3, several histopathological parameters were considered to investigate cellular and structural characteristics of the neodermis: persistence of dermal substitute, vascularization, granulation phase (early/late), re-epithelization, abundance and type of the inflammatory infiltrate, presence of foreign-body reaction. Histopathological specimens were evaluated and quantified in three random sections of each sample by a panel of three blinded experts. Histological images were obtained using a Nikon Labophot-2 light microscope with a DS-5Mc CCD camera. As for vascularization, immunohistochemical staining for CD31 (BK3528S PECAM-1, Cell Signaling, Danvers, MA, USA) was performed on deparaffinized sections obtained at t3 to confirm histological data on vascularization. Immunohistochemical images were obtained using a Nikon Labophot-2 light microscope with a DS-5Mc CCD camera. Finally, vessels were also selected on acquired images (3 sections for each sample) through manual selection of CD31+ vessels, and the vascularized area was calculated through ImageJ software (version 1.45b) and expressed as percentage area occupied by vessels/total area. ## 2.5. Immunofluorescence OCT (Tissue-Teck)-embedded culture samples were cryopreserved at −80 °C and used to prepare 4 μm thick sections using a cryotome (LEICA 1720 rotary cryotome, Nussloch, Germany). An immunofluorescent (IF) stain for alpha-SMA (smooth muscle actin) and both single stain and co-staining for CD90 and Stro-1 were performed on these sections. Immunofluorescent stains were repeated on deparaffinized sections, due to poor quality of the neodermis architecture appreciable after freezing and thawing. After specific site blockage with PBS/BSA, the neodermis sections were incubated with mouse anti-Stro1 and rabbit anti-CD90 antibodies (MAB1038, RD systems; JF-10-09, Invitrogen) for 1 h at room temperature. A signal was visualized with goat polyclonal anti-mouse FITC-conjugated and swine anti-rabbit TRITC-conjugated secondary antibodies (Abcam, Cambridge, UK). Finally, sections were rinsed, permeabilized with $0.1\%$ Tryton X-100 for 5 min at 4 °C and counterstained with 1μg/mL 4′,6-diamidino-2-phenylindole DAPI (DAPI, Sigma Aldrich, St. Louis, MO, USA) for another 5 min at room temperature. Sections were rinsed and mounted on glass slides before evaluation under a Nikon A1 confocal laser scanning microscope. The confocal serial sections were processed with ImageJ software and image rendering was performed using Adobe Photoshop Software. ## 2.6. ELISA Test Frozen samples taken at t3 were preserved at −80 °C for 4–8 weeks. Tissue samples were then processed for protein extraction with lysis buffer according to our previously protocol [32]. The ELISA test for type-I collagen and fibronectin was performed according to manufacturer’s instructions (Col 1 kit—Cloud-Clone Corporation, Katy, TX, USA; Human Fibronectin kit ab219046, Abcam, UK). ELISA tests were repeated at two different dilutions and mean values were considered for statistical analyses. ## 2.7. Statistical Analysis Statistical analysis was performed using STATA® software version 14 (StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX, USA: StataCorp LP.). Numerical data were expressed as mean, standard deviation, and range. Qualitative data were expressed as frequency and percentage. Chi-square test (Fisher’s exact test) and Student’s t-test was used to examine the relation between qualitative variables and continuous variables. A p-value < 0.05 was considered significant. ## 3.1. Baseline Patient Characteristics Seven of the enrolled patients were men and three were women. Age ranged between 58 and 95 years (mean 84.2). None of our patients were active smokers (See Table 1). Cardiovascular risk factors and heart disease were found to be common comorbidities in our cohort (including type-II diabetes, vasculopathy and arterial hypertension). As for the types of excised lesions, four were basal-cell carcinomas (BCCs), four squamous-cell carcinomas (SCCs) and two atypical fibroxanthomas (AFXs) (Table 1). ## 3.2. Clinical Outcomes No significant differences between the two substitutes in terms of global VSS were observed. However, more evident wound bed contracture was evident in four cases in areas treated with Matriderm ($p \leq 0.05$; see Figure 2). One patient experienced a stroke after hospital dismission and required subsequent hospitalization in the ICU, but such an episode was not considered to be related to the use of ADMs. No other major adverse events were recorded during the entire study duration. With regard to minor adverse events, signs of surgical wound infection were detected in five subjects, with Integra being more easily colonized by bacteria compared to Matriderm ($$p \leq 0.05$$). All the infections occurred between t1 and t2. Of these, only three cases were considered to be critical colonization, and they completely resolved after partial silicone layer removal and topical antibiotics and/or antiseptics. Microbiological swabs were performed in all cases refractory to local treatment, and results were positive for Pseudomonas aeruginosa and Proteus mirabilis. Two patients required total silicone layer removal and systemic oral antibiotic therapy for complete resolution of the local infection. No significant infectious sequelae were detected during the observation period. In one case only, tissue samples obtained at t3 were not adequate for performing all the punch biopsies required for further laboratory analyses due to severe tissue damage secondary to infection. ## 3.3. Neodermis Architecture and Composition Histopathological evaluation of cutaneous specimens at t3 led to the detection of significant differences in terms of ADM persistence and quality of the granulation tissue (see Table 2). Matriderm was shown to be re-absorbed more quickly than Integra ($p \leq 0.005$, see Figure 3, panels A,B). Faster maturation of the granulation tissue was also observed in Matriderm-treated areas ($p \leq 0.05$), with some of these also being re-epithelized (Figure 3, panel C). No significant differences between the two dermal scaffolds were detected in terms of inflammatory infiltrate, mostly being composed of both neutrophils and lymphocytes, but occasionally also containing eosinophils. Despite not being statistically significant, granulomatous reactions with giant multinucleated cells were more frequently observed within Integra-treated wound beds (Figure 3 panel D). Vascularization of the wound bed was detected in all cases. The presence of vessel-like structures in the neodermis observed with classical HE stains was confirmed by CD31 immunostaining (Figure 4, panels A,B). Quantitative assessment of vascularized areas confirmed similar vascularization of the wound bed with the use of the two ADMs (Figure 4, panel C,D). From a quantitative point of view, no significant differences in terms of collagen and fibronectin content were detected between the two substitutes (see Table 2). ## 3.4. Cellular Colonization of the ADM Red blood cells and granulocytes were the most prominent infiltrating cells initially found to colonize both ADMs (Figure 5). IF stain for alpha-SMA demonstrated that myo-fibroblasts were already present at t1-t2 in both ADMs. However, a more consistent presence of alpha-SMA positive cells was detected at t3: not only were myofibroblasts present in the neodermis, but dermal vessels also demonstrated a strong positivity for α-SMA, probably due to its expression by capillary pericytes [33] (Figure 6). As for MSC-specific markers, CD90 was already expressed by some cells colonizing the ADM at t1. On the contrary, Stro1 expression was delayed compared to CD90, with Stro1 positive cells only being present from t2. However, MSCs co-expressing both CD90 and Stro1 were only evident in the neodermis at t3 (Figure 7). ## 4. Discussion Despite being widely used in reconstructive surgery, ADMs lacked standardized randomized controlled trials supporting their efficacy in the dermatological setting [34,35,36,37,38]. More data are currently available on their use in breast reconstruction after mastectomy, where ADMs are becoming routinely employed [39]. However, conflicting data are currently emerging from updated observations of ADM-specific side effects, therefore suggesting careful patient selection for ADM-based reconstructive surgery [40,41]. Recently, Lohmander et al. [ 42,43] published the results of a milestone study aimed at assessing the differences in breast reconstruction after mastectomy with and without the use of ADMs. The authors found no significant differences between immediate implant-based breast reconstruction and reconstruction with the use of ADMs in terms of reinterventions or surgical complications, health-related quality of life or patient-reported aesthetic outcome. To date, however, it is impossible to draw similar conclusions regarding the use of ADMs for post-oncological surgery skin wound healing. Various studies have already widely explored the mechanisms of action of dermal substitutes in animal models of wound healing [44,45]. ADMs have already proven not only to provide a collagenic scaffold that increases the dermal thickness, thereby limiting cicatricial depression in healing skin, but also promoting the secretion of endogenous type I and type III collagen in a more physiological manner compared to standard secondary-intention wound healing [44,46]. Only a few comparative studies on ADMs have been published so far on ADM efficacy and potential morphological differences in the neoformed tissue [47,48]. Some of those data specifically focus on Matriderm and Integra, which currently represent two widely used dermal regenerative templates. A study by Joerg Schneider and collaborators [45] aimed to assess differences between Matriderm- and Integra-induced skin regeneration in a rat model of wound healing. The authors found no major differences in engraftment rates, quality of neodermis, or vascularization. Those data were confirmed by Bottcher-Haberzeth et al. in 2012, who did not find statistically significant differences between the two ADMs in terms of neodermis thickness [49]. Most of the in-vivo clinical studies in humans available in the dermatological setting have been performed on surgical reconstruction due to burn wounds and in various surgical sites [50]. In 2020 Philips and coauthors [51] performed a retrospective study aimed at comparing the use of two-stage Integra and single-stage Matriderm at their burn referral center. Comparable grafting rates were observed in both groups. Infections were more common in the Integra group, in line with our data. No significant differences were detected in hematoma development, hypertrophic scarring, or need for secondary surgery. The authors concluded that Integra could be recommended for larger burns with limited donor sites, while Matriderm was preferable for smaller burns in cosmetically sensitive areas. A second comparative study was conducted in 2020 by Vana et al. [ 50], who prospectively analyzed clinical and histopathological outcomes in patients treated with two mm-thick Matriderm or Integra, followed by thin skin autografts (two-step procedure) for burn scars healed with sequelae (e.g., with limited mobility and/or bad aesthetic results). Negative pressure therapy was also applied after surgery. Improvement in mobility and skin quality were demonstrated along with graft contraction, in all patients. No intra- or post-operative adverse events were recorded. The authors found that Integra had lower retraction rates and better skin quality compared to Matriderm. In line with our observations, the authors confirmed the tendency of Integra to persist for longer in the newly formed dermis. In contrast with such publications [50], we did not observe significant differences in terms of VSS total scores. However, VSS remains an operator-dependent parameter, thus potentially leading to evaluation bias. Moreover, we confirmed a slight tendency to wound contraction for the Matriderm-treated portion of the wound bed, in line with the Brazilian study mentioned above [50]. Current research suggests that skin pliability and reduced wound contracture are more easily achieved through the use of cross-linked scaffolds [52]. Recently, a prospective randomized controlled clinical trial, including patients with burn contracture treated using autologous skin grafts and ADMs, was performed by Corrêa and collaborators [53]. Interestingly, the control group (patients treated with skin graft only, without previous ADM positioning) displayed lower rates of wound contraction. No significant differences were detected between Integra® and Matriderm®. As for the dermato-oncological setting, an Italian group already assessed the use of the two ADMs after craniofacial surgery, both with single- and two-step surgery [54]. The authors detected better performances by Matriderm in terms of skin thickness when used in two-step surgery, while Integra was shown to achieve better results for engraftment and clinical outcomes when applied directly to the bone. We therefore decided not to enroll patients where the periosteum could not be preserved due to the established superiority of Integra in such cases. The main differences between our study and the one by Torresini and Gareffa reside in the retrospective nature of their observations; the possible selection biases due to interindividual variability in wound healing; the intrinsic heterogeneity of the surgical procedures (both one and two-step interventions were considered); the lack of histological evaluation. On the other hand, the limited number of patients enrolled in our pilot study represents a possible limitation in extending our findings to larger-scale casuistries. Compared to the other available studies, our work has one major strong point: Matriderm and Integra Bilayer are both positioned in a single surgical wound bed, with each single patient enrolled acting both as a case and as a control, thus eliminating any eventual confounding selection bias. Notwithstanding, the main limit of our study resides in the relatively small number of patients enrolled. Some minor concerns regarding our data could also arise due to the use of Matriderm in a two-step procedure. In fact, Matriderm is generally used in one-stage surgery with concomitant skin grafting, with a 3-week interval between ADM positioning and grafting possibly being associated with wound bed retraction. In line with previously published data [51], we observed higher rates of infection in patients treated with Integra. A possible explanation could reside in the occlusion provided by the silicon membrane present in the Integra device, which could favor bacterial proliferation. Several studies postulated specific skin cancer types to be associated with surgical-site infections [55,56], the most recent ones indicating SCC as possibly carrying the higher risk. Our data do not allow us to draw similar conclusions with significant statical strength: of the five cases observed, three were detected in SCCs and two in BCCs. However, no cases of infection were detected among cutaneous neoplasms other than NMSCs, such as AFX. Quantitative analysis though ELISA testing confirmed the lack of significant differences between the two devices in terms of ECM production in the neo-dermis. Histopathologically, the newly formed dermis was well vascularized in both Matriderm and Integra, without relevant differences. Moreover, Matriderm was found to more often be reabsorbed in the first weeks after ADM positioning compared to Integra. Probably collagen crosslinking in such ADM could give a partial reason for the observed variability in reabsorption rates. Furthermore, immunofluorescence analyses confirmed the ongoing rearrangement of cells, including apha-SMA + pericytes, involved in neovascularization processes occurring in wound healing. Also, the recruitment of mesenchymal stromal cells in the regenerating areas was evident, as shown by the immunolabeling against CD90 and STRO-1, demonstrating shared features between the two ADMs. In conclusion, no significative differences have been found between Matriderm and Integra in our prospective comparative study, both in terms of clinical efficacy and histopathological findings. However, more data are needed to extend our results to a larger casuistry, thereby possibly guiding daily clinical practice. ## References 1. Gantwerker E.A., Hom D.B.. Skin: Histology and Physiology of Wound Healing. Facial Plast. Surg. Clin. N. Am. (2011) 19 441-453. DOI: 10.1016/j.fsc.2011.06.009 2. Reinke J.M., Sorg H.. Wound Repair and Regeneration. Eur. Surg. Res. (2012) 49 35-43. DOI: 10.1159/000339613 3. Khan A.A., Khan I.M., Nguyen P.P., Lo E., Chahadeh H., Cerniglia M., Noriega J.A.. Skin Graft Techniques. Clin. Podiatr. Med. Surg. (2020) 37 821-835. DOI: 10.1016/j.cpm.2020.07.007 4. Harn H.I.-C., Ogawa R., Hsu C.-K., Hughes M.W., Tang M.-J., Chuong C.-M.. The Tension Biology of Wound Healing. Exp. Dermatol. (2019) 28 464-471. DOI: 10.1111/exd.13460 5. Ko G.R., Lee J.S.. Engineering of Immune Microenvironment for Enhanced Tissue Remodeling. Tissue Eng. Regen. Med. (2022) 19 221-236. DOI: 10.1007/s13770-021-00419-z 6. Martin P., Nunan R.. Cellular and Molecular Mechanisms of Repair in Acute and Chronic Wound Healing. Br. J. Dermatol. (2015) 173 370-378. DOI: 10.1111/bjd.13954 7. Guillamat-Prats R.. The Role of MSC in Wound Healing, Scarring and Regeneration. Cells (2021) 10. DOI: 10.3390/cells10071729 8. Tomic-Canic M., Burgess J.L., O’Neill K.E., Strbo N., Pastar I.. Skin Microbiota and Its Interplay with Wound Healing. Am. J. Clin. Dermatol. (2020) 21 36-43. DOI: 10.1007/s40257-020-00536-w 9. Greenhalgh D.G.. Wound Healing and Diabetes Mellitus. Clin. Plast. Surg. (2003) 30 37-45. DOI: 10.1016/S0094-1298(02)00066-4 10. Lindholm C., Searle R.. Wound Management for the 21st Century: Combining Effectiveness and Efficiency. Int. Wound. J. (2016) 13 5-15. DOI: 10.1111/iwj.12623 11. Paganelli A., Benassi L., Pastar I., Pellegrini M., Azzoni P., Vaschieri C., Pisciotta A., Carnevale G., Pellacani G., Magnoni C.. In Vitro Engineering of a Skin Substitute Based on Adipose-Derived Stem Cells. Cells Tissues Organs (2019) 207 46-57. DOI: 10.1159/000501071 12. Ireton J.E., Unger J.G., Rohrich R.J.. The Role of Wound Healing and Its Everyday Application in Plastic Surgery: A Practical Perspective and Systematic Review. Plast. Reconstr. Surg. Glob. Open (2013) 1 e10-e19. DOI: 10.1097/GOX.0b013e31828ff9f4 13. Azmat C.E., Council M.. Wound Closure Techniques. StatPearls (2021) 14. Clark R.A.F., Ghosh K., Tonnesen M.G.. Tissue Engineering for Cutaneous Wounds. J. Investig. Dermatol. (2007) 127 1018-1029. DOI: 10.1038/sj.jid.5700715 15. Raepsaet C., Alves P., Cullen B., Gefen A., Lázaro-Martínez J.L., Lev-Tov H., Najafi B., Santamaria N., Sharpe A., Swanson T.. Clinical Research on the Use of Bordered Foam Dressings in the Treatment of Complex Wounds: A Systematic Review of Reported Outcomes and Applied Measurement Instruments. J. Tissue. Viability (2022) 31 514-522. DOI: 10.1016/j.jtv.2022.05.005 16. Liang Y., Li Z., Huang Y., Yu R., Guo B.. Dual-Dynamic-Bond Cross-Linked Antibacterial Adhesive Hydrogel Sealants with On-Demand Removability for Post-Wound-Closure and Infected Wound Healing. ACS Nano (2021) 15 7078-7093. DOI: 10.1021/acsnano.1c00204 17. Esteban-Vives R., Young M., Over P., Schmelzer E., Corcos A., Ziembicki J., Gerlach J.. In Vitro Keratinocyte Expansion for Cell Transplantation Therapy Is Associated with Differentiation and Loss of Basal Layer Derived Progenitor Population. Differentiation (2015) 89 137-145. DOI: 10.1016/j.diff.2015.05.002 18. Neill B., Rickstrew J., Tolkachjov S.. Reconstructing the Glabella and Nasal Root. J. Drugs. Dermatol. (2022) 21 983-988. DOI: 10.36849/JDD.6132 19. Shahwan K.T., Mori W.S., Bakker C., Mattox A.R., Alam M., Maher I.A.. Reconstruction of Perioral Defects After Mohs Micrographic Surgery or Excision: A Systematic Review of the Literature. Dermatol. Surg. (2021) 47 162-166. DOI: 10.1097/DSS.0000000000002810 20. Hamdan S., Pastar I., Drakulich S., Dikici E., Tomic-Canic M., Deo S., Daunert S.. Nanotechnology-Driven Therapeutic Interventions in Wound Healing: Potential Uses and Applications. ACS Cent. Sci. (2017) 3 163-175. DOI: 10.1021/acscentsci.6b00371 21. Brett E., Chung N., Leavitt W.T., Momeni A., Longaker M.T., Wan D.C.. A Review of Cell-Based Strategies for Soft Tissue Reconstruction. Tissue. Eng. Part. B Rev. (2017) 23 336-346. DOI: 10.1089/ten.teb.2016.0455 22. Le J.M., Bosworth J.W., Honeywell B., Ananthasekar S., Collawn S.S.. Adipose Grafting for Volume and Scar Release. Ann. Plast. Surg. (2021) 86 S487-S490. DOI: 10.1097/SAP.0000000000002873 23. Papa G., Pangos M., Renzi N., Ramella V., Panizzo N., Arnez Z.M., Marij A.Z.. Five Years of Experience Using a Dermal Substitute: Indications, Histologic Studies, and First Results Using a New Single-Layer Tool. Dermatol. Surg. (2011) 37 1631-1637. DOI: 10.1111/j.1524-4725.2011.02156.x 24. Latha A., Vaidya K.G., Baiju P.. Advances in Wound Healing and Wound Care Technologies–A review. Int. J. Pharm. Sci. Rev. Res. (2020) 61 183-191 25. Paganelli A., Giordano E., Fiorentini C., Ferrari B., Reggiani C., Garbarino F., Magnoni C.. Surgical Management and Oncological Follow-up of Cutaneous Squamous Cell Carcinomas Arising in Epidermolysis Bullosa Patients. Int. J. Dermatol. (2022) 61 1171-1174. DOI: 10.1111/ijd.16157 26. Magnoni C., De Santis G., Fraccalvieri M., Bellini P., Portincasa A., Giacomelli L., Papa G.. Integra in Scalp Reconstruction After Tumor Excision: Recommendations from a Multidisciplinary Advisory Board. J. Craniofac. Surg. (2019) 30 2416-2420. DOI: 10.1097/SCS.0000000000005717 27. Abbas Khan M.A., Chipp E., Hardwicke J., Srinivasan K., Shaw S., Rayatt S.. The Use of Dermal Regeneration Template (Integra. J. Plast. Reconstr. Aesthet. Surg. (2010) 63 2168-2171. DOI: 10.1016/j.bjps.2010.03.017 28. Min J.H., Yun I.S., Lew D.H., Roh T.S., Lee W.J.. The Use of Matriderm and Autologous Skin Graft in the Treatment of Full Thickness Skin Defects. Arch. Plast. Surg. (2014) 41 330-336. DOI: 10.5999/aps.2014.41.4.330 29. Pauchot J., Elkhyat A., Rolin G., Mac S., Grumblat A., Fotso A., Humbert P., Tropet Y.. Dermal Equivalents in Oncology: Benefit of One-Stage Procedure. Dermatol. Surg. (2013) 39 43-50. DOI: 10.1111/dsu.12044 30. You H.-J., Choi Y.-S., Kim D.-W.. Use of a Multilayered Acellular Dermal Substitute with Simultaneous Full-Thickness Skin Graft for the One-Stage Coverage of Nasal Skin Defects. J. Cosmet. Dermatol. (2020) 19 3014-3019. DOI: 10.1111/jocd.13342 31. Bourne D.A., James I., Wang S., Bliley J., Grahovac T., Mitchell R.T., Brown S.A., Ambrosio F., Ho J., Lannau B.. Treatment of Burn Contractures with Allogeneic Human Dermal Fibroblasts Improves Vancouver Scar Scale: A Phase I/II Trial. J. Plast. Reconstr. Aesthet. Surg. (2021) 74 3443-3476. DOI: 10.1016/j.bjps.2021.08.018 32. Rossi E., Della Casa L., Piana S., Iannone A.. Conjugated Linoleic Acid Isomers Modulate Protein Expression Profile in Rat Hepatocytes. Genes Nutr. (2012) 7 511-527. DOI: 10.1007/s12263-012-0291-9 33. Alarcon-Martinez L., Yilmaz-Ozcan S., Yemisci M., Schallek J., Kılıç K., Can A., Di Polo A., Dalkara T.. Capillary Pericytes Express α-Smooth Muscle Actin, Which Requires Prevention of Filamentous-Actin Depolymerization for Detection. Elife (2018) 7 e34861. DOI: 10.7554/eLife.34861 34. Álvaro-Afonso F.J., García-Álvarez Y., Lázaro-Martínez J.L., Kakagia D., Papanas N.. Advances in Dermoepidermal Skin Substitutes for Diabetic Foot Ulcers. Curr. Vasc. Pharmacol. (2020) 18 182-192. DOI: 10.2174/1570161117666190408170144 35. Paganelli A., Reggiani C., Fiorentini C., Lando M., Cesinaro A.M., Magnoni C.. Surgical Management of Squamous Cell Carcinoma Arising in Patients Affected by Epidermolysis Bullosa: A Comparative Study. Int. Wound J. (2020) 17 519-521. DOI: 10.1111/iwj.13306 36. Campagnari M., Jafelicci A.S., Carneiro H.A., Brechtbühl E.R., Bertolli E., Duprat Neto J.P.. Dermal Substitutes Use in Reconstructive Surgery for Skin Tumors: A Single-Center Experience. Int. J. Surg. Oncol. (2017) 2017 9805980. DOI: 10.1155/2017/9805980 37. Marcasciano M., Mazzocchi M., Kaciulyte J., Spissu N., Casella D., Ribuffo D., Dessy L.A.. Skin Cancers and Dermal Substitutes: Is It Safe? Review of the Literature and Presentation of a 2-Stage Surgical Protocol for the Treatment of Non-Melanoma Skin Cancers of the Head in Fragile Patients. Int. Wound J. (2018) 15 756-768. DOI: 10.1111/iwj.12924 38. Tufaro A.P., Buck D.W., Fischer A.C.. The Use of Artificial Dermis in the Reconstruction of Oncologic Surgical Defects. Plast. Reconstr. Surg. (2007) 120 638-646. DOI: 10.1097/01.prs.0000270298.68331.8a 39. Onesti M.G., Di Taranto G., Ribuffo D., Scuderi N.. ADM-Assisted Prepectoral Breast Reconstruction and Skin Reduction Mastectomy: Expanding the Indications for Subcutaneous Reconstruction. J. Plast. Reconstr. Aesthet. Surg. (2020) 73 673-680. DOI: 10.1016/j.bjps.2019.11.004 40. Beier L., Faridi A., Neumann C., Paepke S., Mau C., Keller M., Strittmatter H.J., Gerber-Schäfer C., Bauer L., Karsten M.M.. Human Acellular Dermal Matrix (Epiflex. Breast Care (2021) 16 461-467. DOI: 10.1159/000512201 41. Viezel-Mathieu A., Alnaif N., Aljerian A., Safran T., Brabant G., Boileau J.-F., Dionisopoulos T.. Acellular Dermal Matrix-Sparing Direct-to-Implant Prepectoral Breast Reconstruction: A Comparative Study Including Cost Analysis. Ann. Plast. Surg. (2020) 84 139-143. DOI: 10.1097/SAP.0000000000001997 42. Lohmander F., Lagergren J., Johansson H., Roy P.G., Brandberg Y., Frisell J.. Effect of Immediate Implant-Based Breast Reconstruction After Mastectomy with and Without Acellular Dermal Matrix Among Women with Breast Cancer: A Randomized Clinical Trial. JAMA Netw. Open (2021) 4 e2127806. DOI: 10.1001/jamanetworkopen.2021.27806 43. Lohmander F., Lagergren J., Johansson H., Roy P.G., Frisell J., Brandberg Y.. Quality of Life and Patient Satisfaction after Implant-Based Breast Reconstruction with or without Acellular Dermal Matrix: Randomized Clinical Trial. BJS Open (2020) 4 811-820. DOI: 10.1002/bjs5.50324 44. da Conceição Carvalho-Júnior J., Zanata F., Aloise A.C., Ferreira L.M.. Acellular Dermal Matrix in Skin Wound Healing in Rabbits—Histological and Histomorphometric Analyses. Clinics (2021) 76 e2066. DOI: 10.6061/clinics/2021/e2066 45. Schneider J., Biedermann T., Widmer D., Montano I., Meuli M., Reichmann E., Schiestl C.. Matriderm versus Integra: A Comparative Experimental Study. Burns (2009) 35 51-57. DOI: 10.1016/j.burns.2008.07.018 46. Goodarzi P., Falahzadeh K., Nematizadeh M., Farazandeh P., Payab M., Larijani B., Tayanloo Beik A., Arjmand B.. Tissue Engineered Skin Substitutes. Adv. Exp. Med. Biol. (2018) 1107 143-188. DOI: 10.1007/5584_2018_226 47. Lynch M.P., Chung M.T., Rinker B.D.. Dermal Autografts as a Substitute for Acellular Dermal Matrices (ADM) in Tissue Expander Breast Reconstruction: A Prospective Comparative Study. J. Plast. Reconstr. Aesthet. Surg. (2013) 66 1534-1542. DOI: 10.1016/j.bjps.2013.07.002 48. Huang X., Ding Y., Pan W., Lu L., Jin R., Liang X., Chang M., Wang Y., Luo X.. A Comparative Study on Two Types of Porcine Acellular Dermal Matrix Sponges Prepared by Thermal Crosslinking and Thermal-Glutaraldehyde Crosslinking Matrix Microparticles. Front. Bioeng. Biotechnol. (2022) 10 938798. DOI: 10.3389/fbioe.2022.938798 49. Böttcher-Haberzeth S., Biedermann T., Schiestl C., Hartmann-Fritsch F., Schneider J., Reichmann E., Meuli M.. Matriderm. Pediatr. Surg. Int. (2012) 28 171-177. DOI: 10.1007/s00383-011-2990-5 50. Vana L.P.M., Battlehner C.N., Ferreira M.A., Caldini E.G., Gemperli R., Alonso N.. Comparative Long-Term Study between Two Dermal Regeneration Templates for the Reconstruction of Burn Scar Contractures in Humans: Clinical and Histological Results. Burns (2020) 46 596-608. DOI: 10.1016/j.burns.2019.09.005 51. Phillips G.S.A., Nizamoglu M., Wakure A., Barnes D., El-Muttardi N., Dziewulski P.. The Use of Dermal Regeneration Templates for Primary Burns Surgery in A UK Regional Burns Centre. Ann. Burns Fire Disasters (2020) 33 245-252. PMID: 33304216 52. Maitz J., Wang Y., Fathi A., Ximena Escobar F., Parungao R., Zuijlen P., Maitz P., Li Z.. The Effects of Cross-linking a Collagen-elastin Dermal Template on Scaffold Bio-stability and Degradation. J. Tissue Eng. Regen. Med. (2020) 14 1189-1200. DOI: 10.1002/term.3082 53. Corrêa F.B., Castro J.C.D., Almeida I.R., Farina-Junior J.A., Coltro P.S.. Evaluation of Contraction of the Split-Thickness Skin Graft Using Three Dermal Matrices in the Treatment of Burn Contractures: A Randomised Clinical Trial. Wound. Repair Regen. (2022) 30 222-231. DOI: 10.1111/wrr.13002 54. Torresini G., Garreffa E.. Reconstruction with Artificial Dermis of Oncological Surgical Defects in the Craniofacial District—Matriderm. Eur. J. Plast. Surg. (2018) 41 521-528. DOI: 10.1007/s00238-018-1419-y 55. Heal C.F., Buettner P.G., Drobetz H.. Risk Factors for Surgical Site Infection after Dermatological Surgery. Int. J. Dermatol. (2012) 51 796-803. DOI: 10.1111/j.1365-4632.2011.05189.x 56. Nakamura Y., Sasaki K., Ishizuki S., Inoue S., Okune M., Kubota N., Okiyama N., Furuta J., Fujisawa Y.. Invasive and in Situ Lesions of Squamous Cell Carcinoma Are Independent Factors for Postoperative Surgical-Site Infection after Outpatient Skin Tumors Surgery: A Retrospective Study of 512 Patients. J. Dermatol. (2021) 48 497-501. DOI: 10.1111/1346-8138.15782
--- title: Serum Metabolome Adaptations Following 12 Weeks of High-Intensity Interval Training or Moderate-Intensity Continuous Training in Obese Older Adults authors: - Layale Youssef - Mélanie Bourgin - Sylvère Durand - Fanny Aprahamian - Deborah Lefevre - Maria Chiara Maiuri - Vincent Marcangeli - Maude Dulac - Guy Hajj-Boutros - Fanny Buckinx - Eva Peyrusqué - Pierrette Gaudreau - José A. Morais - Gilles Gouspillou - Guido Kroemer - Mylène Aubertin-Leheudre - Philippe Noirez journal: Metabolites year: 2023 pmcid: PMC9967246 doi: 10.3390/metabo13020198 license: CC BY 4.0 --- # Serum Metabolome Adaptations Following 12 Weeks of High-Intensity Interval Training or Moderate-Intensity Continuous Training in Obese Older Adults ## Abstract Physical activity can be effective in preventing some of the adverse effects of aging on health. High-intensity interval training (HIIT) and moderate-intensity continuous training (MICT) are beneficial interventions for the quality of life of obese older individuals. The understanding of all possible metabolic mechanisms underlying these beneficial changes has not yet been established. The aim of this study was to analyze changes in the serum metabolome after 12 weeks of HIIT and MICT in obese older adults. Thirty-eight participants performed either HIIT ($$n = 26$$) or MICT ($$n = 12$$) three times per week for 12 weeks. Serum metabolites as well as clinical and biological parameters were assessed before and after the 12-week intervention. Among the 364 metabolites and ratio of metabolites identified, 51 metabolites changed significantly following the 12-week intervention. Out of them, 21 significantly changed following HIIT intervention and 18 significantly changed following MICT. Associations with clinical and biological adaptations revealed that changes in acyl-alkyl-phosphatidylcholine (PCae) (22:1) correlated positively with changes in handgrip strength in the HIIT group ($r = 0.52$, $p \leq 0.01$). A negative correlation was also observed between 2-oxoglutaric acid and HOMA-IR (r = −0.44, $p \leq 0.01$) when considering both groups together (HIIT and MICT). This metabolite also correlated positively with quantitative insulin-sensitivity check index (QUICKI) in both groups together ($r = 0.46$, $p \leq 0.01$) and the HIIT group ($r = 0.51$, $p \leq 0.01$). Additionally, in the MICT group, fumaric acid was positively correlated with triglyceride levels ($r = 0.73$, $p \leq 0.01$) and acetylcarnitine correlated positively with low-density lipoprotein (LDL) cholesterol ($r = 0.81$, $p \leq 0.01$). These four metabolites might represent potential metabolites of interest concerning muscle strength, glycemic parameters, as well as lipid profile parameters, and hence, for a potential healthy aging. Future studies are needed to confirm the association between these metabolites and a healthy aging. ## 1. Introduction Exercise, well known for its major health benefits, is considered an effective non-pharmacological strategy for the elderly [1]. Obesity, whose prevalence increases with aging [2], has also become a major public health issue. Thus, when low muscle mass and function are combined with excess adiposity, being obese and old becomes a double health burden [3]. In terms of clinical and biological adaptations, high-intensity interval training (HIIT) and moderate-intensity continuous training (MICT) are effective interventions to improve the health status of obese older individuals [4,5]. More specifically, it was previously demonstrated that MICT decreased relative gynoid fat mass and increased lower limb muscle strength, whereas HIIT increased functional capacities and lean mass in obese older adults [4]. Since the impacts of HIIT and MICT on serum metabolites in obese older adults remain largely unexplored, it becomes highly interesting to evaluate the possible mechanisms underlying the changes following these interventions. With advances in mass spectrometry, the number of metabolomics studies has increased. The metabolomic approach has recently become an area of interest in the field of exercise science [6,7]. Physical activity interventions can induce metabolic adaptations through changes in serum metabolomic profile [8]. However, there is still little information regarding changes after different types of exercise interventions and in different types of populations. With respect to mode of exercise, it has been shown that changes in the plasma metabolome can be both mode-dependent and mode-independent, meaning that the molecular response must be interpreted in the context of the mode by which the biological changes occur [9]. In participants with an average age of 62 years, enrichment of pathways involving connective tissue metabolism and lipid signaling was found in plasma samples after treadmill exercise (3 times per week for 50 min), whereas branched-chain amino acids (BCAA), tryptophan, tyrosine, and urea cycle were altered after stretching (twice per week for 50 min) [10]. One study examined the effect of HIIT and MICT on metabolism and counterregulatory stress hormones in well-trained male cyclists and triathletes. The results revealed that carbohydrate oxidation was higher, and fat oxidation was lower following HIIT [11]. Furthermore, metabolomic analysis revealed that tricarboxylic acid (TCA) intermediates and monounsaturated fatty acids increased after HIIT and BCAA decreased after both HIIT and MICT [11]. Therefore, different adaptations were observed following different training modalities. In addition, metabolic adaptations are not the same for all types of populations (age and health status). In healthy young soccer players, analysis of urine sample metabolites was performed before and after a HIIT intervention. Following a 2-day running HIIT session, metabolomic changes included downregulation of steroid hormone metabolites and upregulation during recovery, suggesting increased muscle growth after HIIT [12]. In middle-aged obese men and women, after an exercise intervention (walking/jogging on a treadmill at multiple intensities 5 times per week for 24 weeks), several changing metabolites, such as TCA cycle intermediate (isocitric acid), BCAA, gluconeogenic amino acids, as well as xanthurenic acid, were associated with changes in cardiometabolic risk traits [13]. In a study of participants (aged 30–60 years) with higher risk factors for metabolic syndrome and greater adiposity, serine and glycine were found in lower concentrations. However, when activity measures (measured by estimated calculations) were increasing, their concentrations were higher [14]. Additionally, in patients with pre-diabetes and type 2 diabetes, HIIT was found to be a more time-efficient strategy than MICT as serum ficolin-3 levels (associated with diabetes) decreased after 3 weeks of HIIT training [15]. In addition, in sedentary overweight and obese women (with an average age of 27 years), changes in glucose tolerance were predicted using personalized metabolomics after a 6-week HIIT intervention (cycling for 30 min twice a week) [16]. To our knowledge, no study has explored adaptations in the serum metabolome after HIIT and MICT in obese older men and women. Therefore, the objective of this study was to analyze changes in serum metabolites after 12 weeks of HIIT or MICT in obese older adults. ## 2.1. Study Design This study, which combined two randomized controlled trials, is a sub-analysis of our previous studies [4,5]. The ethics committee of the “Université du Québec à Montréal (UQAM)” approved all procedures (#2014_e_1018_475). After being informed about the study’s purpose, aim, procedures and associated risks, the participants provided their informed written consent. ## 2.2. Participants Participants were recruited from the community via social communication (flyers and meetings in community centers) in the Greater Montreal area. The list of inclusion and exclusion criteria, which was previously described [4,5], is detailed in the Supplementary Material. To be considered as having completed the intervention, participants had to complete at least $80\%$ of the training sessions (minimum: $\frac{29}{36}$ sessions). Among the participants included in the previous analysis [4], only those with pre- and post-intervention serum metabolites analyses (HIIT: $$n = 26$$ vs. MICT: $$n = 12$$) were included in this study. ## 2.3. Exercise Interventions All the participants performed 3 supervised training sessions per week (HIIT or MICT) during 12 consecutive weeks. ## 2.3.1. High-Intensity Interval Training (HIIT) Under the supervision of trained kinesiologists (i.e., certified exercise instructors), participants performed the HIIT training on an elliptical device (TechnoGym Synchro Exc 700, Technogym, NJ, USA) to reduce impacts. The HIIT session lasted 30 min and was divided as follows: [1] five-minute warm-up at a low intensity (50–$60\%$ maximal heart rate (MHR) and/or 8–12 on Borg’ scale); [2] twenty minutes of HIIT consisting of multiple 30-s high-intensity sprints (80–$85\%$ MHR or >17 on Borg’ scale) alternated with 90 s at moderate intensity ($65\%$ MHR or 13–16 on Borg’ scale); and [3] five-minute cool down period (50–$60\%$ MHR or 8–12 on Borg’ scale). To determine the intensity of each cycle, MHR percentage and/or perceived exertion (Borg scale; relying exclusively on perceived exertion for participants using anti-arrhythmic and inotropic agents) were used. The MHR was determined using the following equation: [((220 − age) − Heart Rate rest) × % Heart Rate target] + Heart Rate rest. Continuous adjustment of treadmill speed and resistance during the HIIT intervention ensured that MHR was always above $80\%$ during high intensity intervals. ## 2.3.2. Moderate-Intensity Continuous Training (MICT) Under the supervision of trained kinesiologists (i.e., certified exercise instructors), participants underwent MICT by walking on a treadmill (Precor C936i, Precor, WA, USA). The session lasted 1 h and the MICT was performed at a moderate intensity (60–$70\%$ MHR or 13–14 on the Borg scale). Continuous adjustment of treadmill speed and resistance ensured that MHR was maintained between 60–$70\%$ MHR or 13–14 on the Borg scale. ## 2.4. Clinical Parameters The detailed assessment of the clinical parameters was previously described in Youssef et al., 2022 [4]. To evaluate physical performance and muscle function, the validated tests used were previously described in Buckinx et al., 2018 [17]. The five tests used to assess the physical performance were the 6-min walk test [18,19], the walking speed [20,21], the unipodal balance [22], the timed up and go [23], the chair stand test [24] and the step test [25]. Three tests were used to assess muscle function: grip strength, lower limb muscle power, and lower muscle strength. These measures of muscle function were expressed in absolute units (kg or W or N, respectively) and normalized to body weight and lean limb mass. To evaluate anthropometric characteristics, body weight (kg) and height (m) were determined, which allowed calculation of the BMI (body mass (kg)/height (m2)). To assess the body composition, fat masses (total, android, gynoid, arm and leg; %) and total lean masses (total, arm and leg; kg) were quantified in fasted state. Muscle area, subcutaneous, and intramuscular fat content were used to assess thigh composition. ## 2.5. Biological Parameters Fifteen milliliters of blood were collected from each participant to assess fasting serum levels of biochemical and hormonal markers following a 12-h overnight fast. The detailed assessment of the biological parameters was previously described in Marcangeli et al., 2022 [5]. Briefly, the lipid profile was assessed through total, HDL- and LDL-cholesterol, as well as triglyceride levels. Adipose tissue metabolites and adipokines (free fatty acids, adiponectin and leptin levels, adiponectin/leptin ratio), members of the insulin-like growth factor (IGF) family (IGF1; IGFBP3 and IGFB3/IGF1 molar ratio) and glucose-insulin homeostasis (glucose and insulin levels but also HOMA and QUICKI indices) were assessed. ## 2.6. Metabolomic Profiling Blood samples were collected by a physician in the fasting state before and after the 12-week intervention. Serum samples were used to blindly analyze serum metabolites at the Gustave Roussy Cancer Campus facility (Villejuif, France) using mass spectrometers coupled to multiple different liquid or gas phase chromatography methods. Bile acids metabolomics were obtained using a UHPLC/MS—RRLC 1260 system (Agilent Technologies, Waldbronn, Germany) coupled to a Triple Quadrupole 6410 (Agilent Technologies). Short chain fatty acids, oxylipin, and lipids metabolomics were assessed by a UHPLC/QUAD+—RRLC 1260 system (Agilent Technologies, Waldbronn, Germany) coupled to a 6500+ QTRAP (Sciex, Darmstadt, Germany). Polyamines metabolomics were quantified using a UHPLC/QQQ—RRLC 1260 system (Agilent Technologies, Waldbronn, Germany) coupled to a Triple Quadrupole 6410 (Agilent Technologies). Concerning the level of identification, for the targeted metabolomics it is identified (level 1, validated by standards injections on the Orbitrap, and multiple reaction monitoring MRM developments for the LCQQQ and GCQQQ), and for the metabolomic profiling it is putative. Intra batch correction was performed based on quality control pool and processed on R software. The details of each method were previously described [26,27]. ## 2.7. Statistical Analyses Quantitative results are expressed as mean ± SD. The delta changes (%) were calculated as (post − pre)/pre × 100. Statistical significance tests of the measured metabolites were performed using multivariate and univariate analyses. For the multivariate analysis, principal component analysis (PCA) was conducted to reduce the dimensionality of the data (R-packages FactoMineR and factoextra) and volcano plots were performed for data visualization (R-package tidyverse and ggrepe1). The Levene’s test was used to assess the homogeneity of variances. For the univariate analysis, a linear mixed-models approach (R-package nlme) with a two-factor repeated measures ANOVA was then used to test the intervention effect (Time effect), and the interaction effect (Time × Training effect) on serum metabolites adaptations. Post-hoc analyses were then done using simultaneous tests for general linear hypotheses (R-package emmeans) with a Bonferroni correction. Results were considered statistically significant when p-value < 0.05. A large number of comparisons were made, so the Benjamini Hochberg (BH) false discovery rate (FDR) was performed to correct for multiple comparisons (R-package FSA). The threshold for the FDR was set at 0.05. To assess the association between delta changes of the metabolites and the delta changes of the clinical and biological parameters, Pearson’s correlation coefficient analysis was then performed in the HIIT group. However, as the number of subjects was low in the MICT group, Spearman’s correlation coefficient analysis was used. Only correlations with Pearson’s and Spearman’s correlation coefficient above 0.35 and with p-value < 0.01 were considered. All statistical analyses were performed using the software R (3.6.2) (foundation for statistical computing, Vienna, Austria). Heatmaps of the metabolites were generated with the R-packages ggplot2 and tidyr, and the correlation graphs were drawn using the R-package ggpubr. ## 3.1. Clinical and Biological Characteristics The characteristics of the participants at baseline are present in Table 1. The clinical and biological characteristics of these participants were previously described [4]. However, for this sub-analysis, only those with pre- and post-intervention serum metabolites analyses were included (Tables S1–S3 present in the Supplementary Material). The significance of the clinical and biological parameters was almost similar to the previous study. However, due to the lower number of participants, the total lean mass and the gynoid fat mass, which appeared significant in the previous study, were not significant in the current sub-analysis (Table S2). ## 3.2. Overall Metabolomic Profile Adaptations Following the 12-Week Intervention Following the 12-week HIIT or MICT intervention, 364 serum metabolites and t ratio of metabolites were identified and quantified. The PCA analysis, which was performed following the overall 12-week intervention (Figure 1), revealed no significant changes. ## 3.3. Alterations of Energetic Metabolisms Following the 12-Week HIIT and MICT Intervention To identify the best metabolites of interest, we generated a two-way repeated measures ANOVA. This analysis revealed significant changes ($p \leq 0.05$) for fifty-one metabolites involved in different metabolic pathways (Table 2 and Table 3): six belong to the tricarboxylic acid (TCA) cycle, five to the carbohydrate metabolism, eleven to the amino acid metabolism, and twenty-nine to the fat metabolism. Considering significant changes for each type of intervention alone, 21 metabolites significantly varied after the 12-week HIIT intervention (Figure 2A; abundance was normalized between 0 and 1), and 18 metabolites significantly varied after the 12-week MICT intervention (Figure 2B; abundance was normalized between 0 and 1). For the TCA cycle (Table 2), following the 12-week intervention, five metabolites significantly changed following the HIIT intervention, and one metabolite significantly changed following the MICT intervention. More specifically, following HIIT, two metabolites increased (2-oxoglutaric acid and fumaric acid) while the three others (3-methylhistidine, aspartic acid and the ratio aspartate/malate) significantly decreased. Following MICT, 2-oxoglutaric acid significantly increased. For the carbohydrate metabolism (Table 2), following the 12-week intervention, two metabolites significantly changed following the HIIT intervention, and two metabolites changed following the MICT intervention. More specifically, acetic acid and glyceric acid significantly decreased after HIIT, whereas xylitol and xylose significantly increased after MICT. For the amino acid metabolism (Table 2), following the 12-week intervention, four metabolites significantly changed following the HIIT intervention, and five metabolites significantly changed following the MICT intervention. More specifically, three metabolites decreased (2 hydroxybutyric acid, 2-oxovaleric acid and inosine) while ketoisovaleric acid increased following HIIT. On the other side, the five metabolites that significantly changed after MICT (2-aminoadipic acid, hypotaurine, ornithine, uric acid and xanthine) significantly increased. For the fat metabolism (Table 3), 10 metabolites significantly changed following the HIIT intervention, and 10 metabolites significantly changed following the MICT intervention. More specifically, following HIIT, two metabolites increased (Carnitine C18:0 and margaric acid), whereas eight metabolites significantly decreased (acetylcarnitine, Ceramide (C18:$\frac{1}{24}$:0), Diglyceride (DG) (18:$\frac{1}{18}$:3), DG (20:$\frac{4}{18}$:2), isobutyric acid, linoleic acid, acyl-alkyl-phosphatidylcholine (PCae) (16:0) and PCae (22:1)). Following MICT, an increase was observed for eight metabolites (DG(18:$\frac{1}{18}$:3), margaric acid, panthothenic acid, PCae(15:0), triglyceride (TG) (16:$\frac{1}{18}$:$\frac{1}{18}$:0), TG (16:$\frac{1}{18}$:$\frac{3}{18}$:2), TG (16:$\frac{2}{18}$:$\frac{2}{18}$:2) and undecanoic acid) and a decrease was observed for two metabolites (ether-phosphatidylethanolamine (PEee) (19:1) and TG (12:$\frac{0}{14}$:$\frac{0}{16}$:0)). To strengthen our feature selection criteria, we undertook a method-based false discovery rate (FDR)-adjusted (Tables S4a,b and S5a,b in the Supplementary Material). Following this analysis, two metabolites revealed significant differences due to the 12-week intervention (Time effect), one metabolite due to HIIT effect, and eight metabolites due to MICT effect. More specifically, when the general intervention was considered, significant changes (FDR < 0.05) were observed for only two metabolites, which are 2-oxoglutaric acid and glyceric acid. For the HIIT group alone, only glyceric acid significantly changed (FDR < 0.01). For the MICT group alone, a significant change was revealed for eight metabolites (FDR < 0.05): 2-oxoglutaric acid, xylitol, ketoisovaleric acid, uric acid, DG (18:$\frac{3}{18}$:3), pantothenic acid, TG (16:$\frac{1}{18}$:$\frac{3}{18}$:2) and TG (16:$\frac{2}{18}$:$\frac{2}{18}$:2). To identify the differences in metabolite abundance before and after the intervention, volcano plots were drawn (Figure 3). This shows that four metabolites were upregulated before MICT compared to five metabolites that were upregulated after MICT (Figure 3A). Before HIIT, one metabolite was upregulated compared to two metabolites that were upregulated after HIIT (Figure 3B). Additionally, twelve metabolites were upregulated after MICT compared to five metabolites that were upregulated after HIIT (Figure 3C). ## 3.4. Associations between Changes in Serum Metabolites and Changes in Clinical and Biological Parameters Correlations between serum metabolites delta changes (%) and clinico-biological parameters delta changes (%) were obtained. Significant correlations ($p \leq 0.01$) resulting from the overall intervention, the HIIT intervention and the MICT intervention were obtained (Table 4, Table 5 and Table 6). The impacts of HIIT and MICT on the metabolites that were among the interesting correlations are presented in Figure 4. The delta change in the metabolite PCae (22:1) significantly correlated with the delta change in handgrip strength, a parameter of muscular strength, and an important clinical parameter in the elderly. Following the FDR analysis, the delta change in PCae (22:1) did not appear significant. However, this metabolite was considered since interesting correlations were observed between PCae (22:1) and muscle strength parameters in the HIIT group (Table 4). More specifically, delta change of PCae (22:1) significantly correlated with delta changes of the handgrip strength (rp = 0.52, $p \leq 0.01$; Figure 5A), the handgrip strength relative to body weight (rp = 0.54, $p \leq 0.01$), and the handgrip strength relative to arms lean mass (rp = 0.54, $p \leq 0.01$; Figure 5B). The delta changes of the three metabolites 2-oxoglutaric acid, fumaric acid and acetylcarnitine were significantly correlated with biological parameters delta changes (Table 6). The change in 2-oxoglutaric acid appeared significant following the FDR analysis (FDR < 0.05). When the entire group of participants was considered, this metabolite delta changed significantly, correlated with delta changes of both HOMA-IR (rp = −0.44, $p \leq 0.01$; Figure 5C) and QUICKI (rp = 0.46, $p \leq 0.01$; Figure 5D), which are insulin sensitivity parameters [28]. Interestingly, 2-oxoglutaric acid delta change also correlated with QUICKI delta change (rp = 0.51, $p \leq 0.01$; Figure 5E) in the HIIT group alone. The changes of both fumaric acid and acetylcarnitine did not appear significant following the FDR analysis, however interesting correlations with lipid profile parameters delta changes were observed in the MICT group. More specifically, fumaric acid delta change correlated well with triglycerides level delta change (rs = 0.73, $p \leq 0.01$; Figure 5F), and acetylcarnitine delta change correlated well with LDL cholesterol (rs = 0.81, $p \leq 0.01$; Figure 5G). ## 4.1. Metabolome Adaptations The metabolomic analysis was conducted on a population of obese older men and women following a 12-week HIIT or MICT intervention. Our aim was to identify possible serum metabolites of interest underlying the adaptations in clinical and biological parameters in obese older adults. Given that these HIIT and MICT modalities in this type of population have previously been shown to induce changes in clinical and biological parameters [4,5,17], metabolites changes could then be considered as molecular signatures of clinico-biological adaptations. We observed that several metabolites increased or decreased significantly as a result of the overall 12-week intervention. In addition, some of these metabolites were altered after HIIT or MICT, meaning that the metabolome undergoes specific changes after each type of exercise. Moreover, the metabolite DG (18:$\frac{1}{18}$:3) significantly decreased after HIIT but significantly increased after MICT. This counter change could allow us to speculate that the different metabolisms may not only be altered after exercise but may also be altered differently depending on different types of exercise. A study performed on athletes after a single bout of exercise showed that the serum metabolome of endurance and bodybuilding athletes was affected differently [28]. This result suggests that the serum metabolome may be altered differently depending on the different types of exercise usually performed by the individual, as well as on different specific factors (athletic or non-athletic individuals). The fact that the serum metabolome underwent different adaptations is in line with our results, as the metabolome of our participants was also affected differently after different types of exercise. Interesting correlations between significantly modified metabolites and several clinico-biological parameters were obtained. To the best of our knowledge, this is the first study in which an analysis of the serum metabolites was performed following a 12-week HIIT or MICT intervention in obese older adults. Our intervention consisted of two different exercise modalities with different intensities, and some metabolites changes were elicited by HIIT or by MICT. Interestingly, it was previously found that changes in the metabolome occurred following non-pharmacological interventions. Among these interventions, exercise [29] or nutrition [30] were able to induce changes in the metabolome. In a study evaluating the metabolome changes in adult men following a 3-h marathon [31], glyceric acid increased. In our study, a chronic intervention in older adults, this metabolite decreased following HIIT. This discrepancy could be caused by several factors, like the different age of the participants and the different frequency of intervention, with the marathon being an acute intervention and the HIIT being a chronic intervention over 12 weeks. Therefore, the metabolome is able to encounter different changes depending on different interventions and different factors (age, sex, and type of intervention modality). ## 4.2. PCae (22:1) as a Possible Metabolite of Interest for Muscle Strength One of the major adverse health effects of aging is the loss of muscle strength. This loss becomes much faster than the loss of muscle mass in older individuals, resulting in a decline in muscle quality [32]. In addition, age and body fat have been shown to be inversely associated with muscle strength and quality [33]. Therefore, since obese individuals will experience an even faster decrease in muscle strength and quality, preventing fat gain and preserving lean mass becomes an important issue. As a standard clinical measure, it has been proposed that handgrip strength may be a marker of frailty assessment [34]. Following the 12-week HIIT intervention, a significant positive correlation was observed between the metabolite PCae (22:1) delta change and the handgrip strength delta change of our obese elderly participants. This metabolite is included in the category of fat metabolism metabolites. In a previous study aiming to identify metabolites of interest associated to handgrip strength decline in old participants (aged 55 years and above), it was proposed that acetylcarnitine might reflect specific perturbations occurring in the mitochondrial fat metabolism during aging [35]. In our sub-analysis, we did not find any significant correlation between acetylcarnitine and handgrip strength, which could be due first to the obesity status of our participants and second to the exercise intervention. However, our result could be partially in line with that of Siang Ng and colleagues [35], since the metabolite which correlated well with handgrip strength is also part of fat metabolism. More specifically, PCae is an abundant phospholipid in the cell membranes and plays a role in the regulation of the lipid metabolism [36]. Therefore, fat metabolism could play a major role in the decline of handgrip strength, and metabolites of this metabolism could be of interest for muscular strength. ## 4.3. Possible Metabolites of Interest for Glycemic and Lipid Profile Parameters in Obese Older Adults In a study evaluating metabolic and hormonal responses to HIIT and MICT in well-trained male cyclists and triathletes, distinct changes in specific metabolites were found, suggesting that differences in metabolic demand affect the serum metabolome [11]. Interestingly, similarly to our results, they found that the majority of the TCA cycle intermediates increased after HIIT. Following our 12-week intervention, the metabolite 2-oxoglutaric acid, a TCA cycle intermediate, correlated well with HOMA-IR and QUICKI, two parameters of the glycemic control. Additionally, this metabolite correlated well with QUICKI in the HIIT group. Peake and colleagues did not report any significant change for this metabolite [11], and this discrepancy could be due to different factors such as different physical activity levels between the two cohorts of participants, obesity status and age. In the MICT group, interesting correlations were observed between delta changes of fumaric acid and triglycerides level as well as between delta changes of acetylcarnitine and LDL cholesterol, which are important lipid profile parameters. However, further research is needed to confirm these associations with a larger number of obese older participants undergoing MICT. ## 4.4. Limitations and Future Perspectives This study presents some limitations. The number of participants between the two groups was not high (26 for HIIT and 12 for MICT) since not all participants from the a priori study had extra blood samples. Concerning the participants characteristics, our participants were moderately obese, and this might not reflect the more severely obese population. As only volunteer subjects were included in the 12-week intervention, a selection bias is also possible. Most importantly, different devices were used for the exercise intervention (elliptical trainer for HIIT and treadmill for MICT). As our participants might be osteoporotic, and HIIT might cause high surface impacts, an elliptical trainer was used to prevent joint injuries on the lower limbs. Hence, adding one group performing MICT on the elliptical trainer might be of interest in the future studies. For the metabolomic analyses, intermediate metabolomic analyses (for example in the middle of the 12-week intervention) are missing. Additionally, whole-body metabolisms cannot be represented by the serum metabolome, therefore additional analyses from different samples like plasma, urine, or saliva would be of interest. A replication of our results in an independent validation cohort undergoing the similar 12-week intervention is required. Our results were presented both with and without FDR analysis, to avoid missing any metabolite or ratio of metabolites that could present a potential interest. We tried to identify new putative metabolites of interest specific to HIIT and MICT in obese older adults that could be used to treat obesity and the decline in age-related functions. However, future studies are needed to confirm our results. ## 5. Conclusions To the best of our knowledge, our sub-analysis study is the first to compare the adaptations in the serum metabolites following a 12-week HIIT or MICT in obese older adults. Our results showed that 51 metabolites from different biochemical pathways were affected by the 12-week intervention. Several of these metabolites correlated well with clinical and biological adaptations. Interesting associations were revealed between PCae (22:1) and handgrip strength, as well as between 2-oxoglutaric acid and glycemic parameters (HOMA-IR and QUICKI). Additionally, fumaric acid and acetylcarnitine correlated well with lipid profile parameters (triglycerides level and LDL cholesterol respectively). These metabolites were identified as possible metabolites of interest for muscle strength, glycemic parameters, and lipid profile parameters, and hence, for a potential healthy ageing. ## References 1. Izquierdo M., Merchant R.A., Morley J.E., Anker S.D., Aprahamian I., Arai H., Aubertin-Leheudre M., Bernabei R., Cadore E.L., Cesari M.. International Exercise Recommendations in Older Adults (ICFSR): Expert Consensus Guidelines. J. Nutr. Health Aging (2021) 25 824-853. DOI: 10.1007/s12603-021-1665-8 2. Batsis J.A., Villareal D.T.. Sarcopenic obesity in older adults: Aetiology, epidemiology and treatment strategies. Nat. Rev. Endocrinol. (2018) 14 513-537. DOI: 10.1038/s41574-018-0062-9 3. Barazzoni R., Bischoff S., Boirie Y., Busetto L., Cederholm T., Dicker D., Toplak H., Van Gossum A., Yumuk V., Vettor R.. Sarcopenic Obesity: Time to Meet the Challenge. Obes. Facts (2018) 11 294-305. DOI: 10.1159/000490361 4. Youssef L., Granet J., Marcangeli V., Dulac M., Hajj-Boutros G., Reynaud O., Buckinx F., Gaudreau P., Morais J.A., Mauriege P.. Clinical and Biological Adaptations in Obese Older Adults Following 12-Weeks of High-Intensity Interval Training or Moderate-Intensity Continuous Training. Healthcare (2022) 10. DOI: 10.3390/healthcare10071346 5. Marcangeli V., Youssef L., Dulac M., Carvalho L.P., Hajj-Boutros G., Reynaud O., Guegan B., Buckinx F., Gaudreau P., Morais J.A.. Impact of high-intensity interval training with or without l-citrulline on physical performance, skeletal muscle, and adipose tissue in obese older adults. J. Cachexia Sarcopenia Muscle (2022) 13 1526-1540. DOI: 10.1002/jcsm.12955 6. Duft R.G., Castro A., Chacon-Mikahil M.P.T., Cavaglieri C.R.. Metabolomics and Exercise: Possibilities and perspectives. Mot. Rev. De Educ. Física (2017) 23 e101634. DOI: 10.1590/s1980-6574201700020010 7. Bongiovanni T., Pintus R., Dessì A., Noto A., Sardo S., Finco G., Corsello G., Fanos V.. Sportomics: Metabolomics applied to sports. The new revolution?. Eur. Rev. Med. Pharm. Sci. (2019) 23 11011-11019. DOI: 10.26355/eurrev_201912_19807 8. Castro A., Duft R.G., de Oliveira-Nunes S.G., de Andrade A.L.L., Cavaglieri C.R., Chacon-Mikahil M.P.T.. Association Between Changes in Serum and Skeletal Muscle Metabolomics Profile With Maximum Power Output Gains in Response to Different Aerobic Training Programs: The Times Study. Front. Physiol. (2021) 12 756618. DOI: 10.3389/fphys.2021.756618 9. Morville T., Sahl R.E., Moritz T., Helge J.W., Clemmensen C.. Plasma Metabolome Profiling of Resistance Exercise and Endurance Exercise in Humans. Cell Rep. (2020) 33 108554. DOI: 10.1016/j.celrep.2020.108554 10. Serra M.C., Accardi C.J., Ma C., Park Y., Tran V., Jones D.P., Hafer-Macko C.E., Ryan A.S.. Metabolomics of Aerobic Exercise in Chronic Stroke Survivors: A Pilot Study. J. Stroke Cereb. Dis. (2019) 28 104453. DOI: 10.1016/j.jstrokecerebrovasdis.2019.104453 11. Peake J.M., Tan S.J., Markworth J.F., Broadbent J.A., Skinner T.L., Cameron-Smith D.. Metabolic and hormonal responses to isoenergetic high-intensity interval exercise and continuous moderate-intensity exercise. Am. J. Physiol. Endocrinol. Metab. (2014) 307 E539-E552. DOI: 10.1152/ajpendo.00276.2014 12. Zhao J., Wang Y., Zhao D., Zhang L., Chen P., Xu X.. Integration of metabolomics and proteomics to reveal the metabolic characteristics of high-intensity interval training. Analyst (2020) 145 6500-6510. DOI: 10.1039/D0AN01287D 13. Brennan A.M., Benson M., Morningstar J., Herzig M., Robbins J., Gerszten R.E., Ross R.. Plasma Metabolite Profiles in Response to Chronic Exercise. Med. Sci. Sport. Exerc. (2018) 50 1480-1486. DOI: 10.1249/MSS.0000000000001594 14. Palmnas M.S.A., Kopciuk K.A., Shaykhutdinov R.A., Robson P.J., Mignault D., Rabasa-Lhoret R., Vogel H.J., Csizmadi I.. Serum Metabolomics of Activity Energy Expenditure and its Relation to Metabolic Syndrome and Obesity. Sci. Rep. (2018) 8 3308. DOI: 10.1038/s41598-018-21585-6 15. Liu X., Wang G.. The Effect of High-Intensity Interval Training on Physical Parameters, Metabolomic Indexes and Serum Ficolin-3 Levels in Patients with Prediabetes and Type 2 Diabetes. Exp. Clin. Endocrinol. Diabetes (2021) 129 740-749. DOI: 10.1055/a-1027-6511 16. Kuehnbaum N.L., Gillen J.B., Gibala M.J., Britz-McKibbin P.. Personalized Metabolomics for Predicting Glucose Tolerance Changes in Sedentary Women After High-Intensity Interval Training. Sci. Rep. (2014) 4 6166. DOI: 10.1038/srep06166 17. Buckinx F., Gouspillou G., Carvalho L.P., Marcangeli V., El Hajj Boutros G., Dulac M., Noirez P., Morais J.A., Gaudreau P., Aubertin-Leheudre M.. Effect of High-Intensity Interval Training Combined with L-Citrulline Supplementation on Functional Capacities and Muscle Function in Dynapenic-Obese Older Adults. J. Clin. Med. (2018) 7. DOI: 10.3390/jcm7120561 18. Rikli R.E., Jones C.J.. The Reliability and Validity of a 6-Minute Walk Test as a Measure of Physical Endurance in Older Adults. J. Aging Phys. Act. (1998) 6 363-375. DOI: 10.1123/japa.6.4.363 19. ATS statement: Guidelines for the six-minute walk test. Am. J. Respir. Crit. Care Med. (2002) 166 111-117. DOI: 10.1164/ajrccm.166.1.at1102 20. Guralnik J.M., Ferrucci L., Pieper C.F., Leveille S.G., Markides K.S., Ostir G.V., Studenski S., Berkman L.F., Wallace R.B.. Lower extremity function and subsequent disability: Consistency across studies, predictive models, and value of gait speed alone compared with the short physical performance battery. J. Gerontol. A Biol. Sci. Med. Sci. (2000) 55 M221-M231. DOI: 10.1093/gerona/55.4.M221 21. Ostir G.V., Volpato S., Fried L.P., Chaves P., Guralnik J.M.. Reliability and sensitivity to change assessed for a summary measure of lower body function: Results from the Women’s Health and Aging Study. J. Clin. Epidemiol. (2002) 55 916-921. DOI: 10.1016/S0895-4356(02)00436-5 22. Choi Y.M., Dobson F., Martin J., Bennell K.L., Hinman R.S.. Interrater and intrarater reliability of common clinical standing balance tests for people with hip osteoarthritis. Phys. Ther. (2014) 94 696-704. DOI: 10.2522/ptj.20130266 23. Christopher A., Kraft E., Olenick H., Kiesling R., Doty A.. The reliability and validity of the Timed Up and Go as a clinical tool in individuals with and without disabilities across a lifespan: A systematic review. Disabil. Rehabil. (2021) 43 1799-1813. DOI: 10.1080/09638288.2019.1682066 24. Csuka M., McCarty D.J.. Simple method for measurement of lower extremity muscle strength. Am. J. Med. (1985) 78 77-81. DOI: 10.1016/0002-9343(85)90465-6 25. Chung M.M., Chan R.W., Fung Y.K., Fong S.S., Lam S.S., Lai C.W., Ng S.S.. Reliability and validity of Alternate Step Test times in subjects with chronic stroke. J. Rehabil. Med. (2014) 46 969-974. DOI: 10.2340/16501977-1877 26. Danlos F.X., Grajeda-Iglesias C., Durand S., Sauvat A., Roumier M., Cantin D., Colomba E., Rohmer J., Pommeret F., Baciarello G.. Metabolomic analyses of COVID-19 patients unravel stage-dependent and prognostic biomarkers. Cell Death Dis. (2021) 12 258. DOI: 10.1038/s41419-021-03540-y 27. Tran B.-T.T., Gelin A., Durand S., Texier M., Daste A., Toullec C., Benihoud K., Breuskin I., Gorphe P., Garic F.. Plasma galectins and metabolites in advanced head and neck carcinomas: Evidence of distinct immune characteristics linked to hypopharyngeal tumors. OncoImmunology (2023) 12 2150472. DOI: 10.1080/2162402X.2022.2150472 28. Borai A., Livingstone C., Kaddam I., Ferns G.. Selection of the appropriate method for the assessment of insulin resistance. BMC Med. Res. Methodol. (2011) 11 158. DOI: 10.1186/1471-2288-11-158 29. Kistner S., Rist M.J., Kruger R., Doring M., Schlechtweg S., Bub A.. High-Intensity Interval Training Decreases Resting Urinary Hypoxanthine Concentration in Young Active Men-A Metabolomic Approach. Metabolites (2019) 9. DOI: 10.3390/metabo9070137 30. Rebholz C.M., Lichtenstein A.H., Zheng Z., Appel L.J., Coresh J.. Serum untargeted metabolomic profile of the Dietary Approaches to Stop Hypertension (DASH) dietary pattern. Am. J. Clin. Nutr. (2018) 108 243-255. DOI: 10.1093/ajcn/nqy099 31. Shi R., Zhang J., Fang B., Tian X., Feng Y., Cheng Z., Fu Z., Zhang J., Wu J.. Runners’ metabolomic changes following marathon. Nutr. Metab. (2020) 17 19. DOI: 10.1186/s12986-020-00436-0 32. Goodpaster B.H., Park S.W., Harris T.B., Kritchevsky S.B., Nevitt M., Schwartz A.V., Simonsick E.M., Tylavsky F.A., Visser M., Newman A.B.. The loss of skeletal muscle strength, mass, and quality in older adults: The health, aging and body composition study. J. Gerontol. A Biol. Sci. Med. Sci. (2006) 61 1059-1064. DOI: 10.1093/gerona/61.10.1059 33. Newman A.B., Haggerty C.L., Goodpaster B., Harris T., Kritchevsky S., Nevitt M., Miles T.P., Visser M.. Strength and muscle quality in a well-functioning cohort of older adults: The Health, Aging and Body Composition Study. J. Am. Geriatr. Soc. (2003) 51 323-330. DOI: 10.1046/j.1532-5415.2003.51105.x 34. Syddall H., Cooper C., Martin F., Briggs R., Aihie Sayer A.. Is grip strength a useful single marker of frailty?. Age Ageing (2003) 32 650-656. DOI: 10.1093/ageing/afg111 35. Ng T.K.S., Kovalik J.P., Ching J., Chan A.W., Matchar D.B.. Novel metabolomics markers are associated with pre-clinical decline in hand grip strength in community-dwelling older adults. Mech. Ageing Dev. (2021) 193 111405. DOI: 10.1016/j.mad.2020.111405 36. van der Veen J.N., Kennelly J.P., Wan S., Vance J.E., Vance D.E., Jacobs R.L.. The critical role of phosphatidylcholine and phosphatidylethanolamine metabolism in health and disease. Biochim. Biophys. Acta Biomembr. (2017) 1859 1558-1572. DOI: 10.1016/j.bbamem.2017.04.006
--- title: 'Molecular Characterizations of the Coagulase-Negative Staphylococci Species Causing Urinary Tract Infection in Tanzania: A Laboratory-Based Cross-Sectional Study' authors: - Shukrani Phillip - Martha F. Mushi - Arun Gonzales Decano - Jeremiah Seni - Blandina T. Mmbaga - Happiness Kumburu - Eveline T. Konje - Joseph R. Mwanga - Benson R. Kidenya - Betrand Msemwa - Stephen Gillespie - Antonio Maldonado-Barragan - Alison Sandeman - Wilber Sabiti - Mathew T. G. Holden - Stephen E. Mshana journal: Pathogens year: 2023 pmcid: PMC9967252 doi: 10.3390/pathogens12020180 license: CC BY 4.0 --- # Molecular Characterizations of the Coagulase-Negative Staphylococci Species Causing Urinary Tract Infection in Tanzania: A Laboratory-Based Cross-Sectional Study ## Abstract Background: There is a growing body of evidence on the potential involvement of coagulase-negative Staphylococci (CoNS) in causing urinary tract infections (UTIs). The aim of this study was to delineate virulence potential, antimicrobial resistance genes, and sequence types of CoNS isolated from patients with UTI symptoms and pyuria in Tanzania. Methods: CoNS from patients with UTI symptoms and more than 125 leucocytes/μL were retrieved, subcultured, and whole-genome sequenced. Results: Out of 65 CoNS isolates, 8 species of CoNS were identified; Staphylococcus haemolyticus, $$n = 27$$ ($41.5\%$), and Staphylococcus epidermidis, $$n = 24$$ ($36.9\%$), were predominant. The majority of S. haemolyticus were sequence type (ST) 30, with 8 new ST138-145 reported, while the majority of S. epidermidis were typed as ST490 with 7 new ST1184-1190 reported. Sixty isolates ($92.3\%$) had either one or multiple antimicrobial resistance genes. The most frequently detected resistance genes were 53 ($21\%$) dfrG, 32 ($12.9\%$) blaZ, and 26 ($10.5\%$) mecA genes conferring resistance to trimethoprim, penicillin, and methicillin, respectively. Out of 65 isolates, 59 ($90.8\%$) had virulence genes associated with UTI, with a predominance of the icaC 47 ($46.5\%$) and icaA 14 ($13.9\%$) genes. Conclusion: S. haemolyticus and S. epidermidis harboring icaC, dfrG, blaZ, and mecA genes were the predominant CoNS causing UTI in Tanzania. Laboratories should carefully interpret the significant bacteriuria due to CoNS in relation to UTI symptoms and pyuria before labeling them as contaminants. Follow-up studies to document the outcome of the treated patients is needed to add more evidence that CoNS are UTI pathogens. ## 1. Introduction Globally, urinary tract infection (UTI) affects about 150 million patients annually with a recurrence rate of $27\%$ among women within 6 months of the first episode [1]. In Tanzania, UTI is reported to affect about 2 in every 10 pregnant women [2,3] and 2–3 in every 10 children [4,5,6] annually. Given the decline in malarial prevalence and incidence, UTI is the third most common type of illness affecting children of five years old and above and second most common illness affecting patients visiting outpatient departments (OPDs) in Tanzania [6]. Gram-negative bacteria of the order Enterobacterales have commonly been implicated in causing UTI [7,8]. Furthermore, studies have found associations between UTI and coagulase-negative staphylococci (CoNS) [8,9,10]. The CoNS reported to be involved in causing UTIs include: Staphylococcus haemolyticus, Staphylococcus saprophyticus, Staphylococcus epidermidis, Staphylococcus hominis, Staphylococcus xylosus, Staphylococcus simulans, and Staphylococcus cohnii [8,11,12]. CoNS possess genes encoding virulence factors associated with causation of UTI such as: biofilm formation (icaABC, aap, bhp, atlE, fbe, and embp) [13], hemolysis of red cells (hlg and hla), and adhesion and cell wall-anchored proteins (fnbA, ebpS and sasA, sasF, sasH) [14,15,16,17]. CoNS species have been documented to have diverse sequence types implicated in different clinical conditions. For example, S. epidermidis ST1, ST2, ST5, and ST215 have been associated with both hospital-acquired and community-acquired UTI [18]. S. lugdunensis ST3, ST2, and ST1 have been associated with human clinical isolates (skin or soft tissue, osteoarticular, blood, and material device isolates) [19]. S. epidermidis ST2, ST54, ST28, ST59, ST490, and ST596 have been frequently isolated from blood [20]. In Tanzania, significant bacteriuria cases due to CoNS have been reported to range from $6.2\%$ to $16.7\%$ among women with diabetic and febrile children [21,22], without delineating the specific species involved and their antimicrobial resistance or virulence genes determinants. A recent study in Tanzania using VITEK MS identified S. haemolyticus to be the second most common Gram-positive uropathogen causing community UTIs [23]. In most clinical microbiology laboratories, CoNS are regarded as skin contaminants [24,25]. This practice might have contributed to the mismanagement of patients. In line with this gap, the current study used whole-genome sequencing (WGS), to determine species, sequence types, virulence potential, and antimicrobial resistance genes of CoNS isolated from patients who were clinically and microbiologically confirmed to have UTI caused by CoNS. ## 2.1. Bacteria Speciation and Distribution of CoNS Species from Urine Samples Of the 79 isolates recovered, 65 ($82.3\%$) were confirmed to be CoNS species by WGS, and were identified to 8 different CoNS species (Figure 1). The predominant CoNS species detected were S. haemolyticus $$n = 27$$ ($41.5\%$) and S. epidermidis $$n = 24$$ ($36.9\%$). ## 2.2. Multilocus Sequence Types of CoNS Species Causing UTI Out of 65 CoNS isolates, 24 ($36.9\%$) revealed 11 different sequence types (STs) in the first attempt. Out of the 27 S. haemolyticus isolates, 18 were assigned to 6 different STs predominated by ST30, and 9 isolates were assigned to 8 new ST138-145. Among the 24 S. epidermidis isolates, 9 were assigned to 3 different STs predominated by ST490, and 7 isolates were assigned to 7 new ST 1184-1190 while 8 isolates with mutation had unknown ST and have been submitted to the S. epidermidis pubMLST database (https://pubmlst.org/organisms/staphylococcus-epidermidis; accessed on 10 January 2023) (Table 1). Majority of the isolates with unknown ST had three-point mutation in arcC_8, Supplementary Material File S1. ## 2.3. Virulence Genes of CoNS Species Causing UTI A total of 5 different virulence genes associated with UTIs were identified in 59 ($$n = 65$$, $90.8\%$) isolates (Table 2). Genes responsible for the formation of polysaccharide intercellular adhesin produced by a gene cluster at the intercellular adhesion (ica) locus were predominantly detected, led by icaC, which was present in 47 isolates ($46.5\%$). The highest virulence genes combination was 5 (icaA, icaB, icaC, icaD, fbe), identified in 15 isolates. A total of 6 CoNS isolates had no known virulence genes associated with UTIs. ## 2.4. Antimicrobial Resistance Genes Identified among CoNS Species Out of 65 CoNS isolates, 60 ($92.3\%$) had one or multiple genes coding for antimicrobial resistance (AMR). Most of the resistance genes were identified among S. haemolyticus and S. epidermidis isolates (Table 3). The most frequently identified gene was dfrG ($$n = 53$$, $21.4\%$), which confers resistance to trimethoprim. Other frequently identified genes included blaZ ($$n = 32$$, $12.9\%$) and mecA ($$n = 26$$, $10.5\%$), which confer resistance to penicillin and methicillin, respectively. Totals of 24 ($88.9\%$), 21 ($77.8\%$), and 21 ($77.8\%$) S. haemolyticus had dfrG, blaZ, and mecA genes, respectively, while dfrG was identified in 22 ($91.7\%$) isolates of Staphylococcus epidermidis. ## 3. Discussion Coagulase-negative staphylococci (CoNS) are usually opportunistic pathogens, but have become important pathogens in clinical microbiology laboratories associated with various clinical conditions, such as UTI, skin and soft tissue infections, septicemia, and osteoarticular infections [8,19,26,27,28,29]. This study documents that the varieties of CoNS species causing UTI were predominantly S. haemolyticus and S. epidermidis. The CoNS species investigated were endowed with virulence genes, mainly icaC, which has been reported to facilitate the pathogenesis of UTI. Furthermore, these CoNS isolates harbored AMR genes with the predominance of dfrG, blaZ, and mecA genes. As reported recently by Vitus et al. [ 23], S. haemolyticus was the most predominant CoNS detected as a uropathogen in the current study. S. haemolyticus is among the normal skin microbiota commonly found in the perineum and inguinal area, making it easy for them easy to ascend and cause UTI [30,31]. Furthermore, S. haemolyticus is known to harbor over 82 insertion sequences [32], antibiotic-resistance genes, and some virulence factors, which confer adaptability to different environments and highlight unusual plasticity [30]. The findings also concur with other studies outside Tanzania that reported $49.4\%$ of Gram-positive bacteria causing UTI were Staphylococcus haemolyticus [33,34]. These findings indicate the possibility of S. haemolyticus being among the leading uropathogenic CoNS. The second most identified species was S. epidermidis ($39.6\%$), which was more common compared to $15.2\%$ and $9.14\%$ documented in Libya and China, respectively [11,34]. The differences across the three countries may indicate varying geographical conditions that favor or inhibit their growth. S. epidermidis is one of the most common skin opportunistic pathogens. Other CoNS species identified in the current study included S. saprophyticus, S. hominis, S. lugdunensis, S. simulans, S. warneri, and S. cohnii. These isolates have also been isolated from symptomatic UTI patients in previous studies [11,34,35]. S. saprophyticus has been widely documented as one of the commonest causes of UTI among CoNS species; to our surprise, this pathogen was not commonly detected in the current study despite the study population including young women, emphasizing the need to update our epidemiological knowledge, since S. saprophyticus has been reported to account for $5\%$ to $15\%$ of uncomplicated lower UTI cases in young women [36,37,38]. Multilocus sequence typing can discriminate between strains of the same species irrespective of their clinical conditions and sites [39]. S. haemolyticus was assigned to five different sequence types (ST30, ST1, ST38, ST49, and ST56) including three ST (ST30, ST1, ST38) previously detected from different clinical samples such as eyes, blood, pus, and sputum [40]. Furthermore, $37\%$ of S. haemolyticus were assigned to new ST138-145, indicating difference in epidemiological distribution of these isolates between developed and developing countries because most of known ST were from studies in developed countries. Additionally, in the current study, a total of $33.3\%$ of S. epidermidis were not assigned to known STs. Three types of STs (ST150, ST329, and ST490) assigned to S. epidermidis were detected, as in previous studies but from other human infections apart from UTI [20,41], 7 new ST1184-1190 have been reported for the first time in this study indicating population epidemiological differences of S. epeidermidis. ST3 S. lugdunensis detected in the current study was previously documented from skin or soft tissue, osteoarticular, and blood isolates [19]. The diversity of the STs in this study and other studies elsewhere indicate geographical variations that may dictate inheritance; hence, connoting multiple sources or niches of CoNS. Studies have reported virulence genes such as icaADBC, hla, hla_yidD, hld, and hlb to be implicated in UTI [14,15,17]. In a current study, $90.8\%$ of CoNS were found to harbor UTI virulence genes. This is relatively high compared to previous study reported in India, where $56.1\%$ of CoNS possessed either one or multiple virulence genes [42]. The difference could be explained by the selection of patients, whereby our patients had clinical infections (we only included patients with signs and symptoms and pyuria) contrasting the previous study that had different clinical samples, including exudates, urine, blood, endotracheal, catheter tips, and sputum. The production of poly-N-acetylglucosamine (PNAG) is crucial for biofilm formation in CoNS species and is coded in the icaADBC genes cluster. Biofilm formation protects these bacteria against the antibacterial drugs and the immune system defenses [43]. In the current study, $95.8\%$ of S. epidermidis were detected to have genes coding for biofilm formations (either of icaC, icaD, icaA, and icaB single or in combination). The proportion of biofilm encoding genes detected in the current study is much higher than $22.5\%$ reported by Solati et al. [ 44]. The fbe gene, encoding for fibrinogen binding protein, mediates initial attachment to cell walls during biofilm formation was detected in $12.9\%$ among CoNS, which was relatively lower compared to a $20\%$ and $40\%$ among strong and moderate biofilm forming CoNS, respectively, as reported previously [42]. The dfrG was the most frequently identified ($21.4\%$) AMR determinant gene in the collection of CoNS isolates in the current study. The dfrG gene encodes for a dihydrofolate reductase that confers resistance to trimethoprim. The resistance to trimethoprim is escalating over the time; this might be propagated by overuse of this antibiotic as it is cheap, highly accessible over the counter in our settings, and has been extensively used as a prophylaxis against Pneumocystis jirovecii pneumonia among HIV patients [45]. The blaZ gene, which is usually found in plasmid and confers resistance to penicillin, was identified in $49.2\%$ of isolates, which is low compared to $86.8\%$ reported among CoNS in India [46]. The blaZ gene encodes for a β-lactamase that is synthesized when staphylococci are exposed to β-lactam antibiotics, thus cleaving the β-lactam ring, rendering the penicillin inactive [47]. The current study documents the prevalence of the mecA gene (encoding for an altered penicillin-binding protein (PBP 2a), which confers resistance to methicillin [48]) to be $40\%$ among CoNS isolates causing UTI. This is low compared to $70.7\%$ documented by Shrestha et al. [ 49] in Nepal. The difference in the findings from these two studies might be due to the type of patients studied; Shrestha et al. used patients in intensive care units who had received several courses of antibiotics from primary care hospitals while the majority of patients in the current study are from community-acquired UTIs. The findings reported from this study emphasize the need to strengthen the diagnosis capacity in microbiology laboratory, especially in lower resource settings, to include tests that can correctly identify CoNS species. Additionally, in patients with clinical signs and symptoms, the study findings indicate the need to consider them for treatment and not neglect them as skin contaminant due to their virulence potential elaborated here. Regardless of all their usefulness, the study findings are limited by the small number of isolates involved, which could lead to the absence of some important findings such as S. saprophyticus. ## 4.1. Study Design and Area This was a laboratory based cross sectional study designed to characterize of CoNS isolates collected by HATUA project 2018 to 2020, from three sites in Tanzania i.e., Mwanza, Kilimanjaro and Mbeya. For this study, the laboratory work was conducted from February 2021 to August 2021 in three different laboratories at the Microbiology Research Laboratory of the Catholic University Health and Allied Sciences—Bugando, Mwanza, and Kilimanjaro Clinical Research Institute (KCRI) Biotechnology Laboratory, Moshi-Kilimanjaro. Genome sequencing of isolates was carried out at MicrobesNG, University of Birmingham, UK. ## 4.2. Study Population Isolates collected in this study were part of the HATUA study [50]. Archived CoNS isolates from patients with UTI-like symptoms, pyuria of above 125 leucocytes/μL, and significant bacteriuria obtained as described in the HATUA protocol [50]. In Tanzania, the HATUA project recruited patients with clinical diagnosis of UTI from 10 different health facilities in Mwanza, Kilimanjaro, and Mbeya. Clinically suspected UTI patients were those with signs/symptoms (e.g., fever, burning/irritation during urination, dysuria, pyuria) and microbiologically confirmed were those with significant bacterial growth of >105 CFU/mL in quantitative urine culture [50]. ## 4.3. Sample Size and Isolate Recovery The study used all CoNS isolates that could be recovered from the HATUA biorepository at CUHAS. A total of 101 CoNS isolates previously identified during HATUA were retrieved from the biorepository. On subsequent subculture onto $5\%$ sheep blood agar (SBA) and incubated for 18 to 24 h aerobically at 37 °C, a total of $$n = 79$$ isolates were fully recovered. Recovered isolates were ($$n = 79$$) transported in sterile microbiological transport swab containing Stuart medium (Guangzhou Improved Medical Instruments Co., Ltd., Guangzhou, China) to MicrobesNG, University of Birmingham, UK ($$n = 59$$) and KCRI Biotechnology Laboratory Moshi-Tanzania ($$n = 20$$), where subculture was performed, followed by DNA extraction and WGS. ## 4.4. DNA Extraction and Library Preparation Pure cultures of each strain were grown in plates or broth. Cells were harvested and suspended in a tube with cryopreservative (MicrobankTM, Pro-Lab Diagnostics UK, London, UK) or with DNA/RNA Shield (Zymo Research, Irvine, CA, USA) following MicrobesNG strain submission procedures. First, 5 to 40 μL of the suspension were lysed with 120 μL of TE buffer containing lysozyme (final concentration 0.1 mg/mL) and RNase A (ITW Reagents, Barcelona, Spain) at a final concentration 0.1 mg/mL, incubated for 25 min at 37 °C. Proteinase K (VWR Chemicals, Aurora, OH, USA) at a final concentration of 0.1 mg/mL and SDS (Sigma-Aldrich, St. Louis, MI, USA) at a final concentration of $0.5\%$ v/v were added and incubated for 5 min at 65 °C. Genomic DNA was purified using an equal volume of SPRI beads and resuspended in EB buffer (Qiagen, Hilden, Germany). DNA was quantified with the Quant-iT dsDNA HS kit (ThermoFisher Scientific, Waltham, MA, USA) assay in an Eppendorf AF2200 plate reader (Eppendorf UK Ltd., Stevenage, UK). Extracted DNA was eluted in 10 mM Tris-HCl pH 8.0 or nuclease free water and sent to MicrobesNG for sequencing. ## 4.5. Illumina Sequencing Genomic DNA libraries were prepared using the Nextera XT Library Prep Kit (Illumina, San Diego, CA, USA) following the manufacturer’s protocol with the following modifications: input DNA was increased 2-fold, and PCR elongation time increased to 45 s. DNA quantification and library preparation were carried out on a Hamilton Microlab STAR automated liquid handling system (Hamilton Bonaduz AG, Bonaduz, Switzerland). Pooled libraries were quantified using the Kapa Biosystems Library Quantification Kit for Illumina. Libraries were sequenced using Illumina sequencers (HiSeq/NovaSeq) using a 250 bp paired-end protocol. ## 4.6. Genome Quality Check and Assembly Raw reads quality was assessed using FastQC version 0.7.2 [51]. Sequencing adapters of raw fastq data were trimmed using Trimmomatic version 0.38.1 [52] with a sliding window quality cutoff of Phred score Q15 [53]. Cleaned reads were de novo assembled using SPAdes version 3.12.0 [54,55] and contigs were annotated using Prokka 1.11 [56]. All these tools were run using Galaxy\|Europe (https://usegalaxy.eu/; [accessed on 15 October 2021]). ## 4.7. Bacterial Speciation and Multilocus Sequence Typing Species identification and multilocus sequence typing of CoNS sequences were performed using Speciator and MLST tools, respectively; both are part of Pathogenwatch platform (https://pathogen.watch/ [accessed on 15 September 2021]). ## 4.8. Antimicrobial Resistance Genes Identification and Virulence Genes Annotation Staramr version 0.7.2 [57] was used to search for antimicrobial resistance (AMR)-conferring genes. Prokka version 1.14.6 [56], with the help of “Advanced cut” and “Select line that matches an expression” tools, was used to annotate virulence genes. All these tools were run using Galaxy\|Europe (https://usegalaxy.eu/; [accessed on 15 October 2021]). ## 4.9. Statistical Data Analysis All the data were imported into STATA software version 13.0 (College Station, TX, USA) for analysis where categorical variables (species, sequence types, virulence genes, and antimicrobial resistance genes) were presented as frequencies and percentages. ## 4.10. Ethical Consideration The HATUA project received research and ethical approval from various regulatory bodies within and outside Tanzania [50]. This sub-study further received approval from the Joint CUHAS/BMC Research Ethics and Review Committee (CRE/$\frac{490}{2021}$). ## 5. Conclusions S. haemolyticus and S. epidermidis harboring icaC virulence genes and dfrG, blaZ, and mecA AMR genes were the predominant CoNS causing UTI in Tanzania. New ST of S. haemolyticus and S. epidermidis were most frequently detected indicating population structure geographical variation of S. haemolyticus and S. epidermidis. Laboratories should carefully interpret the significant bacteriuria due to CoNS in relation to UTI symptoms and pyuria before labeling them as contaminants, as they may be potential UTI pathogens. Further follow-up studies to document the outcome of the treated patients is needed to add more evidence that CoNS are probable UTI pathogens. ## References 1. Öztürk R.A.. Murt, Epidemiology of urological infections: A global burden. World J. Urol. (2020) 38 2669-2679. DOI: 10.1007/s00345-019-03071-4 2. Seni J., Tito J., Makoye S., Mbena H., Alfred H., van der Meer F., Pitout J., Mshana S., DeVinney R.. Multicentre evaluation of significant bacteriuria among pregnant women in the cascade of referral healthcare system in North-western Tanzania: Bacterial pathogens, antimicrobial resistance profiles and predictors. J. Glob. Antimicrob. Resist. (2019) 17 173-179. DOI: 10.1016/j.jgar.2018.12.024 3. Chaula T., Seni J., Ng’Walida N., Kajura A., Mirambo M.M., DeVinney R., Mshana S.E.. Urinary Tract Infections among HIV-Positive Pregnant Women in Mwanza City, Tanzania, Are High and Predicted by Low CD4. Int. J. Microbiol. (2017) 2017 4042686. DOI: 10.1155/2017/4042686 4. Mushi M.F., Alex V.G., Seugendo M., Silago V., Mshana S.E.. C—reactive protein and urinary tract infection due to Gram-negative bacteria in a pediatric population at a tertiary hospital, Mwanza, Tanzania. Afr. Health Sci. (2019) 19 3217-3224. DOI: 10.4314/ahs.v19i4.45 5. Msaki B.P., Mshana S.E., Hokororo A., Mazigo H.D., Morona D.. Prevalence and predictors of urinary tract infection and severe malaria among febrile children attending Makongoro health centre in Mwanza city, North-Western Tanzania. Arch. Public Health (2012) 70 4. DOI: 10.1186/0778-7367-70-4 6. Festo E., Kidenya B.R., Hokororo A., Mshana S.E.. Predictors of urinary tract infection among febrile children attending at Bugando Medical Centre Northwestern, Tanzania. Arch. Clin. Microbiol. (2011) 2 5 7. Koneman E.W., Allen S.D., Janda W.M., Schreckenberger P.C., Winn W.C.. Diagnostic microbiology. The Nonfermentative Gram-Negative Bacilli (1997) 253-320 8. Flores-Mireles A.L., Walker J.N., Caparon M., Hultgren S.J.. Urinary tract infections: Epidemiology, mechanisms of infection and treatment options. Nat. Rev. Microbiol. (2015) 13 269-284. DOI: 10.1038/nrmicro3432 9. Hashmi A., Abdullah F.E., Abdullah N.E., Kazmi S.U.. Species Identification and Antibiotic Susceptibilities of Coagulase- Negative Staphylococci Isolated from Urinary Tract Infection Specimens. J. Coll. Physicians Surg. Pak. (2016) 26 581-584. PMID: 27504549 10. Stamm W.E., Norrby S.R.. Urinary Tract Infections: Disease Panorama and Challenges. J. Infect. Dis. (2001) 183 S1-S4. DOI: 10.1086/318850 11. Qiao L.-D., Chen S., Yang Y., Zhang K., Zheng B., Guo H.-F., Yang B., Niu Y.-J., Wang Y., Shi B.-K.. Characteristics of urinary tract infection pathogens and their in vitro susceptibility to antimicrobial agents in China: Data from a multicenter study. BMJ Open (2013) 3 e004152. DOI: 10.1136/bmjopen-2013-004152 12. John J.F., Gramling P.K., O’dell N.M.. Species identification of coagulase-negative staphylococci from urinary tract isolates. J. Clin. Microbiol. (1978) 8 435-437. DOI: 10.1128/jcm.8.4.435-437.1978 13. Cho S.-H., Naber K., Hacker J., Ziebuhr W.. Detection of the. Int. J. Antimicrob. Agents (2002) 19 570-575. DOI: 10.1016/S0924-8579(02)00101-2 14. Iorio N., Azevedo M.. Methicillin-resistant. Int. Microbiol. (2011) 14 13-17. DOI: 10.2436/20.1501.01.130 15. Nasaj M., Saeidi Z., Asghari B., Roshanaei G., Arabestani M.R.. Identification of hemolysin encoding genes and their association with antimicrobial resistance pattern among clinical isolates of coagulase-negative Staphylococci. BMC Res. Notes (2020) 13 1-6. DOI: 10.1186/s13104-020-4938-0 16. Mbanga J., Masuku S., Luphahla S.. Antibiotic Resistance Patterns and Virulence Factors of Coagulase Negative Staphylococcus Associated with Urinary Tract Infections in Bulawayo Province, Zimbabwe. Br. J. Med. Med. Res. (2016) 11 1-9. DOI: 10.9734/BJMMR/2016/19912 17. Chen J., Fang Q.. Whole-Genome Sequencing Analysis of Methicillin-Resistant. Genome Announc. (2016) 4 e00555-16. DOI: 10.1128/genomeA.00555-16 18. Widerström M.. Molecular Epidemiology of Coagulase-Negative Staphylococci in Hospitals and in the Community (2010) 19. Chassain B., Lemée L., Didi J., Thiberge J.-M., Brisse S., Pons J.-L., Pestel-Caron M.. Multilocus Sequence Typing Analysis of. J. Clin. Microbiol. (2012) 50 3003-3009. DOI: 10.1128/JCM.00988-12 20. Ehlers M.M., Strasheim W., Lowe M., Ueckermann V., Kock M.M.. Molecular Epidemiology of. Front. Microbiol. (2018) 9 417. DOI: 10.3389/fmicb.2018.00417 21. Eligius F.L., Sabrina J.M., Ewaldo V.K., Martin H.. Prevalence, antimicrobial resistance and associated risk factors for bacteriuria in diabetic women in Dar es Salaam, Tanzania. Afr. J. Microbiol. Res. (2011) 5 683-689 22. Francis F., Joel M.F., Maulidi F., Samuel Y.M.. Aetiology, antimicrobial susceptibility and predictors of urinary tract infection among febrile under-fives at Muhimbili National Hospital, Dar es Salaam-Tanzania. Afr. J. Microbiol. Res. (2013) 7 1029-1034 23. Silago V., Moremi N., Mtebe M., Komba E., Masoud S., Mgaya F.X., Mirambo M.M., Nyawale H.A., Mshana S.E., Matee M.I.. Multidrug-Resistant Uropathogens Causing Community Acquired Urinary Tract Infections among Patients Attending Health Facilities in Mwanza and Dar es Salaam, Tanzania. Antibiotics (2022) 11. DOI: 10.3390/antibiotics11121718 24. Hagler S., Dobkin D.. Urinary Tract Infection in the Male Caused by. Clin. Pediatr. (1990) 29 527-528. DOI: 10.1177/000992289002900909 25. Kline K.A., Lewis A.L.. Gram-Positive Uropathogens, Polymicrobial Urinary Tract Infection, and the Emerging Microbiota of the Urinary Tract. Microbiol. Spectr. (2016) 4 459-502. DOI: 10.1128/microbiolspec.UTI-0012-2012 26. Megged O.. Coagulase-negative Staphylococci: A rare cause of urinary tract infections in children with consequences on clinical practice. Eur. J. Pediatr. (2022) 181 1099-1104. DOI: 10.1007/s00431-021-04308-4 27. Kumari N., Rai A., Jaiswal C.P., Xess A., Shahi S.K.. Coagulase negative Staphylococci as causative agents of urinary tract infections-prevalence and resistance status in IGIMS, Patna. Indian J. Pathol. Microbiol. (2001) 44 415-419. PMID: 12035353 28. Crude N., Tveten Y., Kristiansen B.-E.. Urinary tract infections in Norway: Bacterial etiology and susceptibility. A retrospective study of clinical isolates. Clin. Microbiol. Infect. (2001) 7 543-547. DOI: 10.1111/j.1469-0691.2001.00306.x 29. Caierão J., Superti S., Dias C.A., D’Azevedo P.A.. Automated systems in the identification and determination of methicillin resistance among coagulase negative staphylococci. Memórias do Instituto Oswaldo Cruz (2006) 101 277-280. DOI: 10.1590/S0074-02762006000300009 30. Takeuchi F., Watanabe S., Baba T., Yuzawa H., Ito T., Morimoto Y., Kuroda M., Cui L., Takahashi M., Ankai A.. Whole-Genome Sequencing of. J. Bacteriol. (2005) 187 7292-7308. DOI: 10.1128/JB.187.21.7292-7308.2005 31. Tristan A., Lina G., Etienne J., Vandenesch F.. Biology and Pathogenicity of Staphylococci Other than. Gram-Posit. Pathog. (2006) ch47 572-586. DOI: 10.1128/9781555816513.ch47 32. Bouchami O., de Lencastre H., Miragaia M.. Impact of Insertion Sequences and Recombination on the Population Structure of. PLoS ONE (2016) 11. DOI: 10.1371/journal.pone.0156653 33. Al-Naqshbandi A.A., Chawsheen M.A., Abdulqader H.H.. Prevalence and antimicrobial susceptibility of bacterial pathogens isolated from urine specimens received in rizgary hospital—Erbil. J. Infect. Public Health (2019) 12 330-336. DOI: 10.1016/j.jiph.2018.11.005 34. Haider J., Hasan A., Bin-Tahir K.. Frequency of urinary tract bacterial infection and their susceptibility patterns among hemodialysis patients in Zliten hospital. J. Microbiol. Exp. (2016) 3 93-97. DOI: 10.15406/jmen.2016.03.00093 35. Orrett F.A., Shurland S.M.. Significance of coagulase-negative staphylococci in urinary tract infections in a developing country. Connect. Med. (1998) 62 199-203 36. Widerström M., Wiström J., Sjöstedt A., Monsen T.. Coagulase-negative staphylococci: Update on the molecular epidemiology and clinical presentation, with a focus on. Eur. J. Clin. Microbiol. Infect. Dis. (2012) 31 7-20. DOI: 10.1007/s10096-011-1270-6 37. Raz R., Colodner R., Kunin C.M.. Who are you—. Clin. Infect. Dis. (2005) 40 896-898. DOI: 10.1086/428353 38. Latham R.H., Running K., Stamm W.E.. Urinary tract infections in young adult women caused by Staphylococcus saprophyticus. JAMA (1983) 250 3063-3066. DOI: 10.1001/jama.1983.03340220031028 39. Maiden M.C.J., Bygraves J.A., Feil E., Morelli G., Russell J.E., Urwin R., Zhang Q., Zhou J., Zurth K., Caugant D.A.. Multilocus sequence typing: A portable approach to the identification of clones within populations of pathogenic microorganisms. Proc. Natl. Acad. Sci. USA (1998) 95 3140-3145. DOI: 10.1073/pnas.95.6.3140 40. Panda S., Jena S., Sharma S., Dhawan B., Nath G., Singh D.V.. Identification of Novel Sequence Types among. PLoS ONE (2016) 11. DOI: 10.1371/journal.pone.0166193 41. Miragaia M., De Lencastre H., Perdreau-Remington F., Chambers H.F., Higashi J., Sullam P.M., Lin J., Wong K.I., King K.A., Otto M.. Genetic Diversity of Arginine Catabolic Mobile Element in. PLoS ONE (2009) 4. DOI: 10.1371/journal.pone.0007722 42. Soumya K.R., Philip S., Sugathan S., Mathew J., Radhakrishnan E.K.. Virulence factors associated with Coagulase Negative Staphylococci isolated from human infections. 3 Biotech (2017) 7 140. DOI: 10.1007/s13205-017-0753-2 43. Fey P.D., Olson M.E.. Current concepts in biofilm formation of. Futur. Microbiol. (2010) 5 917-933. DOI: 10.2217/fmb.10.56 44. Solati S.M., Tajbakhsh E., Khamesipour F., Gugnani H.C.. Prevalence of virulence genes of biofilm producing strains of. AMB Express (2015) 5 134. DOI: 10.1186/s13568-015-0134-3 45. Dworkin M.S., Williamson J., Jones J.L., Kaplan J.E.. Prophylaxis with Trimethoprim-Sulfamethoxazole for Human Immunodeficiency Virus–Infected Patients: Impact on Risk for Infectious Diseases. Clin. Infect. Dis. (2001) 33 393-398. DOI: 10.1086/321901 46. Duran N., Ozer B., Duran G.G., Onlen Y., Demir C.. Antibiotic resistance genes & susceptibility patterns in staphylococci. Indian J. Med. Res. (2012) 135 389. PMID: 22561627 47. Zhang H.Z., Hackbarth C.J., Chansky K.M., Chambers H.F.. A Proteolytic Transmembrane Signaling Pathway and Resistance to β-Lactams in Staphylococci. Science (2001) 291 1962-1965. DOI: 10.1126/science.1055144 48. Chambers H.F.. Methicillin resistance in staphylococci: Molecular and biochemical basis and clinical implications. Clin. Microbiol. Rev. (1997) 10 781-791. DOI: 10.1128/CMR.10.4.781 49. Shrestha L.B., Bhattarai N.R., Rai K., Khanal B.. Antibiotic Resistance and mecA Gene Characterization of Coagulase-negative Staphylococci Isolated from Clinical Samples in Nepal. Infect. Drug Resist. (2020) 13 3163. DOI: 10.2147/IDR.S274163 50. Asiimwe B.B., Kiiru J., Mshana S., Neema S., Keenan K., Kesby M., Mwanga J.R., Sloan D.J., Mmbaga B.T., Smith V.A.. Protocol for an interdisciplinary cross-sectional study investigating the social, biological and community-level drivers of antimicrobial resistance (AMR): Holistic Approach to Unravel Antibacterial Resistance in East Africa (HATUA). BMJ Open (2021) 11 e041418. DOI: 10.1136/bmjopen-2020-041418 51. Micci F., Thorsen J., Panagopoulos I., Nyquist K.B., Zeller B., Tierens A., Heim S.. High-throughput sequencing identifies an NFIA/CBFA2T3 fusion gene in acute erythroid leukemia with t(1;16)(p31;q24). Leukemia (2013) 27 980-982. DOI: 10.1038/leu.2012.266 52. Bolger A.M., Lohse M., Usadel B.. Trimmomatic: A flexible trimmer for Illumina sequence data. Bioinformatics (2014) 30 2114-2120. DOI: 10.1093/bioinformatics/btu170 53. Kang H.J., Kawasawa Y.I., Cheng F., Zhu Y., Xu X., Li M., Sousa A.M.M., Pletikos M., Meyer K.A., Sedmak G.. Spatio-temporal transcriptome of the human brain. Nature (2011) 478 483-489. DOI: 10.1038/nature10523 54. Antipov D., Hartwick N., Shen M., Raiko M., Lapidus A., Pevzner P.A.. plasmidSPAdes: Assembling plasmids from whole genome sequencing data. Bioinformatics (2016) 32 3380-3387. DOI: 10.1093/bioinformatics/btw493 55. Bankevich A., Nurk S., Antipov D., Gurevich A.A., Dvorkin M., Kulikov A.S., Pevzner P.A.. SPAdes: A new genome assembly algorithm and its applications to single-cell sequencing. J. Comput. Biol. (2012) 19 455-477. DOI: 10.1089/cmb.2012.0021 56. Seemann T.. Prokka: Rapid Prokaryotic Genome Annotation. Bioinformatics (2014) 30 2068-2069. DOI: 10.1093/bioinformatics/btu153 57. De Koning W., Miladi M., Hiltemann S., Heikema A., Hays J.P., Flemming S., Stubbs A.P.. NanoGalaxy: Nanopore long-read sequencing data analysis in Galaxy. GigaScience (2020) 9 giaa105. DOI: 10.1093/gigascience/giaa105
--- title: Retroviral Infection and Commensal Bacteria Dependently Alter the Metabolomic Profile in a Sterile Organ authors: - Jessica Spring - Vera Beilinson - Brian C. DeFelice - Juan M. Sanchez - Michael Fischbach - Alexander Chervonsky - Tatyana Golovkina journal: Viruses year: 2023 pmcid: PMC9967258 doi: 10.3390/v15020386 license: CC BY 4.0 --- # Retroviral Infection and Commensal Bacteria Dependently Alter the Metabolomic Profile in a Sterile Organ ## Abstract Both viruses and bacteria produce “pathogen associated molecular patterns” that may affect microbial pathogenesis and anti-microbial responses. Additionally, bacteria produce metabolites, while viruses could change the metabolic profiles of the infected cells. Here, we used an unbiased metabolomics approach to profile metabolites in spleens and blood of murine leukemia virus-infected mice monocolonized with Lactobacillus murinus to show that viral infection significantly changes the metabolite profile of monocolonized mice. We hypothesize that these changes could contribute to viral pathogenesis or to the host response against the virus and thus open a new avenue for future investigations. ## 1. Introduction The intestinal commensal microbiota is a key factor that mediates host health by providing nutrients and vitamins, supporting the development and shaping of the secondary lymphoid organs in the intestine, and conferring resistance of colonization by pathogenic microorganisms [1]. Many of these and other microbiota-dependent processes are mediated by a number of commensal bacteria-derived high- and low-abundance molecules. Various commensal small molecules exhibit potent biological activities that can influence the host’s cellular activities [2,3,4]. Of these small molecules, metabolites have been shown to serve as an energy source, modulate diseases and psychiatric disorders, promote intestinal barrier function, and modulate the immune system [5,6,7]. Metabolites have both positive and negative effects on pathogenic bacteria. These effects include enhanced biofilm formation, increased expression of virulence factors [8,9,10], disruption of bacterial cell structures, suppression of bacterial growth, and stimulation of innate immune cell proliferation [8,11,12,13,14]. Viral infections are also impacted by metabolites. Metabolites can be used as energy sources to promote viral replication [15,16], whereas other metabolites promote the immune response and augment interferon (IFN) expression in a variety of viral infection models [17,18]. While the effect of commensal microbe-derived metabolites on cellular activities and pathogen infections is clear, a reciprocal impact of the commensal microbiota and viral pathogens on metabolites has yet to be investigated. We previously showed that some commensal bacteria, including Lactobacillus murinus, significantly enhanced virally-induced leukemia development without affecting virus replication [19]. We used a similar model coupled with an unbiased metabolomics approach to investigate the influence of cross-talk between commensal bacterial and viral infection on metabolites. Here, we show viral infection and commensal bacterial colonization dependently alter the metabolomic profile in a sterile organ, the spleen. ## 2.1. Mice Breeding and maintenance of mice used in this study were conducted at the animal facility of The University of Chicago. BALB/cJ mice were bought from The Jackson Laboratory (TJL). Females and males were used at a ratio of ~50:50. The Animal Care and Use Committee at The University of Chicago reviewed and approved the studies described here. ## 2.2. Monitoring Sterility in GF Isolators BALB/cJ mice were re-derived as germ-free (GF) at Taconic farms and subsequently housed in sterile isolators at the gnotobiotic facility at the University of Chicago. GF isolator sterility was determined as previously described [20]. Fecal pellets from isolators were collected weekly and quickly frozen. A bead-beating/phenol-chloroform protocol was used to extract DNA. DNA was amplified using primers that widely hybridize to bacterial 16S rRNA gene sequences. In addition, microbiological cultures were established with GF fecal pellets, specific-pathogen-free (SPF) fecal pellets (positive control), sterile saline (sham), and sterile culture medium (negative control). Samples were inoculated into BHI, Nutrient, and Sabbaroud Broth media and incubated aerobically and anaerobically at 37 °C and 42 °C. Cultures were maintained for five days until deemed negative. ## 2.3. Colonization of GF Mice Lactobacillus murinus (L. murinus ASF361) was isolated from the Altered Schaedler Flora (ASF) consortium [21]. Accordingly, fecal matter from ASF-colonized gnotobiotic mice was suspended in sterile PBS and plated onto a selective media for Lactobacilli, namely de Man, Rogosa, and Sharpe agar (MRS) (ThermoFisher Scientific, Waltham, MA, USA). The identity of L. murinus was confirmed by sequencing of the 16S RNA ribosomal amplicon generated via PCR from a bacteria colony. GF BALB/cJ mice were monocolonized with L. murinus by oral gavage of 200 μL of overnight liquid culture grown from a single colony. Monocolonization was validated by sequencing PCR products generated from fecal DNA and 16S rRNA primers and was verified at closing of the experiment. ## 2.4. Virus Isolation and Infection Rauscher-like murine leukemia virus (RL-MuLV) is composed of N, B tropic ecotropic and mink cell focus forming (MCF) virus [22]. RL-MuLV was isolated from tissue culture supernatant of chronically infected SC-1 cells (ATCC CRL-1404). Titers of ecotropic virus within the RL-MuLV mixture were determined via an XC plaque assay [23]. 1 × 103 pfus were intraperitoneally (i.p.) injected into 6–8-week-old BALB/cJ female mice (G0 mice). G0 mice were bred to produce the progeny G1 mice. Spleens from G1 mice were used as a source of virus. Homogenized spleens from 2–3-month-old G1 mice were centrifuged at 4 °C for 15 min at 2000 rpm. Supernatant was collected and layered onto a $30\%$ PBS/sucrose cushion and spun at 31,000 rpm for 1 h at 4 °C in a TW55.1 bucket rotor. The pellet fraction was re-suspended in PBS. Insoluble material was removed by spinning the resuspended fraction at 4 °C at 10,000 rpm. Supernatant was collected and aliquoted, titered via plaque assay, and stored at −80 °C. RL-MuLV isolated from spleens was diluted in sterile PBS followed by filtering through a sterile 0.22 μm membrane in a laminar flow hood. GF and L. murinus-monocolonized BALB/cJ females were injected i.p. with 1 × 103 pfus (G0 mice). G0 females were bred to produce G1 mice, which were used for the metabolomic analysis. Uninfected and infected GF and L. murinus-monocolonized mice were sacrificed at 2 months. Plasma and spleens were isolated and stored at −80 °C. Mice were confirmed to be infected via PCR using primers specific for the LTR of ecotropic virus. Forward primer: 5′ATGAACGACCCCACCAAGT3′ and reverse primer: 5′GAGACCCTCCCAAGGAACAG3′. Spleen weights ranged from 0.07 to 0.09 g among uninfected GF mice; 0.07–0.09 g among uninfected L. murinus-monocolonized; and 0.15–0.19 g among infected L. murinus-monocolonized. Infected mice were non-leukemic as evaluated by histology of H&E-stained spleen sections. ## 2.5. FITC Permeability Assay Permeability of the mouse gut was assessed using a FITC permeability assay. Food and water were removed from mouse cages for four hours. After four hours, mice were gavaged with 60 mg of FITC-dextran (MW 4000, Sigma-Aldrich, St. Louis, MO, USA) per 100 g of mouse [24]. Three hours post gavage, mice were bled into 100 μL heparin, and plasma was separated by spinning the blood at 2000 rpm for ten minutes at 4 °C. Then, 50 μL of each sample was loaded into a 96-well plate in duplicate to measure FITC concentration at emission and excitation wavelengths of 520 nm and 490 nm, respectively, using a TECAN fluorescence spectrophotometer and Magellan software. FITC-dextran standards were diluted in plasma from unmanipulated mice. Fluorescence from the negative control samples (plasma from unmanipulated mice) was subtracted from fluorescence of the standards and experimental samples. ## 2.6. Metabolomics Spleens and sera from six GF uninfected, five GF infected, seven colonized uninfected, and twenty-two colonized infected mice were harvested and maintained at −80 °C prior to metabolite extraction. About 10 mg per sample was used for the following analysis. Extraction solvent (1:2:2 water:acetonitrile:methanol containing stable isotope-labeled internal standards) was added to each spleen sample at a ratio of 400 µL per 10 mg of tissue. Four to six 2.3 mm stainless steel beads were added to each sample tube and method blanks. Samples were homogenized by bead beating at 20 Hz for 15 min. Homogenized samples were placed at −20 °C for 1 h to maximize protein precipitation, followed by brief vortexing and centrifugation for 5 min at 4 °C and 14,000× g. Next, 120 µL of supernatant was taken and passed through a 0.2 µm polyvinylidene fluoride centrifugal filter at 4 °C for 3 min at 6000× g. Collected extract was stored at 4 °C during analysis. Metabolites were extracted from serum samples as follows. Frozen specimens were thawed on ice and inverted five times to mix. From each specimen, 50 µL of serum was aliquoted into new 1.5 mL Eppendorf tubes. Then, 150 µL of 1:1 acetonitrile:methanol, containing stable isotope-labeled internal standards, was added to each 50 µL aliquot, followed by vortexing for 20 s. All tubes were stored at −20 °C for 1 h and then were promptly vortexed for 20 s and centrifuged at −10 °C for 10 min at 21,130× g. Supernatant was transferred to a new 1.5 mL Eppendorf tube and dried at room temperature in a LabConco Speedvac. Dried extracts were resuspended in 50 µL 4:1 acetonitrile:water and stored at 4 °C during analysis. Technical replicates of serum samples were made in the same fashion as described for spleen samples. Data processing, chromatography, and tandem mass spectral data collection methods have been previously described elsewhere [25]. Briefly, hydrophilic interaction liquid chromatography (HILIC) method was used for analysis of the polar metabolites. Prepared samples were injected onto a Waters Acquity UPLC BEH Amide column (150 mm length × 2.1 mm id; 1.7 μm particle size) with an additional Waters Acquity VanGuard BEH Amide pre-column (5 mm × 2.1 mm id; 1.7 μm particle size) maintained at 45 °C coupled to an Thermo Vanquish UPLC. The mobile phases were prepared with 10 mM ammonium formate and $0.125\%$ formic acid in either $100\%$ LC-MS-grade water for mobile phase (A) or 95:5 v/v acetonitrile:water for mobile phase (B). Gradient elution was performed from $100\%$ (B) at 0–2 min to $70\%$ (B) at 7.7 min, $40\%$ (B) at 9.5 min, $30\%$ (B) at 10.25 min, and $100\%$ (B) at 12.75 min, isocratic until 16.75 min with a column flow of 0.4 mL/min. Spectra were collected using a Thermo Q Exactive HF Hybrid Quadrupole-Orbitrap mass spectrometer in both positive and negative mode ionization (separate injections). Full MS-ddMS2 data were collected, and an inclusion list was used to prioritize MS2 selection of metabolites from the in-house “local” library; when additional scan bandwidth was available; MS2 data were collected in a data-dependent manner. The mass range was 60–900 mz, resolution was 60 k (MS1) and 15 k (MS2), centroid data were collected, loop count was 4, and isolation window was 1.2 Da. Noise was defined as peak height below 100,000. Aligned peaks were retained if present in at least $10\%$ of all samples and at least $60\%$ of samples within at least one treatment group. Blank samples, generated by extracting water in place of biological material, were used to remove features not originating from serum or spleen. To be retained, all signals were required to have at least 10-fold increase in signal in one or more samples compared to the blank average. All annotations were assigned based on, at minimum, accurate mass and MS2 spectral matching against the largest freely available spectra repository, MassBank of North America [26]. Technical replicates from each sample matrix (spleen and serum, independently) were generated by pooling all samples of each matrix. Metabolites that were not analytically reproducible based on >$30\%$ reproducibility standard deviation in the technical replicates were removed prior to statistical analysis. Metabolites with RSDs less than $30\%$ were normalized with Log10, quantile normalization, followed by determining the z-score using ANOVA. ## 3. Results Bacterial colonization and viral infection drastically alter the peripheral metabolite landscape. To identify metabolites whose presence and abundance were dictated by both colonization and viral infection, an unbiased metabolomics approach was undertaken. Accordingly, we compared metabolites within the sera and spleens of murine leukemia virus (MuLV)-infected and uninfected BALB/cJ mice that were either germ-free (GF) or monocolonized with mouse commensal Lactobacillus murinus (L. murinus). We used Rauscher-like MuLV variant, which is spread via blood and replicates within the proliferating erythroid progenitor cells, subsequently causing erythroid leukemia [22]. GF and L. murinus-colonized ex-GF mice were injected with 0.22 μM sterilized virus and bred to produce offspring (G1 mice) infected via a natural route. Sera and spleens from four groups of 2-month-old G1 mice, namely uninfected GF, ex-GF L. murinus-colonized, MuLV-infected GF, and MuLV-infected ex-GF L. murinus-colonized, were subjected to coupled liquid chromatography/tandem mass spectrometry (LC-MS/MS). Mass spectrometry spectra were collected under positive and negative ionization modes. Metabolites whose frequencies had a relative standard deviation (RSD) greater than $30\%$ between technical replicates were discarded, and the remaining metabolites were subjected to further analyses (Figure 1A–C). Serum samples collected from uninfected and infected germ-free mice (Figure 1A) were subjected to LC-MS/MS separately from serum samples of uninfected and infected colonized mice (Figure 1B); therefore, it was impossible to match unknown metabolites between these two groups. Consequently, unknown metabolites from the sera only were discarded in later comparative analyses. Only identifiable metabolites found between all four groups of mice were kept for subsequent analyses. This includes 73 and 55 known metabolites identified under positive and negative ionization, respectively (Figure 1A,B and Tables S1 and S2). Z-scores for known metabolites can be found in Spreadsheet S1. MuLV infection of GF mice marginally altered the overall metabolomic profile (Figure 2A) and the level of metabolites within the sera (Figure 2B). In contrast, L. murinus colonization led to the greatest shift in metabolites (Figure 2A,B). MuLV infection of L. murinus-colonized mice induced further minimal changes in the overall metabolomic landscape and quantity of metabolites within the sera (Figure 2A,B). From the spleens of all mouse groups, 60 known metabolites in total were identified under positive ionization and 52 under negative ionization (Figure 1C and Tables S2 and S3 and Spreadsheet S1). Similar to the sera, the vast majority of changes were found among over 1500 unknown metabolites (Figure 1C). Contrary to the sera, splenic metabolites were altered most drastically in the presence of both L. murinus colonization and viral infection compared to either condition alone (Figure 2C,D). Therefore, viral infection in monocolonized mice led to negligible shifts in the metabolomic profile within the sera but greatly influenced metabolites within the spleen. These data indicate that bacterial colonization of the gut together with MuLV infection promotes metabolite accumulation or production in the spleen. Through our analysis, we identified three groups of metabolites that, compared to GF mice, were altered upon MuLV infection or L. murinus colonization or both infection and colonization. Group 1 includes metabolites that were significantly altered in L. murinus-colonized mice; group 2 includes metabolites that were significantly altered in virally infected mice; and group 3 consists of metabolites that were significantly altered only in colonized and infected mice. As we were interested in metabolites whose presence and quantity were mediated by colonization and viral infection, we further investigated metabolites identified in group 3. Considering bacterial colonization alone conferred the largest shift in the metabolomic landscape within the sera (Figure 3A), it was not surprising that only 29 identifiable metabolites were found to be significantly altered in the sera of colonized and infected mice compared to the sera from the other three groups of mice (Figure 3A). As sera from GF uninfected and infected mice were subjected to LC-MS/MS separately from the sera of monocolonized uninfected and infected mice, unknown metabolites from the sera could not be compared between these groups. In contrast, 271 known and unknown metabolites were found to be impacted by both bacterial colonization and viral infection in the spleen (Figure 3C,D). However, the vast majority of them were unknown, and only 21 metabolites were identifiable (Figure 3D). Viral infection has been shown to alter the intestinal barrier, enabling microbial translocation [28,29]. Based on the observation that MuLV infection of colonized mice greatly enhanced the abundance of metabolites within the spleen, we sought to determine whether MuLV increased the permeability of the intestines. To address this possibility, intestinal permeability was compared between uninfected and infected, conventionally housed, specific-pathogen-free (SPF) mice. Mice were orally gavaged with MW 4000 FITC-dextran, a large molecule that will only cross from the intestine into the periphery if the intestinal barrier has been compromised, and fluorescence within the plasma was assayed four hours later. Concentration of FITC detected within the plasma was similar between uninfected and infected SPF mice, suggesting MuLV does not alter intestinal permeability (Figure 4). Thus, enhanced permeability of the intestinal barrier is not the cause of increased metabolites within the sera and spleen in infected mice. ## 4. Discussion The commensal microbiota has been demonstrated to modulate replication and pathogenesis of viruses from various families [30,31,32,33]. Furthermore, commensal-derived metabolites have exhibited both pro- and anti-viral effects. Conversely, any impact of viral infection on microbially-derived metabolites remained unknown. LC-MS/MS metabolomics has been successfully utilized to identify metabolites [34,35,36]. Therefore, we took a similar approach to discover bacterially-derived or bacterially-dependent host-derived metabolites that are influenced by viral infection. Abundances of 271 metabolites were found to be affected by both bacterial colonization and viral infection in the spleen (Figure 3C). The vast majority of these metabolites [250] are unknown, and consequently, their relationship to known metabolites remains to be determined. Known metabolites within the sera and spleen, influenced by the combination of bacterial colonization and viral infection, were grouped by chemical type in an attempt to identify trends in the abundances of similar metabolites (Figure 3B,D). The only discernable trend determined by both the virus and bacterium was the decreased abundance of cholines in the sera and increased abundance of amino acids (both essential and non-essential) in the spleen (Figure 3D). Mechanistically, how MuLV infection in L. murinus-colonized mice alters the metabolomic landscape in the spleen remains to be determined. MuLV infection results in increased extramedullary hematopoiesis within the spleen [22], expanding the population of target cells the virus can infect. MuLV readily replicates within the proliferating erythroid progenitor cells, augmenting the chance of pro-viral integration near a cellular proto-oncogene and leading to the generation of pre-cancerous cells. The presence of the microbiota during the generation of these cells may alter the transcriptional landscape, leading to the production of metabolites within the spleen. Permeability of the intestinal barrier did not increase in SPF mice upon MuLV infection (Figure 4), indicating the observed increase of metabolites within the spleens of monocolonized infected mice may not be due to significant influx of metabolites originating from the gut. Thus, the origin of certain microbially dependent, virally induced metabolites may be within the spleen. However, due to the diffusible nature of small, gut-derived metabolites, we cannot rule out the gut as a source of these metabolites. In summary, our study has demonstrated for the first time that the gut commensal bacterium and the virus together alter the metabolomic landscape of the spleen, a sterile organ not directly connected with the gut. ## References 1. Lee Y.K., Mazmanian S.K.. Has the microbiota played a critical role in the evolution of the adaptive immune system?. Science (2010.0) 330 1768-1773. DOI: 10.1126/science.1195568 2. Gasaly N., de Vos P., Hermoso M.A.. Impact of Bacterial Metabolites on Gut Barrier Function and Host Immunity: A Focus on Bacterial Metabolism and Its Relevance for Intestinal Inflammation. Front. Immunol. (2021.0) 12 658354. DOI: 10.3389/fimmu.2021.658354 3. Donia M.S., Fischbach M.A.. Human Microbiota. Small molecules from the human microbiota. Science (2015.0) 349 1254766. DOI: 10.1126/science.1254766 4. Buffie C.G., Pamer E.G.. Microbiota-mediated colonization resistance against intestinal pathogens. Nat. Rev. Immunol. (2013.0) 13 790-801. DOI: 10.1038/nri3535 5. Lee W.-J., Hase K.. Gut microbiota–generated metabolites in animal health and disease. Nat. Chem. Biol. (2014.0) 10 416-424. DOI: 10.1038/nchembio.1535 6. Sittipo P., Shim J.-W., Lee Y.K.. Microbial Metabolites Determine Host Health and the Status of Some Diseases. Int. J. Mol. Sci. (2019.0) 20. DOI: 10.3390/ijms20215296 7. Averina O.V., Zorkina Y.A., Yunes R.A., Kovtun A.S., Ushakova V.M., Morozova A.Y., Kostyuk G.P., Danilenko V.N., Chekhonin V.P.. Bacterial Metabolites of Human Gut Microbiota Correlating with Depression. Int. J. Mol. Sci. (2020.0) 21. DOI: 10.3390/ijms21239234 8. Li Z., Quan G., Jiang X., Yang Y., Ding X., Zhang D., Wang X., Hardwidge P.R., Ren W., Zhu G.. Effects of Metabolites Derived From Gut Microbiota and Hosts on Pathogens. Front. Cell. Infect. Microbiol. (2018.0) 8 314. DOI: 10.3389/fcimb.2018.00314 9. Kirby D.T., Savage J.M., Plotkin B.J.. Menaquinone (Vitamin K2) enhancement of. J. Biosci. Med. (2014.0) 2014 43928 10. Yang Y., Yao F., Zhou M., Zhu J., Zhang X., Bao W., Wu S., Hardwidge P.R., Zhu G.. F18ab Escherichia coli flagella expression is regulated by acyl-homoserine lactone and contributes to bacterial virulence. Vet. Microbiol. (2013.0) 165 378-383. DOI: 10.1016/j.vetmic.2013.04.020 11. Hochbaum Allon I., Kolodkin-Gal I., Foulston L., Kolter R., Aizenberg J., Losick R.. Inhibitory Effects of d-Amino Acids on. J. Bacteriol. (2011.0) 193 5616-5622. DOI: 10.1128/JB.05534-11 12. Araki S., Suzuki M., Fujimoto M., Kimura M.. Enhancement of Resistance to Bacterial Infection in Mice by Vitamin B. J. Vet. Med. Sci. (1995.0) 57 599-602. DOI: 10.1292/jvms.57.599 13. Jacobson A., Lam L., Rajendram M., Tamburini F., Honeycutt J., Pham T., Van Treuren W., Pruss K., Stabler S.R., Lugo K.. A Gut Commensal-Produced Metabolite Mediates Colonization Resistance to Salmonella Infection. Cell Host Microbe (2018.0) 24 296-307.e7. DOI: 10.1016/j.chom.2018.07.002 14. Rowland I., Gibson G., Heinken A., Scott K., Swann J., Thiele I., Tuohy K.. Gut microbiota functions: Metabolism of nutrients and other food components. Eur. J. Nutr. (2018.0) 57 1-24. DOI: 10.1007/s00394-017-1445-8 15. Sánchez-García F.J., Pérez-Hernández C.A., Rodríguez-Murillo M., Moreno-Altamirano M.M.B.. The Role of Tricarboxylic Acid Cycle Metabolites in Viral Infections. Front. Cell. Infect. Microbiol. (2021.0) 11 725043. DOI: 10.3389/fcimb.2021.725043 16. Fontaine Krystal A., Sanchez Erica L., Camarda R., Lagunoff M.. Dengue Virus Induces and Requires Glycolysis for Optimal Replication. J. Virol. (2015.0) 89 2358-2366. DOI: 10.1128/JVI.02309-14 17. Alwin A., Karst S.M.. The influence of microbiota-derived metabolites on viral infections. Curr. Opin. Virol. (2021.0) 49 151-156. DOI: 10.1016/j.coviro.2021.05.006 18. Trompette A., Gollwitzer E.S., Pattaroni C., Lopez-Mejia I.C., Riva E., Pernot J., Ubags N., Fajas L., Nicod L.P., Marsland B.J.. Dietary Fiber Confers Protection against Flu by Shaping Ly6c− Patrolling Monocyte Hematopoiesis and CD8+ T Cell Metabolism. Immunity (2018.0) 48 992-1005.e8. DOI: 10.1016/j.immuni.2018.04.022 19. Spring J., Khan A.A., Lara S., O’Grady K., Wilks J., Gurbuxani S., Erickson S., Fischbach M., Jacobson A., Chervonsky A.. Gut commensal bacteria enhance pathogenesis of a tumorigenic murine retrovirus. Cell Rep. (2022.0) 40 111341. DOI: 10.1016/j.celrep.2022.111341 20. Wilks J., Beilinson H., Theriault B., Chervonsky A., Golovkina T.. Antibody-mediated immune control of a retrovirus does not require the microbiota. J. Virol. (2014.0) 88 6524-6527. DOI: 10.1128/JVI.00251-14 21. Sarma-Rupavtarm R.B., Ge Z., Schauer D.B., Fox J.G., Polz M.F.. Spatial distribution and stability of the eight microbial species of the altered schaedler flora in the mouse gastrointestinal tract. Appl. Environ. Microbiol. (2004.0) 70 2791-2800. DOI: 10.1128/AEM.70.5.2791-2800.2004 22. Hook L.M., Jude B.A., Ter-Grigorov V.S., Hartley J.W., Morse H.C., Trainin Z., Toder V., Chervonsky A.V., Golovkina T.V.. Characterization of a novel murine retrovirus mixture that facilitates hematopoiesis. J. Virol. (2002.0) 76 12112-12122. DOI: 10.1128/JVI.76.23.12112-12122.2002 23. Rowe W.P., Pugh W.E., Hartley J.W.. Plaque assay techniques for murine leukemia viruses. Virology (1970.0) 42 1136-1139. DOI: 10.1016/0042-6822(70)90362-4 24. Volynets V., Reichold A., Bárdos G., Rings A., Bleich A., Bischoff S.C.. Assessment of the Intestinal Barrier with Five Different Permeability Tests in Healthy C57BL/6J and BALB/cJ Mice. Dig. Dis. Sci. (2016.0) 61 737-746. DOI: 10.1007/s10620-015-3935-y 25. Noecker C., Sanchez J., Bisanz J.E., Escalante V., Alexander M., Trepka K., Heinken A., Liu Y., Dodd D., Thiele I.. Systems biology illuminates alternative metabolic niches in the human gut microbiome. bioRxiv (2022.0). DOI: 10.1101/2022.09.19.508335:2022.09.19.508335 26. Mass Bank of North America 27. Chen S.Y., Feng Z., Yi X.. A general introduction to adjustment for multiple comparisons. J. Thorac. Dis. (2017.0) 9 1725-1729. DOI: 10.21037/jtd.2017.05.34 28. Brenchley J.M., Price D.A., Schacker T.W., Asher T.E., Silvestri G., Rao S., Kazzaz Z., Bornstein E., Lambotte O., Altmann D.. Microbial translocation is a cause of systemic immune activation in chronic HIV infection. Nat. Med. (2006.0) 12 1365-1371. DOI: 10.1038/nm1511 29. Labarta-Bajo L., Nilsen S.P., Humphrey G., Schwartz T., Sanders K., Swafford A., Knight R., Turner J.R., Zúñiga E.I.. Type I IFNs and CD8 T cells increase intestinal barrier permeability after chronic viral infection. J. Exp. Med. (2020.0) 217 e20192276. DOI: 10.1084/jem.20192276 30. Baldridge M.T., Nice T.J., McCune B.T., Yokoyama C.C., Kambal A., Wheadon M., Diamond M.S., Ivanova Y., Artyomov M., Virgin H.W.. Commensal microbes and interferon-lambda determine persistence of enteric murine norovirus infection. Science (2015.0) 347 266-269. DOI: 10.1126/science.1258025 31. Kane M., Case L.K., Kopaskie K., Kozlova A., MacDearmid C., Chervonsky A.V., Golovkina T.V.. Successful transmission of a retrovirus depends on the commensal microbiota. Science (2011.0) 334 245-249. DOI: 10.1126/science.1210718 32. Kuss S.K., Best G.T., Etheredge C.A., Pruijssers A.J., Frierson J.M., Hooper L.V., Dermody T.S., Pfeiffer J.K.. Intestinal microbiota promote enteric virus replication and systemic pathogenesis. Science (2011.0) 334 249-252. DOI: 10.1126/science.1211057 33. Uchiyama R., Chassaing B., Zhang B., Gewirtz A.T.. Antibiotic treatment suppresses rotavirus infection and enhances specific humoral immunity. J. Infect. Dis. (2014.0) 210 171-182. DOI: 10.1093/infdis/jiu037 34. Yoshimoto S., Loo T.M., Atarashi K., Kanda H., Sato S., Oyadomari S., Iwakura Y., Oshima K., Morita H., Hattori M.. Obesity-induced gut microbial metabolite promotes liver cancer through senescence secretome. Nature (2013.0) 499 97-101. DOI: 10.1038/nature12347 35. Sharon G., Cruz N.J., Kang D.-W., Gandal M.J., Wang B., Kim Y.-M., Zink E.M., Casey C.P., Taylor B.C., Lane C.J.. Human Gut Microbiota from Autism Spectrum Disorder Promote Behavioral Symptoms in Mice. Cell (2019.0) 177 1600-1618.e17. DOI: 10.1016/j.cell.2019.05.004 36. Hsiao E.Y., McBride S.W., Hsien S., Sharon G., Hyde E.R., McCue T., Codelli J.A., Chow J., Reisman S.E., Petrosino J.F.. Microbiota modulate behavioral and physiological abnormalities associated with neurodevelopmental disorders. Cell (2013.0) 155 1451-1463. DOI: 10.1016/j.cell.2013.11.024
--- title: Impact of Vancomycin Treatment and Gut Microbiota on Bile Acid Metabolism and the Development of Non-Alcoholic Steatohepatitis in Mice authors: - Kaichi Kasai - Naoya Igarashi - Yuki Tada - Koudai Kani - Shun Takano - Tsutomu Yanagibashi - Fumitake Usui-Kawanishi - Shiho Fujisaka - Shiro Watanabe - Mayuko Ichimura-Shimizu - Kiyoshi Takatsu - Kazuyuki Tobe - Koichi Tsuneyama - Yukihiro Furusawa - Yoshinori Nagai journal: International Journal of Molecular Sciences year: 2023 pmcid: PMC9967260 doi: 10.3390/ijms24044050 license: CC BY 4.0 --- # Impact of Vancomycin Treatment and Gut Microbiota on Bile Acid Metabolism and the Development of Non-Alcoholic Steatohepatitis in Mice ## Abstract The potential roles of the gut microbiota in the pathogenesis of non-alcoholic fatty liver disease, including non-alcoholic steatohepatitis (NASH), have attracted increased interest. We have investigated the links between gut microbiota and NASH development in Tsumura-Suzuki non-obese mice fed a high-fat/cholesterol/cholate-based (iHFC) diet that exhibit advanced liver fibrosis using antibiotic treatments. The administration of vancomycin, which targets Gram-positive organisms, exacerbated the progression of liver damage, steatohepatitis, and fibrosis in iHFC-fed mice, but not in mice fed a normal diet. F$\frac{4}{80}$+-recruited macrophages were more abundant in the liver of vancomycin-treated iHFC-fed mice. The infiltration of CD11c+-recruited macrophages into the liver, forming hepatic crown-like structures, was enhanced by vancomycin treatment. The co-localization of this macrophage subset with collagen was greatly augmented in the liver of vancomycin-treated iHFC-fed mice. These changes were rarely seen with the administration of metronidazole, which targets anaerobic organisms, in iHFC-fed mice. Finally, the vancomycin treatment dramatically modulated the level and composition of bile acid in iHFC-fed mice. Thus, our data demonstrate that changes in inflammation and fibrosis in the liver by the iHFC diet can be modified by antibiotic-induced changes in gut microbiota and shed light on their roles in the pathogenesis of advanced liver fibrosis. ## 1. Introduction Non-alcoholic fatty liver disease (NAFLD) is characterized by the hepatic manifestation of metabolic disorders. NAFLD encompasses a wide range of liver pathologies, from non-alcoholic fatty liver (NAFL) to non-alcoholic steatohepatitis (NASH) and cirrhosis [1]. The etiology of NAFLD involves both genetic predisposition and epigenetic regulation by environmental factors [1]. The multiple parallel-hit hypothesis, which states that multiple factors play a role in its pathogenesis, has been widely accepted [2]. Besides genetic and environmental factors, the internal environment and various metabolites, such as gut microbiota and bile acid (BA), require further investigation to fully elucidate the pathogenesis of NAFLD [3]. BAs are steroid-based molecules that are synthesized from cholesterol in the liver. They are involved in nutritional lipid absorption and function as signaling molecules to regulate the metabolism via BA-sensing receptors, including the nuclear farnesoid X receptor (FXR) and cell membrane Takeda G protein-coupled receptor 5 (TGR5) (also known as GPBAR1 (G protein-coupled membrane receptor 1)) [4,5,6]. Gut microbiota can contribute to secondary BA synthesis and BA metabolism by structural modulation, including deconjugation, dehydroxylation, oxidation, and desulfation [7]. In the large intestine, microbes modify hydroxyl groups on the steroid ring and produce various secondary BAs [8]. The level and composition of BA play roles in intestinal barrier function [9,10]. The increased expression levels of hepatic BA synthesis genes in rats on a high-fat diet (HFD) elevate 12α-hydroxylated (12αOH) BA concentrations, such as cholic acid (CA), in enterohepatic circulation [11,12]. The microbial baiB gene is involved in 7α-dehydroxylation in the large intestine, which produces deoxycholic acid (DCA), a secondary 12αOH BA [13]. Furthermore, high DCA levels induced leaky gut in mice on an HFD [14]. Of note, the liver is continuously exposed to BAs via the enterohepatic circulation, and this exposure is associated with liver disease, including liver cancer [15]. Many studies have focused on identifying the specific roles of gut microbiota in metabolic disorders. Gut dysbiosis is closely associated with a decrease in beneficial short-chain fatty acid-producing bacteria, changes in the composition of BAs [16], and activation of immune responses against lipopolysaccharides (LPS) [17]. Changes in BA and inflammatory signaling, insulin resistance, and glucose metabolism induced by an HFD were shown to be modified by antibiotic-induced changes in gut microbiota [18]. Changes in gut microbiota that affect the gut–liver axis are also associated with the progression of chronic liver disease, advanced fibrosis, and hepatocellular carcinoma (HCC) [19,20,21,22,23,24]. For example, gut sterilization and germ-free conditions can decrease the formation of HCC in mouse models [15,25]. In humans, Ruminococcaceae and Veillonellaceae are the main microbial taxa associated with significant liver fibrosis in non-obese subjects [26]. In the Ruminococcaceae, Ruminococcus faecis alleviated fibrosis in mouse NAFLD models [26], and Faecalibacterium prausnitzii, also in the Ruminococcaceae, was reported to regulate the hepatic fat content and composition of lipid species and reduce adipose tissue inflammation in HFD-fed mice [27]. Histologically, NASH is characterized by steatosis, lobular inflammation, hepatocellular ballooning, and fibrosis of the liver. In humans, fibrosis in NASH typically first appears in zone 3 as a “chicken-wire” pattern that spreads to the portal area, ultimately leading to portal–portal and portal–central bridging fibrosis [28]. However, currently used animal NASH models, such as the methionine- and choline-deficient diet models and carbon tetrachloride (CCl4)-induced model, do not reflect human dietary habits and are not characterized by histological changes similar to those of human NASH, such as bridging fibrosis. A high-fat/cholesterol/cholate-based (iHFC) diet that is not deficient in specific nutrients induced advanced fibrosis that spreads to the portal area of Tsumura-Suzuki non-obese (TSNO) mice [29]. Furthermore, iHFC-fed TSNO mice exhibited stage 3 bridging fibrosis [29]. Therefore, iHFC diet-induced NASH in TSNO mice may be a more representative model of human NASH than current animal models. In our previous study, we also characterized the dynamics of hepatic macrophage subsets in iHFC-fed TSNO mice and investigated their roles in the development of advanced liver fibrosis [30]. The fluorescence-activated cell sorter analysis showed that the recruited macrophages consisted of two distinct subsets: CD11c+/Ly6C− and CD11c−/Ly6C+ cells [30]. Furthermore, the histological and RNA sequence analyses indicated that CD11c+/Ly6C− cells promoted liver fibrosis and hepatic stellate cell activation, whereas CD11c−/Ly6C+ cells played roles in anti-inflammation and tissue repair [30]. The aim of this study was to dissect the links between gut microbiota, liver inflammation, and fibrosis using two antibiotics that target different organisms. We found that vancomycin, which targets Gram-positive organisms, exacerbated the progression of liver damage, steatohepatitis, and fibrosis in iHFC-fed TSNO mice, but not in mice fed a normal diet. F$\frac{4}{80}$+-recruited macrophages were more abundant in the liver of vancomycin-treated mice. The infiltration of CD11c+-recruited macrophages into the liver, forming hepatic crown-like structures (hCLSs), was enhanced by vancomycin treatment. The co-localization of this macrophage subset with collagen in the liver of iHFC-fed TSNO mice was greatly augmented in vancomycin-treated mice. These changes were rarely seen in iHFC-fed TSNO mice with metronidazole, which targets anaerobic organisms. Finally, vancomycin treatment dramatically modulated the level and balance of BA composition in the iHFC-fed TSNO mice, by decreasing bile salt hydrolase (BSH) and major bacterial BA dehydroxylases, which are mainly produced by Gram-positive bacteria. We concluded that the effects of vancomycin on the exacerbation of NASH development depend on an interaction of the gut microbiome, BA metabolism, and inflammatory responses in the liver of the iHFC-fed TSNO mice. ## 2.1. Modification of Gut Microbiota by Vancomycin Exacerbates iHFC Diet-Induced Liver Damage and Lipid Metabolism in TSNO Mice To investigate the relationship between gut microbiota and the development of NASH, we challenged TSNO mice with an iHFC diet, and modified the microbiome by treatment with either metronidazole (MTZ) or vancomycin (VCM) (Figure S1). MTZ is an absorbable antibiotic that targets anaerobic organisms, whereas VCM is a non-absorbable antibiotic that targets primarily Gram-positive organisms. These antibiotics are frequently used to treat Clostridium difficile infection and are commonly used in patients with inflammatory bowel disease. A 16S rRNA sequence analysis of fecal samples was performed to determine the bacterial composition of TSNO mice after 4 weeks on an iHFC diet, with either placebo or antibiotic treatment. The α-diversity and principal component analyses showed clear differences in the microbial communities between mice fed a normal diet (ND) and iHFC-fed mice with placebo (iHFC + P) or between iHFC + P and iHFC-fed mice with antibiotic treatment (iHFC + MTZ and iHFC + VCM) (Figure S2A,B). Despite initial differences in microbiota among these four groups, the gut microbiota of mice in both the ND and iHFC + P groups were dominated by Firmicutes, whereas in the iHFC + MTZ group, Bacteroidetes accounted for $57\%$ of the bacterial sequences (Figure S2C). In the iHFC + VCM group, Bacteroidetes were reduced to $0\%$, and this reduction was associated with an increased abundance of Proteobacteria and Deferribacteres (Figure S2C). A heatmap of the relative abundance of 16S sequences shows that VCM treatment killed most Gram-positive bacteria, such as Ruminococcus and Lachnospiraceae, short-chain fatty acid-producing bacteria (Figure S3). Conversely, Gram-negative bacteria such as Parasutterella, which are involved in BA metabolism, and Escherichia, typical LPS-possessing bacteria, increased and became dominant by VCM treatment (Figure S3). Clostridium are Gram-positive bacteria that may have survived VCM treatment because they formed spores or were highly resistant (Figure S3). We also examined whether antibiotic treatment affected liver weight and liver damage in iHFC-fed mice. The body mass was significantly decreased by VCM treatment only at 8 weeks of feeding (Figure 1A, left), and no significant differences were found in the average daily food intake between the placebo and antibiotic groups (Figure 1A, right). For liver weight, there was a significant difference between the placebo and antibiotic groups or between the MTZ and VCM groups (Figure 1B). The liver weight of VCM-treated mice was significant decreased at both 4 and 8 weeks of treatment (Figure 1B). The liver from iHFC-fed mice was pale, as reported previously (Figure 1C) [30], and here we found that the VCM-treated mice had much paler livers than the placebo- and MTZ-treated mice (Figure 1C). There were marked increases in the plasma alanine aminotransferase (ALT) activity of the iHFC-fed mice after 4 and 8 weeks of VCM treatment (Figure 1D). The iHFC + VCM mice also had high plasma total cholesterol (T-CHO) and triglyceride (TG) concentrations compared with mice in the other groups (Figure 1D). To determine whether the effects of VCM treatment on the liver were due to antibiotic-induced liver damage, we treated ND-fed mice with placebo or antibiotics. MTZ and VCM affected only the body weight of these mice; no other parameter was affected (Figure S4). Thus, VCM treatment exacerbated liver damage and lipid metabolism in iHFC-fed TSNO mice. ## 2.2. Modification of Gut Microbiota by Vancomycin Exacerbates iHFC Diet-Induced Inflammation, Steatosis, Hepatocyte Ballooning, and Fibrosis in the Liver of TSNO Mice To further investigate these findings, we performed a histopathological analysis of the liver from mice in the iHFC + P, iHFC + MTZ, and iHFC + VCM groups. We observed mild steatosis after 4 weeks in the iHFC + P, and the steatosis worsened as the time on the iHFC diet increased (Figure 2A,B). Lobular inflammation and hepatocyte ballooning were observed from 8 weeks (Figure 2A,B). These histopathological changes were significantly exacerbated after 4 weeks of VCM treatment (Figure 2B), and lobular inflammation and hepatocyte ballooning were significantly improved after 8 weeks of MTZ treatment (Figure 2B). We also examined iHFC diet-induced fibrotic changes in the liver of placebo- or antibiotic-treated mice. The Sirius red-positive areas in the sections from VCM-treated mice were significantly higher than those of placebo- or MTZ-treated mice from 4 weeks of treatment, indicating fibrosis was most apparent after VCM treatment (Figure 2C,D). The fibrosis gradually expanded, with bridging fibrosis becoming apparent after 8 weeks of VCM treatment similar to that after 24 weeks on the iHFC diet (Figure 2C,E) [29,30]. Thus, VCM treatment exacerbated steatohepatitis and fibrosis in the liver of iHFC diet-fed TSNO mice. ## 2.3. Modification of Gut Microbiota by Vancomycin Increases the Levels of Inflammation- or Fibrosis-Related Gene Expression in the Liver of iHFC Diet-Fed TSNO Mice We measured the expression levels of inflammatory and fibrotic genes in the liver of mice in the iHFC + P, iHFC + MTZ, and iHFC + VCM groups. The expression levels of pro-inflammatory genes (Tnf, Il6, and Il1b), chemokine and its receptor genes (Ccl2, (C-C motif) ligand 2, and Ccr2), neutrophil granulocyte marker (Mpo), macrophage marker (Adgre1), and M1 macrophage markers (Itgax and Nos2) were markedly increased by VCM treatment (Figure 3A), whereas the expression level of M2 macrophage marker (Arg1) was significantly decreased by VCM treatment compared with their expression by MTZ treatment (Figure 3A). Consistent with the histological data (Figure 2C,D), the expression levels of collagen 1 (Col1a1) and alpha-smooth muscle actin (αSMA) (Acta2) mRNAs were higher in the liver of VCM-treated mice than they were in the placebo- or MTZ-treated mice (Figure 3B). VCM treatment also markedly increased the mRNA expression levels of the tissue inhibitor of metalloproteinase 1 (Timp1) and matrix metallopeptidase 2 (Mmp2), which regulate extracellular matrix degradation (Figure 3B). The mRNA expression of transforming growth factor beta (Tgfb1), which regulates extracellular matrix production, was high in the liver of VCM-treated mice (Figure 3B), and the expression of a hepatic stellate cell marker gene (Des) was significantly higher in the iHFC + VCM group than it was in the iHFC + P and iHFC + MTZ groups (Figure 3B). Thus, modification of the gut microbiome by VCM treatment exacerbates inflammation and fibrosis in the liver of iHFC-fed TSNO mice. ## 2.4. Modification of Gut Microbiota by Vancomycin Augments the Infiltration of CD45+ Leukocytes and F4/80+ Macrophages in the Liver of iHFC Diet-Fed TSNO Mice To investigate the roles of immune cells in the pathogenesis of the VCM-mediated exacerbation of NASH, we isolated non-parenchymal cells from the liver and counted the numbers of CD45+ leukocytes. No significant differences in the numbers of live non-parenchymal cells were detected between the iHFC + P and the iHFC + MTZ or iHFC + VCM group (Figure S5). A flow cytometric analysis demonstrated that the percentage of CD45+ cells including Kupffer cells (KCs) increased by VCM treatment compared with the percentages for the placebo or MTZ treatment (Figure 4A,B), whereas the percentage and cell number of CD45− cells were lower by VCM treatment than they were for the placebo- or MTZ treatment (Figure 4A,C). We also found that the percentage of F$\frac{4}{80}$+ cells excluding KCs was increased by VCM treatment (Figure 4D,E). These results imply that CD45+ leukocytes, which include F$\frac{4}{80}$+ macrophages, accumulate in the liver with VCM treatment and play key roles in exacerbating inflammation and fibrosis in the liver of iHFC-fed mice. ## 2.5. Modification of Gut Microbiota by Vancomycin Augments the Infiltration of F4/80+-Recruited Macrophages in the Liver of iHFC-Fed TSNO Mice We then focused on the subsets of CD45+ and F$\frac{4}{80}$+ non-parenchymal cells in the liver of the antibiotic-treated TSNO mice. Previously, we reported that at least three populations expressing F$\frac{4}{80}$ and/or CD11b were present in ND-fed mouse liver: F$\frac{4}{80}$−/CD11bHi neutrophils, F$\frac{4}{80}$Int/CD11bInt-Hi-recruited macrophages, and F$\frac{4}{80}$Hi/CD11bInt KCs [30]. We also found that the percentage of F$\frac{4}{80}$Hi/CD11bInt KCs was reduced by feeding mice the iHFC diet [30]. The iHFC diet also induced hepatic infiltration with F$\frac{4}{80}$Int/CD11bInt-Hi macrophages, which constitute hCLSs, suggesting that this macrophage subset might be involved in the development of hepatocyte death-induced liver fibrosis in this mouse NASH model [30]. The percentage of F$\frac{4}{80}$Hi/CD11bInt KCs was significantly reduced by MTZ and VCM treatment after 4 weeks of treatment compared with their percentage in the placebo group (Figure 5A,B). Conversely, the percentage and cell number of F$\frac{4}{80}$Int/CD11bInt-Hi macrophages were increased by VCM treatment (Figure 5C). Immunohistochemical staining showed that the F$\frac{4}{80}$-positive areas of liver sections were gradually increased by iHFC-feeding [30]. In the VCM-treated mice, the F$\frac{4}{80}$-positive areas were more than those in the placebo- and MTZ-treated mice (Figure 5D,E). Interestingly, MTZ treatment significantly reduced the F$\frac{4}{80}$-positive area at 8 weeks compared with the placebo (Figure 5E). These results imply that VCM treatment augmented the infiltration of F$\frac{4}{80}$Int/CD11bInt-Hi-recruited macrophages into the liver, thereby forming hCLSs, suggesting that this macrophage subset may be involved in the VCM exacerbation of liver fibrosis in this mouse NASH model. ## 2.6. Modification of Gut Microbiota by Vancomycin Augments the Accumulation of CD11c+-Recruited Macrophages in the Liver of iHFC-Fed TSNO Mice We previously showed that F$\frac{4}{80}$+-recruited macrophages in the liver of iHFC-fed mice consisted of two macrophage subsets: CD11c+/Ly6C− and CD11c−/Ly6C+ cells [30]. The RNA sequence analysis also showed that CD11c+/Ly6C− cells may promote liver fibrosis and hepatic stellate cell activation, whereas CD11c−/Ly6C+ cells may play an anti-inflammatory role and promote tissue repair [30]. The percentages and numbers of these two subsets were slightly increased by VCM treatment (Figure 6A). Significant increases in the cell numbers of these subsets were found in the iHFC + MTZ and iHFC + VCM groups compared with the number in the iHFC + P group (Figure 6B). An immunohistochemical analysis showed that the accumulation of CD11c+ and Ly6C+ cells when fed the iHFC diet was increased by VCM treatment (Figure 6C–F), whereas MTZ treatment significantly decreased the CD11c-positive area at 8 weeks of treatment (Figure 6D). The Ly6C-positive area was also significantly reduced by MTZ treatment at 4 and 8 weeks compared with the placebo and VCM treatment (Figure 6F). ## 2.7. Modification of Gut Microbiota by Vancomycin Augments iHFC Diet-Induced Accumulation of CD11c+ Cells That Are Co-Localized with Collagen Fibers in the Liver of iHFC-Fed TSNO Mice To elucidate the role of CD11c+-recruited macrophages in the development of progressive liver fibrosis in VCM-treated mice, we performed immunofluorescence staining for CD11c and collagen type 1. Collagen deposition was evident in the liver at 12 weeks of feeding mice the iHFC diet [30]. Mice in the iHFC + P group had few CD11c+ cells and low collagen deposition at 8 weeks of feeding (Figure 7A); however, the co-localization of collagen was evident in the liver at this time point by CD11c immunostaining (Figure 7A,B). Interestingly, VCM-treated mice showed more CD11c+ aggregation, which forms hCLSs, and collagen deposition than the placebo-treated mice (Figure 7A), and the co-localization of collagen with CD11c in the liver was markedly enhanced by VCM treatment (Figure 7A,B). Conversely, no changes were detected between CD11c and collagen in the liver from MTZ-treated mice and placebo-treated mice by immunostaining (Figure 7A,B). These findings suggest that CD11c+-recruited macrophages played key roles in the development of progressive liver fibrosis in the iHFC-fed mice treated with VCM. ## 2.8. The iHFC Diet and Antibiotic Treatment Affect the Level of Fecal Bile Acid Metabolism To further investigate the mechanism involved in the exacerbation of liver inflammation and fibrosis in the VCM-treated mouse NASH model, we assessed the effect of the iHFC diet and antibiotic treatment on BA metabolism. We demonstrated that the iHFC diet and iHFC + VCM had major effects on the levels of fecal BA metabolites (Figure 8). Fecal BA profiles in TSNO mice revealed that an iHFC diet (iHFC + P) increased the relative abundances of cholic acid (CA) and deoxycholic acid (DCA) and decreased that of muricholic acid (MCA) (Figure 8A). Total concentrations of fecal unconjugated BAs in the iHFC + P, iHFC + MTZ, and iHFC + VCM groups were decreased compared with those in the ND group (Figure 8B, left). Notably, VCM treatment markedly reduced the total concentrations of unconjugated BAs (Figure 8B, left). Compared with the ND group, the concentration of CA, which was contained in the iHFC diet, was increased in the iHFC+ P, iHFC + MTZ, and iHFC + VCM groups, whereas the concentration of MCA was severely reduced in all three groups (Figure 8B, left). For conjugated BAs in the feces, the taurocholic acid (TCA) concentration was markedly increased in the iHFC + VCM group compared with its concentration in the other group (Figure 8B, right). An analysis of individual BAs confirmed that the CA, DCA, and taurodeoxycholic acid (TDCA) concentrations were increased in the iHFC + P and iHFC + MTZ groups, whereas the lithocholic acid (LCA) and MCA concentrations were reduced in these groups compared with their levels in the ND group (Figure 8C,D). Interestingly, the iHFC-mediated elevation of the CA concentration was enhanced by VCM; whereas DCA was almost undetectable in the VCM-treated mice (Figure 8C); and TDCA, a taurine-conjugated DCA, was also undetectable (Figure 8D). Notably, the concentrations of secondary 12αOH BAs, including LCA and MCA, in the iHFC + VCM group were severely reduced compared with their concentrations in the other groups (Figure 8C). Because BAs exert their effects through FXR and TGR5 [4], we hypothesized that the changes in BA composition induced by the iHFC diet and antibiotic modification of gut microbiota may affect these receptors in the liver. We found that feeding mice the iHFC diet had no effects on the mRNA levels of these receptors (Figure S7). Interestingly, the expression of TGR5 mRNA was markedly increased in the liver of mice in the iHFC + VCM group compared with its expression in the liver of mice in the iHFC + P and iHFC + MTZ groups after 8 weeks of treatment (Figure S7, right). Enterobacteria, especially Gram-positive bacteria, including Clostridium, Lactobacillus, and Bifidobacterium, have bile salt hydrolase (BSH), which deconjugates BAs by removing amino acids from conjugated BAs [31]. Deconjugated free BAs are dehydroxylated mainly in the large intestine. DCA and LCA are produced by 7α-dehydroxylation of CA and chenodeoxycholic acid (CDCA) by the intestinal bacteria, Clostridium XI and Clostridium XIVa, that express the BA inducible (bai) gene, such as baiE [13]. On the basis of the 16S rRNA sequencing analysis results, we performed a predictive functional profile analysis using the PICRUSt2 software Version 2021.2 to predict the metagenome expression of BSH and bai genes. In the iHFC + P group, the metagenome expression of BSH and bai genes, including baiCD, baiH, and baiI, was significantly reduced compared with their expression in the ND group (Figure S8). MTZ treatment showed similar results to placebo treatment (Figure S8). Notably, the metagenome expressions of BSH and various bai genes were almost undetectable in mice in the iHFC + VCM group (Figure S8). Thus, the iHFC diet and VCM treatment affected the level and balance of BA composition. Moreover, these data indicate that VCM treatment decreased the expression levels of BA-modifying bacterial genes, including BSH and other major bacterial BA dehydroxylases, thereby dramatically reducing the fecal levels of DCA, LCA, and TDCA and elevating the fecal levels of CA and TCA. ## 3. Discussion In the present study, we dissected the links between gut microbiota, liver inflammation and advanced fibrosis using a newly established mouse NASH model with two antibiotics, MTZ and VCM. We showed that VCM, which targets Gram-positive organisms, exacerbated the progression of liver damage, steatohepatitis, and fibrosis in the iHFC-fed TSNO mice. VCM-treated mice had abundant F$\frac{4}{80}$+-recruited macrophages in their livers. Among these macrophages, the infiltration of CD11c+ subsets into the liver, forming hCLSs, was enhanced by VCM treatment. The co-localization of this macrophage subset with collagen in the liver of mice was greatly augmented in mice in the iHFC + VCM group. MTZ, which targets anaerobic organisms, had almost no effect on these changes. Additionally, the level and balance of BA composition were dramatically modulated by the VCM treatment, which decreased the expression of BSH and other major bacterial BA dehydroxylases, that were produced mainly by Gram-positive bacteria. Thus, we report that changes in inflammation and fibrosis in the liver by the iHFC diet can be modified by antibiotic-induced changes in gut microbiota. Their specific roles require further investigation, but the present findings provide novel insight into the importance of the gut microbe in the development of inflammation and advanced fibrosis in NASH by modulating BA metabolism. In obese patients with type 2 diabetes mellitus, the intestinal barrier function is impaired, resulting in a leaky gut that facilitates LPS influx [17]. The role of LPS in the development of NAFLD and NASH has attracted some attention [32,33]. Circulating LPS levels and small intestinal permeability are increased in patients with NALFD, and these factors are associated with the severity of hepatic steatosis [32,34,35]. Akkermansia muciniphila is thought to contribute to improved intestinal barrier function [36]. We found that VCM treatment killed most Gram-positive bacteria (Figure S3), including Ruminococcaceae (Clostridium IV) and Lachnospiraceae (Clostridium XIVa), which are useful for the suppression of NASH development and produce short-chain fatty acids that are important for the gut environment and immunity. Additionally, most of the bacteria that survived the VCM treatment were Gram-negative bacteria that had LPS, which may be linked to liver inflammation. Furthermore, Akkermansia muciniphila was decreased in mice in the iHFC + P, iHFC + MTZ, and iHFC + VCM groups compared with its presence in the ND group (Figure S3), suggesting that leaky gut may be involved in the development of NASH in iHFC-fed mice, and that this is exacerbation by VCM treatment. Furthermore, Alistipes, which is inversely related to fibrosis, was decreased in the iHFC + P and iHFC + VCM groups compared with its presence in the ND group (Figure S3). These mechanisms may be involved in the regulation of intestinal bacteria and the development of NASH triggered by the iHFC diet and VCM treatment. Further studies are required to clarify the details of these mechanisms. VCM treatment dramatically modulated the level and balance of BA composition in the iHFC-fed mice (Figure 8) by decreasing the expression of BSH and other major bacterial BA dehydroxylases (Figure S8) that are mainly produced by Gram-positive bacteria, such as Lactobacillus. The level and balance of BA composition have a crucial role in intestinal barrier function [9,10]; in particular, secondary BAs are key factors responsible for leaky gut [14]. Previous studies showed that the level of primary 12αOH BAs in the enterohepatic circulation was selectively increased by feeding an HFD [11,12], and the level of primary BAs in the small intestine has been shown to be higher than that of secondary BAs in the large intestine [12]. In rats on a CA-supplemented diet, VCM treatment increased the TCA concentration in the ileum [37], whereas DCA was almost undetectable at the same site, and fecal CA concentration was high in rats fed the CA with VCM treatment [37]. These data are similar to our present findings (Figure 8) and confirm that VCM treatment suppressed the 7α-dehydroxylation of CA in the large intestine. Moreover, TCA selectively affects the gut permeability in the ileum, which is the site for reabsorbing conjugated BAs [37]. In addition, the effect of the TCA on ileal permeability is much stronger than that of DCA in the colon for gut leakiness [37]. Therefore, we suggest that the iHFC diet with VCM treatment may disrupt gut permeability by increasing TCA levels in the small intestine and induce liver inflammation by enterohepatic circulation in our mouse NASH model. Future studies are needed to investigate the TCA concentrations and mucosal barrier function in the small intestine using the mouse NASH model. There are two major receptors of BAs: FXR and TGR5 [4]. FXR is known to regulate BA and lipid metabolism [38,39]. TGR5 is located on immune cells, such as macrophages [40,41], and has anti-inflammatory effects [42]. Secondary BAs such as DCA and TDCA have high affinities for TGR5 [43]. DCA and LCA were shown to reduce NAFLD by activating FXR and TGR5 in the intestinal tract and inducing the production of FGF$\frac{19}{21}$ and GLP-1, respectively [5,44]. We found that secondary BAs such as DCA and LCA were markedly decreased by VCM, which suggested that the activation of FXR and TGR5 by these BAs may have been suppressed by the VCM treatment. This could explain the exacerbation of NASH development with VCM treatment. We also found that VCM treatment markedly increased TGR5 mRNA expression in the liver of the iHFC-fed mice (Figure S7, right). TGR5 is highly expressed in macrophages and has an important role in regulating inflammation [40,41]. Because VCM treatment enhanced iHFC diet-induced macrophage infiltration in the liver (Figure 4D,E and Figure 5E), this macrophage infiltration probably reflects the elevated TGR5 mRNA expression. In summary, VCM treatment had profound effects on the gut microbiota that resulted in changes in BA metabolism and the development of liver inflammation and advanced fibrosis in a novel diet-induced mouse NASH model (Figure 9). The present findings may provide information that will aid in the development of therapeutic agents for NASH that target the gut microbiome. On the other hand, this study has several limitations (e.g., the lack of comparison with relevant data sets of mice fed ND+P, ND+MTZ, and ND+VCM; the lack of information on inflammation-related gene expression in the gut; the low numerosity of the sample size). Future investigations will clarify these issues as well as the details of Gram-positive bacteria and BAs involved in the development of liver inflammation and fibrosis. ## 4.1. Mice Used in This Study Male TSNO mice (6-weeks-old) were purchased from the Institute of Animal Reproduction (Ibaraki, Japan) and maintained in microisolator cages under specific pathogen-free conditions in the animal facility of Toyama Prefectural University under standard light conditions (12 h light/dark cycle) and with free access to water and food. Seven-week-old male TSNO mice were fed ad libitum either a normal diet (ND) (MF, Oriental-Yeast, Tokyo, Japan) or an iHFC diet, that was high in fat, cholesterol, and cholate ($69.5\%$ standard chow, $28.75\%$ palm oil, $1.25\%$ cholesterol, and $0.5\%$ cholate) (Hayashi Kasei, Osaka, Japan) for the indicated periods. For the antibiotic treatment, 7-week-old mice were fed the ND or iHFC diet for 4 or 8 weeks, and simultaneously treated with either placebo, VCM (Nacalai Tesque, Kyoto, Japan) (50 µg/body weight), or MTZ (Nacalai Tesque) (50 µg/body weight) by oral gavage for 5 days a week. These antibiotics were diluted with sterilized water and prepared in total to be 200 µL/mouse. The animal care policies and procedures/protocol used in the experiments were approved by the Animal Experiment Ethics Committee of Toyama Prefectural University (Approval No. R1-2 and R3-6). ## 4.2. Plasma Biochemical Analysis Blood samples were collected from the inferior vena cava, and plasma samples were also collected. Plasma levels of alanine aminotransferase (ALT), total cholesterol (T-CHO), and triglyceride (TG) were measured by DRI-CHEM NX700 (Fujifilm, Tokyo, Japan) following the manufacturer’s instruction. ## 4.3. Isolation of Non-Parenchymal Cells from Liver To isolate non-parenchymal cells from the liver, mice were anesthetized (isoflurane) and perfusion was performed with phosphate-buffered saline. The isolation of non-parenchymal cells was performed using a Liver Dissociation Kit (Miltenyi Biotech, Bergisch Gladbach, Germany) following the manufacturer’s instructions. The cell suspension was passed through a cell strainer (100 µm) and used for the flow cytometry analysis. ## 4.4. Flow Cytometry Analysis The non-parenchymal cells (2 × 105) were incubated with anti-mouse FcγR (2.4G2) to block binding of the fluorescence-labeled antibodies to FcγR. After 20 min, the cells were stained with predetermined optimal concentrations of the respective antibodies. Then, 7-amino-actinomycin D (7-AAD) (BD Biosciences, San Diego, CA, USA) was used to exclude dead cells. Flow cytometry analyses were conducted on a FACSCantoII (Becton Dickinson & Co., Mountain View, CA, USA), and the data were analyzed with Flowjo software Version 10.8.1 (BD Biosciences). The antibodies for flow cytometry are listed in Table S1. ## 4.5. Preparation of RNA and cDNA Total RNA was extracted using a NucleoSpin RNA Mini Kit (Macherey-Nagel, Düren, Germany) following the manufacturer’s instructions. RNA was reverse transcribed using a PrimeScript® RT Reagent Kit (Takara Bio Inc., Shiga, Japan) following the manufacturer’s instructions. ## 4.6. Quantitative Real-Time PCR A qRT-PCR was performed with a FastStart Universal Probe Master (Roche Applied Science, Mannheim, Germany) and analyzed with a CFX96 Touch™ Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) following the manufacturers’ instructions. Relative transcript abundance was normalized to that of Hprt mRNA. The information for the TaqMan primer/probes (Applied Biosystems, Carlsbad, CA, USA) used for the qRT-PCR is listed in Table S2. ## 4.7. Histological and Immunohistochemistry Analysis Portions of the liver were excised and fixed immediately with $4\%$ formaldehyde at room temperature. Paraffin-embedded tissue sections were cut into 4 μm slices and placed on slides. Sections were stained with hematoxylin and eosin or Sirius red, according to standard procedures. Anti-F$\frac{4}{80}$, anti-CD11c, and anti-Ly6C antibodies were purchased from Cedarlane Laboratories (Burlington, Ontario, Canada), Invitrogen (Waltham, MA, USA), and Abcam (Cambridge, MA, USA), respectively. Positive areas for F$\frac{4}{80}$ and Sirius red were measured using ImageJ software Version 1.53t [45]. Histologic steatosis, lobular inflammation, and hepatocyte ballooning were assessed according to the criteria proposed by Kleiner et al. [ 46]. All histological analyses were performed by pathologists (K.Tsuneyama and M.I-S.), and the histological scores and grade were determined in a blinded manner. ## 4.8. Fluorescent Immunohistochemistry Analysis We incubated 7 μm frozen sections with anti-CD11c (Invitrogen) and secondary antibody (anti-hamster IgG, Southern Biotech, Birmingham, AL, USA), and sequentially incubated using a TSA Fluorescein System (Akoya Biosciences, Marlborough, MA, USA). The sections were also incubated with anti-collagen type 1 (Novotec, Bron, France) and secondary antibody (anti-rabbit IgG Alexa Fluor 594, Abcam). Finally, the sections were incubated with DAPI (Invitrogen). Images were acquired using a BX50 microscope and its imaging system (Olympus, Tokyo, Japan). The positive signal of each co-localized area was selected according to the method of Tolivia et al. ( Figure S6) [47] and analyzed using Adobe Photoshop CS software, version 8. ## 4.9. Metagenomic 16S rRNA Sequencing One fresh fecal pellet was collected from each mouse and stored at −80 °C. For DNA extraction, the frozen stool samples were transferred to a lysis buffer from a ZymoBIOMICS DNA Miniprep Kit (Zymo Research, Irvine, CA, USA), homogenized using a Disruptor Genie (Scientific Industries, Bohemia, NY, USA; 3000 rpm for 20 min), and extracted using a ZymoBIOMICS DNA Miniprep Kit, according to the manufacturer’s instructions. For sequencing, 16S rRNA gene sequence libraries were prepared according to the Illumina (San Diego, CA, USA) protocol, as described previously [48]. The final libraries were pooled, diluted to 4 nM in 5 mM Tris-HCl buffer, and sequenced using a Miseq System with a 600-Cycle Kit (Illumina). A prediction of functional profiles from the 16S rRNA data sets was conducted using the Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt2) software and the Kyoto Encyclopedia of Genes and Genomes (KEGG) database release 70.0. Pathways involved in human diseases were removed because they apply to human cells or tissues. ## 4.10. Bacterial Community Analysis The Quantitative Insight Into Microbial Ecology 1 (QIIME1) pipeline, an open-source bioinformatics tool for the analysis of raw microbiome DNA sequencing data, was used to investigate the bacterial composition as described previously [48]. The taxonomy assignment was performed using the EzbioCloud reference database. ## 4.11. Bile Acid Analysis Fecal BAs were extracted by homogenizing dried feces (30–60 mg) in a mixture of 0.5 mL methanol, 0.8 mL acetonitrile, and 0.2 mL $28\%$ (w/v) ammonium hydroxide with 100 nM 23-dinor-deoxycholic acid as an internal standard (Steraloids, Newport, RI, USA). The fecal homogenates were centrifuged, and the resultant supernatants were applied to solid-phase extraction columns to obtain fractions containing BAs as described previously [49]. The BA levels in the fecal samples were determined by liquid chromatography-electrospray ionization-mass spectrometry (LC-ESI-MS) as described previously [49]. ## 4.12. Statistical Analysis Statistical significance was evaluated by two-way ANOVA followed by the post hoc Tukey test for multiple comparisons. A statistical analysis was performed using GraphPad Prism 9 software (GraphPad; San Diego, CA, USA). When we analyzed the parametric tests, we first performed an F test on Prism 9 software to confirm that the p-value of the F test was >0.05 and equally distributed. $p \leq 0.05$ was considered statistically significant. The results are presented as the mean ± standard deviation (SD). ## References 1. Wree A., Broderick L., Canbay A., Hoffman H.M., Feldstein A.E.. From NAFLD to NASH to cirrhosis-new insights into disease mechanisms. Nat. Rev. Gastroenterol. Hepatol. (2013) 10 627-636. DOI: 10.1038/nrgastro.2013.149 2. Tilg H., Moschen A.R.. Evolution of inflammation in nonalcoholic fatty liver disease: The multiple parallel hits hypothesis. Hepatology (2010) 52 1836-1846. DOI: 10.1002/hep.24001 3. Giuffre M., Campigotto M., Campisciano G., Comar M., Croce L.S.. A story of liver and gut microbes: How does the intestinal flora affect liver disease? A review of the literature. Am. J. Physiol. Gastrointest. Liver Physiol. (2020) 318 G889-G906. DOI: 10.1152/ajpgi.00161.2019 4. Ferrell J.M., Chiang J.Y.L.. Understanding Bile Acid Signaling in Diabetes: From Pathophysiology to Therapeutic Targets. Diabetes. Metab. J. (2019) 43 257-272. DOI: 10.4093/dmj.2019.0043 5. Finn P.D., Rodriguez D., Kohler J., Jiang Z., Wan S., Blanco E., King A.J., Chen T., Bell N., Dragoli D.. Intestinal TGR5 agonism improves hepatic steatosis and insulin sensitivity in Western diet-fed mice. Am. J. Physiol. Gastrointest. Liver Physiol. (2019) 316 G412-G424. DOI: 10.1152/ajpgi.00300.2018 6. Chiang J.Y., Pathak P., Liu H., Donepudi A., Ferrell J., Boehme S.. Intestinal Farnesoid X Receptor and Takeda G Protein Couple Receptor 5 Signaling in Metabolic Regulation. Dig. Dis. (2017) 35 241-245. DOI: 10.1159/000450981 7. Setchell K.D., Lawson A.M., Tanida N., Sjovall J.. General methods for the analysis of metabolic profiles of bile acids and related compounds in feces. J. Lipid. Res. (1983) 24 1085-1100. DOI: 10.1016/S0022-2275(20)37923-2 8. Griffiths W.J., Sjovall J.. Bile acids: Analysis in biological fluids and tissues. J. Lipid. Res. (2010) 51 23-41. DOI: 10.1194/jlr.R001941 9. Hegyi P., Maleth J., Walters J.R., Hofmann A.F., Keely S.J.. Guts and Gall: Bile Acids in Regulation of Intestinal Epithelial Function in Health and Disease. Physiol. Rev. (2018) 98 1983-2023. DOI: 10.1152/physrev.00054.2017 10. Rohr M.W., Narasimhulu C.A., Rudeski-Rohr T.A., Parthasarathy S.. Negative Effects of a High-Fat Diet on Intestinal Permeability: A Review. Adv. Nutr. (2020) 11 77-91. DOI: 10.1093/advances/nmz061 11. Hori S., Abe T., Lee D.G., Fukiya S., Yokota A., Aso N., Shirouchi B., Sato M., Ishizuka S.. Association between 12alpha-hydroxylated bile acids and hepatic steatosis in rats fed a high-fat diet. J. Nutr. Biochem. (2020) 83 108412. DOI: 10.1016/j.jnutbio.2020.108412 12. Yoshitsugu R., Kikuchi K., Iwaya H., Fujii N., Hori S., Lee D.G., Ishizuka S.. Alteration of Bile Acid Metabolism by a High-Fat Diet Is Associated with Plasma Transaminase Activities and Glucose Intolerance in Rats. J. Nutr. Sci. Vitaminol. (2019) 65 45-51. DOI: 10.3177/jnsv.65.45 13. Ridlon J.M., Kang D.J., Hylemon P.B.. Bile salt biotransformations by human intestinal bacteria. J. Lipid Res. (2006) 47 241-259. DOI: 10.1194/jlr.R500013-JLR200 14. Stenman L.K., Holma R., Eggert A., Korpela R.. A novel mechanism for gut barrier dysfunction by dietary fat: Epithelial disruption by hydrophobic bile acids. Am. J. Physiol. Gastrointest. Liver Physiol. (2013) 304 G227-G234. DOI: 10.1152/ajpgi.00267.2012 15. Yoshimoto S., Loo T.M., Atarashi K., Kanda H., Sato S., Oyadomari S., Iwakura Y., Oshima K., Morita H., Hattori M.. Obesity-induced gut microbial metabolite promotes liver cancer through senescence secretome. Nature (2013) 499 97-101. DOI: 10.1038/nature12347 16. Begley M., Gahan C.G., Hill C.. The interaction between bacteria and bile. FEMS Microbiol. Rev. (2005) 29 625-651. DOI: 10.1016/j.femsre.2004.09.003 17. Cani P.D., Amar J., Iglesias M.A., Poggi M., Knauf C., Bastelica D., Neyrinck A.M., Fava F., Tuohy K.M., Chabo C.. Metabolic endotoxemia initiates obesity and insulin resistance. Diabetes (2007) 56 1761-1772. DOI: 10.2337/db06-1491 18. Fujisaka S., Ussar S., Clish C., Devkota S., Dreyfuss J.M., Sakaguchi M., Soto M., Konishi M., Softic S., Altindis E.. Antibiotic effects on gut microbiota and metabolism are host dependent. J. Clin. Investig. (2016) 126 4430-4443. DOI: 10.1172/JCI86674 19. Loomba R., Seguritan V., Li W., Long T., Klitgord N., Bhatt A., Dulai P.S., Caussy C., Bettencourt R., Highlander S.K.. Gut Microbiome-Based Metagenomic Signature for Non-invasive Detection of Advanced Fibrosis in Human Nonalcoholic Fatty Liver Disease. Cell Metab. (2017) 25 1054-1062.e1055. DOI: 10.1016/j.cmet.2017.04.001 20. Boursier J., Mueller O., Barret M., Machado M., Fizanne L., Araujo-Perez F., Guy C.D., Seed P.C., Rawls J.F., David L.A.. The severity of nonalcoholic fatty liver disease is associated with gut dysbiosis and shift in the metabolic function of the gut microbiota. Hepatology (2016) 63 764-775. DOI: 10.1002/hep.28356 21. Bajaj J.S., Heuman D.M., Hylemon P.B., Sanyal A.J., White M.B., Monteith P., Noble N.A., Unser A.B., Daita K., Fisher A.R.. Altered profile of human gut microbiome is associated with cirrhosis and its complications. J. Hepatol. (2014) 60 940-947. DOI: 10.1016/j.jhep.2013.12.019 22. Caussy C., Tripathi A., Humphrey G., Bassirian S., Singh S., Faulkner C., Bettencourt R., Rizo E., Richards L., Xu Z.Z.. A gut microbiome signature for cirrhosis due to nonalcoholic fatty liver disease. Nat. Commun. (2019) 10 1406. DOI: 10.1038/s41467-019-09455-9 23. Qin N., Yang F., Li A., Prifti E., Chen Y., Shao L., Guo J., Le Chatelier E., Yao J., Wu L.. Alterations of the human gut microbiome in liver cirrhosis. Nature (2014) 513 59-64. DOI: 10.1038/nature13568 24. Kamiya T., Ohtani N.. The role of immune cells in the liver tumor microenvironment: An involvement of gut microbiota-derived factors. Int. Immunol. (2022) 34 467-474. DOI: 10.1093/intimm/dxac020 25. Ma C., Han M., Heinrich B., Fu Q., Zhang Q., Sandhu M., Agdashian D., Terabe M., Berzofsky J.A., Fako V.. Gut microbiome-mediated bile acid metabolism regulates liver cancer via NKT cells. Science (2018) 360 730-735. DOI: 10.1126/science.aan5931 26. Lee G., You H.J., Bajaj J.S., Joo S.K., Yu J., Park S., Kang H., Park J.H., Kim J.H., Lee D.H.. Distinct signatures of gut microbiome and metabolites associated with significant fibrosis in non-obese NAFLD. Nat. Commun. (2020) 11 4982. DOI: 10.1038/s41467-020-18754-5 27. Munukka E., Rintala A., Toivonen R., Nylund M., Yang B., Takanen A., Hanninen A., Vuopio J., Huovinen P., Jalkanen S.. Faecalibacterium prausnitzii treatment improves hepatic health and reduces adipose tissue inflammation in high-fat fed mice. ISME J. (2017) 11 1667-1679. DOI: 10.1038/ismej.2017.24 28. Takahashi Y., Fukusato T.. Histopathology of nonalcoholic fatty liver disease/nonalcoholic steatohepatitis. World J. Gastroenterol. (2014) 20 15539-15548. DOI: 10.3748/wjg.v20.i42.15539 29. Ichimura-Shimizu M., Omagari K., Yamashita M., Tsuneyama K.. Development of a novel mouse model of diet-induced nonalcoholic steatohepatitis-related progressive bridging fibrosis. Biosci. Biotechnol. Biochem. (2021) 85 941-947. DOI: 10.1093/bbb/zbaa107 30. Tada Y., Kasai K., Makiuchi N., Igarashi N., Kani K., Takano S., Honda H., Yanagibashi T., Watanabe Y., Usui-Kawanishi F.. Roles of Macrophages in Advanced Liver Fibrosis, Identified Using a Newly Established Mouse Model of Diet-Induced Non-Alcoholic Steatohepatitis. Int. J. Mol. Sci. (2022) 23. DOI: 10.3390/ijms232113251 31. Begley M., Hill C., Gahan C.G.. Bile salt hydrolase activity in probiotics. Appl. Environ. Microbiol. (2006) 72 1729-1738. DOI: 10.1128/AEM.72.3.1729-1738.2006 32. Carpino G., Del Ben M., Pastori D., Carnevale R., Baratta F., Overi D., Francis H., Cardinale V., Onori P., Safarikia S.. Increased Liver Localization of Lipopolysaccharides in Human and Experimental NAFLD. Hepatology (2020) 72 470-485. DOI: 10.1002/hep.31056 33. Gabele E., Dostert K., Hofmann C., Wiest R., Scholmerich J., Hellerbrand C., Obermeier F.. DSS induced colitis increases portal LPS levels and enhances hepatic inflammation and fibrogenesis in experimental NASH. J. Hepatol. (2011) 55 1391-1399. DOI: 10.1016/j.jhep.2011.02.035 34. Harte A.L., da Silva N.F., Creely S.J., McGee K.C., Billyard T., Youssef-Elabd E.M., Tripathi G., Ashour E., Abdalla M.S., Sharada H.M.. Elevated endotoxin levels in non-alcoholic fatty liver disease. J. Inflamm. (2010) 7 15. DOI: 10.1186/1476-9255-7-15 35. Miele L., Valenza V., La Torre G., Montalto M., Cammarota G., Ricci R., Masciana R., Forgione A., Gabrieli M.L., Perotti G.. Increased intestinal permeability and tight junction alterations in nonalcoholic fatty liver disease. Hepatology (2009) 49 1877-1887. DOI: 10.1002/hep.22848 36. Plovier H., Everard A., Druart C., Depommier C., Van Hul M., Geurts L., Chilloux J., Ottman N., Duparc T., Lichtenstein L.. A purified membrane protein from Akkermansia muciniphila or the pasteurized bacterium improves metabolism in obese and diabetic mice. Nat. Med. (2017) 23 107-113. DOI: 10.1038/nm.4236 37. Liu H., Kohmoto O., Sakaguchi A., Hori S., Tochigi M., Tada K., Lee Y., Kikuchi K., Ishizuka S.. Taurocholic acid, a primary 12alpha-hydroxylated bile acid, induces leakiness in the distal small intestine in rats. Food Chem. Toxicol. (2022) 165 113136. DOI: 10.1016/j.fct.2022.113136 38. Lu T.T., Makishima M., Repa J.J., Schoonjans K., Kerr T.A., Auwerx J., Mangelsdorf D.J.. Molecular basis for feedback regulation of bile acid synthesis by nuclear receptors. Mol. Cell (2000) 6 507-515. DOI: 10.1016/S1097-2765(00)00050-2 39. Jiang C., Xie C., Li F., Zhang L., Nichols R.G., Krausz K.W., Cai J., Qi Y., Fang Z.Z., Takahashi S.. Intestinal farnesoid X receptor signaling promotes nonalcoholic fatty liver disease. J. Clin. Investig. (2015) 125 386-402. DOI: 10.1172/JCI76738 40. Pols T.W., Nomura M., Harach T., Lo Sasso G., Oosterveer M.H., Thomas C., Rizzo G., Gioiello A., Adorini L., Pellicciari R.. TGR5 activation inhibits atherosclerosis by reducing macrophage inflammation and lipid loading. Cell Metab. (2011) 14 747-757. DOI: 10.1016/j.cmet.2011.11.006 41. Keitel V., Donner M., Winandy S., Kubitz R., Haussinger D.. Expression and function of the bile acid receptor TGR5 in Kupffer cells. Biochem. Biophys. Res. Commun. (2008) 372 78-84. DOI: 10.1016/j.bbrc.2008.04.171 42. Perino A., Schoonjans K.. TGR5 and Immunometabolism: Insights from Physiology and Pharmacology. Trends Pharmacol. Sci. (2015) 36 847-857. DOI: 10.1016/j.tips.2015.08.002 43. Kawamata Y., Fujii R., Hosoya M., Harada M., Yoshida H., Miwa M., Fukusumi S., Habata Y., Itoh T., Shintani Y.. A G protein-coupled receptor responsive to bile acids. J. Biol. Chem. (2003) 278 9435-9440. DOI: 10.1074/jbc.M209706200 44. Henriksson E., Andersen B.. FGF19 and FGF21 for the Treatment of NASH-Two Sides of the Same Coin? Differential and Overlapping Effects of FGF19 and FGF21 From Mice to Human. Front. Endocrinol. (2020) 11 601349. DOI: 10.3389/fendo.2020.601349 45. Schneider C.A., Rasband W.S., Eliceiri K.W.. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods (2012) 9 671-675. DOI: 10.1038/nmeth.2089 46. Kleiner D.E., Brunt E.M., Van Natta M., Behling C., Contos M.J., Cummings O.W., Ferrell L.D., Liu Y.C., Torbenson M.S., Unalp-Arida A.. Design and validation of a histological scoring system for nonalcoholic fatty liver disease. Hepatology (2005) 41 1313-1321. DOI: 10.1002/hep.20701 47. Tolivia J., Navarro A., del Valle E., Perez C., Ordonez C., Martinez E.. Application of Photoshop and Scion Image analysis to quantification of signals in histochemistry, immunocytochemistry and hybridocytochemistry. Anal. Quant. Cytol. Histol. (2006) 28 43-53. PMID: 16566279 48. Ishibashi R., Furusawa Y., Honda H., Watanabe Y., Fujisaka S., Nishikawa M., Ikushiro S., Kurihara S., Tabuchi Y., Tobe K.. Isoliquiritigenin Attenuates Adipose Tissue Inflammation and Metabolic Syndrome by Modifying Gut Bacteria Composition in Mice. Mol. Nutr. Food. Res. (2022) 66 e2101119. DOI: 10.1002/mnfr.202101119 49. Watanabe S., Chen Z., Fujita K., Nishikawa M., Ueda H., Iguchi Y., Une M., Nishida T., Imura J.. Hyodeoxycholic Acid (HDCA) Prevents Development of Dextran Sulfate Sodium (DSS)-Induced Colitis in Mice: Possible Role of Synergism between DSS and HDCA in Increasing Fecal Bile Acid Levels. Biol. Pharm. Bull. (2022) 45 1503-1509. DOI: 10.1248/bpb.b22-00373
--- title: Healthy Food Prices Increased More Than the Prices of Unhealthy Options during the COVID-19 Pandemic and Concurrent Challenges to the Food System authors: - Meron Lewis - Lisa-Maree Herron - Mark D. Chatfield - Ru Chyi Tan - Alana Dale - Stephen Nash - Amanda J. Lee journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9967271 doi: 10.3390/ijerph20043146 license: CC BY 4.0 --- # Healthy Food Prices Increased More Than the Prices of Unhealthy Options during the COVID-19 Pandemic and Concurrent Challenges to the Food System ## Abstract Food prices have escalated due to impacts of the COVID-19 pandemic on global food systems, and other regional shocks and stressors including climate change and war. Few studies have applied a health lens to identify the most affected foods. This study aimed to assess costs and affordability of habitual (unhealthy) diets and recommended (healthy, equitable and more sustainable) diets and their components in Greater Brisbane, Queensland, Australia from 2019 to 2022 using the Healthy Diets Australian Standardised Affordability and Pricing protocol. Affordability was determined for reference households at three levels of income: median, minimum wage, and welfare-dependent. The recommended diet cost increased $17.9\%$; mostly in the last year when the prices of healthy foods, such as fruit, vegetables and legumes, healthy fats/oils, grains, and meats/alternatives, increased by $12.8\%$. In contrast, the cost of the unhealthy foods and drinks in the habitual diet ‘only’ increased $9.0\%$ from 2019 to 2022, and $7.0\%$ from 2021 to 2022. An exception was the cost of unhealthy take-away foods which increased by $14.7\%$ over 2019–2022. With government COVID-19-related payments, for the first time recommended diets were affordable for all and food security and diets improved in 2020. However, the special payments were withdrawn in 2021, and recommended diets became $11.5\%$ less affordable. Permanently increasing welfare support and providing an adequate minimum wage, while keeping basic, healthy foods GST-free and increasing GST to $20\%$ on unhealthy foods, would improve food security and diet-related health inequities. Development of a Consumer Price Index specifically for healthy food would help highlight health risks during economic downturns. ## 1. Introduction Poor diet is the leading single preventable risk factor contributing to the burden of disease in Australia and globally [1]. Less than $4\%$ of Australian adults consume a diet consistent with the recommendations of the Australian Dietary Guidelines (ADGs) [2], with over $35\%$ of dietary energy derived from unhealthy discretionary food and drinks [3]. Relatedly, more than two thirds ($67\%$) of Australian adults and $25\%$ of children aged two to 17 years are overweight or obese [4]. In Australia, rates of obesity, poor-quality diet, and diet-related chronic disease (including type 2 diabetes, heart disease, and some cancers [1,5]) follow a socioeconomic gradient [6,7,8]. Ensuring food security is key to reducing prevalence of obesity and chronic disease particularly in low socioeconomic groups. Food security is a fundamental human right [9]. It means everyone is able to obtain at all times a sufficient quantity of quality (safe and nutritious) food that meets their preferences to sustain “an active and healthy life” [10]. Food security is determined by availability, accessibility, affordability, and acceptability of food. The affordability of healthy diets is impacted by both the cost of food and drinks and the household financial resources available. Economic access to food implies that people “have sufficient money to purchase the food they want to eat, to meet cultural and social as well as health and nutritional norms; that this money is not absorbed in other expenditure demands (rent, fuel, debt repayment, etc.); [ and] that people can … obtain food in ways which are dignified and in keeping with social norms” [11]. Income level, income shocks and rising costs of living are key determinants of economic access to healthy food [12,13,14]. As with all diets, recommended diets (which are healthy, equitable, and sustainable consistent with the Australian Dietary Guidelines (ADGs) [2]) are considered unaffordable when they cost more than $30\%$ of household income [15]. When a household needs to spend more than $25\%$ of their disposable income on healthy food, they may experience “food stress” and are vulnerable to food insecurity [16]. Before the COVID-19 pandemic, families relying on government welfare/financial assistance had to spend over a third of their household income to buy a recommended diet, and households in rural and remote areas on low incomes needed to spend an even greater proportion [17]. Prior to COVID-19, estimates of the prevalence of food insecurity in Australia were around 4 to $14\%$ in the general population and up to $82\%$ in low-income groups [18]. Over the past three years, food insecurity has increased significantly [19,20]. According to the latest national Hunger Report from Australian food charity Foodbank, extrapolating responses from a nationally representative survey conducted in July 2022, $21\%$ of Australians had experienced severe food insecurity in the past 12 months (up from $17\%$ in 2021) [19]. Of those experiencing food insecurity, $64\%$ cited increased or high living expenses and $42\%$ reported “reduced or low income or government benefits” as a key cause [19]. Previous studies have assessed diet costs during the COVID-19 pandemic and, opportunistically, the impact on diet affordability of government economic responses that boosted incomes for many low-income households [17,21]. In response to COVID-19 impacts, including increased unemployment, in 2020 the Australian Government introduced “JobKeeper” payments to help businesses pay employees who were stood down, and also lump sum Economic Support Payments and a fortnightly Coronavirus Supplement (hereafter abbreviated as ESP and CS) for eligible recipients of some income support payments, including the unemployment benefit “JobSeeker” [22]. Diet costs increased from 2019 to 2020, largely driven by rising prices of most healthy food groups [21]. However, increased income support meant that for the first time, welfare-dependent families had economic access to recommended diets [17,21]. Food prices have continued to escalate since 2020 due to global, national, and regional shocks and stressors, including the impacts of climate change, increasing frequency and severity of extreme weather events (bushfires and floods), the Russian invasion of Ukraine, and changing demographics (such as reduced immigration and increased internal migration to rural areas [23]), as well as disruptions to food production and supply due to impacts on workforces of COVID-19-related public health restrictions [24,25,26,27]. These factors also contributed to increased costs of fuel, feed, and fertiliser, exacerbating increasing food prices. This study aimed to assess costs of habitual (unhealthy) diets and recommended diets in Greater Brisbane in 2021 and 2022, and compared results with those reported from earlier studies in 2019 and 2020 [21], to explore changing costs and affordability in the context of these shocks and stressors. ## 2. Materials and Methods The Healthy Diets ASAP (Australian Standardised Affordability and Pricing) methods protocol [28] was applied to assess the cost, cost differential, and affordability of habitual (unhealthy) and recommended (healthy, more equitable and sustainable) diets in the Greater Brisbane region of Queensland, Australia annually from 2019 to 2022. Findings from 2019 and 2020 have been previously reported [17,21]. This study collected and analysed data in 2021 and 2022 and compared diet cost, cost differentials, and affordability across the four timepoints. All cost values are provided in Australian dollars ($). The Healthy Diets ASAP protocol is consistent with the International Network for Food and Obesity/non-communicable diseases Research, Monitoring and Action Support (INFORMAS) framework’s ‘optimal’ approach to assess diet price and affordability [29], addressing limitations of earlier efforts to measure food cost and affordability in Australia [30,31]. Details of the background, description, collaborative development process, application, and testing of the protocol have been published previously [28,32]. The protocol has five parts: standardised habitual and recommended diet pricing tools; store location and sampling; calculation of median gross and minimum wage disposable income; food price data collection; and analysis and reporting [28]. ## 2.1. Diet Pricing Tools The diet pricing tools specify the types of foods and drinks in the habitual and recommended diets and quantities for a reference household of four (adult male and female 31–50 years of age, a 14-year-old boy, and an 8-year-old girl) per fortnight [28]. The recommended diet contains healthy food and drinks, in line with the recommendations of the ADGs [2]. The habitual diet is based on reported dietary intake data from the most recent Australian Health Survey National Nutrition and Physical Activity Survey (NNPAS) 2011–2013 [3]. It includes some healthy food and drinks in lower amounts than recommended in the ADGs, and many discretionary foods and drinks (defined by the ADGs as not being necessary for health and high in saturated fat, added sugar, sodium and/or alcohol) [2]. The recommended diet contains slightly less energy (33,610 kJ/day) for the reference household than the habitual diet (33,869 kJ/day), and is more sustainable, requiring less water, protecting biodiversity, and generating $25\%$ lower greenhouse gas emissions in its production [33]. Table 1 lists the types of food and drinks included in each diet pricing tool. Detailed lists of components and quantities have been published previously [28]. ## 2.2. Store Locations and Sampling Food and drink price data for 2019 were sourced from a previous survey using the Healthy Diets ASAP protocol to assess diet costs and affordability in locations throughout Queensland [17,21]. The Australian Bureau of Statistics’ (ABS) Statistical Geography Standard classes medium-sized geographical areas into SA2 locations, where communities “interact together socially and economically” [34]. In 2019, SA2 locations across Queensland were stratified into quintiles of socioeconomic disadvantage based on the Socioeconomic Indexes for Areas (SEIFA) Index of Relative Socioeconomic Disadvantage [35]. Eighteen locations in SEIFA quintiles 1 (most disadvantaged), 3 (median disadvantaged) and 5 (least disadvantaged) were randomly selected for inclusion; the final samples included 10 locations in Greater Brisbane (3, 4, and 3 locations in SEIFA quintiles 1, 3, and 5, respectively). In 2020, restriction of movement implemented as a public health measure in response to the COVID-19 pandemic meant data collection in stores beyond Greater Brisbane was not possible. Hence, only the 10 SA2 locations in Greater Brisbane included in the 2019 sample were re-surveyed for food prices in 2020 [21]. Additionally, because of the pandemic, one of the large supermarket chains did not allow ‘unnecessary’ store visits (e.g., for research purposes) so prices from that supermarket were collected from its website matched to the SA2 locations. Previous studies comparing in-store to online prices have found insignificant price differences [36]. In similar months in 2021 and 2022, food prices were collected in the same 10 locations in Greater Brisbane. At each location, two large supermarkets (one of each major supermarket chain), an independent grocery store, a bakery, a fish and chip shop, two fast food restaurants, and one alcohol outlet were surveyed. As per the Healthy Diets ASAP protocol [28], if a store had closed since the previous survey, a similar, proximate food outlet was surveyed instead. In 2021 and 2022, prices were collected online for the two supermarkets in all locations and collected in-store for the remaining outlets. In 2022, a duplicate set of prices were collected in-store from the large supermarkets in two locations for validation. ## 2.3. Price Data Collection Price data were collected by trained research assistants, following the Healthy Diets ASAP protocol, between August and October each year. Collection of price data in 2019 and 2020 has been detailed elsewhere [17,21]. In 2021, food price data were collected by L.-M.H. and R.C.T. and in 2022 by L.-M.H., M.L., A.D., and S.N. Permission to collect data was requested and received from national head offices of large supermarket chains, and also from store managers in each outlet. The data collection protocol outlines the procedure followed if the stipulated brands and sizes were not available or were on price promotion [28]. ## 2.4. Household Income Calculation Household incomes were calculated as per the Healthy Diets ASAP protocol [28], using publicly available national data from government agencies. This study assessed diet affordability for three categories of income for the reference household. For each timepoint, the median gross household income (before taxation) per fortnight in each SA2 area was sourced from the ABS 2016 Census Community Profile [37] and adjusted by the ABS Wage Price Index [38]. Minimum wage disposable and welfare-dependent household incomes were calculated based on the set of assumptions detailed in the protocol, using payment entitlement data from Services Australia [39]. Calculations of minimum wage and welfare-dependent household incomes for 2020 included the ESP and CS provided between May and September 2020, are detailed elsewhere [21]. ## 2.5. Analysis and Reporting Food and drink price data were entered into the Healthy Diets ASAP data collection web portal [40] by R.C.T., A.D. and S.N. Data were double entered, and any discrepancies were resolved by consensus. Data were cleaned and checked by M.L. and A.J.L. As per the Healthy Diets ASAP protocol, if a value was missing, the mean price of the item in other stores in the same SA2 location was substituted. Spreadsheet algorithms generated results for each location in Microsoft Office Excel files which were cross-checked by M.L. and A.J.L. Diet costs and affordability were calculated for each SA2 area surveyed in Greater Brisbane. The mean costs of the habitual and recommended diets, and the cost and proportion of the total spent on different ADG food groups and components, were calculated for the reference household per fortnight. Results were reported for SA2 SEIFA quintile 1, 3, and 5, and for Greater Brisbane as a whole. Affordability of habitual and recommended diets was calculated for households with the three different income levels described above. The results for each year were compared to relevant findings of previous surveys, to assess changes in diet costs, cost of ADG food groups and components, and affordability of the diets. Consumer Price Index of food and non-alcoholic beverages (CPI-food) data for Brisbane from 2019 to 2022 were sourced from the ABS [41] for comparison with observed changes in food prices. Statistical analysis was conducted by paired t-tests; statistical significance was set at p ≤ 0.05. ## 3.1. Selected Locations and Stores Surveyed In 2019, food prices were collected from 80 outlets in 10 locations in Greater Brisbane [17]; in 2020, price data were collected in-store from 68 outlets and online from 10 outlets (supermarkets) [21]; in 2021 prices were collected from 80 outlets: in-store for 60 and online for 20 (supermarkets); and in 2022, prices were collected from 80 outlets: in-store for 60 and online for 20 (supermarkets). For validation, prices were re-collected in-store for four supermarket outlets across two locations. ## 3.2. Diet Cost Data The mean costs ± standard error (SE) of the habitual and recommended diets and diet components in Greater Brisbane from 2019 to 2022 are reported in Table 2. Total costs of the recommended diets and the healthy and discretionary (unhealthy) components of the habitual diets for the four timepoints are also presented in Figure 1. Diet cost data for each location surveyed (Supplementary Figures S1 and S2) and by SEIFA quintile (Supplementary Table S1) are presented in supplementary files. Cumulative increase in the cost of the habitual and recommended diets and CPI-food is shown in Figure 2. Within the manuscript, data are presented and analysed for Greater Brisbane as a whole, as change in food prices was similar regardless of SEIFA quintiles (Supplementary Table S1). The duplicated diet costs for prices collected in-store were within $1\%$ of diet costs for prices collected online from the large supermarkets in two locations. ## 3.3. Diet Cost in Greater Brisbane, 2022 In 2022, the mean cost of the recommended diet in Greater Brisbane was AUD 729.71 per fortnight for the reference household of two adults and two children (Table 2). The habitual diet was $18.4\%$ more expensive than the recommended (healthy) diet at a cost of AUD 863.93 (Table 2). Shifting from a habitual to recommended diet would save households AUD 134.22 per fortnight on average. In 2022, nearly $58\%$ of the total cost of the habitual diet was required to purchase discretionary food and drinks, including takeaway foods (around $20\%$ of total diet cost), alcoholic drinks ($11\%$ of total diet cost), and sugar sweetened drinks (around $4\%$ of total diet cost) (Table 2). ## 3.4. Changes in Fortnightly Diet Costs over Time Between 2019 and 2022, in Greater Brisbane the cost of the recommended diet increased by $17.9\%$ from AUD 619.04 to AUD 729.71 ($p \leq 0.001$) (Table 2). Around three-quarters of this increase occurred in the 12 months prior to the most recent survey in 2022, during which time the cost of healthy foods increased $12.8\%$ ($p \leq 0.001$) (Table 2, Figure 1). Over the four timepoints, the cost of the habitual diet increased $11.9\%$ from AUD 772.20 to AUD 863.83 ($p \leq 0.001$). Again, most of the rise occurred between 2021 and 2022, during which period the total cost of the habitual diet increased by $7.8\%$ ($p \leq 0.001$). This was around half the rate of increase for the recommended diet; the main reason for this was that the cost of the discretionary (unhealthy) food and drinks in the habitual diet only increased by $7.0\%$ ($p \leq 0.001$) during that year (Table 2; Figure 1). The CPI-food for the period September quarter 2021 to September quarter 2022 in Brisbane was $8.6\%$ [41], so the increase in the cost of the recommended diet ($12.8\%$) was $49\%$ higher than the relevant reported CPI-food and even higher ($64\%$) than the increase in cost of the habitual diet ($7.8\%$), as illustrated in Figure 2. For the two-year period prior to September 2021, the CPI-food for Brisbane increased by $3.8\%$. During this period, the assessed cost increases of the recommended diet and the habitual diet were much closer to CPI-food; however, the cost increases in the recommended diet tended to be higher than those in the habitual (unhealthy) diet. This differential was exacerbated in 2022 (Figure 2). The differential between the cost of the habitual and recommended diets was $24.7\%$ in 2019 and was similar for the following two years, being $23.9\%$ in 2020 ($$p \leq 0.90$$) and $23.8\%$ in 2021 ($$p \leq 0.98$$). However, the cost differential declined significantly to $18.4\%$ from 2021 to 2022 ($p \leq 0.001$). In 2022, the recommended diet was relatively more expensive than it had been compared to the habitual diet in 2019 ($$p \leq 0.006$$). ## 3.5. Changes in Fortnightly Cost of Food Groups over Time In the recommended diet, food groups with the highest cost increases from 2019 to 2022 and from 2021 to 2022, respectively, were: vegetables and legumes ($20.5\%$ and $34.9\%$, both $p \leq 0.001$), grain (cereal—mostly wholegrain) foods ($15.6\%$ and $11.4\%$, both $p \leq 0.001$), fruit ($23.4\%$, $p \leq 0.001$ and $8.2\%$, $$p \leq 0.010$$), and healthy fats and oils ($27.3\%$, $17.5\%$, both $p \leq 0.001$) (Table 2). Notably the cost of ‘vegetables and legumes’ had decreased from 2019 to 2020 (−$12.7\%$, $p \leq 0.001$) before increasing from 2020 to 2021 ($4.8\%$, $p \leq 0.001$) and then increased markedly from 2021 to 2022 ($34.9\%$, $p \leq 0.001$), as noted previously. Costs of products in the milk, cheese, and yoghurt group increased initially from 2019 to 2020 ($7.4\%$, $$p \leq 0.012$$), before declining slightly from 2020 to 2021 (−$3.1\%$, $$p \leq 0.018$$), then rising most significantly from 2021 to 2022 ($10.3\%$, $p \leq 0.001$); an overall increase of $14.7\%$ ($p \leq 0.001$) from 2019 to 2022. The cost increases in the lean meats, poultry, fish, eggs, and plant-based alternatives group were more consistent throughout the three years, increasing by $6.1\%$ from 2019 to 2020 ($$p \leq 0.011$$), $3.5\%$ from 2020 to 2021 ($$p \leq 0.013$$), and $6.4\%$ from 2021 to 2022 ($p \leq 0.001$), leading to an overall increase of $16.9\%$ ($p \leq 0.001$) from 2019 to 2022. From 2019 to 2022, in contrast to the marked increase in the cost of the healthy foods and drinks in the recommended diet ($17.9\%$), the cost of all the discretionary (unhealthy) foods and drinks in the habitual diet increased by ‘only’ $9.0\%$ (Table 2). There was also a large variation in the price changes within individual components of the latter. For example, the price of alcoholic drinks was relatively stable from 2019 to 2020, decreased in the first years of the COVID-19 pandemic (−$6.4\%$, $p \leq 0.001$), then increased by $3.2\%$ from 2021 to 2022 ($p \leq 0.001$), leading to an overall decrease of −$1.8\%$ ($p \leq 0.001$) from 2019 to 2022. The cost of sugar sweetened beverages followed a similar pattern, increasing by $12.3\%$ ($p \leq 0.001$) from 2021–2022, while the cost of artificially sweetened beverages increased throughout the four timepoints by $21.5\%$ ($p \leq 0.001$) (Table 2). Among discretionary (unhealthy) items, the highest price increases from 2019 to 2022 were in take-away foods ($14.7\%$, $p \leq 0.001$); while increasing by $5.7\%$ ($p \leq 0.001$) from 2019 to 2020, the cost of takeaway foods did not change significantly in the first year of the COVID-19 pandemic, then increased from 2021 to 2022 ($6.0\%$, $p \leq 0.001$). ## 3.6. Changes in Diet Affordability over Time Table 3 presents household incomes at the three different levels for the reference household per fortnight and affordability of the recommended diet in Greater Brisbane from 2019 to 2022. Affordability of the recommended diet over time is also presented in Figure 3. Calculations of the minimum wage disposable household incomes and welfare-dependent household incomes at the four timepoints are provided in Supplementary Tables S2 and S3. In 2019, median gross household income per fortnight (AUD 3188.00) was $35\%$ higher than household income for those on minimum wage (AUD 2358.00) and $83\%$ higher than households on welfare income (AUD 1739.68). Due to the ESP and CS payments, both the minimum wage disposable household income and welfare-dependent household income in Greater Brisbane increased markedly between May and September 2020, by $41.5\%$ and $77.3\%$, respectively. After the ESP and CS payments ceased early in 2021, both incomes returned to levels only $17\%$ and $11\%$ higher, respectively, than in 2019 (Table 3). In 2019, the reference household on median household income would have needed to spend $20.8\%$ of their income to purchase the recommended diet in Greater Brisbane, which was affordable. Those households on minimum wage had to pay $26.0\%$ of their household income to purchase the recommended diet, so would have been in food stress. However, welfare-dependent households could not afford the recommended diet, which cost $35.6\%$ of household income (Table 3, Figure 3). In contrast, for the first time, due to the ESP and CS introduced in 2020, the recommended diet was affordable for the reference household on welfare income, costing $20.3\%$ of household income (Table 3, Figure 3). In addition, in 2020 the household on minimum wage was no longer in food stress, with the recommended diet costing $19.3\%$ of household income. Hence, affordability of the recommended diet and economic access to healthy diets was similar for vulnerable households and those on median income in 2020. However, the ESP and CS were withdrawn incrementally from late 2020 and removed in early 2021. In 2021, welfare-dependent households again could not afford recommended diets, which cost $34.6\%$ of household income. Those on minimum wage fared better at this time, narrowly avoiding food stress as the recommended diet cost $24.1\%$ of their household income. However, diet affordability continued to worsen in all households with the increasing cost of food from 2021 to 2022. While approaching the food stress threshold, those on median income could still afford the recommended diet which cost $23.2\%$ of their household income in 2022. Those on minimum wage were again experiencing food stress with the recommended diet costing $26.4\%$ of their household income. Most worryingly, recommended diets were once again unaffordable for welfare-dependent households, costing $37.7\%$ of their household income. Since 2019, economic access to healthy diets had reduced by $11.5\%$ for households on median income, $1.5\%$ for those on minimum income, and $5.0\%$ for those on welfare. ## 4. Discussion The Healthy Diets ASAP protocol was applied to assess the cost, relative cost, and affordability of habitual (unhealthy) diets and recommended (healthy, equitable and sustainable) diets, for a reference family of two adults and two children at three different household income levels in Greater Brisbane once a year from 2019 to 2022. This enabled comparison of diet costs and affordability at timepoints before and during the COVID-19 pandemic and concurrent shocks and stresses at global, national, and regional levels that may have impacted economic aspects of food security in Greater Brisbane. The study found that from 2019 to 2021 prices of food and drinks in greater Brisbane increased gradually consistent with the CPI-food of $3.8\%$ but escalated markedly between 2021 and 2022 (Figure 2). Between September 2021 and 2022, the cost of the healthy food and drinks comprising the recommended diet increased by an average of $12.8\%$, nearly double the rate of the increase in the cost of the discretionary foods and drinks in the habitual diet during the same period ($7.0\%$), and $42\%$ more than the Brisbane CPI-food in that period ($8.6\%$). A major contributor to the increased cost of a healthy diet from 2021 to 2022 was the increase in fruit, vegetable, and grain prices that has been attributed to heavy rainfall and flooding in key Queensland food production areas during that year [42]. Other previous natural disasters, including the wild bushfires of 2019–2020, were likely to have ongoing impacts on supply [21,43]. In October 2022, the Australian Government Treasury predicted fruit and vegetables prices would increase a further $8\%$ in the six months following due to new flood events in Australian food growing regions [42]. Continuing COVID-19-related supply chain disruptions were also reported. These included reduced workforce, such as reduced availability of international seasonal fruit pickers due to Australian border closures and forced isolation of workers in manufacturing, distributing, transport, and retail, including those with COVID-19 and close contacts [44,45]. Other factors included high fuel, and hence high transport costs, and high fertiliser and feed costs due to the Russian invasion of Ukraine [25,27]. The war has also been linked with increased cost of cereal products/bread due to constrained global wheat supply and increased cost of cooking oil [26,46], which grew by $17.5\%$ in the last year of this study. Another key factor was increasing global and national inflation [25,41]. In Australia, this exacerbated the increased price of gas and electricity, which are used for cooking, and of fuel used for shopping [41]. The escalating cost of living also pressured essential spending, such as on rent and mortgages for housing, which impacted spending on food [41]. As cash rates rise to combat inflation, this can lead to increased interest payments on borrowing, also increasing pressure on household budgets [41]. As prices of healthy foods escalate, unhealthy (discretionary and/or ultra-processed) foods have become relatively cheaper, which influences dietary choices, particularly in an economic downturn [47,48]. In times of financial stress, families, particularly in low socio-economic groups, tend to purchase the cheapest and most affordable food; this is often the less healthy products on price promotion [47,48]. This is worrying given that, when last measured nationally in Australia (2011–2012), more than one third of adults’ energy intake, and nearly $40\%$ of children’s, was derived from unhealthy, discretionary food [49]. Total intake of discretionary food and drinks is associated with increased body mass index, and lower consumption of fruit and vegetables [50]. The relatively greater increase in the price of healthy, compared to discretionary, food and drinks during the study likely contributed to reported reductions in the intake of fresh produce, increased intake of unhealthy foods and weight gain during the COVID-19 pandemic [51,52]. Price is just one of many factors influencing dietary choices. Other aspects of the ‘obesogenic’ food environment [53] that drive food choices include the constant availability and promotion of unhealthy food and drinks, and convenience. These factors may explain why the cost of take-away foods increased more than other unhealthy foods and drinks during this study (Table 2). Better promotion of healthy food and drinks and improvement in the nutrient profile of some convenience foods (e.g., reduction of salt content [54]) may contribute to healthier dietary intakes. A common perception that healthy foods are more expensive than unhealthy options is also relevant [2,55,56,57,58]. However, the studies supporting this are not grounded in reported dietary intakes such as the Healthy Diets ASAP protocol [28,29,32]. For example, they often cost arbitrary ‘healthy’ and ‘unhealthy’ food lists and often exclude alcohol and take-away foods, which comprise 20–$25\%$ of the cost of habitual Australian diets [28]. Results are also reported in different units (price per energy or weight unit, serve or nutrient-density) and analysis is frequently spurious [59]. Corresponding with findings of previous research, this study confirmed that healthy diets can be less expensive than habitual diets. Studies applying the Healthy Diets ASAP protocol have found that habitual diets are $14\%$ to $23\%$ more expensive than the recommended diet in Brisbane [28,32], across regional and remote areas of Queensland [17], in Sydney and Canberra [60], in regional Victoria [61], in remote Aboriginal and Torres Strait Islander communities [62,63], and nationally in areas serviced by the two large supermarket chains [64]. However, the cost differential between the diets reduced in 2022, reflecting the differing pricing and relative composition of the habitual and recommended diets. As the recommended diet comprises only healthy food and drinks, its total cost increased at a higher rate than the total cost of the habitual diet, which includes lesser quantities of healthy foods but many unhealthy items. A similar differential decrease was identified also in remote Aboriginal communities in Central Australia; in these communities between May 2021 and June 2022, the cost of the habitual diet increased approximately $5\%$, while the cost of the recommended diet increased by around $10\%$ [65]. Despite the reduced cost differential ($18.4\%$ in 2022 down from $24.7\%$ in 2019), the cost saving for households buying the recommended diet rather than the habitual diet remained significant in 2022—AUD 134.22 per fortnight for the reference household in Greater Brisbane. However, this was approximately AUD 20 per fortnight less than during the preceding years. Especially at a time of increased pressure on household budgets and financial stress [66], this likely reduced the financial incentive for households to improve their diet. Urgent action is needed to further increase the relative affordability of recommended diets compared to habitual diets. The lower cost of recommended diets compared to habitual diets is largely attributable to the exemption of “basic, healthy foods” from the $10\%$ Goods and Services Tax (GST) in Australia, which helps to keep the relative price of healthy food and drinks down. Modelling studies support increasing the GST on unhealthy food and drinks to $20\%$ to increase the incentive for consumers to choose healthier options [32,67]. Targeted levies, such as $30\%$ GST on sugar-sweetened beverages also have been suggested to discourage intake of specific choices harmful to health [68]. Analysis of data collected annually from 2019 to 2022 highlighted the impact on affordability of recommended diets during the temporary increases to income support for the most vulnerable households during the early months of the COVID-19 pandemic. The ‘natural experiment’ created when the Australian Government provided the ESP and CS, effectively doubling welfare income for those on JobSeeker in 2020, demonstrated that increasing household income raised an estimated 646,000 people above the poverty line [69] and made healthy diets affordable for welfare-dependent families for the first time [21]. Affordability of the recommended diet improved dramatically for low-income households in 2020, due to government provision of the ESP and CS from April. Affordability of recommended diets improved by $42\%$ for welfare-dependent households. More than $90\%$ of recipients of the ESP and CS income supplements reported that they could afford more healthy foods, including fruit and vegetables [70]. Findings of other studies and surveys confirm a positive impact of increased income support on food security and households’ capacity to buy healthy food [19,71] and subsequent increases in the prevalence of poverty and food insecurity after the income supplements decreased and then ceased (March 2021) [19,72]. Recommended diets became unaffordable for welfare-dependent households again, requiring $34.6\%$ of their disposable income in 2021, and $37.7\%$ in 2022. Families unable to commit such a large proportion of their disposable income to food because of other escalating living costs (particularly rent, household utilities such as electricity, and fuel/transport) report not being able to buy enough food or skipping meals to save money [73]. Low household income is the most consistent determinant of food insecurity [74]. The results of this study have highlighted that CPI-food is a blunt economic instrument. To help more usefully and transparently identify the likely health consequences of the changing prices of foods, it is recommended that the CPI-healthy food be determined and reported. A potential model for this was the ABS Australian Dietary Guidelines CPI reported in 2015 [75]. ## Limitations For this study it would have been ideal if COVID-19-related ‘JobKeeper’ payments made to businesses to support staff whose working hours were reduced during the pandemic could have been included in addition to ‘JobSeeker’ payments. However, it was too difficult to tease out ‘JobKeeper’ entitlements and payments at the household level; hence, the incomes calculated only reflect those made primarily to low-income households. It should be noted that median income may have varied for some employees ‘stood down’ during the early years of the COVID-19 pandemic. However, this would not have affected the findings relevant to households on minimum wage or dependent on welfare. There are inherent methodological limitations to the Healthy Diets ASAP protocol that have been reported elsewhere [28]. Most pertinent to this analysis, they include assumptions used in income calculations that did not include the possibility of low-income households receiving the JobKeeper supplement in 2020, for example. The most recent available national dietary intake data were collected in Australia in 2011–2012 [3]. It is likely ‘habitual’ diets of Australian families have changed in line with changes in the food supply and environment in recent years, including the rapid growth of meal delivery services [76,77], but recent granular data on diet patterns are not available currently. It is anticipated that updated national dietary intake data will be available in 2025 from the planned 2023 Australian Intergenerational Health and Mental Health Study (incorporating the National Nutrition and Physical Activity Survey) [78]. For the first time, national food security data will also be assessed robustly in this study [78]. The price collection protocol of Healthy Diets ASAP includes collection of the prices of major Australian brands for packaged food and drinks. Some households, particularly low socioeconomic households, may choose to purchase lower priced generic products (‘home’ or ‘own’ brands) and/or shop at ‘budget’ supermarkets as a coping strategy to stretch food budgets [79]. However, when applying this strategy, recommended diets can still be stressful to afford for welfare-dependent households [79], and are likely to be more so given the identified price rises of unpackaged healthy foods (such as fruit and vegetables) in this study. Surveys were conducted only in urban locations in one state capital city in Australia, hence observed changes in food costs are not generalisable directly to other locations nationally, or other countries internationally. However, particularly for low-income groups, they do highlight current challenges in maintaining economic access to food security and healthy diets and help explain diet-related health inequities seen in Australia [79,80]. ## 5. Conclusions Having the ability to adequately feed ourselves is a basic human right; this requires that food must be affordable without compromising any other essential needs, such as secure housing (rent) or medicines [81]. Over the three years since 2019, the cost of food and drinks has escalated, particularly in the most recent year, with the prices of healthy foods and drinks increasing at almost double the rate of increase in prices of unhealthy items. The recommended diet remains less expensive than the habitual diet, but the cost differential reduced significantly between 2021 and 2022, rendering recommended diets relatively less affordable than habitual (unhealthy) diets. The development and publication of a CPI-healthy food index would help better identify health and related economic risks. Findings highlight the potential impact of government policy settings on the affordability of food and economic aspects of food security, and hence dietary choices and diet-related health such as obesity, cardiovascular disease, type 2 diabetes, and some cancers. Despite effective relief conferred by the ESP and CS while paid, recommended diets have become increasingly unaffordable for Australian families on low income, and are less affordable now than before the COVID-19 pandemic. Affordability of healthy food could be guaranteed by government commitments to help families access essential needs, and also by commitments to keep basic, healthy food and drinks GST-free. Additional benefits would be delivered if GST on unhealthy foods and drinks was increased to $20\%$. This could help encourage healthier choices, and the revenue raised could be hypothecated to nutrition and health promotion programs. Permanently increasing welfare support and providing an adequate minimum wage would enable low-income households to meet the costs of living and afford adequate healthy food, to protect their food security and diet-related health. ## References 1. Global burden of 87 risk factors in 204 countries and territories, 1990–2019: A systematic analysis for the Global Burden of Disease Study 2019. Lancet (2020.0) 396 1223-1249. DOI: 10.1016/S0140-6736(20)30752-2 2. 2. National Health and Medical Research Council Australian Dietary Guidelines—Providing the Scientific Evidence for Healthier Australian DietsNational Health and Medical Research CouncilCanberra, Australia2013. Australian Dietary Guidelines—Providing the Scientific Evidence for Healthier Australian Diets (2013.0) 3. 3. Australian Bureau of Statistics Australian Health Survey, National Health Survey, 2011–2012Australian Bureau of StatisticsCanberra, Australia2013. Australian Health Survey, National Health Survey, 2011–2012 (2013.0) 4. Overweight and Obesity 2017–18. (2022.0) 5. 5. Australian Institute of Health and Welfare Australian Burden of Disease Study: Impact and Causes of Illness and Death in Australia 2018AIHWCanberra, Australia2021. Australian Burden of Disease Study: Impact and Causes of Illness and Death in Australia 2018 (2021.0) 6. Harris B., Fetherston H., Calder R.. Australia’s Health Tracker by Socio-Economic Status 2017 (2017.0) 7. Backholer K., Spencer E., Gearon E., Magliano D.J., McNaughton S.A., Shaw J.E., Peeters A.. The association between socio-economic position and diet quality in Australian adults. Public Health Nutr. (2016.0) 19 477-485. DOI: 10.1017/S1368980015001470 8. Lewis M., Lee A.J.. Dietary inequity? A systematic scoping review of dietary intake in low socio-economic groups compared with high socio-economic groups in Australia. Public Health Nutr. (2021.0) 24 393-411. DOI: 10.1017/S1368980020003006 9. 9. The Committee on Economic Social and Cultural Rights General Comment No. 12 on the Right to Adequate FoodUnited NationsGeneva, Switzerland1999. General Comment No. 12 on the Right to Adequate Food (1999.0) 10. 10. Food and Agriculture Organization of the United Nations (FAO) Rome Declaration on World Food Security and World Food Summit Plan of Action, 1996FAORome, Italy1996. Rome Declaration on World Food Security and World Food Summit Plan of Action, 1996 (1996.0) 11. Dowler E.A., O’Connor D.. Rights-based approaches to addressing food poverty and food insecurity in Ireland and UK. Soc. Sci. Med. (2012.0) 74 44-51. DOI: 10.1016/j.socscimed.2011.08.036 12. Temple J.B.. The Association between Stressful Events and Food Insecurity: Cross-Sectional Evidence from Australia. Int. J. Environ. Res. Public Health (2018.0) 15. DOI: 10.3390/ijerph15112333 13. Kleve S., Booth S., Davidson Z.E., Palermo C.. Walking the Food Security Tightrope—Exploring the Experiences of Low-to-Middle Income Melbourne Households. Int. J. Environ. Res. Public Health (2018.0) 15. DOI: 10.3390/ijerph15102206 14. Temple J.B., Booth S., Pollard C.M.. Social Assistance Payments and Food Insecurity in Australia: Evidence from the Household Expenditure Survey. Int. J. Environ. Res. Public Health (2019.0) 16. DOI: 10.3390/ijerph16030455 15. Burns C., Friel S.. It’s time to determine the cost of a healthy diet in Australia. Aust. N. Z. J. Public Health (2007.0) 31 363-365. DOI: 10.1111/j.1753-6405.2007.00088.x 16. Landrigan T.J., Kerr D.A., Dhaliwal S.S., Pollard C.M.. Protocol for the Development of a Food Stress Index to Identify Households Most at Risk of Food Insecurity in Western Australia. Int. J. Environ. Res. Public Health (2018.0) 16. DOI: 10.3390/ijerph16010079 17. Lee A., Patay D., Herron L.-M., Harrison E.P., Lewis M.. Affordability of current, and healthy, more equitable, sustainable diets by area of socioeconomic disadvantage and remoteness in Queensland: Insights into food choice. Int. J. Equity Health (2021.0) 20 153. DOI: 10.1186/s12939-021-01481-8 18. Louie S., Shi Y., Allman-Farinelli M.. The effects of the COVID-19 pandemic on food security in Australia: A scoping review. Nutr. Diet. (2022.0) 79 28-47. DOI: 10.1111/1747-0080.12720 19. 19. Foodbank Australia Foodbank Hunger Report 2022Foodbank Australia LtdNorth Ryde, Australia2022. Foodbank Hunger Report 2022 (2022.0) 20. Kent K., Murray S., Penrose B., Auckland S., Visentin D., Godrich S., Lester E.. Prevalence and Socio-Demographic Predictors of Food Insecurity in Australia during the COVID-19 Pandemic. Nutrients (2020.0) 12. DOI: 10.3390/nu12092682 21. Lee A.J., Patay D., Herron L.-M., Tan R.C., Nicoll E., Fredericks B., Lewis M.. Affordability of Heathy, Equitable and More Sustainable Diets in Low-Income Households in Brisbane before and during the COVID-19 Pandemic. Nutrients (2021.0) 13. DOI: 10.3390/nu13124386 22. Klapdor M.. COVID-19 Economic Response—Social Security Measures Part 1: Temporary Supplement and Improved Access to Income Support. (2020.0) 23. Robertson G.. Australia’s Population: Recent Changes. Parliamentary Library Briefing Book: Key Issues for the 47th Parliament (2022.0) 24. Malik A., Li M., Lenzen M., Fry J., Liyanapathirana N., Beyer K., Boylan S., Lee A., Raubenheimer D., Geschke A.. Impacts of climate change and extreme weather on food supply chains cascade across sectors and regions in Australia. Nat. Food (2022.0) 3 631-643. DOI: 10.1038/s43016-022-00570-3 25. 25. United Nations Global Crisis Response Group Brief No. 2: Global Impact of War in Ukraine: Billions of People Face the Greaest Cost-of-Living Crisis in a GenerationUnited NationsGeneva, Switzerland2022. Brief No. 2: Global Impact of War in Ukraine: Billions of People Face the Greaest Cost-of-Living Crisis in a Generation (2022.0) 26. Glauber J., Hernandez M., Laborde D., Martin W., Rice B., Vos R.. No End in Sight Yet for the Global Food Price Crisis. (2022.0) 27. 27. United Nations Global Crisis Response Group Brief No. 1: Global Impact of War in Ukraine on Food, Energy and Finance SystemsUnited NationsGeneva, Switzerland2022. Brief No. 1: Global Impact of War in Ukraine on Food, Energy and Finance Systems (2022.0) 28. Lee A.J., Kane S., Lewis M., Good E., Pollard C.M., Landrigan T.J., Dick M.. Healthy diets ASAP—Australian Standardised Affordability and Pricing methods protocol. Nutr. J. (2018.0) 17 88. DOI: 10.1186/s12937-018-0396-0 29. Lee A., Mhurchu C.N., Sacks G., Swinburn B., Snowdon W., Vandevijvere S., Hawkes C., L’Abbé M., Rayner M., Sanders D.. Monitoring the price and affordability of foods and diets globally. Obes. Rev. (2013.0) 14 82-95. DOI: 10.1111/obr.12078 30. Lewis M., Lee A.. Costing ‘healthy’ food baskets in Australia—A systematic review of food price and affordability monitoring tools, protocols and methods. Public Health Nutr. (2016.0) 19 2872-2886. DOI: 10.1017/S1368980016002160 31. Seal J.. Monitoring the price and availability of healthy food—Time for a national approach?. Nutr. Diet. (2004.0) 61 197-199 32. Lee A.J., Kane S., Ramsey R., Good E., Dick M.. Testing the price and affordability of healthy and current (unhealthy) diets and the potential impacts of policy change in Australia. BMC Public Health (2016.0) 16. DOI: 10.1186/s12889-016-2996-y 33. Hendrie G.A., Ridoutt B.G., Wiedmann T.O., Noakes M.. Greenhouse Gas Emissions and the Australian Diet—Comparing Dietary Recommendations with Average Intakes. Nutrients (2014.0) 6 289-303. DOI: 10.3390/nu6010289 34. Australian Statistical Geography Standard (ASGS) Edition 3. (2021.0) 35. Socio-Economic Indexes for Areas. (2022.0) 36. Zorbas C., Lee A., Peeters A., Lewis M., Landrigan T., Backholer K.. Streamlined data-gathering techniques to estimate the price and affordability of healthy and unhealthy diets under different pricing scenarios. Public Health Nutr. (2021.0) 24 1-11. DOI: 10.1017/S1368980020001718 37. 2016 Census Community Profiles. (2022.0) 38. 6345.0—Wage Price Index, Australia. (2020.0) 39. Payment and Service Finder. (2019.0) 40. Lee A.J., Lewis M., Goodwin S.. Healthy Diets ASAP Portal. (2020.0) 41. Consumer Price Index, Australia. (2022.0) 42. McLeod C.. Fruit and Vegetable Prices Set to Leap by 8 Per Cent after October Floods. (2022.0) 43. 6401.0—Consumer Price Index, Australia. (2020.0) 44. Sullivan K., Jasper C.. Farm labour shortage continues despite surge of working holiday visa applications. ABC Rural (2022.0) 45. Kelly C.. Australian Farm Workers Entitled to Minimum Wage in Major Industry Shake-Up (2021.0) 46. FAO Food Price Index 47. Zorbas C., Browne J., Chung A., Peeters A., Booth S., Pollard C., Allender S., Isaacs A., Hawkes C., Backholer K.. Shifting the social determinants of food insecurity during the COVID-19 pandemic: The Australian experience. Food Secur. (2022.0) 15 151-170. DOI: 10.1007/s12571-022-01318-4 48. Davies T., Coyle D., Shahid M., Pettigrew S., Wu J.H., Marklund M.. Packaged foods purchased on price promotion in Australia. Appetite (2023.0) 180 106352. DOI: 10.1016/j.appet.2022.106352 49. 4364.0.55.007—Australian Health Survey: Nutrition First Results—Foods and Nutrients, 2011–2012. (2014.0) 50. Sui Z., Wong W.K., Louie J.C.Y., Rangan A.. Discretionary food and beverage consumption and its association with demographic characteristics, weight status, and fruit and vegetable intakes in Australian adults. Public Health Nutr. (2017.0) 20 274-281. DOI: 10.1017/S1368980016002305 51. Bennett G., Young E., Butler I., Coe S.. The Impact of Lockdown During the COVID-19 Outbreak on Dietary Habits in Various Population Groups: A Scoping Review. Front. Nutr. (2021.0) 8 626432. DOI: 10.3389/fnut.2021.626432 52. Neira C., Godinho R., Rincón F., Mardones R., Pedroso J.. Consequences of the COVID-19 Syndemic for Nutritional Health: A Systematic Review. Nutrients (2021.0) 13. DOI: 10.3390/nu13041168 53. Swinburn B., Egger G., Raza F.. Dissecting obesogenic environments: The development and application of a framework for identifying and prioritizing environmental interventions for obesity. Prev. Med. (1999.0) 29 563-570. DOI: 10.1006/pmed.1999.0585 54. Coyle D., Shahid M., Dunford E., Ni Mhurchu C., McKee S., Santos M., Popkin B., Trieu K., Marklund M., Neal B.. Estimating the potential impact of Australia’s reformulation programme on households’ sodium purchases. BMJ Nutr. Prev. Health (2021.0) 4. DOI: 10.1136/bmjnph-2020-000173 55. Ball K., Crawford D., Mishra G.. Socio-economic inequalities in women’s fruit and vegetable intakes: A multilevel study of individual, social and environmental mediators. Public Health Nutr. (2007.0) 9 623-630. DOI: 10.1079/PHN2005897 56. Burns C.. The Vulnerable and the Disadvantaged. Aust. Econ. Rev. (2008.0) 41 90-96. DOI: 10.1111/j.1467-8462.2008.00489.x 57. Thornton L.E., Lamb E.K., Ball K.. Fast food restaurant locations according to socioeconomic disadvantage, urban–regional locality, and schools within Victoria, Australia. SSM—Popul. Health (2016.0) 2 1-9. DOI: 10.1016/j.ssmph.2015.12.001 58. Zorbas C., Palermo C., Chung A., Iguacel I., Peeters A., Bennett R., Backholer K.. Factors perceived to influence healthy eating: A systematic review and meta-ethnographic synthesis of the literature. Nutr. Rev. (2018.0) 76 861-874. DOI: 10.1093/nutrit/nuy043 59. Davis G.C., Carlson A.. The inverse relationship between food price and energy density: Is it spurious?. Public Health Nutr. (2015.0) 18 1091-1097. DOI: 10.1017/S1368980014001098 60. Lee A.J., Kane S., Herron L.-M., Matsuyama M., Lewis M.. A tale of two cities: The cost, price-differential and affordability of current and healthy diets in Sydney and Canberra, Australia. Int. J. Behav. Nutr. Phys. Act. (2020.0) 17 80. DOI: 10.1186/s12966-020-00981-0 61. Love P., Whelan J., Bell C., Grainger F., Russell C., Lewis M., Lee A.. Healthy Diets in Rural Victoria—Cheaper than Unhealthy Alternatives, Yet Unaffordable. Int. J. Environ. Res. Public Health (2018.0) 15. DOI: 10.3390/ijerph15112469 62. Lee A., Lewis M.. Testing the Price of Healthy and Current Diets in Remote Aboriginal Communities to Improve Food Security: Development of the Aboriginal and Torres Strait Islander Healthy Diets ASAP (Australian Standardised Affordability and Pricing) Methods. Int. J. Environ. Res. Public Health (2018.0) 15. DOI: 10.3390/ijerph15122912 63. Lee A.J., Patay D., Summons S., Lewis M., Herron L., Nona F., Canuto C., Ferguson M., Twist A.. Cost and affordability of healthy, equitable and more sustainable diets in the Torres Strait Islands. Aust. N. Z. J. Public Health (2022.0) 46 340-345. DOI: 10.1111/1753-6405.13225 64. Zorbas C., Brooks R., Bennett R., Lee A., Marshall J., Naughton S., Lewis M., Peeters A., Backholer K.. Costing recommended (healthy) and current (unhealthy) diets in urban and inner regional areas of Australia using remote price collection methods. Public Health Nutr. (2022.0) 25 528-537. DOI: 10.1017/S1368980021004006 65. Lewis M., Herron L.-M., Bryce S., Wells L., Balmer L., Rainow S., Lee A.. Cost and Affordability of Healthy, Equitable and more Sustainable Diets, and Store Food Environments, in the APY Lands, June 2022. (2022.0) 66. Biddle N., Gray M.. Economic and Other Wellbeing in Australia—October 2022 (2022.0) 67. Thow A.M., Downs S., Jan S.. A systematic review of the effectiveness of food taxes and subsidies to improve diets: Understanding the recent evidence. Nutr. Rev. (2014.0) 72 551-565. DOI: 10.1111/nure.12123 68. 68. World Health Organization Regional Office for Europe Using Price Policies to Promote Healthier DietsWHO Regional Office for EuropeCopenhagen, Denmark2015. Using Price Policies to Promote Healthier Diets (2015.0) 69. Davidson P., Bradbury B., Wong M.. Poverty in Australia 2022: A Snapshot (2022.0) 70. Survey of 955 People Receiving the New Rate of Jobseeker and other Allowances. (2020.0) 71. Koltai J., Toffolutti V., McKee M., Stuckler D.. Prevalence and changes in food-related hardships by socioeconomic and demographic groups during the COVID-19 pandemic in the UK: A longitudinal panel study. Lancet Reg. Health—Eur. (2021.0) 6 100125. DOI: 10.1016/j.lanepe.2021.100125 72. Grudnoff M.. Opportunity Lost: Half a Million Australians in Poverty without the Coronavirus Supplement. Discussion Paper (2021.0) 73. Baker E., Bentley R., Beer A., Daniel L.. Renting in the Time of COVID-19: Understanding the Impacts, AHURI Final Report No. 340 (2020.0) 74. Loopstra R.. Interventions to address household food insecurity in high-income countries. Proc. Nutr. Soc. (2018.0) 77 270-281. DOI: 10.1017/S002966511800006X 75. 6401.0—Consumer Price Index, Australia, Dec 2015. (2016.0) 76. Dana L.M., Hart E., McAleese A., Bastable A., Pettigrew S.. Factors associated with ordering food via online meal ordering services. Public Health Nutr. (2021.0) 24 5704-5709. DOI: 10.1017/S1368980021001294 77. Meal Delivery Services Uber Eats, Menulog, Deliveroo and DoorDash Experienced Rapid Growth during 2020—A Year of Lockdowns & Work from Home. (2021.0) 78. Intergenerational Health and Mental Health Study (IHMHS). (2022.0) 79. Lewis M., McNaughton S.A., Rychetnik L., Lee A.J.. Cost and Affordability of Healthy, Equitable and Sustainable Diets in Low Socioeconomic Groups in Australia. Nutrients (2021.0) 13. DOI: 10.3390/nu13082900 80. Lewis M., McNaughton S.A., Rychetnik L., Chatfield M.D., Lee A.J.. Dietary Intake, Cost, and Affordability by Socioeconomic Group in Australia. Int. J. Environ. Res. Public Health (2021.0) 18. DOI: 10.3390/ijerph182413315 81. 81. UN Office of the High Commissioner for Human Rights The Right to Adequate Food. Human Rights Fact Sheet No. 34United Nations Human Rights and Food and Agriculture Organization of the United Nations (FAO)Geneva, Switzerland2010. The Right to Adequate Food. Human Rights Fact Sheet No. 34 (2010.0)
--- title: Machine Learning Models to Predict the Risk of Rapidly Progressive Kidney Disease and the Need for Nephrology Referral in Adult Patients with Type 2 Diabetes authors: - Chia-Tien Hsu - Kai-Chih Pai - Lun-Chi Chen - Shau-Hung Lin - Ming-Ju Wu journal: International Journal of Environmental Research and Public Health year: 2023 pmcid: PMC9967274 doi: 10.3390/ijerph20043396 license: CC BY 4.0 --- # Machine Learning Models to Predict the Risk of Rapidly Progressive Kidney Disease and the Need for Nephrology Referral in Adult Patients with Type 2 Diabetes ## Abstract Early detection of rapidly progressive kidney disease is key to improving the renal outcome and reducing complications in adult patients with type 2 diabetes mellitus (T2DM). We aimed to construct a 6-month machine learning (ML) predictive model for the risk of rapidly progressive kidney disease and the need for nephrology referral in adult patients with T2DM and an initial estimated glomerular filtration rate (eGFR) ≥ 60 mL/min/1.73 m2. We extracted patients and medical features from the electronic medical records (EMR), and the cohort was divided into a training/validation and testing data set to develop and validate the models on the basis of three algorithms: logistic regression (LR), random forest (RF), and extreme gradient boosting (XGBoost). We also applied an ensemble approach using soft voting classifier to classify the referral group. We used the area under the receiver operating characteristic curve (AUROC), precision, recall, and accuracy as the metrics to evaluate the performance. Shapley additive explanations (SHAP) values were used to evaluate the feature importance. The XGB model had higher accuracy and relatively higher precision in the referral group as compared with the LR and RF models, but LR and RF models had higher recall in the referral group. In general, the ensemble voting classifier had relatively higher accuracy, higher AUROC, and higher recall in the referral group as compared with the other three models. In addition, we found a more specific definition of the target improved the model performance in our study. In conclusion, we built a 6-month ML predictive model for the risk of rapidly progressive kidney disease. Early detection and then nephrology referral may facilitate appropriate management. ## 1. Introduction Diabetes mellitus (DM) is a major cause of life expectancy reduction and premature death [1,2,3]. The mortality in diabetic patients is significantly increased when the renal function is impaired [4]. With the improvement in treatment, trends in the rates of some diabetic complications have decreased, such as stroke or acute myocardial infarction, although the burden of diabetes is continuously increasing. However, diabetic kidney disease is still the leading cause of end-stage kidney disease (ESKD) [5,6]. According to the report by the United States Renal Data System (USRDS) in 2020 [7], Taiwan has persistently reported the highest incidence and prevalence of end-stage kidney disease worldwide. In the 2020 annual report on kidney disease in Taiwan, Lai et al. [ 8] reported the percentage of diabetes among incident dialysis patients increased from $45.3\%$ in 2010 to $46.2\%$ in 2018, and the percentage of diabetes among prevalent dialysis patients increased from $39.7\%$ in 2010 to $47.8\%$ in 2018. Early detection of rapidly progressive kidney disease and nephrology referral is an important point to decrease complications and mortality [9]. Albuminuria is an important marker of diabetic kidney disease (DKD) and is associated with a poor outcome, but some type 2 diabetes mellitus (T2DM) patients have a GFR decline before the onset of albuminuria [10]. In addition, nondiabetic kidney diseases are also the possible cause of rapidly progressive kidney disease [11], and these patients should be promptly referred to an experienced nephrologist for further surveying and management. With the heterogeneous phenotype of type 2 diabetic renal disease, novel tools are required for the early detection of rapidly progressive kidney disease and the need for nephrology referral in T2DM patients. Artificial intelligence (AI) has been widely applied in medical fields for diagnostic assistance, outcome prediction, and guiding treatment. Machine learning (ML) is a subset of AI. ML models are algorithms that teach a computer to learn from data [12,13]. There have been some studies of AI applications in DKD [14,15,16,17,18], but only a few studies have focused on the prediction of diabetic nephropathy and renal function decline [15,17]. The aim of our study was to construct a 6-month ML predictive model for the risk of rapidly progressive kidney disease and the need for nephrology referral in adult patients with T2DM. ## 2.1. Study Subjects We retrospectively extracted the electronic medical records (EMR) in our hospital from January 2008 to June 2021. Among them, we found 62,360 patients with a diagnosis of type 2 diabetes mellitus (T2DM) according to the International Classification of Diseases codes, and the inclusion criteria of our study were as follows: [1] hospitalized at least once with ICD-9 or ICD10 coding for T2DM, [2] at least two outpatient ICD-9 or ICD10 codings for T2DM, and [3] age at diagnosis of T2DM ≥ 20 years. The exclusion criteria were [1] patients who underwent dialysis before the diagnosis of T2DM and [2] renal transplant patients. Our study was approved by the institutional review board of Taichung Veterans General Hospital (IRB TCVGH No: SE22064A). Patient informed consent was waived because all protected health information was deidentified and the retrospective data analysis nature of this study. This research was funded by grants from the Ministry of Science and Technology of Taiwan (MOST 108-2314-B-005-005-MY3). ## 2.2. Data Extraction All the extracted personal information of the patients was deidentified. The demographic features used for the machine learning models included age, sex, height, weight, and body mass index (BMI). The laboratory features include serum creatinine (Cr), blood urea nitrogen (BUN), fasting glucose, random glucose, glycated hemoglobin (HbA1c), spot urine protein to creatinine ratio (UPCR), spot urine albumin to creatinine ratio (UACR), hemoglobin (HGB), hematocrit (HCT), albumin, total protein, aspartate aminotransferase (AST), alanine transaminase (ALT), creatine phosphokinase (CPK), high-sensitivity C-reactive protein (hsCRP), serum sodium (Na), serum potassium (K), red blood count (RBC), white blood count (WBC), platelet, total bilirubin (Bil-T), uric acid (UA), total cholesterol (CHO), low-density lipoprotein (LDL), and triglyceride (TG). The comorbidities were extracted according to the ICD-9 or ICD-10 codes and included diabetic retinopathy, hypertension, coronary arterial disease (CAD), stroke, peripheral arterial disease (PAD), congestive heart failure (CHF), acute kidney injury (AKI), liver cirrhosis, cancer, bacteremia, sepsis, shock, peritonitis, ascites, and bleeding esophageal varices. ## 2.3. Study Design and Label Definition In this study, we aimed to construct multiple machine learning models to predict the risk of rapidly progressive kidney disease and the need for nephrology referral in diabetes patients. We compared two different prediction outcomes of renal function deterioration and the need for nephrology referral in diabetes patients (Figure 1): [1] the estimated glomerular filtration rate (eGFR) falling below 30 mL/min/1.73 m2 and [2] the eGFR falling below 45 mL/min/1.73 m2. Clinical guidelines [19,20] recommend the referral of DM patients to nephrology when the eGFR falls below 30 mL/min/1.73 m2. However, a previous study showed a GFR < 45 mL/min/1.73 m2 at the time of referral is also a significant risk factor for mortality [21]. Hence, the outcomes of our predictive models were aggravated renal function from eGFR ≥ 60 mL/min/1.73 m2 to [1] GFR < 30 mL/min/1.73 m2 and [2] to <45 mL/min/1.73 m2. We selected adult T2DM patients with pair eGFR records of a 180-day period between the reference point and prediction target point. We first determined the target point for each individual patient and then went back to determine the reference point to select patients who fitted the criteria for the reference point. We labeled patients as being in the “referral” group if the eGFR was persistently lower than our outcomes (eGFR < 45 or <30 mL/min/1.73 m2) at the target point and 90 days after the target point. We confirmed chronic kidney disease if the eGFR did not recover 90 days after the target point in the “referral” group. On the other hand, we labeled patients as being in the “non-referral” group if [1] the eGFR was persistently ≥ 30 mL/min/1.73 m2 at the target point and 90 days after the target point or [2] the eGFR was persistently ≥ 45 mL/min/1.73 m2 at the target point and 90 days after the target point. We further enrolled patients according to the criteria for the reference point as follows: [1] eGFR ≥ 60 mL/min/1.73 m2 at the reference point, [2] 180-day average eGFR ≥ 60 mL/min/1.73 m2 prior to the reference point, and [3] T2DM diagnosis before the reference point. ## 2.4. Data Preprocessing and Machine Learning Models We discussed with the domain experts for outliers of laboratory features. We excluded outliers of laboratory features on the basis of medical knowledge, wherein the error values were obviously inconsistent with the actual situation. Patients in the non-referral group had a more stable condition than patients in the referral group, which resulted in less laboratory examinations among patients in the non-referral group. There were a few patients with more than 12 missing features in the referral group. We excluded patients with more than 12 missing features to deal with the missing data in the non-referral group and the imbalance of the data set. After that, features with more than $40\%$ missing values were excluded, and the mean of this feature was used to interpolate the remaining missing data [22,23]. We chose the “last” and “average” values of each feature in the 180-day period before the reference point as input data (Figure 1). We treated our prediction of referral need as a binary classification problem. The architecture of our prediction models is shown in Figure 2. The study cohort was divided into the following two parts: [1] the data from January 2008 to December 2019 as the training/validation data set, and [2] the data from January 2020 to June 2021 as the testing data set. Then, the training/validation data set was randomly divided, with $80\%$ used for training and $20\%$ for validation. We performed fivefold cross-validation within the training/validation data set to identify the optimal classifier [24,25,26,27,28]. The optimal classifier was then used to predict our outcome for each patient in the testing data set. The testing data set was independent of the training/validation data set. It provided an unbiased final model performance metric. We compared the performance of three classical machine learning algorithms: logistic regression (LR), random forest (RF), and extreme gradient boosting (XGBoost) to develop the predictive models. We further applied an ensemble approach using soft voting classifier to classify the referral group [29,30]. In the ensemble model, LR, RF, and XGBoost classifier were ensembled. We used the soft voting calculated on the predicted probability of the output class. All analyses were performed using Python (version 3.8) [31]. We used the area under the receiver operating characteristic curve (AUROC), precision, recall, and accuracy as the metrics to evaluate the performance between different models. We also calculated the Shapley additive explanations (SHAP) values to evaluate the feature importance that explored the relationship between the outcome and the feature [32,33]. The assessment of normality was conducted using the Kolmogorov–Smirnov test. The continuous variables with normal distribution are shown as mean ± standard deviation, whereas the continuous variables with non-normal distribution are presented as the median (first quartile, third quartile). The categorical variables are reported as numbers (percentage). Tests for the statistical significance were conducted using the chi-squared test for categorical variables and the Mann–Whitney test for non-parametric continuous variables. The level of significance was set at $p \leq 0.05.$ Statistical analyses were performed using MedCalc for Windows, version 20.210 (MedCalc Software, Ostend, Belgium). Medical data are usually unbalanced. Because the imbalance of data were found, we performed a pilot experiment with a new target of outcome (persistent eGFR ≥ 60 mL/min/1.73 m2) to find the optimal method for this problem (see Appendix A). We applied Downsample, the Synthetic Minority Oversampling Technique (SMOTE) algorithm, and Tomek Link [34] to cope with the imbalance of data [35,36]. However, the result (see Table A1 in Appendix A) showed no obvious improvement of performance. Finally, we input the original data set into our machine learning models without any of the abovementioned methods. ## 3. Results A total of 19,892 adult T2DM patients were enrolled in “experiment 1” to predict the rapid renal function decline and nephrology referral when the eGFR was persistently lower than 30 mL/min/1.73 m2. Among these, there were 19,244 adult T2DM patients in the “non-referral” group and 648 adult T2DM patients in the “referral” group. In addition, a total of 16,145 adult T2DM patients were enrolled in “experiment 2” to predict the rapid renal function decline and nephrology referral when the eGFR was persistently lower than 45 mL/min/1.73 m2. Among these, there were 15,159 adult T2DM patients in the “non-referral” group and 986 adult T2DM patients in the “referral” group. ## 3.1. Experiment 1: Predict Rapidly Progressive Kidney Disease and Nephrology Referral When the eGFR Was Persistently Lower than 30 mL/min/1.73 m2 Table 1 reveals the baseline demographic and clinical characteristics of the included patients in experiment 1. The age of the patients was significantly older in the referral group. Patients in the referral group had significantly more comorbidities, higher creatinine, higher BUN, higher HbA1c, lower HGB, lower albumin, higher hsCRP, higher uric acid, higher TG, higher UPCR, and higher UACR. The missing data for each variable in the experiment 1 are shown in Appendix B Table A2. Table 2 demonstrates the three models to predict rapidly progressive kidney disease and nephrology referral when the eGFR was persistently < 30 mL/min/1.73 m2. All three models achieved an accuracy of more than 0.91 and an AUROC of more than 0.96. The XGB model had higher accuracy and relatively higher precision in the referral group as compared with the LR and RF models. However, LR and RF models had higher recall in the referral group. In general, the ensemble voting classifier had relatively higher accuracy, higher AUROC, and higher recall in the referral group as compared with the other three models. Figure 3 shows the confusion matrix and predictive probabilities of the XGBoost model in experiment 1. The plot of the predictive probabilities in Figure 3 revealed this model could distinguish the “referral” from the “non-referral” group in both the training/validation data set (Figure 3A) and testing data set (Figure 3B). Figure 4 demonstrates the SHAP summary plot of the top 15 features for the XGBoost model in experiment 1. The higher the SHAP value of a feature, the higher the probability of rapidly progressive kidney disease. A dot denotes each feature value for the model of each patient. The dots are colored according to the values of the features for the respective patient and accumulate to describe the density. Blue represents the lower feature values, and red represents the higher feature values. ## 3.2. Experiment 2: Predict Rapidly Progressive Kidney Disease and Nephrology Referral When the eGFR Was Persistently Lower than 45 mL/min/1.73 m2 Table 3 shows the baseline demographic and clinical characteristics of the included patients in experiment 2. The age of the patients was significantly older in the referral group. Patients in the referral group had significantly more comorbidities, higher creatinine, higher BUN, lower HGB, lower albumin, higher hsCRP, higher uric acid, higher TG, higher UPCR, and higher UACR. The missing data for each variable in experiment 2 is shown in Appendix B Table A3. Table 4 reveals the three models to predict rapidly progressive kidney disease and nephrology referral when the eGFR was persistently < 45 mL/min/1.73 m2. All three models achieved an accuracy of more than 0.88 and an AUROC more than 0.93. The XGB model had higher accuracy and relatively higher precision in the referral group as compared with the LR and RF models. However, LR and RF models had higher recall in the referral group. In general, the ensemble voting classifier had relatively higher accuracy, higher AUROC, and higher recall in the referral group as compared with the other models. Figure 5 shows the confusion matrix and predictive probabilities of the XGBoost model in experiment 2. The plot of the predictive probabilities of Figure 5 revealed this model could distinguish the “referral” from the “non-referral” group in both the training/validation data set (Figure 5A) and the testing data set (Figure 5B). Figure 6 demonstrates the SHAP summary plot of the top 15 features for the XGBoost model in experiment 2. The first three features were the same in both experiment 1 and experiment 2, and the importance of proteinuria increased in experiment 2 (eGFR was persistently < 45 mL/min/1.73 m2) as compared with experiment 1 (eGFR was persistently < 30 mL/min/1.73 m2). Proteinuria (UPCR or UACR) is also an important predictor for the risk of rapidly progressive kidney disease and the need for nephrology referral. ## 3.3. Additional Experiment with Loose Inclusion and Labeling Criteria for Both Experiments 1 and 2 We conducted an additional experiment with loose inclusion and labeling criteria for both experiments 1 and 2. In this additional experiment, we included T2DM patients with one laboratory result showing an eGFR ≥ 60 mL/min/1.73 m2 at the reference point and a T2DM diagnosis before the reference point. We also labeled patients with only one laboratory result, showing an eGFR < 30 mL/min/1.73 m2 for experiment 1 and an eGFR < 45 mL/min/1.73 m2 for experiment 2 in this additional experiment. We did not confirm patients with a 180-day average eGFR ≥ 60 mL/min/1.73 m2 prior reference point and a persistently lower eGFR 90 days after the target point in this additional experiment. Table 5 reveals that the accuracy and AUROC decreased in all of the three ML models for the additional experiment with loose inclusion and labeling criteria. ## 4. Discussion Due to the heterogeneous phenotype of type 2 diabetic renal disease, the optimal time for the nephrology referral of T2DM patients is still challenging [6]. The American Diabetes Association (ADA) recommends that [1] diabetes patients should be referred for evaluation for RRT if they have an eGFR < 30 mL/min/1.73 m2, and [2] diabetes patients should be referred to a physician experienced in the care of kidney disease for uncertainty about the etiology of kidney disease, difficult management issues, and rapidly progressive kidney disease [19]. However, Pinier et al. [ 21] performed a retrospective survival analysis in DM patients in a 13-year period, and the study showed that both an eGFR < 30 mL/min/1.73 m2 and <45 mL/min/1.73 m2 at the time of referral were powerful risk factors for mortality. Therefore, we performed one experiment with a predictive target of an eGFR < 30 mL/min/1.73 m2 and another one with a predictive target of eGFR < 45 mL/min/1.73 m2. In addition, our study design also predicted rapidly progressive kidney disease in a 6-month period. This is an important indication for nephrology referral in T2DM as well. Few studies have focused on the prediction of diabetic nephropathy and renal function decline [14,15,16,17,18]. Makino et al. [ 17] constructed a logistic regression ML learning model based on big data from the electronic medical records (EMR) of diabetes patients. Their logistic regression model had 3073 features with time series data. The accuracy of their logistic regression ML model to predict DKD aggravation was 0.71. Dong et al. [ 15] built up a 3-year DKD risk predictive model in patients with T2DM and normo-albuminuria, and their study showed the LightGBM model was the best model with an area under curve (AUC) of 0.815. Owing to the different study design and predictive target, all models in our study achieved an accuracy of more than 0.88 and an AUROC more than 0.93. Our study mainly focused on T2DM patients with rapidly progressive kidney disease in the 6-month period, as this condition is an important indication for nephrology referral. Early detection of this condition is a key to improving renal outcome and reducing complications. Additionally, our study design confirmed the target condition with persistent renal function impairment 90 days after the target point. The more specific and strict definition of the predictive target could improve the model performance in our study (Table 5). Our result showed that the XGB model had higher accuracy and relatively higher precision in the referral group as compared with the LR and RF models, but LR and RF models had higher recall in the referral group. The lower precision means that the model had more false alarms, and the false alarms may increase the clinical load of the nephrologist. However, the higher recall may be more important for patient safety because it means that less patients who need nephrology referral (adult T2DM patients with rapidly progressive kidney disease) are neglected. In general, the ensemble voting classifier had relatively higher accuracy, higher AUROC, and higher recall in the referral group as compared with the other three models. Some potential limitations of this study should be acknowledged. First, the nature of the retrospective study may cause some unrecognized confounding factors to bias the findings. Second, we did not analyze the impact of medication in our study, and some medication may be associated with rapidly progressive kidney disease. Third, our study included a small sample size and was conducted at a single hospital. The majority of the population was Taiwanese. Fourth, the data set was highly unbalanced, despite our attempts to deal with this problem. Models trained on imbalanced data may cause the accuracy paradox. Precision and recall may be better metrics in such conditions. Fifth, we excluded patients with more than 12 missing features to deal with the missing data in the non-referral group and the imbalance of dataset, which may introduce bias in analysis. Sixth, only internal validation was performed in our study; external validation using a different data set is needed. Hence, further multicenter and multinational studies are required to confirm the stability of the performance of our predictive model. ## 5. Conclusions In conclusion, we built a 6-month machine learning predictive model for the risk of rapidly progressive kidney disease and the need for nephrology referral in adult patients with T2DM and an initial eGFR ≥ 60 mL/min/1.73 m2. Our result showed that the XGB model had higher accuracy and relatively higher precision in the referral group as compared with the LR and RF models, but LR and RF models had higher recall in the referral group. In general, the ensemble voting classifier had relatively higher accuracy, higher AUROC, and higher recall in the referral group as compared with the other three models. Early detection of rapidly progressive kidney disease is key to improving the renal outcome and reducing complications in adult patients with T2DM. ## References 1. Kianmehr H., Zhang P., Luo J., Guo J., Pavkov M.E., Bullard K.M., Gregg E.W., Ospina N.S., Fonseca V., Shi L.. Potential Gains in Life Expectancy Associated With Achieving Treatment Goals in US Adults With Type 2 Diabetes. JAMA Netw. Open (2022) 5 e227705. DOI: 10.1001/jamanetworkopen.2022.7705 2. Loukine L., Waters C., Choi B.C., Ellison J.. Impact of diabetes mellitus on life expectancy and health-adjusted life expectancy in Canada. Popul. Health Metr. (2012) 10 7. DOI: 10.1186/1478-7954-10-7 3. Rao Kondapally Seshasai S., Kaptoge S., Thompson A., Di Angelantonio E., Gao P., Sarwar N., Whincup P.H., Mukamal K.J., Gillum R.F., Holme I.. Diabetes mellitus, fasting glucose, and risk of cause-specific death. N. Engl. J. Med. (2011) 364 829-841. PMID: 21366474 4. Afkarian M., Sachs M.C., Kestenbaum B., Hirsch I.B., Tuttle K.R., Himmelfarb J., De Boer I.H.. Kidney disease and increased mortality risk in type 2 diabetes. J. Am. Soc. Nephrol. (2013) 24 302-308. DOI: 10.1681/ASN.2012070718 5. Gregg E.W., Li Y., Wang J., Rios Burrows N., Ali M.K., Rolka D., Williams D.E., Geiss L.. Changes in diabetes-related complications in the United States, 1990–2010. N. Engl. J. Med. (2014) 370 1514-1523. DOI: 10.1056/NEJMoa1310799 6. Martínez-Castelao A., Soler M.J., Górriz Teruel J.L., Navarro-González J.F., Fernandez-Fernandez B., de Alvaro Moreno F., Ortiz A.. Optimizing the timing of nephrology referral for patients with diabetic kidney disease. Clin. Kidney J. (2021) 14 5-8. DOI: 10.1093/ckj/sfaa125 7. 7. United States Renal Data System 2022 USRDS Annual Data Report: Epidemiology of Kidney Disease in the United StatesNational Institutes of Health, National Institute of Diabetes and Digestive and Kidney DiseasesBethesda, MD, USA2022. 2022 USRDS Annual Data Report: Epidemiology of Kidney Disease in the United States (2022) 8. Lai T.-S., Hsu C.-C., Lin M.-H., Wu V.-C., Chen Y.-M.. Trends in the incidence and prevalence of end-stage kidney disease requiring dialysis in Taiwan: 2010–2018. J. Formos. Med. Assoc. (2022) 121 S5-S11. DOI: 10.1016/j.jfma.2021.12.013 9. Smart N.A., Dieberg G., Ladhani M., Titus T.. Early referral to specialist nephrology services for preventing the progression to end-stage kidney disease. Cochrane Database Syst. Rev. (2014) 6 CD007333. DOI: 10.1002/14651858.CD007333.pub2 10. Porrini E., Ruggenenti P., Mogensen C.E., Barlovic D.P., Praga M., Cruzado J.M., Hojs R., Abbate M., de Vries A.P.. Non-proteinuric pathways in loss of renal function in patients with type 2 diabetes. Lancet Diabetes Endocrinol. (2015) 3 382-391. DOI: 10.1016/S2213-8587(15)00094-7 11. Kritmetapak K., Anutrakulchai S., Pongchaiyakul C., Puapairoj A.. Clinical and pathological characteristics of non-diabetic renal disease in type 2 diabetes patients. Clin. Kidney J. (2018) 11 342-347. DOI: 10.1093/ckj/sfx111 12. Jiang F., Jiang Y., Zhi H., Dong Y., Li H., Ma S., Wang Y., Dong Q., Shen H., Wang Y.. Artificial intelligence in healthcare: Past, present and future. Stroke Vasc. Neurol. (2017) 2 230-243. DOI: 10.1136/svn-2017-000101 13. Yuan Q., Zhang H., Deng T., Tang S., Yuan X., Tang W., Xie Y., Ge H., Wang X., Zhou Q.. Role of artificial intelligence in kidney disease. Int. J. Med. Sci. (2020) 17 970-984. DOI: 10.7150/ijms.42078 14. Belur Nagaraj S., Pena M.J., Ju W., Heerspink H.L., Consortium B.D.. Machine-learning-based early prediction of end-stage renal disease in patients with diabetic kidney disease using clinical trials data. Diabetes Obes. Metab. (2020) 22 2479-2486. DOI: 10.1111/dom.14178 15. Dong Z., Wang Q., Ke Y., Zhang W., Hong Q., Liu C., Liu X., Yang J., Xi Y., Shi J.. Prediction of 3-year risk of diabetic kidney disease using machine learning based on electronic medical records. J. Transl. Med. (2022) 20 143. DOI: 10.1186/s12967-022-03339-1 16. Huang G.-M., Huang K.-Y., Lee T.-Y., Weng J.T.-Y.. An interpretable rule-based diagnostic classification of diabetic nephropathy among type 2 diabetes patients. BMC Bioinform. (2015) 16. DOI: 10.1186/1471-2105-16-S1-S5 17. Makino M., Yoshimoto R., Ono M., Itoko T., Katsuki T., Koseki A., Kudo M., Haida K., Kuroda J., Yanagiya R.. Artificial intelligence predicts the progression of diabetic kidney disease using big data machine learning. Sci. Rep. (2019) 9 11862. DOI: 10.1038/s41598-019-48263-5 18. Rodriguez-Romero V., Bergstrom R.F., Decker B.S., Lahu G., Vakilynejad M., Bies R.R.. Prediction of nephropathy in type 2 diabetes: An analysis of the ACCORD trial applying machine learning techniques. Clin. Transl. Sci. (2019) 12 519-528. DOI: 10.1111/cts.12647 19. Association A.D.. 11. Microvascular Complications and Foot Care: Standards of Medical Care in Diabetes-2021. Diabetes Care (2020) 44 S151-S167. DOI: 10.2337/dc21-S011 20. Levin A., Stevens P.E., Bilous R.W., Coresh J., De Francisco A.L., De Jong P.E., Griffith K.E., Hemmelgarn B.R., Iseki K., Lamb E.J.. Kidney Disease: Improving Global Outcomes (KDIGO) CKD Work Group. KDIGO 2012 clinical practice guideline for the evaluation and management of chronic kidney disease. Kidney Int. Suppl. (2013) 3 1-150 21. Pinier C., Gatault P., Francois M., Barbet C., Longuet H., Rabot N., Noble J., Bailly E., Buchler M., Sautenet B.. Renal function at the time of nephrology referral but not dialysis initiation as a risk for death in patients with diabetes mellitus. Clin. Kidney J. (2018) 11 762-768. DOI: 10.1093/ckj/sfy032 22. Emmanuel T., Maupong T., Mpoeleng D., Semong T., Mphago B., Tabona O.. A survey on missing data in machine learning. J. Big Data (2021) 8 140. DOI: 10.1186/s40537-021-00516-9 23. Soley-Bori M.. Dealing with missing data: Key assumptions and methods for applied analysis. Boston Univ. (2013) 4 1-19 24. Chauhan N.K., Singh K.. Performance assessment of machine learning classifiers using selective feature approaches for cervical cancer detection. Wirel. Pers. Commun. (2022) 124 2335-2366. DOI: 10.1007/s11277-022-09467-7 25. Kumari V.A., Chitra R.. Classification of diabetes disease using support vector machine. Int. J. Eng. Res. Appl. (2013) 3 1797-1801 26. Neamțu B.M., Visa G., Maniu I., Ognean M.L., Pérez-Elvira R., Dragomir A., Agudo M., Șofariu C.R., Gheonea M., Pitic A.. A decision-tree approach to assist in forecasting the outcomes of the neonatal brain injury. Int. J. Environ. Res. Public Health (2021) 18. DOI: 10.3390/ijerph18094807 27. Prusty S., Patnaik S., Dash S.K.. SKCV: Stratified K-fold cross-validation on ML classifiers for predicting cervical cancer. Front. Nanotechnol. (2022) 4 972421. DOI: 10.3389/fnano.2022.972421 28. Yang T., Zhang L., Yi L., Feng H., Li S., Chen H., Zhu J., Zhao J., Zeng Y., Liu H.. Ensemble Learning Models Based on Noninvasive Features for Type 2 Diabetes Screening: Model Development and Validation. JMIR Med. Inform. (2020) 8 e15431. DOI: 10.2196/15431 29. de Oliveira G.P., Fonseca A., Rodrigues P.C.. Diabetes diagnosis based on hard and soft voting classifiers combining statistical learning models. Braz. J. Biom. (2022) 40 415-427. DOI: 10.28951/bjb.v40i4.605 30. Kumari S., Kumar D., Mittal M.. An ensemble approach for classification and prediction of diabetes mellitus using soft voting classifier. Int. J. Cogn. Comput. Eng. (2021) 2 40-46. DOI: 10.1016/j.ijcce.2021.01.001 31. Pedregosa F., Varoquaux G., Gramfort A., Michel V., Thirion B., Grisel O., Blondel M., Prettenhofer P., Weiss R., Dubourg V.. Scikit-learn: Machine learning in Python. J. Mach. Learn. Res. (2011) 12 2825-2830 32. Lundberg S.M., Erion G.G., Lee S.-I.. Consistent individualized feature attribution for tree ensembles. arXiv (2018) 33. Tseng P.-Y., Chen Y.-T., Wang C.-H., Chiu K.-M., Peng Y.-S., Hsu S.-P., Chen K.-L., Yang C.-Y., Lee O.K.-S.. Prediction of the development of acute kidney injury following cardiac surgery by machine learning. Crit. Care (2020) 24 478. DOI: 10.1186/s13054-020-03179-9 34. Tomek I.. Two Modifications of CNN. IEEE Trans. Syst. Man Cybern (1976) SMC-6 769-772 35. Barros T.M., Souza Neto P.A., Silva I., Guedes L.A.. Predictive models for imbalanced data: A school dropout perspective. Educ. Sci. (2019) 9. DOI: 10.3390/educsci9040275 36. He H., Garcia E.A.. Learning from Imbalanced Data. IEEE Trans. Knowl. Data Eng. (2009) 21 1263-1284
--- title: 'Sugar Molecules Detection via C2N Transistor-Based Sensor: First Principles Modeling' authors: - Asma Wasfi - Sarah Awwad - Mousa Hussein - Falah Awwad journal: Nanomaterials year: 2023 pmcid: PMC9967288 doi: 10.3390/nano13040700 license: CC BY 4.0 --- # Sugar Molecules Detection via C2N Transistor-Based Sensor: First Principles Modeling ## Abstract Real-time detection of sugar molecules is critical for preventing and monitoring diabetes and for food quality evaluation. In this article, a field effect transistor (FET) based on two-dimensional nitrogenated holey graphene (C2N) was designed, developed, and tested to identify the sugar molecules including xylose, fructose, and glucose. Both density functional theory and non-equilibrium Green’s function (DFT + NEGF) were used to study the designed device. Several electronic characteristics were studied, including work function, density of states, electrical current, and transmission spectrum. The proposed sensor is made of a pair of gold electrodes joint through a channel of C2N and a gate was placed underneath the channel. The C2N monolayer distinctive characteristics are promising for glucose sensors to detect blood sugar and for sugar molecules sensors to evaluate food quality. The electronic transport characteristics of the sensor resulted in a unique signature for each of the sugar molecules. This proposed work suggests that the developed C2N transistor-based sensor could detect sugar molecules with high accuracy. ## 1. Introduction Carbohydrates are important organic substances for both people and plants because they play a number of crucial roles in growth and development. Glucose and fructose have significant importance since they are important nutrients in people diet. Moreover, xylose levels are measured to check if there is problem with peoples’ ability to absorb nutrients. They can be found naturally in a variety of foods or additives. The detection of these sugar molecules is highly important to evaluate food quality [1,2]. As well, reliable and quick sugar detection during food production and storage is highly important. The detection of glucose, fructose, and xylose can be utilized to assess food quality since they reveal details about a food product’s nutritional value, flavor, and sweetness. Simple sugars such as glucose are frequently present in fruits, vegetables, and grains. High glucose levels in a food product can be a sign that it is high in carbohydrates and energy, but they can also be a sign that the food is overripe or that it has been processed at high temperatures [3]. Fructose is frequently present in fruits and honey. A food product with high fructose levels may be sweet and have a lot of natural sugars. Fructose content, which is frequently seen in processed foods, can also be an indication that the product has been sweetened with high fructose corn syrup [4]. Dietary carbohydrates contain xylose. Fruits, cereals, bread, and vegetables, including potatoes, peas, and carrots, all include it as part of their sugar composition. Detecting xylose can be used to identify the presence of specific types of fruits or vegetables in a food product. One way to recognize the presence of specific fruits and vegetables in a food product is to look for the sugar xylose, which is frequently present in certain foods. It is possible to identify and confirm the composition of a food product by analyzing the xylose content in a sample. A large number of people worldwide suffer from diabetes [5,6], and it can lead to significant complications, such as heart attack, kidney failure, and blindness [7,8]. Moreover, glucose metabolism anomalies might lead to various diseases and problems [6,9]. Thus, it is highly crucial to monitor glucose levels and to supervise patients with diabetes. Researchers have developed various types of biosensors to help patients track their glucose concentration without the need to go to the hospital [10]. Blood sugar levels detection is a highly critical research area [11,12]. The two main electrochemical glucose identification methods are non-enzymatic or enzymatic [13]. Enzymatic methods utilize detection elements such as glucose oxidase enzyme (GOx). These methods oxidize glucose and generate compounds that can be detected, such as CO2, O2, or H2O2. When glucose oxidase interacts with enzymatic sensors, it releases oxygen, gluconolactone, and hydrogen. The glucose present on the sensor’s surface oxidizes and is expressed in typical current values. These sensors have high selectivity, however, they still face some challenges, such as: (i) degraded sensitivity with time due to enzymatic leaking; (ii) short life-time and low stability; and (iii) reduction in high overpotentials [14]. Until this time, most of the available blood glucose sensors depend on glucose oxidase (GOx) enzyme-based recognition unit [11,15]. Enzymatic glucose-based sensors have remarkable selectivity and sensitivity; however, they have low detection limits and are not stable with temperatures and humidity variations. Additionally, they require costly enzymes. Currently, research work is being focused on glucose biosensors that are cheaper, non-enzymatic, and sensitive to detect body glucose by using various body fluids, such as saliva, tear, or sweat for the purpose of self-monitoring of diabetes [12]. Various popular techniques can be used to detect glucose such as colorimetric and fluorometric methods [16,17,18,19,20,21]. The main idea behind these methods is using a chemical indicator or a fluorescent probe that changes colour or fluorescence intensity due to the addition of glucose target [21]. These methods are suitable for point-of-care applications as they are relatively simple, easy to use, and cheap. However, they lack the sensitivity of other techniques as the colour change or fluorescence signal can be influenced by other factors, such as temperature, pH, and the presence of other analytes. One more popular method to detect glucose is paper-based sensors, which is also known as lateral flow assays [22]. Paper-based methods consist of a strip of paper coated with a reagent that is sensitive toward glucose. The targeted glucose sample is added to the strip, and the presence of glucose is identified by the change in colour and fluorescence [22,23]. These devices are popular as they are easy to use, portable, cheap, and simple. However, their signal can be changed due to the presence of other analytes and they are not as sensitive as other techniques [24]. Recently, there has been an increasing interest in electrochemical sensors [25] and biosensors to detect glucose since they provide high selectivity and sensitivity. Other various methods are being explored for the potential of glucose detection such as Raman spectroscopy [26] and mass spectrometry [27]. Electrochemical biosensing is applied extensively to detect biomolecules to diagnose and detect various diseases [28,29]. The biosensing research field has witnessed huge enhancement due to the development of field effect transistors biosensors. These biosensors have shown great performance due to their reliable detection, high sensitivity, and real time monitoring [30]. The sensing mechanism of transistor-based sensors depends on the change in the channel electrical resistance due to molecular addition and adsorption [31,32,33]. These devices have shown effective identification of molecules, ions, bacteria, and several biological entities [28,31,34,35,36,37]. As the biosensor performance relies on its surface to enhance the charge transfer, two-dimensional graphene including functionalized graphene nanomaterial is considered the best selection for glucose sensors. Platinum-functionalized graphene was utilized to detect glucose with a 0.6 M detection limit [38]. Moreover, gold nanoparticles were explored to detect 0.3 μM concentration of glucose [39]. Several nanomaterial biosensors, such as graphene and carbon nanotubes, were used for glucose detection. However, it poses the challenge of potential toxicity [40,41]. Various technologies were studied to design electrochemical reaction sensors based on non-enzymatic glucometers, including carbon-based materials, such as reduced graphene oxide (GO), graphene, metal nanoparticles [42,43], and carbon nanotube (CNT) [44,45]. Carbon nanomaterials doped with nitrogen have better performance in biosensors compared to pristine carbon. Carbon nanomaterials doped with nitrogen are used in biosensors because of their special characteristics that make them suitable for utilization in these kinds of applications [46]. Because the surface to volume ratios of carbon nanomaterials, such as carbon nanotubes and graphene are high, a lot of biomolecules can be adsorbed onto the surface [47]. Nitrogen atom doping of carbon nanomaterials improves their electrical conductivity, increasing their sensitivity for sensing applications [48]. It is possible to create nitrogen-doped carbon nanomaterials by adding nitrogen to carbon nanomaterials. These materials are more stable and have better electrical conductivity than pristine carbon. The electrical conductivity of carbon nanomaterials can be improved by nitrogen atoms acting as electron acceptors, increasing their sensitivity for biosensing applications [49]. Additionally, nitrogen doping can increase the carbon nanostructures’ chemical stability, strengthening their resistance to degradation. Carbon nanomaterials that have been doped with nitrogen are less toxic and more stable in biological settings, which can increase their biocompatibility. Additionally, compared to pristine carbon, nitrogen-doped carbon nanomaterials have demonstrated enhanced stability and biocompatibility, making them appropriate for application in biosensors [50]. Overall, nitrogen-doped carbon nanomaterials are a desirable option for use in biosensors due to their large surface area, electrical conductivity, and biocompatibility [46,48,50]. The novelty of this work is based on using C2N-FET for the first time as a sensor to recognize each of the sugar molecules. To the best of our knowledge, this is the first research that utilizes FET consisting of C2N channel and a pair of gold electrodes to identify glucose, fructose, and xylose molecules. Within the many carbon nanostructures rich with nitrogen, C2N has been synthesized and computationally studied [51,52]. In this work, first principles modeling was used to study the sensing properties of C2N field effect transistor (FET) for the purpose of non-enzymatic glucose detection. This is the first report that uses C2N FET to detect glucose. In this research, a field effect transistor based on two-dimensional nitrogenated holey graphene (C2N) was developed, designed, and tested to identify the sugar molecules including xylose, fructose, and glucose. Both density functional theory and non-equilibrium Green’s function (DFT + NEGF) were used to study the designed sensor. Various electronic characteristics were studied such as: work function, density of states, electrical current, and transmission spectrum. The proposed sensor is made of a pair of gold electrodes joint through a channel of C2N and a gate was placed underneath the channel. The C2N monolayer distinctive characteristics are promising for glucose sensors to detect blood sugar. Moreover, the detection of the three types of sugar molecules can be used to evaluate food quality. ## 2. Materials and Methods The simulation work was produced using the graphical user interface of Virtual Nanolab and the Quantumwise Atomistix Toolkit (QuantumATK 2018.06 developed by Copenhagen, Denmark). United Arab Emirates University High Performance Computing (HPC) was utilized to run ATK-VNL simulations. Seven nodes with a total of 36 processors each have been used for HPC. As a result, 252 processors were used to complete the task. ## 2.1. Sensor Setup and Configuration The setup and configuration of the C2N based sensor were conducted and investigated via Quantumwise (ATK-VNL). Figure 1 displays the nanoscale system setup. The left and right gold electrodes, the C2N central area which consists of one layer of C2N, and the gate terminal located beneath the central region make up the C2N metal-semiconductor-metal junction system. The gate is formed of two layers: a metallic layer and a 2.9 Å dielectric layer of SiO2 with a dielectric constant of 3.9. The C2N channel width is 13 Å and length is 28 Å, while the gold electrode length is 10 Å. The system consists of 209 atoms. First-principle electronic transport measurements were generated to detect each of the sugar molecules electronic signature. A, B, and C are indictors for A-, B-, and C-direction as displayed in Figure 1. Figure 2 shows the atomic structure for each of the sugar molecules: glucose, fructose, and xylose. Due to their unique electronic and chemical structure, each molecule has a distinct electronic signature. Various electronic transport characteristics, including device density of states, transmission spectrum, work function, and electronic current, are generated for the bare C2N transistor and for the transistor with each of the sugar molecules. Figure 3 shows the C2N transistor structures with fructose. The big hollow site shown in Figure 3, which is the most stable site for xylose, fructose, and glucose for the adsorption of each of the sugar molecules [53]. The gate voltage was fixed at 1V, and finite bias voltage was fixed between right and left electrode and ranged from 0 to 1 V. ## 2.2. Computational Method First-principles method is conducted within the generalized gradient approximation (GGA) exchange correlation function. For the plane-wave basis set, a cut-off energy of 80 Ha is utilized. A 1 × 1 × 1 k-mesh is used to optimize the structure, while a denser mesh of 2 × 2 × 135 is used for the electronic transport calculations. The systems are optimized till the forces on each atom in the supercell are less than 0.05 eV/Å. Each of the sugar molecules was optimized separately. Moreover, the gold atoms were optimized before forming the electrodes. Then, the C2N channel was optimized. At the end, the whole sensor with each of the sugar molecules was optimized. 1 × 1 × 1 k-mesh and Monkhorst-Pack grid, a type of uniform grid that is known to provide good convergence, were used for optimization as conducted by previous studies [54]. For the electronic transport characteristics such as IV a denser k-mesh grid was used. Quantumatk website [55] and other articles [56] recommend using 100 along the transport direction which is represented as the C direction in Figure 1. Thomas et al. used 1 × 1 × 100 k-point samplings along the device transport direction to generate the IV calculations [57]. In this work, a 2 × 2 × 135 k-point was utilized. The electronic transport characteristics are generated by utilizing the density functional theory and non-equilibrium Green’s function (NEGF) approach. The sugar molecules are positioned on the C2N monolayer to investigate the transport characteristics of the C2N monolayer and the sugar molecules. Three areas are included: the left electrode, the right electrode, and the scattering region with each of the sugar molecules. The k-point grid for the electrodes and the scattering region calculation is 2 × 2 × 135. The computed transmission probability of the electrons with energy (E) is generated, as shown in Equation [1]:[1] TE=TrΓREξRΓLEξAE Here, ΓLE and ΓRE are the broadening matrix for the left and right electrodes, respectively. ξA and ξR refer to the advanced and retarded Green’s function, respectively. The zero bias conductance is generated with the relation ξ=ξ0TEF, where ξ0=2e2/h is the quantum conductance. E and h refer to the electron charge and Planck’s constant, respectively. The difference of the Fermi functions is used to calculate the integration of TE,V over the energy window fS, DE=1+expE−EF−eVS,D/kBT−1, which gives the total current displayed in Equation [2]:[2]$I = 2$eh∫−∞∞dE TE,VfSE−fDE QuantumATK generates the density of state based on the following equations [58]: The DeviceDensityOfStates (DDOS) DE is computed via the spectral density matrix σE=σLE+σRE, where L/R refers to the left and right electrodes. The local density of states (LDOS) is computed as:[3]DE,r=∑ijσijE∅ir∅jr The basis set orbitals ∅ir are real functions in QuantumATK through the use of solid harmonics. The device density of state is then obtained by integrating LDOS over all space:[4]DE=∫drDE,r=∑ijσijESij where, Sij=∫∅ir∅jrdr is the overlap matrix. Introducing MiE=∑jσijESij, the equation can be written as [5]DE=∑iMiE where MiE is considered as the contribution of DDOS from orbital i. MiE is a spectral Mulliken Population with:[6]Mi=∫MiEf(E−μkBT) dE ## 3. Results and Discussion The electrical transport properties were generated for the C2N FET to achieve the practical investigation of the designed C2N FET sensor to specifically detect each of the sugar molecules. Density of states, work function, transmission spectrum, current variation, and current-voltage characteristics were generated for the C2N FET, the C2N FET with the presence of glucose molecule, the C2N FET with the presence of fructose molecule, and for the C2N FET with the presence of xylose molecule. ## 3.1. Device Density of States (DDOS) A distinct and significant change in the FET Device DOS have been noticed in the presence of the different sugar molecules. Figure 4 displays a comparison of the DDOS for the bare C2N FET (without any target molecule) and for the C2N FET in the presence of each of the sugar molecules. Figure 4a shows that the bare C2N FET have more energy states than the C2N FET in the presence of glucose molecule, which can be observed at the energy levels of −3.8, −3.6, −3.2, and −2.9 eV. Furthermore, the presence of fructose molecule affected the C2N FET DOS differently, as displayed in Figure 4b, where a new energy spike can be observed at energy level 3.85 eV. Similarly, a significant change in DOS can be noticed in the C2N FET when it is exposed to xylose molecule, as displayed in Figure 4c. Two new energy spikes were noticed at energy levels of 3.7 and 3.9 eV, as shown in Figure 4c. Figure 5 displays the partial DOS, which reflects a closer look and more detailed information about the effect of each of the sugar molecules on the DDOS. It was noticed that, when a target molecule is added to the device, one unique peak is increased in the DDOS due to glucose (Figure 5a) or fructose (Figure 5b) or xylose (Figure 5c). This indicates that adding each of the sugar molecules results in new electronic states within the energy range of that peak. This may indicate that the sugar molecule is interacting with the C2N channel and modifying its electronic structure. The change in the DDOS is caused by the sugar molecule accepting or donating electrons from the channel material or by forming chemical bonds between the target molecule and the C2N channel. The DOS of a material is defined as the measure of the number of available electronic states within a certain energy range. The DOS changes due to the presence of various types of molecules since they can introduce defects of impurities into the material, which leads to a change in the electronic structure of the material. As an example, when the material is exposed to a target molecule, the impurities can result in additional energy levels, which can modify the density of states. Moreover, impurities affect electronic states symmetry, which modifies the DOS. Additionally, the mechanical and chemical properties can be changed leading to a change in the DOS. The variation in the DOS depends on the type and concentration of the defects. ## 3.2. Work Function The C2N FET response to each of the sugar molecules is investigated by calculating the work function displayed in Figure 6. The calculated work function value for the C2N FET is 5.92 eV; for the C2N FET with glucose, it is 6.08 eV; for the C2N FET with fructose, it is 6.05 eV; and for the C2N FET with xylose, it is 6.106. Figure 6 shows an increment in the work function for the C2N FET with each of the sugar molecules in comparison to the bare C2N FET. This increment indicates that the adsorption of each of the sugar molecules leads to a decrement in the electron mobility. The work function increment is caused by the cloud charge transfers from the C2N channel toward the sugar molecules. The study’s findings are in line with previous research work [53]. The increment in the work function of C2N due to presence of each of the sugar molecules is believed to be associated with changes in the electronic characteristics of the C2N material due to the interaction between each of the target molecules and the C2N surface [53]. The energy needed to remove an electron from the surface can increase when the target molecule accepts electrons from the C2N material, increasing the work function. Moreover, it is expected that the movement of charge carriers from the C2N material to the sugar molecules leads to a decrement in the density near Fermi level. Thus, the Fermi level shifts to higher energies leading to an increment in the work function. ## 3.3. Transmission Spectrum Figure 7 shows the transmission spectra T(E) for the C2N FET with and without each of the sugar molecules (glucose, fructose, and xylose) at different biases: (a) $V = 0$ V, (b) $V = 0.2$ V, and $V = 0.4$ V. The figure shows the changes in transmission spectrum when different sugar molecules are added at a varying applied voltage. The transmission spectrum has a low value in the energy range [0.4, 0.9] eV because of the energy window within the band gap of the semiconducting C2N channel. ## 3.4. Current-Voltage The current vs. voltage characteristics for the C2N FET sensor and for each of the sugar molecules adsorbed via the C2N FET sensor are shown in Figure 8. A fixed 1 V gate potential was used while the Vds was set to 0.2, 0.4, 0.6, 0.8, and 1 V. Figure 8 shows the current voltage curves for C2N FET at 0.2, 0.4, 0.6, 0.8, 1 V before and after the addition of each of the sugar molecules. The variation in current reading with the addition of sugar molecules indicates successful detection. The adsorbed target molecule interacts with the C2N-FET and changes its conductivity by changing the carriers’ concentration. C2N is a semiconducting nanomaterial, which has a nonlinear resistance, resulting in a nonlinear IV curve, as shown in Figure 8. The current of the C2N-FET differs noticeably for each sugar molecule. The size, electrical state, and way that each sugar molecule interacts with the C2N-FET channel are all unique. When the gate potential was fixed at 1 V and the bias voltage among the left and right electrodes was fixed at 0.4 V, the created sensor produced the best results. The best sensitivity was achieved by setting the bias voltage at 0.4 V, as shown in Figure 8. This work is a proof of concept that the developed C2N-FET can be utilized to detect the different types of sugar molecules. The sensor showed the best sensitivity at 0.4 V bias voltage. Figure 9 shows the sensor’s response (change in current), where the highest variation in the electrical signal was due to glucose molecule adsorption. These results show that the device has high selectivity for glucose and results in a distinct electrical current for each of the sugar molecules. The current variation is due to the change in the charge and the electrical potential after introducing the target molecules which alters the charge carriers’ density. Thus, the sensor conductivity and current change. This work is a proof of concept that the modeled and studied C2N FET can be utilized as a sensor for sugar molecules detection, such as glucose, fructose, and xylose. This research indicates that each of the sugar molecules have a unique electronic signature that can be identified via the designed C2N FET. After employing the computational methods to detect each of the sugar molecules, the results of the sensor can be utilized to identify the performance in real-time applications. Such computational methods provide valuable results, such as how each of the sugar molecules will interact with the sensor. These results can be utilized to optimize the sensor design and performance in terms of stability, sensitivity, and selectivity. Moreover, the used computational method provides insights into the electronic transport characteristics of the system due to each of the sugar molecules. These electronic properties include electron density, work function, transmission spectrum, and current–voltage measurements. These results can be utilized to understand how the target molecule interacts with the sensor and affects the sensor’s performance. After the validation of the sensor via computational methods, the sensor can be designed, fabricated, and tested in real-time applications. Then, the sensor’s performance can be evaluated by comparing the computational expectations with the experimental findings. In this work, C2N FET was utilized to detect each one of the sugar molecules separately, where each one of them resulted in a unique electronic signature and unique variation in current indicating the possibility of detecting each of them in real-time applications. The highest sensitivity was toward glucose molecule, which can be used to monitor and control diabetes. The addition of a mixture of two or three sugar molecules is also expected to result in a specific variation in current and a unique electronic signature, since each sugar molecule interacts with the C2N channel and modifies its electronic properties in a unique way. In general, the employed computational method results in valuable information about the performance of the designed sensor. However, computational methods do not show the limit of detection of the sensor in real-time applications. The limit of detection can only be identified by experiment by measuring the response of the fabricated sensor toward various concentration of the target analyte. Combining both computational method with experimental data can be used to overcome the limitations of such technology. This comparison leads to identifying the potential sources of error and uncertainty. It is worth mentioning that the employment of computational methods enables researchers to suggest future directions to study to enhance the sensor’s performance and then test it experimentally. Introducing structure variables, such as surface roughness, pores, and alien molecules to a sensor, will result in a significant effect on its electronic properties and performance. In terms of work function, the existence of surface roughness or defects lead to a shift in the work function. Moreover, surface roughness can also affect the amount of charge that can be stored on the device, which affects its sensitivity. In terms of density of states, impurities and defects can generate localized states within the bandgap of the sensor, which can modify electrical current and the conductivity of the device. Moreover, surface roughness and pores can affect the DOS by creating additional pathways for charge carriers to pass through. In terms of current, impurities and defects work as scattering centers for the charge carriers, which might lead to a reduction in the device mobility and current. Introducing structure variables impact the electronic properties and performance of a sensor, affecting its work function, density of states, and current. ## 4. Conclusions Real-time identification of the different sugar molecules is essential for monitoring and preventing diabetes and to evaluate food quality. In this research, a field effect transistor based on two-dimensional nitrogenated holey graphene (C2N) was designed, developed, and tested to identify the sugar molecules, including xylose, fructose, and glucose. To investigate the characteristics of this device, non-equilibrium Green’s function and density functional theory (NEGF + DFT) were utilized. Various electronic properties were studied, including density of states, work function, transmission spectrum, and electrical current. The proposed sensor consists of a pair of gold electrodes connected via a channel of C2N and a gate. The electronic characteristics of the C2N FET changed because of the adsorption of the target molecules. The measurable variations in the electronic characteristics with each sugar molecule validate the potential of the C2N FET sensor in detecting sugar molecules. The C2N monolayer distinctive characteristics are promising for glucose sensors to detect blood sugar. ## References 1. Liu K., Wang X., Luo B., Wang C., Hou P., Dong H., Li A., Zhao C.. Enzyme-Free Electrochemical Sensors for In Situ Quantification of Reducing Sugars Based on Carboxylated Graphene–Carboxylated Multiwalled Carbon Nanotubes–Gold Nanoparticle–Modified Electrode. Front. Plant Sci. (2022.0) 13. DOI: 10.3389/fpls.2022.872190 2. Crespo-Rosa J.R., Foca G., Ulrici A., Pigani L., Zanfrognini B., Cubillana-Aguilera L., Palacios-Santander J.M., Zanardi C.. Simultaneous Detection of Glucose and Fructose in Synthetic Musts by Multivariate Analysis of Silica-Based Amperometric Sensor Signals. Sensors (2021.0) 21. DOI: 10.3390/s21124190 3. Rippe J.M., Angelopoulos T.J.. Sugars and Health Controversies: What Does the Science Say?. Adv. Nutr. (2015.0) 6 493S-503S. DOI: 10.3945/an.114.007195 4. Rippe J.M., Angelopoulos T.J.. Sucrose, High-Fructose Corn Syrup, and Fructose, Their Metabolism and Potential Health Effects: What do We Really Know?. Adv. Nutr. (2013.0) 4 236-245. DOI: 10.3945/an.112.002824 5. Association A.D.. Diagnosis and Classification of Diabetes Mellitus. Diabetes Care (2010.0) 33 S62-S69. DOI: 10.2337/dc10-S062 6. Sehit E., Altintas Z.. Significance of nanomaterials in electrochemical glucose sensors: An updated review (2016–2020). Biosens. Bioelectron. (2020.0) 159 112165. DOI: 10.1016/j.bios.2020.112165 7. Gregg E.W., Sattar N., Ali M.K.. The changing face of diabetes complications. Lancet Diabetes Endocrinol. (2016.0) 4 537-547. DOI: 10.1016/S2213-8587(16)30010-9 8. Nathan D.M.. Long-Term Complications of Diabetes Mellitus. N. Engl. J. Med. (1993.0) 328 1676-1685. DOI: 10.1056/NEJM199306103282306 9. Lee A.K., Warren B., Lee C.J., McEvoy J.W., Matsushita K., Huang E.S., Sharrett A.R., Coresh J., Selvin E.. The Association of Severe Hypoglycemia with Incident Cardiovascular Events and Mortality in Adults with Type 2 Diabetes. Diabetes Care (2017.0) 41 104-111. DOI: 10.2337/dc17-1669 10. Gubala V., Harris L.F., Ricco A.J., Tan M.X., Williams D.E.. Point of Care Diagnostics: Status and Future. Anal. Chem. (2012.0) 84 487-515. DOI: 10.1021/ac2030199 11. Mani V., Devasenathipathy R., Chen S.-M., Wang S.-F., Devi P., Tai Y.. Electrodeposition of copper nanoparticles using pectin scaffold at graphene nanosheets for electrochemical sensing of glucose and hydrogen peroxide. Electrochim. Acta (2015.0) 176 804-810. DOI: 10.1016/j.electacta.2015.07.098 12. Dong X.-C., Xu H., Wang X.-W., Huang Y.-X., Chan-Park M.B., Zhang H., Wang L.-H., Huang W., Chen P.. 3D Graphene–Cobalt Oxide Electrode for High-Performance Supercapacitor and Enzymeless Glucose Detection. ACS Nano (2012.0) 6 3206-3213. DOI: 10.1021/nn300097q 13. State S., Enache L.-B., Potorac P., Prodana M., Enachescu M.. Synthesis of Copper Nanostructures for Non-Enzymatic Glucose Sensors via Direct-Current Magnetron Sputtering. Nanomaterials (2022.0) 12. DOI: 10.3390/nano12234144 14. Sakr M.A., Elgammal K., Delin A., Serry M.. Performance-Enhanced Non-Enzymatic Glucose Sensor Based on Graphene-Heterostructure. Sensors (2020.0) 20. DOI: 10.3390/s20010145 15. Sternberg R., Barrau M.-B., Gangiotti L., Thévenot D.R., Bindra D.S., Wilson G.S., Velho G., Froguel P., Reach G.. Study and development of multilayer needle-type enzyme-based glucose microsensors. Biosensors (1989.0) 4 27-40. DOI: 10.1016/0265-928X(89)80032-X 16. Huang Z., Zheng L., Feng F., Chen Y., Wang Z., Lin Z., Lin X., Weng S.. A Simple and Effective Colorimetric Assay for Glucose Based on MnO. Sensors (2018.0) 18. DOI: 10.3390/s18082525 17. Wang C., Li J., Tan R., Wang Q., Zhang Z.. Colorimetric method for glucose detection with enhanced signal intensity using ZnFe. Analyst (2019.0) 144 1831-1839. DOI: 10.1039/C8AN02330A 18. Wu X., Yin J., Liu J., Gu Y., Wang S., Wang J.. Colorimetric detection of glucose based on the binding specificity of a synthetic cyclic peptide. Analyst (2020.0) 145 7234-7241. DOI: 10.1039/D0AN00211A 19. Klonoff D.. Overview of Fluorescence Glucose Sensing: A Technology with a Bright Future. J. Diabetes Sci. Technol. (2012.0) 6 1242-1250. DOI: 10.1177/193229681200600602 20. Gao J., Zhou J., Qu X.. Fluorometric Method for Quantitative Determination of Glucose and Its Application to Human Serum. Anal. Sci. (2005.0) 21 409-412. DOI: 10.2116/analsci.21.409 21. Wang C., Tan R., Li L., Liu D.. Dual-Modal Colorimetric and Fluorometric Method for Glucose Detection Using MnO. Chem. Res. Chin. Univ. (2019.0) 35 767-774. DOI: 10.1007/s40242-019-9130-5 22. Ilacas G.C., Basa A., Nelms K.J., Sosa J.D., Liu Y., Gomez F.A.. Paper-based microfluidic devices for glucose assays employing a metal-organic framework (MOF). Anal. Chim. Acta (2019.0) 1055 74-80. DOI: 10.1016/j.aca.2019.01.009 23. Hou Y., Lv C.-C., Guo Y.-L., Ma X.-H., Liu W., Jin Y., Li B.-X., Yang M., Yao S.-Y.. Recent Advances and Applications in Paper-Based Devices for Point-of-Care Testing. J. Anal. Test. (2022.0) 6 247-273. DOI: 10.1007/s41664-021-00204-w 24. Kasetsirikul S., Shiddiky M., Nguyen N.-T.. Challenges and perspectives in the development of paper-based lateral flow assays. Microfluid. Nanofluidics (2020.0) 24 17. DOI: 10.1007/s10404-020-2321-z 25. Heller A., Feldman B.. Electrochemical Glucose Sensors and Their Applications in Diabetes Management. Chem. Rev. (2008.0) 108 2482-2505. DOI: 10.1021/cr068069y 26. Scholtes-Timmerman M.J., Bijlsma S., Fokkert M.J., Slingerland R., van Veen S.J.F.. Raman Spectroscopy as a Promising Tool for Noninvasive Point-of-Care Glucose Monitoring. J. Diabetes Sci. Technol. (2014.0) 8 974-979. DOI: 10.1177/1932296814543104 27. Wahjudi P.N., Patterson M.E., Lim S., Yee J.K., Mao C.S., Lee W.N.P.. Measurement of glucose and fructose in clinical samples using gas chromatography/mass spectrometry. Clin. Biochem. (2010.0) 43 198-207. DOI: 10.1016/j.clinbiochem.2009.08.028 28. Wu G., Dai Z., Tang X., Lin Z., Lo P.K., Meyyappan M., Lai K.W.C.. Biosensing: Graphene Field-Effect Transistors for the Sensitive and Selective Detection of. Adv. Healthc. Mater. (2017.0) 6 1700736. DOI: 10.1002/adhm.201700736 29. Piccinini E., Bliem C., Reiner-Rozman C., Battaglini F., Azzaroni O., Knoll W.. Enzyme-polyelectrolyte multilayer assemblies on reduced graphene oxide field-effect transistors for biosensing applications. Biosens. Bioelectron. (2017.0) 92 661-667. DOI: 10.1016/j.bios.2016.10.035 30. Fu W., Jiang L., van Geest E.P., Lima L.M.C., Schneider G.F.. Sensing at the Surface of Graphene Field-Effect Transistors. Adv. Mater. (2017.0) 29 1603610. DOI: 10.1002/adma.201603610 31. Zhan B., Li C., Yang J., Jenkins G., Huang W., Dong X.. Graphene Field-Effect Transistor and Its Application for Electronic Sensing. Small (2014.0) 10 4042-4065. DOI: 10.1002/smll.201400463 32. Fenoy G., Marmisollé W., Azzaroni O., Knoll W.. Acetylcholine biosensor based on the electrochemical functionalization of graphene field-effect transistors. Biosens. Bioelectron. (2019.0) 148 111796. DOI: 10.1016/j.bios.2019.111796 33. Wasfi A., Awwad F., Gelovani J.G., Qamhieh N., Ayesh A.I.. COVID-19 Detection via Silicon Nanowire Field-Effect Transistor: Setup and Modeling of Its Function. Nanomaterials (2022.0) 12. DOI: 10.3390/nano12152638 34. Zhang X., Jing Q., Ao S., Schneider G.F., Kireev D., Zhang Z., Fu W.. Ultrasensitive Field-Effect Biosensors Enabled by the Unique Electronic Properties of Graphene. Small (2020.0) 16 1902820. DOI: 10.1002/smll.201902820 35. Tan X., Yang M., Zhu L., Gunathilaka G., Zhou Z., Chen P.Y., Zhang Y., Cheng M.M.C.. Ultrasensitive and Selective Bacteria Sensors Based on Functionalized Graphene Transistors. IEEE Sens. J. (2022.0) 22 5514-5520. DOI: 10.1109/JSEN.2022.3147229 36. Verhulst A.S., Ruić D., Willems K., Dorpe P.V.. Boosting the Sensitivity of the Nanopore Field-Effect Transistor to Translocating Single Molecules. IEEE Sens. J. (2022.0) 22 5732-5742. DOI: 10.1109/JSEN.2022.3149345 37. Cherik I.C., Mohammadi S.. Dielectric Modulated Doping-Less Tunnel Field-Effect Transistor, a Novel Biosensor Based on Cladding Layer Concept. IEEE Sens. J. (2022.0) 22 10308-10314. DOI: 10.1109/JSEN.2022.3163475 38. Wu H., Wang J., Kang X., Wang C., Wang D., Liu J., Aksay I.A., Lin Y.. Glucose biosensor based on immobilization of glucose oxidase in platinum nanoparticles/graphene/chitosan nanocomposite film. Talanta (2009.0) 80 403-406. DOI: 10.1016/j.talanta.2009.06.054 39. Mishra A.K., Jarwal D.K., Mukherjee B., Kumar A., Ratan S., Tripathy M.R., Jit S.. Au nanoparticles modified CuO nanowire electrode based non-enzymatic glucose detection with improved linearity. Sci. Rep. (2020.0) 10 11451. DOI: 10.1038/s41598-020-67986-4 40. Taguchi M., Ptitsyn A., McLamore E.S., Claussen J.C.. Nanomaterial-mediated Biosensors for Monitoring Glucose. J. Diabetes Sci. Technol. (2014.0) 8 403-411. DOI: 10.1177/1932296814522799 41. Kwak Y.H., Choi D.S., Kim Y.N., Kim H., Yoon D.H., Ahn S.-S., Yang J.-W., Yang W.S., Seo S.. Flexible glucose sensor using CVD-grown graphene-based field effect transistor. Biosens. Bioelectron. (2012.0) 37 82-87. DOI: 10.1016/j.bios.2012.04.042 42. Scandurra A., Ruffino F., Sanzaro S., Grimaldi M.G.. Laser and thermal dewetting of gold layer onto graphene paper for non-enzymatic electrochemical detection of glucose and fructose. Sens. Actuators B Chem. (2019.0) 301 127113. DOI: 10.1016/j.snb.2019.127113 43. Juſík T., Podešva P., Farka Z., Kováſ D., Skládal P., Foret F.. Nanostructured gold deposited in gelatin template applied for electrochemical assay of glucose in serum. Electrochim. Acta (2016.0) 188 277-285. DOI: 10.1016/j.electacta.2015.12.009 44. Kwon S.-Y., Kwen H.-D., Choi S.-H.. Fabrication of Nonenzymatic Glucose Sensors Based on Multiwalled Carbon Nanotubes with Bimetallic Pt-M (M = Ru and Sn) Catalysts by Radiolytic Deposition. J. Sens. (2012.0) 2012 784167. DOI: 10.1155/2012/784167 45. Tortorich R.P., Shamkhalichenar H., Choi J.-W.. Inkjet-Printed and Paper-Based Electrochemical Sensors. Appl. Sci. (2018.0) 8. DOI: 10.3390/app8020288 46. Hwang H.S., Jeong J.W., Kim Y.A., Chang M.. Carbon Nanomaterials as Versatile Platforms for Biosensing Applications. Micromachines (2020.0) 11. DOI: 10.3390/mi11090814 47. Rauti R., Musto M., Bosi S., Prato M., Ballerini L.. Properties and behavior of carbon nanomaterials when interfacing neuronal cells: How far have we come?. Carbon (2019.0) 143 430-446. DOI: 10.1016/j.carbon.2018.11.026 48. Lobov I.A., Davletkildeev N.A., Nesov S.N., Sokolov D.V., Korusenko P.M.. Effect of Nitrogen Atoms in the CNT Structure on the Gas Sensing Properties of PANI/CNT Composite. Appl. Sci. (2022.0) 12. DOI: 10.3390/app12147169 49. Dandu N.K., Chandaluri C.G., Ramesh K., Saritha D., Mahender Reddy N., Ramesh G.V., Dave S., Das J., Ghosh S.. Chapter 11—Carbon nanomaterials: Application as sensors for diagnostics. Advanced Nanomaterials for Point of Care Diagnosis and Therapy (2022.0) 211-248 50. Yang C., Denno M.E., Pyakurel P., Venton B.J.. Recent trends in carbon nanomaterial-based electrochemical sensors for biomolecules: A review. Anal. Chim. Acta (2015.0) 887 17-37. DOI: 10.1016/j.aca.2015.05.049 51. Xu B., Xiang H., Wei Q., Liu J.Q., Xia Y.D., Yin J., Liu Z.G.. Two-dimensional graphene-like C. Phys. Chem. Chem. Phys. (2015.0) 17 15115-15118. DOI: 10.1039/C5CP01789K 52. Hussain T., Sajjad M., Singh D., Bae H., Lee H., Larsson J.A., Ahuja R., Karton A.. Sensing of volatile organic compounds on two-dimensional nitrogenated holey graphene, graphdiyne, and their heterostructure. Carbon (2020.0) 163 213-223. DOI: 10.1016/j.carbon.2020.02.078 53. Panigrahi P., Sajjad M., Singh D., Hussain T., Andreas Larsson J., Ahuja R., Singh N.. Two-dimensional Nitrogenated Holey Graphene (C. Appl. Surf. Sci. (2022.0) 573 151579. DOI: 10.1016/j.apsusc.2021.151579 54. Chang P.-H., Liu H., Nikolić B.K.. First-principles versus semi-empirical modeling of global and local electronic transport properties of graphene nanopore-based sensors for DNA sequencing. J. Comput. Electron. (2014.0) 13 847-856. DOI: 10.1007/s10825-014-0614-8 55. Why Are So Many k-Points Needed in the Transport Direction in a Device Calculation? 56. Kaur J., Kumar R., Vohra R., Sawhney R.S.. Density functional theory investigations on the interaction of uracil with borospherene. Bull. Mater. Sci. (2022.0) 45 22. DOI: 10.1007/s12034-021-02595-z 57. Thomas S., Kumar V., Roy D.R., Zaeem M.A.. Two-Dimensional Boron–Phosphorus Monolayer for Reversible NO. ACS Appl. Nano Mater. (2020.0) 3 10073-10081. DOI: 10.1021/acsanm.0c02072 58. Device Density of State
--- title: Characterization and microRNA Expression Analysis of Serum-Derived Extracellular Vesicles in Severe Liver Injury from Chronic HBV Infection authors: - Min Liu - Xionghao Liu - Mengmeng Pan - Yu Zhang - Xiangling Tang - Wanxi Liu - Mingri Zhao - Jing Ma - Ning Zhou - Yongfang Jiang - Wenlong Wang - Mujun Liu journal: Life year: 2023 pmcid: PMC9967308 doi: 10.3390/life13020347 license: CC BY 4.0 --- # Characterization and microRNA Expression Analysis of Serum-Derived Extracellular Vesicles in Severe Liver Injury from Chronic HBV Infection ## Abstract Background: Extracellular vesicle (EV) microRNAs have been documented in several studies to have significantly different expressions in hepatitis B virus (HBV)-related liver diseases, such as hepatocellular carcinoma (HCC). The current work aimed to observe the characteristics of EVs and EV miRNA expressions in patients with severe liver injury chronic hepatitis B (CHB) and patients with HBV-associated decompensated cirrhosis (DeCi). Methods: The characterization of the EVs in the serum was carried out for three different groups, namely, patients with severe liver injury-CHB, patients with DeCi, and healthy controls. EV miRNAs were analyzed using miRNA-seq and RT-qPCR arrays. Additionally, we assessed the predictive and observational values of the miRNAs with significant differential expressions in serum EVs. Results: Patients with severe liver injury-CHB had the highest EV concentrations when compared to the normal controls (NCs) and patients with DeCi ($p \leq 0.001$). The miRNA-seq of the NC and severe liver injury-CHB groups identified 268 differentially expressed miRNAs (\|FC\| > 2, $p \leq 0.05$). In this case, 15 miRNAs were verified using RT-qPCR, and it was found that novel-miR-172-5p and miR-1285-5p in the severe liver injury-CHB group showed marked downregulation in comparison to the NC group ($p \leq 0.001$). Furthermore, compared with the NC group, three EV miRNAs (novel-miR-172-5p, miR-1285-5p, and miR-335-5p) in the DeCi group showed various degrees of downregulated expression. However, when comparing the DeCi group with the severe liver injury-CHB group, only the expression of miR-335-5p in the DeCi group decreased significantly ($p \leq 0.05$). For the severe liver injury-CHB and DeCi groups, the addition of miR-335-5p improved the predictive accuracy of the serological levels, while miR-335-5p was significantly correlated with ALT, AST, AST/ALT, GGT, and AFP. Conclusions: The patients with severe liver injury-CHB had the highest number of EVs. The combination of novel-miR-172-5p and miR-1285-5p in serum EVs helped in predicting the progression of the NCs to severe liver injury-CHB, while the addition of EV miR-335-5p improved the serological accuracy of predicting the progression of severe liver injury-CHB to DeCi. ## 1. Introduction Chronic hepatitis B (CHB) is a persistent inflammatory disease of the liver caused by hepatitis B virus (HBV) infection. Following infection with HBV, inflammation induces hepatocyte necrosis, which is a crucial pathophysiological process of disease progression. CHB will progress to severe liver damage if not managed promptly and effectively. Liver diseases constitute an enormous healthcare burden and are a growing cause of concern [1]. The World Health Organization (WHO) estimates that, in 2019, 296 million people worldwide were living with CHB. Further, hepatitis B caused approximately 820,000 deaths, mostly attributable to cirrhosis liver decompensation (DeCi) and hepatocellular carcinoma (HCC) (https://www.who.int/news-room/fact-sheets/detail/hepatitis-b, accessed on 24 June 2022). The five-year mortality rate of patients with decompensated cirrhosis without liver transplantation is up to $85\%$ [2]. Chronic HBV infection can be classified into five phases: (I) HBeAg-positive chronic infection, (II) HBeAg-positive chronic hepatitis, (III) HBeAg-negative chronic infection, (IV) HBeAg-negative chronic hepatitis, and (V) HBsAg-negative phase [3]. HBeAg-positive chronic hepatitis B is characterized by the presence of serum HBeAg, high levels of HBV DNA, and elevated ALT. In the liver, there is moderate or severe liver necroinflammation and the accelerated progression of fibrosis [4]. Patients with compensated cirrhosis are asymptomatic or have mild clinical symptoms, with the appearance of complications with portal hypertension, ascites, hepatic encephalopathy, varicose bleeding, etc., indicating that the disease has progressed to the decompensated stage [5,6]. Serological tests, liver aspiration, and imaging can be used to help determine the course of the disease during this period. Blood tests include measurements of transaminases (ALT/AST), fibrosis-related factors, HBV DNA levels, and the presence or absence of HBV antigens [7]. They have limited effectiveness in indicating the progression of liver disease [8]. Liver biopsy is invasive and has the potential to cause serious complications, including bleeding, bile peritonitis, and pain, and it has the potential to cause diagnostic mistakes attributable to the heterogeneous distribution [9,10]. Liver stiffness with transient elastography (TE, FibroScan) and magnetic resonance elastography (MRE) are proven methods to assess liver fibrosis and cirrhosis. However, they are expensive, available only at certain limited liver health centers, and are inappropriate for patients with MR contraindications [11,12,13]. Therefore, the identification of a new non-invasive biomarker or scoring system to predict the severity of liver disease and improve current methods is urgently needed. The presence of extracellular vesicles (EVs) with bilayer membrane structures in the circulatory system can protect target molecules and reflect the pathological or physiological conditions of the original cells or tissues [14,15,16]. EVs and their inclusions vary with physiological states and physiological conditions in the organism [17,18]. According to their biogenesis process, EVs can be further divided into exosomes, micro-vesicles, and apoptotic bodies [19]. Exosomes and the nucleic acids and proteins that they contain are regarded as types of liquid biopsies, which boast the advantages of non-invasiveness and dynamic monitoring, and they are able to overcome the limitations of tumor spatiotemporal heterogeneity [15,16]. EVs are attractive biomarkers for estimating the severity and prognosis of liver diseases, including chronic viral hepatitis B infections, cirrhosis, primary liver cancers, non-alcoholic and alcoholic steatohepatitis, and acute liver failure [20]. A study examining the potential diagnostic utility of EVs in NAFLD reported that 65 patients with NAFLD had higher concentrations of T-cell and monocyte-derived EVs than healthy controls [21]. In various studies, it has been found that several subpopulations of leukocyte-derived EVs (CD45+, CD11a+, and CD4+) are more abundant in patients with cirrhosis than in healthy individuals [22,23,24]. This may be related to the systemic inflammation of cirrhosis rather than to the severity of cirrhosis. In this case, miRNA is a non-coding RNA molecule with a length of approximately 22 nucleotides. It inhibits gene expression by binding to the 3′ untranslated region (UTR) of a target mRNA [25]. Extracellular miRNAs are mainly secreted by specific tissues or cells and packaged into micro-vesicles, protecting miRNAs from degradation by RNases in blood [26]. EV miRNAs can act as excellent non-invasive biomarkers to evaluate disease processes, progression, and treatment responses [27,28,29]. Several studies indicated that the expression levels of some EV miRNAs in blood are significantly altered in adults with liver diseases, such as liver injury, liver cancer, and viral hepatitis [30,31]. Two miRNAs, miRNA-122 (a major hepatic miRNA) and miRNA-192, have been found to increase with fibrosis stage in serum EVs [32,33,34]. Of note, as fewer than 10 patients were included in each study, this result requires further validation. miRNA-21 expression has been found to be higher in the circulating serum EVs of patients with HCC ($$n = 30$$) than in patients with CHB without cirrhosis and healthy controls ($$n = 30$$ in each group) [35]. The expression levels of miRNA-122 in circulating serum EVs can distinguish early-stage HCC from liver cirrhosis (LC) [36]. The search for EV miRNAs that can aid in predicting whether patients with CHB will rapidly progress to DeCi disease is essential. In this study, we collected samples from patients with severe liver injury-CHB and DeCi and performed a systematic observation and a comparative analysis of the EVs and their miRNAs for both diseases to provide ideas for subsequent research. ## 2.1. Participants and Sample Collection A total of 58 patients with severe liver injury-CHB (HBeAg-positive chronic hepatitis B) and 33 patients with DeCi were enrolled in the present study. All patients were hospitalized in the Infectious Diseases Department of the Second Xiangya Hospital of Central South University (Changsha, China) from April 2015 to October 2020, and they were diagnosed according to the guidelines of Hepatitis B Virus-related Cirrhosis Clinical Management [37]. We established exclusion criteria to reduce the effects of other physical or pathological factors on the production of serum EV miRNAs. The exclusion criteria were as follows: patients with alcoholic cirrhosis, cirrhosis in the compensatory phases, chronic hepatitis C or D, other chronic liver diseases (such as autoimmune hepatitis, hepatolenticular degeneration, severe organ dysfunction, and hematological diseases), autoimmune diseases, malignant tumor diseases, and mental disorders. The clinical data of the patients are summarized in Table 1. We also selected 58 normal controls (NCs) who had visited the Second Xiangya Hospital of Central South University as the parallel controls. Blood samples were collected from all patients and healthy controls. To remove cell debris, serum separation was accomplished via centrifugation at 300× g for 5 min at 4 °C within 2 h, and then the samples were stored at −80 °C until the EVs separated. ## 2.2. Serum EV Isolation The serum from 5 mL of the blood sampled from each patient was centrifuged at 1500× g for 10 min to eliminate cell precipitation and debris, and then the cell-free supernatant was centrifuged at 15,000× g for 30 min to remove large EVs. Five pools, each from the NC group and the severe liver injury-CHB group, were used for NGS sequencing, and every pool contained a mixture of samples from 5 random individuals (400 µL/sample) of the same group. For the RT qPCR analysis, 1 mL of serum was taken from each sample for detection. For NTA, TEM, and WB analyses, 0.5 mL of serum was taken for detection. Each pool or sample was filtered through a 0.22 μm pore filter (EMD Millipore, Billerica, MA, USA) to remove any larger particles. Next, each sample or pool of samples was centrifuged in a 100 KD centrifuge concentrator tube (EMD Millipore) at 3000× g for 30 min at 4 °C. Eventually, the supernatant from the previous procedure was enriched with $\frac{1}{4}$ volume of ExoQuick-5A™ (System Biosciences, Inc., Palo Alto, Santa Clara, CA, USA) according to the manufacturer’s instructions. Finally, the EVs extracted were resuspended with a specific volume of phosphate-buffered saline (PBS) for different purposes and stored at −80 °C. ## 2.3. Characterization and Quantification of EVs A nanoparticle tracking analysis (NTA) was carried out to track EV diameter and concentration. All samples were diluted with PBS before the NTA analysis, and then 100 µL of the sample was loaded into the EV analysis chamber of the Zetaview equipment PMX 110 (Particle Metrix, Meerbusch, Germany). Microsoft Excel 2013 (Microsoft Corp., Seattle, WA, USA) was used to handle the data obtained from Zetaview 8.04.02 SP2, whose parameters were optimized appropriately during the experiment. For the transmission electron microscopy (TEM), 20 µL of the EV suspension was dropped into a carbon-coated 200 mesh copper grid in the form of a droplet for a period of time (more than 1 min). Subsequently, $2\%$ uranyl acetate solution was added to yield negative staining, which was fixed in 10 min. The unnecessary liquid was removed using filter paper at room temperature, and then a Tecnai biological transmission electron microscope (model: Tecnai G2 Spirit, Thermo Fisher Scientific, Waltham, MA, USA) was used to obtain micrographs at 80 KV. Further, we extracted EV proteins using RIPA Lysis Buffer (Beyotime, Shanghai, China) and a Protease Inhibitor Cocktail (Sigma-Aldrich, Darmstadt, Germany) for the Western blotting analysis. The EV proteins were then quantified using a Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific). Next, 15 µg of the EV proteins was loaded into each well of a $10\%$ sodium dodecyl sulphate polyacrylamide gel for electrophoresis (SDS-PAGE) (Beyotime, Shanghai, China), followed by separation under the condition of 180 V (Tannon, Shanghai, China). Next, The EV proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Merck KGaA, Darmstadt, Germany) for 1.5 h with a 252 mA current. We purchased the following anti-bodies from Santa Cruz (Santa Cruz, CA, USA): anti-CD63 (sc-5275), CD9 (sc-13118), TSG101 (sc-7964), and calnexin (sc-23954). After blocking with $5\%$ non-fat dry milk in 1 × TBST (TBS, $0.1\%$ Tween 20) buffer for 1 h, the membrane was incubated in a primary antibody solution dissolved with 1 × TBST containing $5\%$ non-fat dry milk (CD63, 1:2000; CD9, 1:2000; TSG101, 1:2000; and calnexin, 1:2000) at 4 °C overnight. The membrane was washed three times with 1 × TBST for 10 min. Next, the secondary antibody at a dilution of 1:10,000 was incubated at room temperature for 1.5 h and rewashed three times. The target protein was detected with an enhanced chemiluminescence kit (Super-Signal™ West Femto Maximum Sensitivity Substrate, Thermo Fisher Scientific, MA, USA) using a Bio-Rad imaging system according to the manufacturer’s instructions. ## 2.4. RNA Extraction from Serum EVs According to the manufacturer’s instructions, the NGS sequencing sample’s total RNA was extracted from the serum EVs using an miRNeasy Mini Kit (Qiagen, Valencia, CA, USA). As for the RT-qPCR analysis samples, cel-miR-39 (RiBoBioInc., Guangzhou, China) was used as an external reference during extraction, normalizing the process of RNA extraction and PCR. Briefly, 1000 fmol of external cel-miR-39-3p was added to the EVs and mixed well. Next, the miRNA was extracted from the serum-isolated EVs utilizing a miRNeasy Mini Kit and an RNeasy MinElute Cleanup Kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions. All RNAs were stored at −80 °C until further analysis. ## 2.5. RNA Sequencing and Data Analysis Equal amounts of total RNA were taken for library preparation after identification and quantification using an Agilent 2100 bioanalyzer (Thermo Fisher Scientific, MA, USA) and then purified by electrophoretic separation on a $15\%$ PAGE gel. Small RNAs of 18–30 nt were recovered and sequentially ligated to 3’ and 5’ RNA adapters. After transcription and amplification, a fragment of 110–130 bp was selected and purified. The library was quality assessed and quantitated using the Agilent 2100 bioanalyzer and QPCR, and then it was sequenced using the BGISEQ-500 platform (BGI, Shenzhen, China). After filtering, clean tags were mapped to the reference genome with Bowtie 2 [38]. miRDeep2 was used to predict novel miRNA by exploring the secondary structure [39], and Cmsearch was performed for Rfam mapping [40]. DE miRNAs between two groups were performed using DEGseq. Next, we plotted volcano plots of DE miRNAs and the expression heatmap of 15 miRNAs online (https://www.bioinformatics.com.cn/, accessed on 20 March 2021). ## 2.6. RT-qPCR Stem-loop primers and specific primers for miRNAs were designed using miRNA Design software (http://www.vazyme.com/companyfile/7/, accessed on 20 March 2021), as shown in Table S1. Reverse transcription for cDNA production was performed using a HiScript III 1st Strand cDNA Synthesis Kit (Vazyme, Nanjing, China). Next, we measured miRNA levels according to the manufacturer’s instructions with miRNA Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China). In short, using RNA as a template, reverse transcription was carried out with the HiScript III 1st Strand cDNA Synthesis Kit and miRNA-specific RT primers to obtain cDNA. The cDNA was then amplified with miRNA Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China) and specific primers with the following thermocycler protocol: 95 °C for 5 min + (95 °C for 10 s; 56 °C for 30 s; 72 °C for 30 s) for 40 cycles. The qPCR was run on a Bio-Rad CFX96 touch qPCR system, and a data analysis was performed using Bio-Rad CFX Manager software. For this study, we calculated the expression levels of EV miRNAs in the samples by using the relative quantification method. For homogenization, equal quantities of cel-miR-39-3p were added as an internal control to each sample prior to RNA extraction. ΔCt was calculated as Ct (miRNA of interest) − Ct (cel-miR-39-3p), and the relative expression values for the target microRNA was calculated as 2-∆Ct. All reactions were run in triplicate. ## 2.7. Statistical Analysis GraphPad Prism 8.3 and SPSS 23.0 were utilized for data processing and statistical analyses. After calculating whether it was a normal distribution, the differences in clinical characteristics between the severe liver injury-CHB and DeCi groups were assessed using nonparametric tests and t-tests, respectively. We used mean (SE) to present miRNA expression values. The differences in expression levels between the two groups were compared using the nonparametric Mann-Whitney test. A receiver operating characteristic curve (ROC) analysis was carried out to assess the predictive value of the EV miRNAs and their different combinations. The predictive values of each combination were obtained by performing binary logistic regression before the ROC curve analysis, and area under the curve (AUC) > 0.7 and $p \leq 0.05$ were considered the criteria for evaluating the predictive markers. Spearman’s analysis was used to evaluate the correlation between the miRNAs and clinical practices. A two-tailed t-test was used for p values, and $p \leq 0.05$ was regarded as statistically significant. After calculating whether it was a normal distribution, the differences in clinical characteristics between the severe liver injury-CHB and DeCi groups were assessed using nonparametric tests and t-tests, respectively. ## 3.1. Study Population In this study, 149 patients (severe liver injury-CHB = 58, DeCi = 33, and NC = 58) from the Second Xiangya Hospital were enrolled. There were 25 individuals in each group (NC and severe liver injury-CHB) for NGS, and there were 33 persons in each group for training and validation. The primary characteristics and a comparison between the patients with severe liver injury-CHB and the patients with DeCi are presented in Table 1. Even though all the patients were randomly selected, there were more male patients than female patients. The vast majority of the patients with DeCi were either on or had been on antiviral therapy prior to hospital admission. The patients from the severe liver injury-CHB, DeCi, and NC groups were aged 42 (interquartile range (IQR): 29–54), 53 (IQR: 48–61), and 37 (IQR: 30.5–49) years, respectively. As expected, the patients with severe liver injury-CHB had higher levels of ALT, AST, TBIL, DBIL, GGT, and AFP than the patients with DeCi. The serum levels of WBC, L, PLT, ALB, and CHE were lower, and INR, lgA levels were higher in patients with DeCi than in patients with severe liver injury-CHB. Additionally, there were no considerable differences between the patients with severe liver injury-CHB and the patients with DeCi in other clinical features. The patients who were enrolled were not diagnosed as having alcoholic cirrhosis, hepatitis C, hepatitis D, etc. ## 3.2. Characterization of Serum EVs We visualized the serum-derived EVs by using TEM and NTA, and then we characterized them by using a Western blot for EV markers. In the representative images, the isolated EVs with diameters of 50–150 nm and their characteristic rounded membrane-bound morphology was observed at high-magnification views (Figure S1 and Figure 1B). Next, as shown in Figure 1A, the isolated EVs displayed markedly detectable expressions of EV markers, such as CD63, CD9, and TSG101, and had no expression of the endoplasmic reticulum calnexin, which is consistent with the previously reported EV characteristics. Additionally, in our study, the patients with severe liver injury-CHB displayed the highest EV concentrations, which were approximately 2.6-fold greater than those of the patients in the DeCi group ($p \leq 0.001$, Figure 1D). The EV diameter significantly differed between the NC vs. severe liver injury-CHB and severe liver injury-CHB vs. DeCi groups, with sizes of 100.23 nm [83.24–112.70], 126.77 nm [122.30–134.57], and 115.07 nm [108.78–115.95] for the NC, severe liver injury-CHB, and DeCi groups, respectively (Figure 1C). No difference in the concentrations and sizes of the EVs were found in the NTA analysis after stratification by gender (Figure S2). ## 3.3. Serum-Derived Exosome miRNA Screening We sequenced NC and severe liver injury-CHB samples in order to screen for differences in the serum EV miRNAs. First, we constructed sample pools for the different groups. For each group, we took five samples and randomly pooled them into a single pool and finally obtained five pools. The sample pools of each group were sequenced and analyzed using a high-throughput sequencing technique (NC = 25 and severe liver injury-CHB = 25). A total of 364 miRNAs were identified as differentially expressed in the sequencing, of which, 268 miRNAs were significantly up- or down-regulated (\|FC\| > 2, $p \leq 0.05$; Figure 2A). Based on sequencing results and literature, 11 miRNAs were selected for detection in our training set (\|FC\| > 2, $p \leq 0.0001$, at least ≥50 copies; Table S1). In addition, four miRNAs previously reported to be associated with disease-related molecules (hsa-miR-1285-5p, hsa-miR-204-3p, hsa-miR-335-5p, hsa-miR-877-5p) were also selected for detection (\|FC\| > 2, $p \leq 0.0001$). Heat maps of the 15 miRNAs are shown in Figure 2B. ## 3.4. The Differences in the Expressions of miRNAs between the NC Group and the Severe Liver Injury-CHB Group In this study, we focused on the 15 most viable candidate miRNAs for follow-up validation. Among these, miR-172-5p was a novel miRNA that was identified and sequenced (aggcuggagugcaguggcg). The primers are shown in Table S2. Ten samples each from the NC and severe liver injury-CHB groups were randomly enrolled, and serum EV RNA was extracted for an RT-qPCR assay. In all the groups screened, a total of six miRNAs were shortlisted based on valid Ct values (Ct < 35): miR-172-5p, miR-411-5p, miR-1285-5p, miR-335-5p, miR-877-5p, and miR-4433a-3p. For a further evaluation of the potential of these miRNAs, the levels of these six individual miRNAs were assessed by RT-qPCR in an independent cohort of NCs ($$n = 23$$) and patients with severe liver injury-CHB ($$n = 23$$). The miRNA levels were normalized by cel-miR-39-3p. Interestingly, only miR-172-5p and miR-1285-5p were found to be significantly differentially expressed (Figure 3A,B and Figure S3). In order to assess the miRNAs, we performed a receiver operator characteristic (ROC) curve analysis. Comparing the severe liver injury-CHB group with the NC group, the areas under the curve (AUCs) of miR-172-5p and miR-1285-5p were 0.808 ($95\%$ confidence interval (CI): 0.7003–0.9159, $p \leq 0.001$) and 0.794 ($95\%$ CI: 0.664–0.924, $p \leq 0.001$), respectively (Figure 3C). MiR-172-5p showed the highest sensitivity and specificity, with percentages of 84.9 and $72.7\%$ (Table S3), respectively. Next, we evaluated the predictive values of various combinations of these miRNAs [24]. The combination of two miRNAs (miR-172-5p and miR-1285-5p) showed an AUC area of 0.798 ($95\%$ CI: 0.6661–0.9291, $p \leq 0.001$) (Figure 3C), and its sensitivity and specificity were $68.2\%$ and $90.5\%$, respectively (Table S3). These data suggest that low expressions of miR-172-5p and miR-1285-5p in serum EVs predict a higher risk of chronic mild CHB developing into severe liver injury. ## 3.5. The Difference in the Expressions of miRNAs between the NC Group and the DeCi Group We simultaneously examined the expressions of the six shortlisted miRNAs in the patients with DeCi ($$n = 33$$) and obtained results similar to those of patients with severe liver injury-CHB. Compared to the NC group, miR-172-5p and miR-1285-5p showed significantly decreased expression levels (based on the RT-qPCR results) in the DeCi group. Additionally, miR-335-5p was significantly downregulated in the DeCi group compared to that in the NC group. The ROC curve analysis showed that the AUCs of miR-172-5p, miR-1285-5p, and miR-335-5p were 0.752 ($95\%$ CI: 0.6345–0.8697, $p \leq 0.001$), 0.827 ($95\%$ CI: 0.7094–0.9453, $p \leq 0.001$), and 0.696 ($95\%$ CI: 0.5627–0.8292, $p \leq 0.01$), respectively (Figure 4). The largest AUC was 0.855 ($95\%$ CI: 0.7345–0.9758, $p \leq 0.0001$) for a combination of three miRNAs (miR-172-5p, miR-1285-5p, and miR-335-5p) compared to those of single miRNAs or other miRNA panels. The sensitivity and specificity of this combination were $83.3\%$ and $81.0\%$, respectively (Figure 4 and Table S3). This combination in serum EVs displayed a high predictive accuracy for DeCi vs. NCs. ## 3.6. The Differences in the Expressions of miRNAs between the Severe Liver Injury-CHB Group and the DeCi Group Furthermore, we performed a statistical analysis of the data from the severe liver injury-CHB and DeCi groups, and we found that miR-335-5p was further reduced in the DeCi group compared to that in the severe liver injury-CHB group ($p \leq 0.05$) (Figure 5). The ROC curve analysis of miR-335-5p showed an AUC area of 0.660 ($95\%$ CI: 0.5307–0.7893, $$p \leq 0.023$$). The low expression of miR-335-5p in serum EVs predicts a greater risk of further development of DeCi in patients with CHB and severe liver injury. Additionally, we attempted to perform a joint analysis based on the available clinical indicators. We systematically analyzed the differences in the serological levels between the severe liver injury-CHB and DeCi groups and calculated their predictive sensitivity and specificity. The sensitivity and specificity of CHE, Ca2+, and AST/ALT exceeded expectations and showed better predictive properties (Table S3). The highest sensitivity and specificity ($91.7\%$ and $91.3\%$, respectively) were achieved after combining miR-335-5p with the aforementioned clinical indicators. Additionally, to detect the correlation between miR-335-5p and disease, we calculated the correlation between miRNA concentrations and serological detection levels using Spearman’s correlation analysis. The levels of miR-335-5p were remarkably related to the degree of inflammation (ALT and AST), GGT, and the index of hepatocyte regeneration (AFP) in a wide range of serological parameters (Table 2). However, miR-172-5p and miR-1285-5p displayed no statistical relevance to serological levels. ## 4. Discussion HBV infection is the primary cause of progressive liver disease. Liver biopsy has the potential to cause diagnostic mistakes attributable to heterogeneous distribution [9]. Blood tests include measurements of transaminases (ALT/AST), fibrosis-related factors, HBV DNA levels, and the presence or absence of HBV antigens. They have limited effectiveness in indicating the progression of liver disease [7]. In the past few decades, identifying biological phenotypes at the histological level has shown great potential for disease diagnosis. For example, clinical glycomic techniques used to distinguish patients with compensated cirrhosis from those with non-cirrhotic chronic liver disease have shown a sensitivity of $79\%$ and a specificity of $86\%$ [41]. Using genome-wide miRNA microarrays, miR-106b and miR-181b in serum were identified as specific biomarkers with an AUC of 0.774 for chronic HBV-associated liver cirrhosis (HBV-LC) and 0.915 for non-chronic HBV-LC [42]. Additionally, miR-101 in serum was reported to discriminate HBV-HCC from HBV-LC with a sensitivity of $95.5\%$ and a specificity of $90.2\%$ [43]. Serum microfibrillar-associated protein 4 (MFAP-4) can help distinguish patients with LC from individuals without liver disease (AUC = 0.97) using a proteomic approach [44]. EVs, as well as EV RNAs and EV proteins, are regarded as types of liquid biopsy, which boasts the advantages of non-invasiveness and dynamic monitoring, and they are able to overcome the limitations of tumor spatiotemporal heterogeneity. miRNAs are short non-coding RNA oligonucleotides that can increase mRNA degradation or decrease their translation by targeting specific mRNAs [45]. EVs with bilayer membranes in the circulatory system can protect their miRNA cargos and influence the target cells’ behavior, making EVs a popular candidate for studying the mechanisms of diseases and intercellular communication [15,16]. The role of EV miRNAs in HCC has been studied in several articles. For instance, serum exosome miR-122 levels are lower in patients with HCC than in patients with CHB or DeCi [46,47]. In contrast to normal hepatocyte-derived exosomes, 49 significantly over-expressed miRNAs have been observed in HCC [48]. In this study, EVs were isolated from serum based on their physical size and the principle that hydrophobic protein and lipid molecules can bind to PEGs. EVs were identified based on their physical shape, diameter, and protein markers, such as CD63 and CD9 [16]. In the Western blot results, we observed a gradual increase in the expressions of CD63 and CD9 with disease progression. ScRNA-seq demonstrated the expansion of macrophages in a mouse model of NASH with a pro-inflammatory phenotype, and CD9 is overexpressed in macrophages [49]. Furthermore, Sabine et al. confirmed that macrophage-mediated inflammation is critical in the pathogenesis of non-alcoholic steatohepatitis (NASH), and a higher level of CD63 and CD9 expression was detected in monocyte-infiltrating macrophages entering the liver [50]. The high expression of CD63 and CD9 found in the current research may be associated with macrophages, and more profound validation studies will be necessary shortly. We observed that patients with severe liver injury-CHB had the largest EVs size, which became smaller when progressing to DeCi. It has been reported that the expression level of TSG101 can affect the quantity and size of EVs [51]. In the present study, we also detected higher expression levels of TSG101 in patients with severe liver injury-CHB than in patients with DeCi, which may contribute to the change in EV size. In addition, changes in the physiological environment may also affect the size of EVs. For example, increased micro-environment acidity led to significantly smaller EV sizes in patients with prostate cancer (PCA) compared to individuals with no signs of urological disease [52]. Many studies have demonstrated that higher EV concentrations are detected in patients with diseases when comparing multiple diseases with normal groups [53,54,55,56,57]. In the current study, we also confirmed that patients with severe liver injury-CHB had the most numerous EVs. Interestingly, in our study, it was also found that the Ca2+ levels in peripheral blood were significantly higher in patients with severe liver injury-CHB than in patients with DeCi ($p \leq 0.001$), with a similar trend to the concentration of EVs. It has been demonstrated that exosomes are released in a Ca2+-dependent manner [58,59]. The production and shedding of red blood cell (RBC)-derived MV are associated with calcium and calcium carriers [60,61]. Increased intracellular Ca2+ concentrations in erythrocytes and platelets promote micro-vesicular release [62,63]. In addition, Ca2+ influx promotes cellular repair when a certain level of cellular damage has occurred, and the size or shape of the specific membrane damage prompts its removal via lysosomal cytosolic spitting or micro-vesicular shedding [64]. We speculate that the distribution of the number of EVs in the patients in this study may have some correlation with Ca2+ levels. The ROC analysis of the NC group compared with both the severe liver injury-CHB and DeCi groups indicated that the novel microRNA miR-172-5p was the best predictor, followed by miR-1285-5p. miR-335-5p might predict the risk of the further development of DeCi in patients with CHB and severe liver injury. The novel-miR-172-5p was a newly identified miRNA in this study. miR-172-5p was predicted to target multiple disease-related genes using bioinformatics software. CAMP response element-binding protein 1 (CREB1) is a core transcription factor, and it may be a promising therapeutic target for liver diseases [65]. HBV DNA polymerase attenuates HBV replication by activating the CREB1-HOTTIP-HOXA13 axis [66]. However, in this study, Spearman’s analysis did not find it to be associated with any clinical features (such as HBV DNA). Insulin-like growth factor binding protein 5 (IGFBP5) is also noteworthy, as it induced effects similar to those induced by TGFB1 [67]. IGFBP5 is a secretory protein associated with cell proliferation, adhesion, migration, systemic inflammatory response, and fibrosis. Furthermore, IGFBP is involved in the evolution of liver fibrosis, where an increase in IGFBP5 likely leads to a higher production of collagen type I in fibroblasts and enhanced tissue fibrosis [68]. novel-miR-172-5p is likely to be a helpful marker for liver diseases. Decreased expression levels of novel-miR-172-5p may promote the liver fibrosis process by affecting its target gene IGFBP, for which a more systematic study of this view is essential. There is no clear evidence of a correlation between miR-1285 expression levels and chronic liver diseases, such as HCV infection, non-alcoholic steatohepatitis, and autoimmune liver disease. Several papers have reported their association with tumor processes. As a regulatory miRNA of p53, miR-1285-5p is a tumor suppressor that inhibits cell proliferation and migration. One study showed that miR-1285-5p expression was lower in tumor tissues than in normal tissues [69,70]. DAPK2, a known tumor-suppressor, mediates both the anti-proliferative and the pro-apoptotic effects of miR-1285 depletion [71]. Moreover, the over-expression of miR-1285 weakens the TGF-β2-induced EMT process [72]. Considering these reports, we speculate that miR-1285-5p may be essential in regulating the development of liver parenchymal cells into fibrotic cells. More functional studies on miR-1285-5p and its targets are expected. miR-335 regulates numerous genes and plays multifunctional roles. Through exosomes, miR-335-5p completes the transfer from donor to recipient cells and promotes the migration, invasion, and EMT of cancer cells [73]. Transcription Factor EB (TFEB), a major regulator of lysosomal biogenesis and autophagy, increases insulin receptor substrate 1 (IRS1) protein and modulates glucose tolerance by decreasing miR-335 levels [74]. Additionally, in our research, the expression of miR-335-5p was strongly correlated with the degree of inflammation (ALT, AST), GGT, and the index of hepatocyte regeneration (AFP) levels, which may facilitate an intensive study of the mechanisms of disease progression. The diagnosis of HBV-related cirrhosis, which is the result of the development of CHB, involves the diagnosis of etiology, the assessment of compensated or decompensated status, and the evaluation of complication profiles. We observed the characteristics of the EVs in the disease phenotypes in the severe liver injury-CHB and DeCi samples in this study. The patients with severe liver injury-CHB had the highest number of EVs, which might be closely associated with Ca2+ concentrations. The expression level of novel-miR-172-5p in serum EVs differed significantly between the healthy individuals and the patients with severe liver injury-CHB or DeCi. The expression level of miR-335-5p decreased gradually with the progression of HBV-associated liver disease. Although the difference in miR-335-5p used in this study could not satisfactorily distinguish between the severe liver injury-CHB and DeCi groups, the inclusion of EV miRNA improved its predictive accuracy. A more comprehensive serological analysis of the different phases of HBV disease progression holds promise for predicting disease progression and for diagnosing compensated cirrhosis. Furthermore, studies on the correlation of miRNAs with liver fibrosis progression, inflammation, and hepatocyte regeneration would be greatly appreciated. ## 5. Conclusions We successfully isolated EVs and analyzed their concentrations and miRNA components. Patients with severe liver injury-CHB had the highest number of EVs, which might be closely related to Ca2+ concentrations. Furthermore, this study revealed the differential expression profiles of serum EV miRNAs in NC, severe liver injury-CHB, and DeCi groups, and it verified the candidate miRNA expression differences in the different groups using RT-qPCR. novel-miR-172-5p and miR-1285-5p in serum EVs provided a positive predictive accuracy for severe liver injury-CHB vs. NC. A combination of three miRNAs (novel-miR-172-5p, miR-1285-5p, and miR-335-5p) on a panel had the highest detection accuracy when differentiating the DeCi group from the normal group. Additionally, miR-335-5p was statistically different compared to the DeCi and severe liver injury-CHB groups, which may indicate a positive correlation with the level of inflammation. The addition of EV miR-335-5p improved the serological level of accuracy in predicting severe liver injury-CHB progression to DeCi. In summary, our research demonstrated the differences in EV characteristics and EV miRNA expressions between NC, severe liver injury-CHB, and DeCi groups, providing ideas for subsequent studies of predictive markers and studies of disease progression mechanisms. ## References 1. Asrani S.K., Hall L., Reddy V., Ogola G., Izzy M.. Comorbid Chronic Diseases and Survival in Compensated and Decompensated Cirrhosis: A Population-Based Study. Off. J. Am. Coll. Gastroenterol. \| ACG (2022) 117 2009-2016. DOI: 10.14309/ajg.0000000000001909 2. Hu L., Zhu Y., Zhang J., Chen W., Li Z., Li L., Zhang L., Cao D.. Potential Circulating Biomarkers of Circulating Chemokines CCL5, MIP-1β and HA as for Early Detection of Cirrhosis Related to Chronic HBV (Hepatitis B Virus) Infection. BMC Infect. Dis. (2019) 19. DOI: 10.1186/s12879-019-4130-0 3. Lampertico P., Agarwal K., Berg T., Buti M., Janssen H.L.A., Papatheodoridis G., Zoulim F., Tacke F.. EASL 2017 Clinical Practice Guidelines on the Management of Hepatitis B Virus Infection. J. Hepatol. (2017) 67 370-398. DOI: 10.1016/j.jhep.2017.03.021 4. EASL Clinical Practice Guidelines: Management of Chronic Hepatitis B Virus Infection. J. Hepatol. (2012) 57 167-185. DOI: 10.1016/j.jhep.2012.02.010 5. Tsochatzis E.A., Bosch J., Burroughs A.K.. Liver Cirrhosis. Lancet (2014) 383 1749-1761. DOI: 10.1016/S0140-6736(14)60121-5 6. Marcellin P., Kutala B.K.. Liver Diseases: A Major, Neglected Global Public Health Problem Requiring Urgent Actions and Large-Scale Screening. Liver Int. (2018) 38 2-6. DOI: 10.1111/liv.13682 7. 7.World Health Organisation Guideline for the Prevention, Care and Treatment of Persons with Chronic Hepatitis B InfectionWorld Health OrganizationGeneva, Switzerland2015bookshelf ID:NBK305553. World Health Organisation Guideline for the Prevention, Care and Treatment of Persons with Chronic Hepatitis B Infection (2015) 8. Sonneveld M.J., Brouwer W.P., Chan H.L.-Y., Piratvisuth T., Jia J.-D., Zeuzem S., Liaw Y.-F., Hansen B.E., Choi H., Wat C.. Optimisation of the Use of APRI and FIB-4 to Rule out Cirrhosis in Patients with Chronic Hepatitis B: Results from the SONIC-B Study. Lancet Gastroenterol. Hepatol. (2019) 4 538-544. DOI: 10.1016/S2468-1253(19)30087-1 9. Calvopina D.A., Coleman M.A., Lewindon P.J., Ramm G.A.. Function and Regulation of MicroRNAs and Their Potential as Biomarkers in Paediatric Liver Disease. Int. J. Mol. Sci. (2016) 17. DOI: 10.3390/ijms17111795 10. Bedossa P., Dargère D., Paradis V.. Sampling Variability of Liver Fibrosis in Chronic Hepatitis C. Hepatology (2003) 38 1449-1457. DOI: 10.1016/j.hep.2003.09.022 11. Yang G., Zhuang L., Sun T., Yeo Y.H., Tao L., Zhang W., Ma W., Wu L., Yang Z., Yang Y.. Serum Glial Cell Line-Derived Neurotrophic Factor (SGDNF) Is a Novel Biomarker in Predicting Cirrhosis in Patients with Chronic Hepatitis B. 2021, 2022, 1048104. Can. J. Gastroenterol. Hepatol. (2021) 2022 1048104. DOI: 10.1155/2022/1048104 12. Poynard T., de Ledinghen V., Zarski J.P., Stanciu C., Munteanu M., Vergniol J., France J., Trifan A., Le Naour G., Vaillant J.C.. Relative Performances of FibroTest, Fibroscan, and Biopsy for the Assessment of the Stage of Liver Fibrosis in Patients with Chronic Hepatitis C: A Step toward the Truth in the Absence of a Gold Standard. J. Hepatol. (2012) 56 541-548. DOI: 10.1016/j.jhep.2011.08.007 13. Imajo K., Kessoku T., Honda Y., Tomeno W., Ogawa Y., Mawatari H., Fujita K., Yoneda M., Taguri M., Hyogo H.. Magnetic Resonance Imaging More Accurately Classifies Steatosis and Fibrosis in Patients with Nonalcoholic Fatty Liver Disease than Transient Elastography. Gastroenterology (2016) 150 626-637. DOI: 10.1053/j.gastro.2015.11.048 14. Chen T., Wang C., Yu H., Ding M., Zhang C., Lu X., Zhang C.-Y., Zhang C.. Increased Urinary Exosomal MicroRNAs in Children with Idiopathic Nephrotic Syndrome. EBioMedicine (2019) 39 552-561. DOI: 10.1016/j.ebiom.2018.11.018 15. Ratajczak J., Miekus K., Kucia M., Zhang J., Reca R., Dvorak P., Ratajczak M.Z.. Embryonic Stem Cell-Derived Microvesicles Reprogram Hematopoietic Progenitors: Evidence for Horizontal Transfer of MRNA and Protein Delivery. Leukemia (2006) 20 847-856. DOI: 10.1038/sj.leu.2404132 16. Théry C., Witwer K.W., Aikawa E., Alcaraz M.J., Anderson J.D., Andriantsitohaina R., Antoniou A., Arab T., Archer F., Atkin-Smith G.K.. Minimal Information for Studies of Extracellular Vesicles 2018 (MISEV2018): A Position Statement of the International Society for Extracellular Vesicles and Update of the MISEV2014 Guidelines. J. Extracell. Vesicles (2018) 7 1535750. DOI: 10.1080/20013078.2018.1535750 17. Quesenberry P.J., Aliotta J.M., Deregibus M.C., Camussi G.. Role of Extracellular RNA-Carrying Vesicles in Cell Differentiation and Reprogramming. Stem Cell Res. Ther. (2015) 6 153. DOI: 10.1186/s13287-015-0150-x 18. de Paredes A.G.G., Manicardi N., Tellez L., Ibañez L., Royo F., Bermejo J., Blanco C., Fondevila C., Lanza V.F., Garcia-Bermejo L.. Molecular Profiling of Decompensated Cirrhosis by a Novel MicroRNA Signature. Hepatol. Commun. (2021) 5 309-322. DOI: 10.1002/hep4.1642 19. Lemoinne S., Thabut D., Housset C., Moreau R., Valla D., Boulanger C.M., Rautou P.-E.. The Emerging Roles of Microvesicles in Liver Diseases. Nat. Rev. Gastroenterol. Hepatol. (2014) 11 350-361. DOI: 10.1038/nrgastro.2014.7 20. Thietart S., Rautou P.-E.. Extracellular Vesicles as Biomarkers in Liver Diseases: A Clinician’s Point of View. J. Hepatol. (2020) 73 1507-1525. DOI: 10.1016/j.jhep.2020.07.014 21. Kornek M., Lynch M., Mehta S.H., Lai M., Exley M., Afdhal N.H., Schuppan D.. Circulating Microparticles as Disease-Specific Biomarkers of Severity of Inflammation in Patients with Hepatitis C or Nonalcoholic Steatohepatitis. Gastroenterology (2012) 143 448-458. DOI: 10.1053/j.gastro.2012.04.031 22. Rautou P.-E., Bresson J., Sainte-Marie Y., Vion A.-C., Paradis V., Renard J., Devue C., Heymes C., Letteron P., Elkrief L.. Abnormal Plasma Microparticles Impair Vasoconstrictor Responses in Patients With Cirrhosis. Gastroenterology (2012) 143 166-176.e6. DOI: 10.1053/j.gastro.2012.03.040 23. Payancé A., Silva-Junior G., Bissonnette J., Tanguy M., Pasquet B., Levi C., Roux O., Nekachtali O., Baiges A., Hernández-Gea V.. Hepatocyte Microvesicle Levels Improve Prediction of Mortality in Patients with Cirrhosis. Hepatology (2018) 68 1508-1518. DOI: 10.1002/hep.29903 24. Campello E., Zanetto A., Spiezia L., Radu C.M., Gavasso S., Ferrarese A., Farinati F., Senzolo M., Simioni P.. Hypercoagulability Detected by Circulating Microparticles in Patients with Hepatocellular Carcinoma and Cirrhosis. Thromb. Res. (2016) 143 118-121. DOI: 10.1016/j.thromres.2016.05.021 25. Friedman R.C., Farh K.K.-H., Burge C.B., Bartel D.P.. Most Mammalian MRNAs Are Conserved Targets of MicroRNAs. Genome Res. (2009) 19 92-105. DOI: 10.1101/gr.082701.108 26. Diehl P., Fricke A., Sander L., Stamm J., Bassler N., Htun N., Ziemann M., Helbing T., El-Osta A., Jowett J.B.M.. Microparticles: Major Transport Vehicles for Distinct MicroRNAs in Circulation. Cardiovasc. Res. (2012) 93 633-644. DOI: 10.1093/cvr/cvs007 27. Vallabhajosyula P., Korutla L., Habertheuer A., Yu M., Rostami S., Yuan C.X., Reddy S., Liu C., Korutla V., Koeberlein B.. Tissue-Specific Exosome Biomarkers for Noninvasively Monitoring Immunologic Rejection of Transplanted Tissue. J. Clin. Investig. (2017) 127 1375-1391. DOI: 10.1172/JCI87993 28. Habertheuer A., Korutla L., Rostami S., Reddy S., Lal P., Naji A., Vallabhajosyula P.. Donor Tissue-Specific Exosome Profiling Enables Noninvasive Monitoring of Acute Rejection in Mouse Allogeneic Heart Transplantation. J. Thorac. Cardiovasc. Surg. (2018) 155 2479-2489. DOI: 10.1016/j.jtcvs.2017.12.125 29. Szabo G., Bala S.. MicroRNAs in Liver Disease. Nat. Rev. Gastroenterol. Hepatol. (2013) 10 542-552. DOI: 10.1038/nrgastro.2013.87 30. Farid W.R.R., Pan Q., van der Meer A.J.P., de Ruiter P.E., Ramakrishnaiah V., de Jonge J., Kwekkeboom J., Janssen H.L.A., Metselaar H.J., Tilanus H.W.. Hepatocyte-Derived MicroRNAs as Serum Biomarkers of Hepatic Injury and Rejection after Liver Transplantation. Liver Transplant. (2012) 18 290-297. DOI: 10.1002/lt.22438 31. Ohno S., Ishikawa A., Kuroda M.. Roles of Exosomes and Microvesicles in Disease Pathogenesis. Adv. Drug Deliv. Rev. (2013) 65 398-401. DOI: 10.1016/j.addr.2012.07.019 32. Lee Y.-S., Kim S.Y., Ko E., Lee J.-H., Yi H.-S., Yoo Y.J., Je J., Suh S.J., Jung Y.K., Kim J.H.. Exosomes Derived from Palmitic Acid-Treated Hepatocytes Induce Fibrotic Activation of Hepatic Stellate Cells. Sci. Rep. (2017) 7 3710. DOI: 10.1038/s41598-017-03389-2 33. Pirola C.J., Gianotti T.F., Castaño G.O., Mallardi P., San Martino J., Ledesma M.M.G.L., Flichman D., Mirshahi F., Sanyal A.J., Sookoian S.. Circulating MicroRNA Signature in Non-Alcoholic Fatty Liver Disease: From Serum Non-Coding RNAs to Liver Histology and Disease Pathogenesis. Gut (2015) 64 800-812. DOI: 10.1136/gutjnl-2014-306996 34. Akuta N., Kawamura Y., Watanabe C., Nishimura A., Okubo M., Mori Y., Fujiyama S., Sezaki H., Hosaka T., Kobayashi M.. Impact of Sodium Glucose Cotransporter 2 Inhibitor on Histological Features and Glucose Metabolism of Non-alcoholic Fatty Liver Disease Complicated by Diabetes Mellitus. Hepatol. Res. (2019) 49 531-539. DOI: 10.1111/hepr.13304 35. Wang H., Hou L., Li A., Duan Y., Gao H., Song X.. Expression of Serum Exosomal MicroRNA-21 in Human Hepatocellular Carcinoma. Biomed Res. Int. (2014) 2014 864894. DOI: 10.1155/2014/864894 36. Xue X., Zhao Y., Wang X., Qin L., Hu R.. Development and Validation of Serum Exosomal MicroRNAs as Diagnostic and Prognostic Biomarkers for Hepatocellular Carcinoma. J. Cell. Biochem. (2019) 120 135-142. DOI: 10.1002/jcb.27165 37. The Guidelines of Prevention and Treatment for Chronic Hepatitis B (2019 Version). Zhonghua Gan Zang Bing Za Zhi = Zhonghua Ganzangbing Zazhi = Chin. J. Hepatol. (2019) 27 938. DOI: 10.3760/cma.j.issn.1007-3418.2019.12.007 38. Langmead B., Trapnell C., Pop M., Salzberg S.L.. Ultrafast and Memory-Efficient Alignment of Short DNA Sequences to the Human Genome. Genome Biol. (2009) 10 R25. DOI: 10.1186/gb-2009-10-3-r25 39. Friedländer M.R., Chen W., Adamidi C., Maaskola J., Einspanier R., Knespel S., Rajewsky N.. Discovering MicroRNAs from Deep Sequencing Data Using MiRDeep. Nat. Biotechnol. (2008) 26 407-415. DOI: 10.1038/nbt1394 40. Nawrocki E.P., Eddy S.R.. Infernal 1.1: 100-Fold Faster RNA Homology Searches. Bioinformatics (2013) 29 2933-2935. DOI: 10.1093/bioinformatics/btt509 41. Callewaert N., Van Vlierberghe H., Van Hecke A., Laroy W., Delanghe J., Contreras R.. Noninvasive Diagnosis of Liver Cirrhosis Using DNA Sequencer–Based Total Serum Protein Glycomics. Nat. Med. (2004) 10 429-434. DOI: 10.1038/nm1006 42. Chen Y.-J., Zhu J.-M., Wu H., Fan J., Zhou J., Hu J., Yu Q., Liu T.-T., Yang L., Wu C.-L.. Circulating MicroRNAs as a Fingerprint for Liver Cirrhosis. PLoS ONE (2013) 8. DOI: 10.1371/journal.pone.0066577 43. Xie Y., Yao Q., Butt A.M., Guo J., Tian Z., Bao X., Li H., Meng Q., Lu J.. Expression Profiling of Serum MicroRNA-101 in HBV-Associated Chronic Hepatitis, Liver Cirrhosis, and Hepatocellular Carcinoma. Cancer Biol. Ther. (2014) 15 1248-1255. DOI: 10.4161/cbt.29688 44. Mölleken C., Sitek B., Henkel C., Poschmann G., Sipos B., Wiese S., Warscheid B., Broelsch C., Reiser M., Friedman S.L.. Detection of Novel Biomarkers of Liver Cirrhosis by Proteomic Analysis. Hepatology (2009) 49 1257-1266. DOI: 10.1002/hep.22764 45. Guo H., Ingolia N.T., Weissman J.S., Bartel D.P.. Mammalian MicroRNAs Predominantly Act to Decrease Target MRNA Levels. Nature (2010) 466 835-840. DOI: 10.1038/nature09267 46. Sohn W., Kim J., Kang S.H., Yang S.R., Cho J.-Y., Cho H.C., Shim S.G., Paik Y.-H.. Serum Exosomal MicroRNAs as Novel Biomarkers for Hepatocellular Carcinoma. Exp. Mol. Med. (2015) 47 e184. DOI: 10.1038/emm.2015.68 47. Csak T., Bala S., Lippai D., Satishchandran A., Catalano D., Kodys K., Szabo G.. MicroRNA-122 Regulates Hypoxia-Inducible Factor-1 and Vimentin in Hepatocytes and Correlates with Fibrosis in Diet-Induced Steatohepatitis. Liver Int. (2015) 35 532-541. DOI: 10.1111/liv.12633 48. Cho H.J., Eun J.W., Baek G.O., Seo C.W., Ahn H.R., Kim S.S., Cho S.W., Cheong J.Y.. Serum Exosomal MicroRNA, MiR-10b-5p, as a Potential Diagnostic Biomarker for Early-Stage Hepatocellular Carcinoma. J. Clin. Med. (2020) 9. DOI: 10.3390/jcm9010281 49. Saviano A., Henderson N.C., Baumert T.F.. Single-Cell Genomics and Spatial Transcriptomics: Discovery of Novel Cell States and Cellular Interactions in Liver Physiology and Disease Biology. J. Hepatol. (2020) 73 1219-1230. DOI: 10.1016/j.jhep.2020.06.004 50. Daemen S., Gainullina A., Kalugotla G., He L., Chan M.M., Beals J.W., Liss K.H., Klein S., Feldstein A.E., Finck B.N.. Dynamic Shifts in the Composition of Resident and Recruited Macrophages Influence Tissue Remodeling in NASH. Cell Rep. (2021) 34 108626. DOI: 10.1016/j.celrep.2020.108626 51. Colombo M., Moita C., van Niel G., Kowal J., Vigneron J., Benaroch P., Manel N., Moita L.F., Théry C., Raposo G.. Analysis of ESCRT functions in exosome biogenesis, composition and secretion highlights the heterogeneity of extracellular vesicles. J. Cell Sci. (2013) 126 5553-5565. DOI: 10.1242/jcs.128868 52. Logozzi M., Mizzoni D., Di Raimo R., Giuliani A., Maggi M., Sciarra A., Fais S.. Plasmatic Exosome Number and Size Distinguish Prostate Cancer Patients From Healthy Individuals: A Prospective Clinical Study. Front. Oncol. (2021) 11 4258. DOI: 10.3389/fonc.2021.727317 53. Burrello J., Gai C., Tetti M., Lopatina T., Deregibus M.C., Veglio F., Mulatero P., Camussi G., Monticone S.. Characterization and Gene Expression Analysis of Serum-Derived Extracellular Vesicles in Primary Aldosteronism. Hypertension (2019) 74 359-367. DOI: 10.1161/HYPERTENSIONAHA.119.12944 54. Bernal-Mizrachi L., Jy W., Jimenez J.J., Pastor J., Mauro L.M., Horstman L.L., De Marchena E., Ahn Y.S.. High Levels of Circulating Endothelial Microparticles in Patients with Acute Coronary Syndromes. Am. Heart J. (2003) 145 962-970. DOI: 10.1016/S0002-8703(03)00103-0 55. Amabile N., Guérin A.P., Leroyer A., Mallat Z., Nguyen C., Boddaert J., London G.M., Tedgui A., Boulanger C.M.. Circulating Endothelial Microparticles Are Associated with Vascular Dysfunction in Patients with End-Stage Renal Failure. J. Am. Soc. Nephrol. (2005) 16 3381-3388. DOI: 10.1681/ASN.2005050535 56. Sabatier F., Darmon P., Hugel B., Combes V., Sanmarco M., Velut J.-G., Arnoux D., Charpiot P., Freyssinet J.-M., Oliver C.. Type 1 And Type 2 Diabetic Patients Display Different Patterns of Cellular Microparticles. Diabetes (2002) 51 2840-2845. DOI: 10.2337/diabetes.51.9.2840 57. González-Quintero V.H., Jiménez J.J., Jy W., Mauro L.M., Hortman L., O’Sullivan M.J., Ahn Y.. Elevated Plasma Endothelial Microparticles in Preeclampsia. Am. J. Obstet. Gynecol. (2003) 189 589-593. DOI: 10.1067/S0002-9378(03)00469-1 58. Wang L., Wang J., Wang Z., Zhou J., Zhang Y.. Higher Urine Exosomal MiR-193a Is Associated with a Higher Probability of Primary Focal Segmental Glomerulosclerosis and an Increased Risk of Poor Prognosis among Children with Nephrotic Syndrome. Front. Cell Dev. Biol. (2021) 9 727370. DOI: 10.3389/fcell.2021.727370 59. Savina A., Furlán M., Vidal M., Colombo M.I.. Exosome Release Is Regulated by a Calcium-Dependent Mechanism in K562 Cells. J. Biol. Chem. (2003) 278 20083-20090. DOI: 10.1074/jbc.M301642200 60. Prudent M., Crettaz D., Delobel J., Seghatchian J., Tissot J.D., Lion N.. Differences between Calcium-Stimulated and Storage-Induced Erythrocyte-Derived Microvesicles. Transfus. Apher. Sci. (2015) 53 153-158. DOI: 10.1016/j.transci.2015.10.012 61. Lange S., Gallagher M., Kholia S., Kosgodage U.S., Hristova M., Hardy J., Inal J.M.. Peptidylarginine DeiminasesRoles in Cancer and Neurodegeneration and Possible Avenues for Therapeutic Intervention via Modulation of Exosome and Microvesicle (EMV) Release?. Int. J. Mol. Sci. (2017) 18. DOI: 10.3390/ijms18061196 62. Choi D.-Y., Park J.-N., Paek S.-H., Choi S.-C., Paek S.-H.. Detecting Early-Stage Malignant Melanoma Using a Calcium Switch-Enriched Exosome Subpopulation Containing Tumor Markers as a Sample. Biosens. Bioelectron. (2021) 198 113828. DOI: 10.1016/j.bios.2021.113828 63. Bucki R., Bachelot-Loza C., Zachowski A., Giraud F., Sulpice J.C.. Calcium Induces Phospholipid Redistribution and Microvesicle Release in Human Erythrocyte Membranes by Independent Pathways. Biochemistry (1998) 37 15383-15391. DOI: 10.1021/bi9805238 64. Draeger A., Schoenauer R., Atanassoff A.P., Wolfmeier H., Babiychuk E.B.. Dealing with Damage: Plasma Membrane Repair Mechanisms. Biochimie (2014) 107 66-72. DOI: 10.1016/j.biochi.2014.08.008 65. Luo W.-J., Cheng T.-Y., Wong K.-I., Fang W., Liao K.-M., Hsieh Y.-T., Su K.-Y.. Novel Therapeutic Drug Identification and Gene Correlation for Fatty Liver Disease Using High-Content Screening: Proof of Concept. Eur. J. Pharm. Sci. (2018) 121 106-117. DOI: 10.1016/j.ejps.2018.05.018 66. Zhao X., Fan H., Chen X., Zhao X., Wang X., Feng Y., Liu M., Li S., Tang H.. Hepatitis B Virus DNA Polymerase Restrains Viral Replication Through the CREB1/HOXA Distal Transcript Antisense RNA Homeobox A13 Axis. Hepatology (2021) 73 503-519. DOI: 10.1002/hep.31284 67. Allan G.J., Beattie J., Flint D.J.. Epithelial Injury Induces an Innate Repair Mechanism Linked to Cellular Senescence and Fibrosis Involving IGF-Binding Protein-5. J. Endocrinol. (2008) 199 155-164. DOI: 10.1677/JOE-08-0269 68. Lecca M.R., Maag C., Berger E.G., Hennet T.. Fibrotic Response in Fibroblasts from Congenital Disorders of Glycosylation. J. Cell. Mol. Med. (2011) 15 1788-1796. DOI: 10.1111/j.1582-4934.2010.01187.x 69. Hironaka-Mitsuhashi A., Otsuka K., Gailhouste L., Calle A.S., Kumazaki M., Yamamoto Y., Fujiwara Y., Ochiya T.. MiR-1285-5p/TMEM194A Axis Affects Cell Proliferation in Breast Cancer. Cancer Sci. (2020) 111 395-405. DOI: 10.1111/cas.14287 70. Wang X., Yan M., Zhao L., Wu Q., Wu C., Chang X., Zhou Z.. Low-Dose Methylmercury-Induced Genes Regulate Mitochondrial Biogenesis via MiR-25 in Immortalized Human Embryonic Neural Progenitor Cells. Int. J. Mol. Sci. (2016) 17. DOI: 10.3390/ijms17122058 71. Villanova L., Barbini C., Piccolo C., Boe A., De Maria R., Fiori M.E.. MiR-1285-3p Controls Colorectal Cancer Proliferation and Escape from Apoptosis through DAPK2. Int. J. Mol. Sci. (2020) 21. DOI: 10.3390/ijms21072423 72. Pao S.I., Lin L.T., Chen Y.H., Chen C.L., Chen J.T.. Repression of Smad4 by MicroRNA-1285 Moderates TGF-β-Induced Epithelial–Mesenchymal Transition in Proliferative Vitreoretinopathy. PLoS ONE (2021) 16. DOI: 10.1371/journal.pone.0254873 73. Sun X., Lin F., Sun W., Zhu W., Fang D., Luo L., Li S., Zhang W., Jiang L.. Exosome-Transmitted MiRNA-335-5p Promotes Colorectal Cancer Invasion and Metastasis by Facilitating EMT via Targeting RASA1. Mol. Ther. Nucleic Acids (2021) 24 164-174. DOI: 10.1016/j.omtn.2021.02.022 74. Sun J., Lu H., Liang W., Zhao G., Ren L., Hu D., Chang Z., Liu Y., Garcia-Barrio M.T., Zhang J.. Endothelial TFEB (Transcription Factor EB) Improves Glucose Tolerance via Upregulation of IRS (Insulin Receptor Substrate) 1 and IRS2. Arterioscler. Thromb. Vasc. Biol. (2021) 41 783-795. DOI: 10.1161/ATVBAHA.120.315310
--- title: Factors Associated with Adherence to Treatment in Patients with HIV and Diabetes Mellitus authors: - Cristina Rivera-Picón - María Hinojal Benavente-Cuesta - María Paz Quevedo-Aguado - Juan Luis Sánchez-González - Pedro Manuel Rodríguez-Muñoz journal: Journal of Personalized Medicine year: 2023 pmcid: PMC9967318 doi: 10.3390/jpm13020269 license: CC BY 4.0 --- # Factors Associated with Adherence to Treatment in Patients with HIV and Diabetes Mellitus ## Abstract We aim to identify the factors that influence the therapeutic adherence of subjects with chronic disease. The design followed in this work was empirical, not experimental, and cross-sectional with a correlational objective. The sample consisted of a total of 400 subjects (199 patients with HIV and 201 patients with diabetes mellitus). The instruments applied for data collection were a sociodemographic data questionnaire, the 4-item Morisky Medication Adherence Scale (MMAS-4) and the Coping Strategies Questionnaire. In the group of subjects with HIV, that the use of emotional coping strategies was related to lower adherence to treatment. On the other hand, in the group of subjects with diabetes mellitus, the variable related to compliance with treatment was the duration of illness. Therefore, the predictive factors of adherence to treatment were different in each chronic pathology. In the group of subjects with diabetes mellitus, this variable was related to the duration of the disease. In the group of subjects with HIV, the type of coping strategy used predicted adherence to treatment. As a result of these results, it is possible to develop health programmes to promote, from nursing consultations to adherence to treatment of patients with HIV and diabetes mellitus. ## 1. Introduction Currently, we are facing an increase in life expectancy and a clear ageing of the world’s population. This has led to a significant increase in the incidence of chronic diseases, which represent a public health problem [1]. In this way, subjects with chronic pathologies must have good management of their disease to minimize its impact, improve health outcomes, prevent further disability and reduce healthcare costs. In this context of study, the concept of therapeutic regimen management is presented as a key component in the management of these diseases [2]. However, a high number of subjects with chronic pathologies do not comply with treatment adequately, and the rate of adherence to it in this type of patient is low [3]. Based on these assumptions, this study wishes to determine the factors that are related to better drug adherence in certain chronic diseases. In addition, knowing these factors is essential to achieve adequate management of the therapeutic regimen. This is justified because, in order to favor the management of the therapeutic regimen, it must be taken into account that health must be centered on the patient, for which reason their active collaboration is needed, taking into account their fears, their will or their difficulties in the treatment [4]. However, it should be noted that each chronic pathology has clinical characteristics and a particular therapeutic plan, which is why we cannot include all of them under the same heading of “chronicity”. Consequently, in this study we have considered it necessary to specifically analyse adherence to treatment in two specific chronic diseases: HIV and diabetes mellitus. HIV remains a relevant chronic disease worldwide. This infection, has had an enormous impact on morbidity, the demography and economy of the most affected countries. The latest data recorded in Spain show that 3244 new cases of HIV were diagnosed since 2019 [5]. The introduction of highly active antiretroviral therapy (HAART) has modified the natural history of HIV infection, despite the adverse reactions associated with such treatment [6,7]. The antiretroviral treatments currently used allow for reducing transmission rates, as well as increasing the survival of patients diagnosed with HIV [8]. In this way, this infection has become a chronic disease [9]. There is evidence that the incorrect taking of antiretroviral treatment is associated with the appearance of viral strains resistant to it, increasing the risk of disease progression. Therefore, despite the clear advantages of the treatment, it is essential to maintain correct management of the therapeutic regimen [10,11,12]. However, most research on subjects with HIV indicates that adherence to treatment in this group of patients is low [9,13,14]. On the other hand, diabetes mellitus is one of the most prevalent chronic diseases diagnosed today. In 2002, the WHO announced a global prevalence of diabetes of $3\%$, which corresponds to 170 million people in the world diagnosed with this pathology [15]. It was even estimated that this figure would more than double by the year 2025 [16]. Currently, these forecasts have already been exceeded, since the latest figures provided by the International Diabetes Federation (IDF), corresponding to 2019, state that $9.3\%$ of adults between 20 and 79 years old have diabetes, which corresponds to a total of 463 million people. Based on this data, the IDF estimates that 578 million adults will live with the disease in 2030. In 2045, they estimate that the figure will rise to 700 million [17]. Through the National Health Survey/European Health Survey in Spain (ENS/EESN), estimates of the prevalence of diabetes in Spain have also been obtained. The last one, corresponds to the year 2017, and states that the prevalence has almost doubled in Spain between 1993 ($4.1\%$) and 2017 ($7.8\%$). In subjects with diabetes, the importance of adherence to treatment to achieve adequate metabolic control is highlighted, which allows for delaying and preventing the development of complications associated with this disease [18,19]. However, it has been described that pharmacological compliance in diabetic subjects is also low [20,21]. Therefore, there is no doubt that incorrect adherence to treatment is a major health problem. Thus, the work of professionals is important to guarantee individualized and comprehensive patient care, allowing their active participation and improving the therapeutic relationship [22]. In addition, both in diabetic subjects and in those with a diagnosis of HIV, the adherence to treatment is a highly complex phenomenon, in which different variables may intervene [23]. Among these, factors related to the treatment itself stand out, such as the duration of illness, number of doses or associated side effects [24]. On the other hand, psychosocial factors are reflected, among which coping strategies have stood out. Certain coping strategies have been described as possible health protective factors, indicating their relationship with adherence to pharmacological treatment. Thus, it is stated that low levels of adherence to treatment are related to the use of emotional or avoidant strategies, while rational coping strategies can predict high adherence to treatment in certain chronic pathologies [25,26]. On the other hand, in relation to the sociodemographic variables, there was more heterogeneity of evidence, finding different results on the relationship between sex, marital status, educational level and the age of the patient with the level of adherence to treatment [27,28,29]. Consequently, the objective of this study was to identify those factors associated with adherence to drug treatment in subjects with diabetes mellitus or HIV. This could guide health teams in planning strategies that favor the management of the therapeutic regimen, through a more active collaboration of the patient. ## 2.1. Design The study had a non-experimental cross-sectional design, with a correlational objective. ## 2.2. Participants Subjects diagnosed with Diabetes or HIV+. The sample was collected at the University Hospital of Salamanca (HUS) and at different Primary Care centers in that province. In the first place, the sample of subjects with HIV was obtained through an incidental sampling, in the infectious diseases unit of the Salamanca clinical hospital. Subsequently, we obtained the sample of subjects with diabetes. A quota sampling was applied in order to obtain homogeneous subsamples. Said subsample of diabetic subjects was defined as having a size equivalent to that of subjects with HIV. The collection of the subsample of diabetic patients was carried out from different Internal Medicine floors of the Clinical Hospital of Salamanca and different Salamanca health centres (namely, “Periurbana Sur” and “Capuchinos” Health Centres). The following inclusion criteria were having a confirmed diagnosis of the disease (HIV or diabetes), regardless of its stage. They also had to be in current treatment, be of legal age and voluntarily participate in the study. As exclusion criteria, any illness or disorder that would prevent the patient from signing the informed consent was considered or completing the study. ## 2.3. Data Collection The data was collected taking a total of two years to collect them (2018–2020). The questionnaires used were: ## 2.3.1. Sociodemographic Data Questionnaire The variables collected through this instrument were of a sociodemographic nature and information related to health: Type of disease (HIV or diabetes) Duration of diagnosis of the disease (1–5 years, 5–10 years or more than 10 years) Sociodemographic variables studied (age, sex, marital status and educational level) ## 2.3.2. 4-Item Morisky Medication Adherence Scale (Morisky-Green-Levine) The questionnaire originally developed by Morisky et al. [ 1986] under the name “4-item Morisky Medication Adherence Scale” (MMAS-4). In Spanish, this questionnaire was validated by Val-Jiménez et al. [ 1992]. The scale is made up of four questions with a dichotomous response format (Yes/No), which reflects the patient’s behaviour regarding compliance. The patient shows good adherence to treatment when he/she correctly answers the four questions (No/Yes/No/No), corresponding to poor adherence if he/she answers three or fewer questions adequately. Therefore, through this questionnaire, patients are categorized as adherent to treatment or nonadherent to treatment [30]. Originally, the Morisky-Green-Levine treatment adherence questionnaire had adequate psychometric properties with high discrimination (ρbp = 0.43), high sensitivity (sensitivity = 0.81) and low specificity (specificity = 0.44). In Spain, this test was validated by Val-Jiménez et al. [ 1992] with a sample of hypertensive patients. The sensitivity was somewhat lower, while the specificity value was practically the same (sensitivity = 0.52, specificity = 0.44). ## 2.3.3. Coping Strategies Questionnaire (Sandín and Chorot) Coping styles were measured by means of the structured self-assessment known as the Coping Strategies Questionnaire (Cuestionario de Afrontamiento del Estrés (CAE)). The scale assesses seven coping factors: social support seeking, religious coping, overt emotional expression, avoidance coping, problem-solving coping, positive reappraisal and negative auto-focused coping. Also, it allows measuring the two most general dimensions, evaluated in our study: Rational Coping and Emotional Coping. The scale is made up of 42 items with a response range that goes from 0, never, to 4, almost always The total variance explained by the two general dimensions (Rational and Emotional) was $49.3\%$ [31]. ## 2.4. Ethical Considerations This study obtained a favourable report from the Ethics Committee “Research with Medicines in the Salamanca Health Area” (CEIC code: PI$\frac{02}{01}$/2018). In addition, it received authorization by the Healthcare Complex of the University of Salamanca and by the Salamanca Primary Care Directorate. ## 2.5. Data Analysis Data analysis was performed using the Statistics Package for the Social Sciences (SPSS) version 25 (IBM Corp, Armonk, NY, USA). A descriptive analysis of the sociodemographic variables was carried out, analyzing the differences between the subsamples. For this, Pearson’s χ2 test was applied, determining the size of the effect through Cramer’s V. To identify the factors that influence or help predict therapeutic adherence in subjects diagnosed with HIV and diabetes mellitus, logistic regression analysis was used. In this case, for each group of subjects with diabetes and HIV, a logistic regression analysis was carried out. The dependent variable was adherence to the treatment on the part of patients. This variable had two values, presence or absence of adherence. The independent variables were the type of coping strategy, sex, educational level, age, duration of illness and marital status. The level of statistical significance used throughout the study was 0.05, with a $95\%$ confidence interval. ## 3.1.1. Variables Related to the Type of Disease A total of 400 patients participated in the present study. Depending on the type of disease of the subjects, two groups are differentiated: HIV patients ($$n = 199$$) and diabetics ($$n = 201$$). Marital status was the only significant variable (χ2 = 42.484; $p \leq 0.01$). However, Cramer’s V value ($V = 0.322$) reflects that the effect is moderate. Most of the participants are male ($$n = 294$$), with mainly secondary or lower education ($$n = 328$$) and with a mean age of between 44 and 50 years ($$n = 124$$). Also, we observe that separated or divorced subjects are a minority ($$n = 38$$) (Table 1). ## 3.1.2. Variables Related to Adherence to Treatment The final sample consisted of 400 subjects, with $66\%$ showing nonadherent behaviour. Table 2 shows the descriptions of the sociodemographic variables based on adherence. No significant differences were found in any of the sociodemographic variables regarding adherence to treatment. ## 3.2.1. Subjects with HIV Table 3 shows the values of the covariates before becoming part of the model. The only variables that are related to adherence to treatment in subjects with HIV were emotional coping and rational coping ($p \leq 0.05$). The regression by specific steps showed that the best model was the one that only included the covariate use of emotional coping strategies (B = −0.065, ET = 0.024, Wald = 7.501, $$p \leq 0.006$$). Table 4 shows the values of the final model. It can be observed that the greater the use of emotional coping strategies in patients with HIV was, the lower the adherence to treatment (B = −0.065, Exp(B) = 0.937). Table 5 shows the results of different tests to assess the adequacy of the model. With a cut-off point of 0.6, about $66\%$ of patients were correctly classified as adherent or nonadherent. Compared with the null model, the mismatch was reduced by $5.3\%$ when the emotional coping variable was included. Finally, the Hosmer-Lemeshow test was not significant ($$p \leq 0.789$$). All these results led to the idea of a good fit of the model. ## 3.2.2. Subjects with Diabetes Table 6 shows the independent variable values in the null model. The only variable that improved the prediction of the null model was the duration of disease ($p \leq 0.05$). Table 7 shows the results of the independent variables that have intervened in the equation. The only variable that was significant was the duration of disease. More specifically, the proportion of patients who were adherent was the same between people with 1–5 years of disease and 5–10 years of disease ($p \leq 0.05$). However, the proportion of patients who were adherent was higher among patients with more than 10 years of disease duration compared to that among patients with 1–5 years of disease duration ($B = 0.275$, ET = 0.358, Wald = 0.590, $$p \leq 0.443$$). The odds of showing adherence with more than 10 years of disease duration was 2.65 times the odds of adherence with 1–5 years of disease duration. Table 8 shows the results of different tests to assess the adequacy of the model. When interpreting the goodness of fit through Nagelkerke’s R2, we observed that the mismatch was reduced by $4.8\%$. Furthermore, $60.5\%$ of well-predicted cases were detected in the alternative model, using a 0.6 cut-off point for classification. ## 4. Discussion In the present study, it was observed that a greater use of emotional coping strategies was related to lower adherence to treatment among subjects diagnosed with HIV. However, this variable was not relevant to determine adherence to treatment in subjects with diabetes. On the other hand, the duration of disease was related to the presence of adherence to treatment in subjects with diabetes, observing that adherence was more likely among patients with more than 10 years of disease duration than among patients with 1–5 years of disease duration. These results were not found in the sample of patients with HIV. In addition, the sociodemographic variables studied, such as sex, marital status, age or educational level of the subjects, did not show a significant relationship with compliance with pharmacological treatment in either of the two subsamples. In comparison with other studies, as we have indicated in the background of the article, some authors have pointed out the relationship between active or rational coping strategies with high adherence to treatment. In a complementary manner, others highlight that low levels of adherence are related to avoidant and emotional coping strategies, which are also called passive, palliative or maladaptive Strategies [32]. We highlight that these data have been confirmed in different investigations with certain chronic pathologies, such as kidney disease, multiple sclerosis or HIV [25,26,33]. In relation to the pathologies addressed in our study, we highlight the research of Weaver et al. [ 2005] carried out on subjects with HIV, which showed that coping strategies aimed at problem-solving were not related to adherence to HAART [34]. These results are similar to those found in our research, where only emotional coping strategies predicted the level of adherence. However, Delmas et al. [ 2008] did detect that those subjects with HIV who used active coping strategies experienced higher levels of adherence to treatment. It should be noted that, although the coping variable focused on the solution of the problem was significant considered in isolation in our results, it was not included in the final model [35]. Different investigations carried out in subjects with HIV have confirmed that the use of emotional, passive or maladaptive coping strategies are associated with a lack of adherence to treatment [36,37,38]. These results are consistent with those obtained in our research. In parallel, the results identified in various projects involving patients with diabetes also coincided with those obtained in our study. In this way, different authors defend that there is no statistically significant relationship between the use of different types of coping strategies and adherence [39,40,41,42]. In relation to sociodemographic variables, there is more heterogeneity of results, both in subjects with diabetes mellitus and in those with a diagnosis of HIV. Based on age, there is no consensus as to its role in adherence, since some studies confirm that adherence increases as age increases [29,43,44] and other investigations affirm the opposite [45]. The level of studies has also been proposed as a factor related to adherence in chronic diseases, highlighting less adherence at a lower level of studies [45,46,47]. However, other investigations present opposite data, justifying that subjects with a high level of education present with multiple responsibilities and occupations that require a lot of time and attention, which negatively interferes with their adherence [48]. Other research with diabetic subjects indicates that the titration does not predict adherence to treatment [44]. The study by Quiñones et al. [ 2018] only found a relationship between said variable and adherence in those diabetic patients who also suffered kidney damage [49]. Similarly, we have found studies on patients with HIV that also indicated the level of education is not related to adherence [29]. Studies that relate gender and adherence also do not offer conclusive results on the relationship between both variables. Thus, while some projects defend that it is women who have the greatest adherence [44], others affirm the opposite [50]. Based on marital status, we found authors who observe that married or partnered subjects have better adherence to treatment [51]. However, many others defend that it is not a variable associated with the adherence of patients with chronic diseases [52,53]. By way of summary, there are studies that affirm that sociodemographic variables are not related to the level of adherence to treatment. These results are consistent with those of our study. Thus, investigations carried out with samples of subjects with HIV found that age, sex, marital status, educational level and treatment time did not have a significant relationship with adherence [27,28,54]. Also, projects carried out with diabetic subjects coincided in stating that neither age, sex, marital status nor level of study are related to the level of adherence to treatment [55,56]. The results obtained in our research are in line with those obtained in these studies. Finally, in relation to the duration of illness, we can affirm that our results also coincide with those of some studies carried out with patients diagnosed with diabetes mellitus, which show that adherence to treatment is higher in those subjects with a greater duration of illness [57,58]. Ramos et al. [ 2017] more specifically detected that there were more behaviours adherent to treatment after 10 years of illness [44]. He even stated that this increase in adherence also occurred during the first 2 years of illness. In line with the results of our work, in the investigation of Kirkman et al. [ 2015], it is stated that patients who were newly diagnosed with diabetes were significantly less likely to be adherent to treatment [43]. In our research, an increased probability of showing adherent behaviour was found when patients had been diagnosed for more than 10 years but not when they had been diagnosed for 2 years. Thus, as we have presented in the introduction, these results would be useful so that, as a final objective, nursing could enhance the therapeutic management of patients. This is essential because inadequate management of the therapeutic regimen becomes a threat to the health, well-being and quality of life of the subjects [59]. Limitations: Different authors defend that adherence to treatment is a highly complex phenomenon, in which multiple factors intervene. Thus, there are studies that have demonstrated the relationship of other variables with adherence to treatment in chronic diseases that have not been included in our study, which is their weakness. Among these variables, the secondary effects of treatment, doctor-patient relationship, social support and comorbidities with other pathologies stand out [24,60]. On the other hand, we highlight as strengths the use of a larger sample than those used in other studies [61,62,63,64,65]. Future research should take into account other variables not considered in this study and which may be relevant but which, in order to seek parsimony in the research, we did not include. ## 5. Conclusions Taking into account the results of the aforementioned analysis, we can affirm that the subjects with a diagnosis of HIV who use emotional coping strategies less frequently have greater adherence to treatment. Therefore, emotional coping strategies are presented as a significant variable, which allows for predicting the level of adherence to treatment. However, these data were only evidenced in subjects with HIV but not in subjects with diabetes. In parallel, only in patients with diabetes did the duration of illness have an influence on the level of adherence to treatment. Thus, it has been confirmed that diabetic subjects who have had more than 10 years of disease duration have greater adherence than patients who have had the disease 1–5 years. However, the sociodemographic characteristics studied (sex, age, marital status and educational level) are not variables related to the level of adherence to treatment. On the other hand, rational coping strategies, or those based on the use of religion and avoidance, do not appear as significant variables that allow us to estimate adherence to treatment. These results coincide with the two subsamples studied, subjects with diabetes and subjects with HIV. We consider it essential to know the factors related to adherence to treatment in chronic diseases. In this way, it is possible to predict and anticipate those subjects who will have poor adherence to treatment. Thus, the conclusions obtained are of great interest among nursing staff for the development of health programs that increase the patient’s skills to promote and maintain the management of the therapeutic regimen. However, the results of this study show that the factors that predict adherence to treatment are different for each chronic disease. Therefore, it is essential to continue expanding the present study with other chronic pathologies. ## References 1. Fernández-Lázaro C.I., García-González J.M., Adams D.P., Fernandez-Lázaro D., Mielgo-Ayuso J., Caballero-García A., Moreno F., Córdova A., Miron-Canelo J.A.. Adherence to treatment and related factors among patients with chronic conditions in primary care: A cross-sectional study. BMC Fam. Pract. (2019) 14. DOI: 10.1186/s12875-019-1019-3 2. Dios-Guerra C., Pérula L.A.. Factors related to Ineffective Handle of Therapeutic Regimen (IHTR) in chronic patients of nursing consultation. Index Enferm. (2012) 21 14-17 3. Danielson E., Melin-Johansso C., Modanloo M.. Adherence to treatment in patients with chronic diseases: From alertness to persistence. Int. J. Community Based Nurs. Midwifery (2019) 7 248-257. PMID: 31641674 4. Iancu M.A., Mateiciuc I.I., Stanescu A.A., Matei D., Diaconu C.C.. Therapeutic Compliance of Patients with Arterial Hypertension in Primary Care. Medicina (2020) 56. DOI: 10.3390/medicina56110631 5. 5. Ministerio de Sanidad Consumo y Bienestar Social Vigilancia Epidemiológica del VIH y Sida en España 2018: Sistema de Información sobre Nuevos Diagnósticos de VIH y Registro Nacional de Casos de SidaPlan Nacional Sobre el Sida, D.G. de Salud Pública, Calidad e Innovación/Centro Nacional de EpidemiologíaISCIIIMadrid, Spain2019. Vigilancia Epidemiológica del VIH y Sida en España 2018: Sistema de Información sobre Nuevos Diagnósticos de VIH y Registro Nacional de Casos de Sida (2019) 6. Sax P.E.. Acute and Early HIV Infection. Clinical Manifestations and Diagnosis. (2019) 7. Varela-Arévalo M., Hoyos-Hernández P.. Adherence to treatment for HIV/AIDS: Beyond the uptake of antiretrovirals. Rev. Salud Pública (2015) 17 528-540. PMID: 28453073 8. Mitzel L.D., Vanable P.A.. Necessity and concerns beliefs and HIV medication adherence: A systematic review. J. Behav. Med. (2019) 43 1-5. DOI: 10.1007/s10865-019-00089-2 9. Bezabhe W.M., Chalmers L., Bereznicki L.R., Peterson G.M.. Adherence to Antiretroviral Therapy and Virologic Failure. Medicine (2016) 95 e3361. DOI: 10.1097/MD.0000000000003361 10. Kangethe A., Polson M., Lord T.C., Evangelatos T., Oglesby A.. Real-world health plan data analysis: Key trends in medication adherence and overall costs in patients with HIV. J. Manag. Care Spec. Pharm. (2019) 25 88-93. DOI: 10.18553/jmcp.2019.25.1.088 11. Hines D.M., Ding Y., Wade R.L., Beaubrun A., Cohen J.P.. Treatment adherence and persistence among HIV-1 patients newly starting treatment. Patient Prefer. Adherence (2019) 13 1927-1939. DOI: 10.2147/PPA.S207908 12. Sánchez-Rivero I., Madoz-Gúrpide A., Parro-Torres C., Hernández-Huerta D., Mangado E.O.. Coverage and Adherence of Antiretroviral Therapy Among Chinese HIV-positive Men Who Have Sex with Men With High CD4 Counts in the Era of “Treat All”. Adicciones (2020) 32 7-18. DOI: 10.20882/adicciones.1025 13. Costa J.D.M., Torres T.S., Esteves L., Mendes P.. Adherence to antiretroviral therapy for HIV/AIDS in Latin America and the Caribbean: Systematic review and meta-analysis. J. Int. AIDS Soc. (2018) 21 e25066. DOI: 10.1002/jia2.25066 14. Gutiérrez-Gabriel I., Godoy-Guinto J., Lucas-Alvarado H., Pineda-Germán B., Vázquez-Cruz E., Hernández-De a Rosa M.C., Sosa-Jurado F.. Quality of life and psychological variables affecting adherence to antiretroviral treatment in Mexican patients with HIV/AIDS. Rev. Chil. Infectología (2019) 36 331-339. DOI: 10.4067/S0716-10182019000300331 15. López M.J., Docampo M.. Change over time in prevalence of diabetes mellitus (DM) in Spain (1999–2014). Endocrinol. Diabetes Nutr. (2018) 65 515-523. DOI: 10.1016/j.endinu.2018.06.006 16. Ramirez S., Villa-Ruano N., García D.. Genetic epidemiology on causal theories and pathogenesis of type 2 diabetes mellitus. Gac. Médica México (2017) 153 864-874 17. Saeedi P., Petersohn I., Salpea P., Malanda B., Karuranga S., Unwin N., Colagiuri S., Guariguata L., Motala A., Ogurtsova K.. Global and regional diabetes prevalence estimates for 2019 and projections for 2030 and 2045: Results from the International Diabetes Federation Diabetes Atlas, 9 th edition. Diabetes Res. Clin. Pract. (2019) 157 107843. DOI: 10.1016/j.diabres.2019.107843 18. Orozco-Beltrán D., Mata-Cases M., Artola S., Conthe P., Mediavilla J., Miranda C.. Adherence of Type 2 Diabetes Mellitus approach: Current situation and possible solutions. Aten. Primaria (2016) 48 406-420. DOI: 10.1016/j.aprim.2015.09.001 19. Ting C.Y., Ahmad Zaidi Adruce S., Lim C.J., Abd Jabar A.H.A., Ting R.S.K., Ting H., Ting H., Osman N.A., Ngau E., Talin B.A.. Effectiveness of a pharmacist-led structured group-based intervention in improving medication adherence and glycaemic control among type 2 diabetes mellitus patients: A randomized controlled trial. Res. Soc. Adm. Pharm. (2020) 17 344-355. DOI: 10.1016/j.sapharm.2020.03.026 20. Shrivastava S., Shrivastava P., Ramasamy J.. Role of self-care in management of diabetes mellitus. J. Diabetes Metab. Disord. (2013) 12 1. DOI: 10.1186/2251-6581-12-14 21. Krass I., Schieback P., Dhippayom T.. Systematic Review or Meta-analysis. Adherence to diabetes medication: A systematic review. Diabet. Med. (2015) 32 725-737. DOI: 10.1111/dme.12651 22. Iniesta-Sánchez J., Abad-Corpa E., Royo-Morales T., Sáez A., Rodriguez J.J., Carrillo A.. Impact assessment of a plan of nursing care of patients diagnosed with COPD nurse “Ineffective management of therapeutic regimen,” in terms of improving the nursing outcomes (NOC) “Knowledge of therapeutic regime”. Enferm. Glob. (2016) 15 39-48 23. Blanco J.A., Hernández S., Botas P., Rodríguez-Rodero S., Morales P., Naya L.D., Menéndez-Torre E., Delgado E.. Gender differences in the mortality of people with type 2 diabetes: Asturias Study 2018. Gac. Sanit. (2019) 34 442-448. DOI: 10.1016/j.gaceta.2019.02.014 24. Menditto E., Cahir C., Aza-Pascual M., Bruzzese D., Poblador-Plou B., Malo S., Costa E., González-Rubio F., Gimeno-Miguel A., Orlando V.. Adherence to chronic medication in older populations: Application of a common protocol among three European cohorts. Patient Prefer. Adherence (2018) 12 1975-1987. DOI: 10.2147/PPA.S164819 25. Corallo P.S.Y., Bonanno M.C.S., Marcella D.C., Carmela R., Edoard S., GianGaetano D.A., Viviana L.B., Giuseppo V., Plácido B., Silvia M.. Therapeutic adherence and coping strategies in patients with multiple sclerosis. Medicine (2019) 98 e16532. DOI: 10.1097/MD.0000000000016532 26. Hwang H.C., Kim H.R., Han D.H., Hong J.S., Jeong S., Shin J., Kim S.M., Hwang J.H., Kim S.. Influence of Major Coping Strategies on Treatment Non-adherence and Severity of Comorbid Conditions in Hemodialysis Patients. J. Korean Med. Sci. (2018) 33 e148. DOI: 10.3346/jkms.2018.33.e148 27. Abadiga M., Hasen T., Mosisa G., Abdisa E.. Adherence to antiretroviral therapy and associated factors among Human immunodeficiency virus positive patients accessing treatment at Nekemte referral. PLoS ONE (2020) 15. DOI: 10.1371/journal.pone.0232703 28. Orellana-Zanabria G., Morales-Rezza E.. Factors associated with adherence to HAART in patients with HIV/AIDS in the central hospital of the police forces. An. Fac. Med. (2019) 19 46-54 29. Sabino T.E., Avelino-Silva V.I., Cavalcantte C., Goulart S.P., Luiz O.C., Fonseca L.A.M., Casseb J.S.. Adherence to antiretroviral treatment and quality of life among transgender women living with HIV/AIDS in São Paulo, Brazil. AIDS Care (2021) 33 31-38. DOI: 10.1080/09540121.2019.1710449 30. Morisky D.E., Green L.W., Levine D.M.. Concurrent and predictive validity of a self-reported measure of medication adherence. Med. Care (1986) 24 67-74. DOI: 10.1097/00005650-198601000-00007 31. Sandín y Chorot B.. The Coping Strategies Questionnaire: Development and preliminary validation. Rev. Piscopatoogia Piscologia Clin. (2002) 8 39-54 32. Khechane N., Mwaba K.. Treatment adherence and doping with stress among black South African hemodialysis patients. Soc. Behav. Personal. (2004) 32 777-782. DOI: 10.2224/sbp.2004.32.8.777 33. Sampaio de Brito D.C., Oliveira E., Dos Santos F.R., Basile Colugnati F., Lucchetti G., Sanders-Pinheiro H.. Stress, coping and adherence to immunosuppressive medications in kidney transplantation: A comparative study. Sao Paulo Med. J. (2016) 134 292-299. DOI: 10.1590/1516-3180.2015.01071008 34. Weaver K.E., Llabre M., Durán R.E., Antoni M.H., Ironson G., Penedo F.J., Schneiderman N.A.. Stress and Coping Model of Medication Adherence and Viral Load in HIV-Positive Men and Women on Highly Active Antiretroviral Therapy (HAART). Health Psychol. (2005) 24 385-392. DOI: 10.1037/0278-6133.24.4.385 35. Delmas P., Delpierre C., Côté J., Lauwers-Cancès V., Delon S.. Study of the Promosud Cohort. Predictors of Adherence to Treatment Plans by French Patients Living With HIV. Perspect. Enfermo (2008) 5 28-35 36. Ahumada M., Escalante E., Santiago I.. Preliminary study of relations between coping strategies, social support and adherence to treatment in people living with HIV/AIDS. Subj. Procesos Cogn. (2011) 15 55-70 37. Lyimo R.A., Stutterheim S.E., Hospers H.J., Glee T.. Stigma, Disclosure, Coping, and Medication Adherence Among People Living with HIV/AIDS in Northern Tanzania 1. AIDS Patient Care STDS (2014) 28 98-105. DOI: 10.1089/apc.2013.0306 38. Malow R., Devieux J., Stein J., Rosenberg R., Jean-Gilles M., Attonito J., Koenig S.P., Raviola G., Sévère P., Pape J.. Depression, Substance Abuse and Other Contextual Predictors of Adherence to Antiretroviral Therapy (ART) Among Haitians. AIDS Behav. (2013) 17 1221-1230. DOI: 10.1007/s10461-012-0400-1 39. Pedraza G.L., Vega C.Z.. Stress, coping, emotions and therapeutic adherence in diabetic patients. Eureka (2018) 15 173-185 40. Ortiz M.. Stress, Copying Style, and Treatment Adherence in Adolescents Sufferingfrom Type I Diabetes. Ter. Psicol. (2006) 24 139-147 41. Ortiz M., Ortiz E., Gómez D.. Psychosocial Factors Associated with Adherence to Type 2 Diabetes Mellitus Treatment. Ter. Psicológica (2011) 29 5-11. DOI: 10.4067/S0718-48082011000100001 42. Smalls B.L., Walker R.J., Hernandez-Tejada M.A., Campbell J.A., Davis K.S., Egede L.E.. Associations between coping, diabetes knowledge, medication adherence and self-care behaviors in adults with type 2 diabetes. Gen. Hosp. Psychiatry (2012) 34 385-389. DOI: 10.1016/j.genhosppsych.2012.03.018 43. Kirkman M.S., Rowan M.T., Levin R., Fonseca V.A., Schmittdiel J.A., Herman W.H., Aubert R.E.. Determinants of Adherence to Diabetes Medications: Findings from a Large Pharmacy Claims Database. Diabetes Care (2015) 38 604-609. DOI: 10.2337/dc14-2098 44. Ramos Y., Morejón R., Gómez M., Reina M.E., Rangel C., Cabrera Y.. Therapeutic adherence in patients with type 2 diabetes mellitus. Rev. Final. (2017) 7 89-98 45. Canales S., Barra E.. Self-efficacy, social support and adherence to treatment in adult patients with type 2 diabetes. Psicol. Salud (2014) 24 167-173 46. Domínguez E.. Social inequalities and diabetes mellitus. Rev. Cuba. Endocrinol. (2013) 24 200-213 47. Fornos-Pérez J.A., Andrés-Rodríguez N.F., Andrés-Iglesias J.C., Mera-Gallego R., Mera-Gallego I., Penín-Álvarez Ó., Brizuela-Rodicio L.. Assessment of compliance with hypoglycemic and antihypertensive treatments in Galicia. Farm. Comuniarios (2017) 9 5-13. DOI: 10.5672/FC.2173-9218.(2017/Vol9).004.02 48. Martín L.A., Bayarre V.H., Corugedo R.M.C., Vento I.F., La Rosa M.Y., Orbay A.M.C.. Adherence to treatment observed in hypertensive patients from health areas of three cuban provinces. Rev. Cuba. Salud Publica (2015) 41 33-45 49. Quiñones Á., Ugarte C., Chávez C., Mañalich J.. Psychological variables associated with adherence to treatment and complications in patients with type 2 diabetes mellitus. Rev. Med. Chile (2018) 146 1151-1158. DOI: 10.4067/S0034-98872018001001151 50. Camargo C., Cavassan N., Tasca K., Meneguin S., Miot H., Souza L.. Depression and coping are associated with failure of adherence to antiretroviral therapy among people living with HIV/AIDS. Res. Hum. Retrovir. (2019) 35 1181-1188. DOI: 10.1089/aid.2019.0050 51. Zhao H., Zhang N., Ho V., Ding M., He W., Niu J., Yang M., Du X.L., Zorzi D., ChavezMacGregor M.. Adherence to treatment guidelines and survival for older patients with stage II o III colon cancer in Texas from 2001 to 2011. Cancer (2018) 124 679-687. DOI: 10.1002/cncr.31094 52. Camp Y., Vrijens B., Elseviers M.M.. Adherence to phosphate binders in hemodialysis patients: Prevalence and determinants. J. Nephrol. (2014) 27 673-679. DOI: 10.1007/s40620-014-0062-3 53. Ossareh S., Tabrizian S., Zebarjadi M., Joodat R.. Prevalence of Depression in Maintenance Hemodialysis Patients with Adherence to Medications. Iran J. Kidney Dis. (2014) 8 467-474. PMID: 25362222 54. Lemos L.D.A., Teles M.L., Reis R.K., Ferrer A.C., Gir E., Gimeniz M.. Adherence to antiretrovirals in people coinfected with the human immunodeficiency virus and tuberculosis. Rev. Lat. Am. Enferm. (2016) 24 e2691. DOI: 10.1590/1518-8345.0537.2691 55. Pereira A., De Sousa B., Borges D.S., García J., Silva F.V., Coelho M.M., Júnior R.. Adherence to the treatment with oral antidiabetic medications in primary health care. Rev. Rene (2015) 16 425-433 56. Leites-Docío A., García-Rodríguez P., Fernández-Cordeiro M., Tenorio-Salgueiro L., Fornos-Pérez J.A., Andrés-Rodríguez N.F.. Evaluation of non-adherence to hipoglucemiary treatment in community pharmacy. Farm. Comuniarios (2019) 11 5-13. DOI: 10.5672/FC.2173-9218.(2019/Vol11).001.02 57. Klinovszky A., Kiss I.M., Papp-Zipernovszky O., Lengyel C., Buzás N.. Associations of different adherences in patients with type 2 diabetes mellitus. Patient Prefer. Adherence (2019) 13 395-407. DOI: 10.2147/PPA.S187080 58. Marinho F.S., Moram C.B.M., Rodrigues P.C., Leite N.C., Salles G.F., Cardoso C.R.L.. Treatment Adherence and Its Associated Factors in Patients with Type 2 Diabetes: Results from the Rio de Janeiro Type 2 Diabetes Cohort Study. J. Diabetes Res. (2018) 2018 8970196. DOI: 10.1155/2018/8970196 59. Bakker R.H., Kastermans M.C., Dassen T.W.. An analysis of the nursing diagnosis ineffective management of therapeutic regimen compared to noncompliance and Orem’s self-care deficit theory of nursing. Nurs. Diagn. (1995) 6 161-166. DOI: 10.1111/j.1744-618X.1995.tb00287.x 60. Escandón-Nagel N., Azócar B., Pérez C., Matus V.. Adherence to treatment in type 2 diabetes and itsassociation with quality of life and depression. Rev. Psicoter. (2015) 26 125-138 61. Ceylan E., Koç A., Inkaya A.Ç., Ünal S.. Determination of medication adherence and related factors among people living with HIV/AIDS in a Turkish university hospital. Turk. J. Med. Sci. (2019) 49 198-205. DOI: 10.3906/sag-1802-137 62. Chirambo L., Valeta M., Banda T.M., Nyondo-Mipando A.L.. Factors influencing adherence to antiretroviral treatment among adults accessing care from private health facilities in Malawi. BMC Public Health (2019) 19. DOI: 10.1186/s12889-019-7768-z 63. Ghanbari A., Khiaban M.O., Aslani A., Faraji A.R., Mohammadi M., Kazemi A.F.. Assessment of adherence to therapy and exploring of barriers and facilitators in HIV positive patients in Tabriz-Iran: A mixed method study protocol. Reprod. Health (2019) 17 4-9. DOI: 10.1186/s12978-019-0766-x 64. Hushie M.. Exploring the barriers and facilitators of dietary self-care for type 2 diabetes: A qualitative study in Ghana. Health Promot. Perspect. (2019) 9 223-232. DOI: 10.15171/hpp.2019.31 65. Radwan M., Sari A.A., Rashidian A., Takian A., Elsous A., Abou-Dagga S.. Factors hindering the adherence to clinical practice guideline for diabetes mellitus in the Palestinian primary healthcare clinics: A qualitative study. BMJ Open (2018) 8 e021195. DOI: 10.1136/bmjopen-2017-021195
--- title: 'Association between Consumption of Dietary Supplements and Chronic Kidney Disease Prevalence: Results of the Korean Nationwide Population-Based Survey' authors: - Yina Fang - Hwasun Lee - Serhim Son - Sewon Oh - Sang-Kyung Jo - Wonyong Cho - Myung-Gyu Kim journal: Nutrients year: 2023 pmcid: PMC9967330 doi: 10.3390/nu15040822 license: CC BY 4.0 --- # Association between Consumption of Dietary Supplements and Chronic Kidney Disease Prevalence: Results of the Korean Nationwide Population-Based Survey ## Abstract Despite the enormous global market of dietary supplements, the impact of dietary supplements on kidney disease is still unclear. Based on the National Health and Nutrition Examination Survey from 2015 to 2017, this study evaluated the association between dietary supplement and chronic kidney disease (CKD) in 13,271 Korean adults. Among the dietary supplements, vitamin and mineral intake was the highest at $61.41\%$, followed by omega-3 fatty acids at $11.85\%$, and ginseng at $7.99\%$. The prevalence of CKD was significantly higher in those who consumed amino acids and proteins, ginseng and red ginseng, and herbal medicine (plant extract)-berries than in those who did not. Conversely, patients who consumed probiotic supplements had a significantly lower prevalence of CKD than those who did not. In the population without CKD risk factors or history of CKD, the prevalence of CKD was high in the group consuming ginseng and red ginseng. After adjusting for covariates, the herbal medicine (plant extract)-berry group showed an independent association with CKD incidence. In conclusion, it is suggested that dietary supplements may affect kidney function. Further large-scale cohort studies are required to elucidate the exact effects of each dietary supplement on CKD. ## 1. Introduction Over the past decades, chronic kidney disease (CKD) has affected 10–$15\%$ of the population worldwide, and paralleling epidemics of hypertension and diabetes, the number of CKD patients has increased rapidly in recent years, making a significant impact on the global health burden [1,2,3,4]. Owing to the high prevalence, morbidity rates, and medical costs of CKD, prevention and optimal management of the disease is an important public health issue. However, since there is no effective kidney-targeting drug that can inhibit the progression of CKD other than treatment of underlying diseases such as hypertension, hyperglycemia, and dyslipidemia and monitoring of complications, interest in the effect of nutrition or dietary supplements on kidney disease is increasing. Diets such as protein restriction diet, the Mediterranean diet, and plant-based diets are currently being investigated for their potential roles in delaying CKD progression. Today, there are thousands of dietary supplements available on the market, including vitamins and minerals, plant ingredients and extracts, proteins and amino acids, omega-3 fatty acids, probiotics, and prebiotics [5,6,7]. Some research has indicated that dietary supplements can compensate for the nutritional deficit derived from unbalanced diets and can assist in prolonging the lifespan and provide some benefits in diseases, although evidence of direct effects is still insufficient [7,8,9]. In the last decade, the prevalence of dietary supplement use has increased dramatically [6,7], and the global market size of dietary supplements was valued at nearly 121 billion USD in 2018 [6]. The use of dietary supplements in modern society is not limited to the middle-aged and elderly, and the interest in dietary supplements is growing among young people as well. This is because most dietary supplements are not classified as drugs by the Food and Drug Administration and are easily obtainable with unrestricted exposure to advertisements. Dietary supplement intake data from the Korean National Health and Nutrition Examination Survey (KNHANES) also showed that approximately $42\%$ of Korean adults aged ≥19 years have taken a dietary supplement at least once in their lifetime [10]. However, the role of dietary supplements is not well understood. Although the beneficial effects of dietary supplements have been shown in some studies, systematic studies on their positive and negative effects have not been reported. In particular, if there is a problem with the function of the liver or kidney, which is the path through which dietary metabolites are absorbed and excreted, it may affect the dysfunction of these organs, and studies on the effect and safety of dietary supplements are needed. The possibility of drug–supplement interactions should also be considered. Guidelines for CKD management recommend that patients with CKD should avoid using nutritional protein supplements and herbal remedies and should use them only under the supervision of a physician or pharmacist [11]. Extensive animal and clinical research have been conducted on the role of vitamin/mineral supplements and probiotics/prebiotics in kidney disease [12,13,14]. For example, vitamin D may contribute to improved clinical outcomes of CKD patients [12], and probiotics can lower the levels of inflammatory mediators and have clinical benefits in patients at different stages of CKD [13,14]. However, the association between CKD and numerous dietary supplements (e.g., ginseng and red ginseng, methyl sulfonyl methane, lutein containing supplements, propolis, and milk thistle) has not been well studied. Thus, the present study aimed to characterize the extent of dietary supplement use in a nationally representative sample and to explore the possible consequences of dietary supplements in patients with CKD. ## 2.1. Data Source and Study Population This study used data from the KNHANES conducted from 2015 to 2017, which is a nationwide cross-sectional sample survey conducted by the Korea Centers for Disease Control and Prevention. KNHANES includes an annual survey sample of approximately 10,000 people and consists of biochemical and clinical profiles for diseases, socioeconomic status, anthropometric measures, health-related behaviors, quality of life, health screenings, and nutritional surveys, of which the nutrition survey collects data on dietary supplements through a 24-h recall method [15]. Of the 23,657 survey participants from 2015 to 2017, 20,837 who participated in the dietary supplements survey were selected for this study (Figure 1). In addition, we excluded subjects that aged <19 years ($$n = 4378$$), those with missing laboratory data ($$n = 1586$$), and those consuming two or more dietary supplements simultaneously ($$n = 1602$$). The remaining 13,271 subjects were selected for this study, including 2752 subjects who used dietary supplements and 733 subjects with CKD. We defined CKD as dipstick-positive proteinuria or estimated glomerular filtration rate (eGFR) ≤ 60 mL/min/1.73 m2, calculated using the Chronic Kidney Disease Epidemiology Collaboration equation. ## 2.2. Different Types of Dietary Supplements Dietary supplement data were assessed by using the following questions: “Did you consume dietary supplements over the past 24 h?” and “What is the brand name and manufacturer name of the dietary supplement that you used over the past 24 h?” According to the Health Fictional Food Code issued by the Ministry of Food and Drug Safety Notification (No. 2020-92) [16], dietary supplements were classified as aloe, amino acids and protein, gamma-linolenic acid, ginseng and red ginseng, glucosamine, herbal medicine and plant extract, lutein containing supplements, methyl sulfonyl methane, milk thistle, chlorella/spirulina, omega-3 fatty acid, probiotics (pre-, post-), propolis, and vitamin and mineral (Figure 2). Owing to the many different types of herbal medicines and plant extract used, we divided herbs into herbal medicine (plant extract) Asian, herbal medicine (plant extract) berry, herbal medicine (plant extract) ginkgo biloba, and herbal medicine (plant extract) others. Although we found 46 different dietary supplements, only the top 17 dietary supplements have been used in this study, and the rest of the dietary supplements were classified into “OTHER category (e.g., honey extract, linolenic acid dietary fiber, hyaluronic acid, placenta and collagen, squalene and alkoxyglycerol, and ursodeoxycholic acid). ## 2.3. Assessment of Covariates The covariates for this study included sociodemographic variables (age, sex, and education level), body mass index (BMI) (calculated as weight/height2), smoking status, drinking status, physical activity, systolic blood pressure (SBP), diastolic blood pressure (DBP), and medical history (hypertension and diabetes). Serum creatinine (Cr), fasting blood glucose (FBG), and triglyceride (TG) levels were measured. ## 2.4. Statistical Analyses All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Differences between continuous variables were analyzed using the independent t-test or Mann–Whitney U test, while categorical variables were examined using the chi-square test. Multivariable logistic regression analysis was used to analyze the association between dietary supplement consumption and CKD. Model 1 is a crude model without adjustment. Model 2 was adjusted for age and gender. Model 3 was adjusted for all variables in Model 2, as well as smoking status (smoker, ex-smoker, and non-smoker), drinking status (never, less than once a month, 1–4 times a month, and ≥5 times a month), education level (less than elementary school graduation, middle School graduation, high School graduation and college graduate or higher), physical activity (≥2 days/week, <2 days/week), SBP, DBP, BMI (<25 kg/m2, ≥25 kg/m2), FBG level, and TG level. ## 3.1. General Characteristics of the Study Subjects Table 1 shows the characteristics of 13,271 study subjects, including 2752 in the dietary supplement group. The participants in the dietary supplement group were more likely to be women ($p \leq 0.001$), older adults ($p \leq 0.001$), non-smoker ($p \leq 0.001$), and highly educated individuals ($p \leq 0.001$). In addition, the DBP ($$p \leq 0.048$$) and serum Cr ($$p \leq 0.001$$) levels were high in the non-dietary supplement group. Among those who taking dietary supplements (Figure 2), most participants consumed vitamin and mineral ($61.41\%$, $$n = 1690$$), followed by omega-3 fatty acid ($11.85\%$, $$n = 326$$) and ginseng and red ginseng ($7.99\%$, $$n = 210$$). ## 3.2. Association between Different Dietary Supplements and CKD Table 2 displays dietary supplements statuses of the participants. Overall, amino acids and protein ($p \leq 0.031$), ginseng and red ginseng ($p \leq 0.009$), and herbal medicine (plant extract)-berry ($p \leq 0.046$) users had a higher prevalence of CKD compared to non-users. However, the CKD prevalence was lower in the group consuming probiotics than in non-user group ($p \leq 0.017$). Age is an important confounder; therefore, we redefined the CKD group according to the age-adapted eGFR threshold proposed in a recent study [17], and found that the prevalence of CKD was also high in the group that consumed ginseng and red ginseng ($p \leq 0.03$) (Supplementary Table S1). However, the above results cannot rule out the possibility that CKD patients could be overusing dietary supplements. Therefore, to clarify the direct association of dietary supplements with the development of CKD, we further analyzed dietary supplement intake in healthy individuals without a history or risk factors for CKD such as diabetes, hypertension, hyperlipidemia, and cardiovascular diseases. The results showed that the group consuming ginseng and red ginseng had a higher prevalence of CKD compared with the non-users ($p \leq 0.035$) (Supplementary Table S2). In subgroup analysis according to sex, we observed that female participants who consumed amino acids and proteins ($p \leq 0.009$), ginseng and red ginseng ($p \leq 0.007$), and herbal medicine (plant extracts)-berries ($p \leq 0.009$) had a higher incidence of CKD than those who did not (Supplementary Table S3). Logistic regression analysis was used to explore the association between the different dietary supplements and CKD (Table 3). After adjusting for age and sex, we observed that the prevalence of CKD was significantly high only in the herbal medicine (plant extract)-berry group (odds ratio (OR): 4.598, $95\%$ confidence interval (CI): 1.123–18.821). Based on Model 2, we added smoking status, drinking status, education level, activity level, BMI, SBP, DBP, fasting glucose, and triglyceride for adjustment (Model 3). This result was similar to that of Model 2 (OR, 4.809; $95\%$ CI, 1.077–21.473). ## 4. Discussion The present large cross-sectional study assessed the extent of dietary supplement use in Korea and found that participants who consumed some dietary supplements (e.g., acids and protein, ginseng and red ginseng, and herbal medicine-berry) had a higher CKD prevalence than those who did not. In addition, the group consuming probiotics had a lower prevalence of CKD than the group that did not consume probiotics. To our knowledge, this is the first study to examine the association between dietary supplements and CKD in an Asian cohort of over 10,000 participants. Similar to those previously reported in other studies [5,10,18,19], the most commonly consumed dietary supplements in Korea were vitamin and mineral ($61.41\%$), omega-3 fatty acid ($11.85\%$) and probiotics (pre-, post-) ($4.72\%$). However, the important difference is that Koreans prefer to consume ginseng and red ginseng compared with people in other countries ($7.99\%$). Ginseng is mainly grown in East Asian countries, and Korea is one of the world’s leading producers of ginseng [20,21]. It has been valued for its remarkable therapeutic properties. Increasing evidence has demonstrated that ginsenoside, the main component of ginseng, has antioxidant, anti-apoptotic, and inhibitory effects on inflammatory cytokines. It can also modulate blood pressure and metabolism [21,22]. The association between ginseng and red ginseng consumption and kidney disease is controversial and not fully understood. Karunasagara et al. [ 23] demonstrated that red ginseng showed renoprotective effects in streptozotocin-induced diabetic rats and suppressed renal inflammation and fibrosis by blocking TGF-β1 activation. Sun et al. [ 24] revealed that ginsenoside Rb1 can reduce renal apoptosis and alleviate renal dysfunction by activating the Nrf2/ARE signaling pathway and enhancing heme oxygenase expression. However, clinical studies have reported conflicting results. A randomized controlled trial found that long-term ginseng intake did not affect renal function in patients [25]. Additionally, other studies have reported the negative effects of ginseng consumption on the kidneys [26,27,28]. In the present study, we found that participants who consumed ginseng and red ginseng had a high prevalence of CKD. The results were the same even in a healthy population with no history or risk factors for CKD, suggesting that ginseng consumption may have a direct effect on kidney function. Although the mechanism is unknown, it is possible that ginseng has been consumed in excessive quantities, or that interactions between ginseng and drugs have had an effect. A recent study reported that ginseng has a strong interaction effect, especially in patients taking anticoagulants [29]. Considering that dietary supplementation users’ insufficient recognition of the herb-drug interactions, standards and guidelines for the safety of dietary supplements based on information about the medications taken are needed. Probiotics, prebiotics, and synbiotics are important dietary supplements whose market has rapidly increased in recent years. Several studies have revealed that a bidirectional relationship exists between the gut microbiota and the kidney, and CKD itself can trigger dysbiosis and an altered intestinal environment. This substantial derangement is mainly caused by the decreased consumption of dietary fibers, frequent use of antibiotics, intestinal wall edema, metabolic acidosis, and uremia [30,31,32]. Recent studies have shown that dysbiosis and leaky gut in CKD are associated with an altered mucosal immune response through activation of intestinal immune cells with inflammatory cytokine production, potentially resulting in systemic inflammation and exacerbated cardiovascular/renal complications [33,34]. Therefore, improving the gut environment is considered one of the interventions to slow down the progression of CKD and prevent CKD-related complications, and probiotic consumption has attracted much attention as an adjuvant therapy to modulate gut dysbiosis. Several animal studies have demonstrated that probiotic supplements improve renal inflammation and fibrosis progression in animals with CKD [35,36]. In addition, some clinical studies have demonstrated that synbiotics have beneficial effects on improving intestinal dysbiosis and reducing serum p-cresyl sulfate levels in pre-dialysis CKD patients, and probiotic supplementation improves glucose homeostasis and systemic inflammation in dialysis patients [37,38,39]. However, owing to the limited number of studies and small sample sizes, they do not provide strong evidence for the efficacy of prebiotics or probiotics in treating CKD patients [39]. In our study, the group that consumed probiotics (pre-and post-) had a lower prevalence of CKD than the group that did not consume probiotics (pre-and post-). Given the abundance of evidence indicating the importance of kidney-gut interactions in patients with kidney disease, clinical studies are needed to demonstrate the effectiveness of microbiome-modifying therapies in large-scale CKD patients. Modifying protein intake is an important dietary strategy for slowing CKD. Several RCTs have evaluated the effect of dietary protein restriction on renal outcomes, and overall, they suggested the benefit of dietary protein restriction. The 2020 Kidney Disease Outcomes Quality Initiative guidelines recommend dietary protein restriction in patients with metabolically stable pre-dialysis CKD to reduce disease progression or mortality [40]. According to a prospective cohort study based on the Korean Genome and Epidemiology Study conducted between 2001 and 2014, high total protein intake-induced renal hyperfiltration causes fast eGFR decline in healthy adults with normal renal function [41]. A large Italian general-population study also evaluated the effects of protein intake on serum creatinine and eGFR through questionnaire [42]. Results showed an association between protein intake and decreased renal function, and also confirmed the association between higher protein diets and eGFR levels, even excluding participants with known diabetes, hypertension or CKD. Similarly, we observed a higher prevalence of CKD among participants consuming amino acid or protein supplements. According to a recent study, the average dietary protein intake of Koreans is almost twice the estimated average requirement [43]. Therefore, additional protein supplementation can lead to an excessive protein load on the body and renal damage in several ways through increased glomerular pressure and an additional acid load on the kidney [42,44,45]. Therefore, protein supplementation should be prescribed with particular caution in patients with kidney disease, and further evaluation of the effects of prolonged exposure to high protein intake on renal function in healthy subjects is warranted. Interestingly, we observed that consumption of herbal supplements, including berries, was associated with CKD incidence, even after adjusting for the major risk factors for the development of CKD, such as age, BMI, SBP, DBP, and fasting glucose level; this relationship was prevalent among participants with a history of CKD and CKD risk factors. Berries have conquered the global market as dietary supplements rich in vitamins, dietary prebiotic fibers, and micronutrients (e.g., zinc and iron) [46]. Morsy. et al. [ 47] showed that prophylactic administration of açaí berry extract in an ischemia-reperfusion animal model can improve renal function parameters and suppress the expression of renal proinflammatory cytokines and endothelin-1. Nair et al. [ 48] found that blueberries can protect rats with metabolic syndrome from chronic kidney injury by inhibiting Toll-like receptor 4 and attenuating mitogen-activated protein kinase activity. This renoprotective effect is attributed to the high content of flavonoids, polyphenols, and other bioactive compounds with powerful antioxidant and anti-inflammatory properties [46,49]. However, high concentrations of flavonoids in berries, especially anthocyanins, have been reported to have a cyclooxygenase (COX) inhibitory action similar to non-steroidal anti-inflammatory drugs (NSAIDs) [26,50]. NSAID-induced COX inhibition is known to be associated with CKD progression; therefore, we do not exclude the possibility that chronic use of berries may produce a similar clinical phenotype [51,52]. Another noteworthy relationship between berry consumption and kidney disease is that some berries contain a significant amount of potassium, such as 100 g of blackcurrants contain 322 mg of potassium [53], and hyperkalemia may be associated with the risk of arrhythmia or worsening of heart failure and CKD [54,55]. However, in this study, further analysis such as potassium levels or possible toxicants in berry consumers could not be performed and berries were not analyzed by type. Considering that some types of berries have a positive effect on kidney health, additional research is needed to analyze changes in kidney function after consumption of different berry types. This study has several important advantages. Despite the steady growth of the global supplement market over the past decade, there is a dearth of information regarding the association between dietary supplements and kidney disease. This study did not simply examine the association between dietary supplements and CKD but explored as many as 17 different dietary supplements. In addition, we analyzed the general Korean population of more than 10,000 individuals using KNHANES data that are representative of the entire population. Third, to reduce the effects of the interaction between different dietary supplements, we excluded participants who consumed two or more dietary supplements from the study design. However, this study also has some limitations. First, because this was a cross-sectional study, the results cannot prove a causal relationship between CKD and dietary supplements. Secondly, since these data are only for Korean adults, the results may not be extended to other ethnic groups, and future studies in other countries and races should be conducted. ## 5. Conclusions Our study was a large cross-sectional study that examined the association between dietary supplementation and CKD in >10,000 individuals. The study found that high incidence of CKD in the groups consuming dietary supplements, indicating that not all dietary supplements were safe for kidneys. Due to the growth of the dietary supplement market, we require special attention to the potential risk of dietary supplements to the kidney health. Furthermore, future large cohort studies are needed to provide knowledge about the precise role of each dietary supplement in CKD progression and to establish safety recommendations and guidelines for dietary supplements. ## References 1. Levin A., Tonelli M., Bonventre J., Coresh J., Donner J.A., Fogo A.B., Fox C.S., Gansevoort R.T., Heerspink H.J.L., Jardine M.. Global kidney health 2017 and beyond: A roadmap for closing gaps in care, research, and policy. Lancet (2017.0) 390 1888-1917. DOI: 10.1016/S0140-6736(17)30788-2 2. Chen J.H., Tsai C.C., Liu Y.H., Wu P.Y., Huang J.C., Chung T.L., Su H.M., Chen S.C.. Sex Difference in the Associations among Hyperuricemia with New-Onset Chronic Kidney Disease in a Large Taiwanese Population Follow-Up Study. Nutrients (2022.0) 14. DOI: 10.3390/nu14183832 3. Luo S., Grams M.E.. Epidemiology research to foster improvement in chronic kidney disease care. Kidney Int. (2020.0) 97 477-486. DOI: 10.1016/j.kint.2019.11.010 4. Lin P.C., Chou C.L., Ou S.H., Fang T.C., Chen J.S.. Systematic Review of Nutrition Supplements in Chronic Kidney Diseases: A GRADE Approach. Nutrients (2021.0) 13. DOI: 10.3390/nu13020469 5. Ronis M.J.J., Pedersen K.B., Watt J.. Adverse Effects of Nutraceuticals and Dietary Supplements. Annu. Rev. Pharmacol. Toxicol. (2018.0) 58 583-601. DOI: 10.1146/annurev-pharmtox-010617-052844 6. Tan E.C.K., Eshetie T.C., Gray S.L., Marcum Z.A.. Dietary Supplement Use in Middle-aged and Older Adults. J. Nutr. Health Aging (2022.0) 26 133-138. DOI: 10.1007/s12603-022-1732-9 7. Wierzejska R.E.. Dietary Supplements-For Whom? The Current State of Knowledge about the Health Effects of Selected Supplement Use. Int. J. Environ. Res. Public Health (2021.0) 18. DOI: 10.3390/ijerph18178897 8. Crawford C., Brown L.L., Costello R.B., Deuster P.A.. Select Dietary Supplement Ingredients for Preserving and Protecting the Immune System in Healthy Individuals: A Systematic Review. Nutrients (2022.0) 14. DOI: 10.3390/nu14214604 9. Kim M., Lee Y., Park K.. Vitamin and Mineral Supplement Use among Korean Adults: Baseline Data from the Trace Element Study of Korean Adults in Yeungnam Area. Nutrients (2018.0) 10. DOI: 10.3390/nu10010050 10. Choi J.. The Association between Health Conditions, Consciousness, Involvement, and Knowledge and Dietary Supplement Intake among University Students in South Korea. Int. J. Environ. Res. Public Health (2019.0) 16. DOI: 10.3390/ijerph16204028 11. Rovin B.H., Adler S.G., Barratt J., Bridoux F., Burdge K.A., Chan T.M., Cook H.T., Fervenza F.C., Gibson K.L., Glassock R.J.. Executive summary of the KDIGO 2021 Guideline for the Management of Glomerular Diseases. Kidney Int. (2021.0) 100 753-779. DOI: 10.1016/j.kint.2021.05.015 12. Junarta J., Jha V., Banerjee D.. Insight into the impact of vitamin D on cardiovascular outcomes in chronic kidney disease. Nephrology (2019.0) 24 781-790. DOI: 10.1111/nep.13569 13. Tian N., Li L., Ng J.K., Li P.K.. The Potential Benefits and Controversies of Probiotics Use in Patients at Different Stages of Chronic Kidney Disease. Nutrients (2022.0) 14. DOI: 10.3390/nu14194044 14. Zheng H.J., Guo J., Wang Q., Wang L., Wang Y., Zhang F., Huang W.J., Zhang W., Liu W.J., Wang Y.. Probiotics, prebiotics, and synbiotics for the improvement of metabolic profiles in patients with chronic kidney disease: A systematic review and meta-analysis of randomized controlled trials. Crit. Rev. Food Sci. Nutr. (2021.0) 61 577-598. DOI: 10.1080/10408398.2020.1740645 15. The Homepage of Korea National Health and Nutrition Examination Survey 16. Health Functional Food Code 17. Delanaye P., Jager K.J., Bökenkamp A., Christensson A., Dubourg L., Eriksen B.O., Gaillard F., Gambaro G., van der Giet M., Glassock R.J.. CKD: A Call for an Age-Adapted Definition. J. Am. Soc. Nephrol. JASN (2019.0) 30 1785-1805. DOI: 10.1681/ASN.2019030238 18. Alsaleem S.A., Asiri M.M., Alsaleem M.A., AlShahrani A.N., Alamer K.A., Mahfouz A.A.. Dietary Supplement Use among Primary Health Care Attendants in Abha City, Southwestern Saudi Arabia. Nutrients (2021.0) 13. DOI: 10.3390/nu13092968 19. Brown A.C.. An overview of herb and dietary supplement efficacy, safety and government regulations in the United States with suggested improvements. Part 1 of 5 series. Food Chem. Toxicol. (2017.0) 107 449-471. DOI: 10.1016/j.fct.2016.11.001 20. Xu W., Choi H.K., Huang L.. State of Panax ginseng Research: A Global Analysis. Molecules (2017.0) 22. DOI: 10.3390/molecules22091518 21. Zhang H., Abid S., Ahn J.C., Mathiyalagan R., Kim Y.J., Yang D.C., Wang Y.. Characteristics of Panax ginseng Cultivars in Korea and China. Molecules (2020.0) 25. DOI: 10.3390/molecules25112635 22. Park S.H., Chung S., Chung M.Y., Choi H.K., Hwang J.T., Park J.H.. Effects of Panax ginseng on hyperglycemia, hypertension, and hyperlipidemia: A systematic review and meta-analysis. J. Ginseng Res. (2022.0) 46 188-205. DOI: 10.1016/j.jgr.2021.10.002 23. Karunasagara S., Hong G.L., Park S.R., Lee N.H., Jung D.Y., Kim T.W., Jung J.Y.. Korean red ginseng attenuates hyperglycemia-induced renal inflammation and fibrosis via accelerated autophagy and protects against diabetic kidney disease. J. Ethnopharmacol. (2020.0) 254 112693. DOI: 10.1016/j.jep.2020.112693 24. Sun Q., Meng Q.T., Jiang Y., Liu H.M., Lei S.Q., Su W.T., Duan W.N., Wu Y., Xia Z.Y., Xia Z.Y.. Protective effect of ginsenoside Rb1 against intestinal ischemia-reperfusion induced acute renal injury in mice. PLoS ONE (2013.0) 8. DOI: 10.1371/journal.pone.0080859 25. Stavro P.M., Woo M., Leiter L.A., Heim T.F., Sievenpiper J.L., Vuksan V.. Long-term intake of North American ginseng has no effect on 24-hour blood pressure and renal function. Hypertension (2006.0) 47 791-796. DOI: 10.1161/01.HYP.0000205150.43169.2c 26. Grubbs V., Plantinga L.C., Tuot D.S., Hedgeman E., Saran R., Saydah S., Rolka D., Powe N.R.. Americans’ use of dietary supplements that are potentially harmful in CKD. Am. J. Kidney Dis. (2013.0) 61 739-747. DOI: 10.1053/j.ajkd.2012.12.018 27. Lee J.H., Yoon J.H., Kim S.S., Ma S.K., Kim S.W., Bae E.H.. Panax Ginseng Induces Toxic Hepatitis and Acute Kidney Injury. Chonnam Med. J. (2017.0) 53 168-169. DOI: 10.4068/cmj.2017.53.2.168 28. Paik D.J., Lee C.H.. Review of cases of patient risk associated with ginseng abuse and misuse. J. Ginseng Res. (2015.0) 39 89-93. DOI: 10.1016/j.jgr.2014.11.005 29. Mohammadi S., Asghari G., Emami-Naini A., Mansourian M., Badri S.. Herbal Supplement Use and Herb-drug Interactions among Patients with Kidney Disease. J. Res. Pharm. Pract. (2020.0) 9 61-67. DOI: 10.4103/jrpp.JRPP_20_30 30. Mahmoodpoor F., Rahbar Saadat Y., Barzegari A., Ardalan M., Zununi Vahed S.. The impact of gut microbiota on kidney function and pathogenesis. Biomed. Pharmacother. Biomed. Pharmacother. (2017.0) 93 412-419. DOI: 10.1016/j.biopha.2017.06.066 31. Feng Z., Wang T., Dong S., Jiang H., Zhang J., Raza H.K., Lei G.. Association between gut dysbiosis and chronic kidney disease: A narrative review of the literature. J. Int. Med. Res. (2021.0) 49 3000605211053276. DOI: 10.1177/03000605211053276 32. Kim M.G., Yang J., Jo S.K.. Intestinal microbiota and kidney diseases. Kidney Res. Clin. Pract. (2021.0) 40 335-343. DOI: 10.23876/j.krcp.21.053 33. Yang J., Lim S.Y., Ko Y.S., Lee H.Y., Oh S.W., Kim M.G., Cho W.Y., Jo S.K.. Intestinal barrier disruption and dysregulated mucosal immunity contribute to kidney fibrosis in chronic kidney disease. Nephrol Dial Transpl. (2019.0) 34 419-428. DOI: 10.1093/ndt/gfy172 34. Ramezani A., Raj D.S.. The gut microbiome, kidney disease, and targeted interventions. J. Am. Soc. Nephrol. (2014.0) 25 657-670. DOI: 10.1681/ASN.2013080905 35. Wang I.K., Yen T.H., Hsieh P.S., Ho H.H., Kuo Y.W., Huang Y.Y., Kuo Y.L., Li C.Y., Lin H.C., Wang J.Y.. Effect of a Probiotic Combination in an Experimental Mouse Model and Clinical Patients with Chronic Kidney Disease: A Pilot Study. Front. Nutr. (2021.0) 8 661794. DOI: 10.3389/fnut.2021.661794 36. Zhu H., Cao C., Wu Z., Zhang H., Sun Z., Wang M., Xu H., Zhao Z., Wang Y., Pei G.. The probiotic L. casei Zhang slows the progression of acute and chronic kidney disease. Cell Metab. (2021.0) 33 1926-1942. DOI: 10.1016/j.cmet.2021.06.014 37. Rossi M., Johnson D.W., Morrison M., Pascoe E.M., Coombes J.S., Forbes J.M., Szeto C.C., McWhinney B.C., Ungerer J.P., Campbell K.L.. Synbiotics Easing Renal Failure by Improving Gut Microbiology (SYNERGY): A Randomized Trial. Clin. J. Am. Soc. Nephrol. CJASN (2016.0) 11 223-231. DOI: 10.2215/CJN.05240515 38. Soleimani A., Zarrati Mojarrad M., Bahmani F., Taghizadeh M., Ramezani M., Tajabadi-Ebrahimi M., Jafari P., Esmaillzadeh A., Asemi Z.. Probiotic supplementation in diabetic hemodialysis patients has beneficial metabolic effects. Kidney Int. (2017.0) 91 435-442. DOI: 10.1016/j.kint.2016.09.040 39. Choi E., Yang J., Ji G.E., Park M.S., Seong Y., Oh S.W., Kim M.G., Cho W.Y., Jo S.K.. The effect of probiotic supplementation on systemic inflammation in dialysis patients. Kidney Res. Clin. Pract. (2022.0) 41 89-101. DOI: 10.23876/j.krcp.21.014 40. Ikizler T.A., Burrowes J.D., Byham-Gray L.D., Campbell K.L., Carrero J.J., Chan W., Fouque D., Friedman A.N., Ghaddar S., Goldstein-Fuchs D.J.. KDOQI Clinical Practice Guideline for Nutrition in CKD: 2020 Update. Am. J. Kidney Dis. Off. J. Natl. Kidney Found. (2020.0) 76 S1-S107. DOI: 10.1053/j.ajkd.2020.05.006 41. Jhee J.H., Kee Y.K., Park S., Kim H., Park J.T., Han S.H., Kang S.W., Yoo T.H.. High-protein diet with renal hyperfiltration is associated with rapid decline rate of renal function: A community-based prospective cohort study. Nephrol. Dial. Transplant. (2020.0) 35 98-106. DOI: 10.1093/ndt/gfz115 42. Vukovic V., Hantikainen E., Raftopoulou A., Gögele M., Rainer J., Domingues F.S., Pramstaller P.P., Garcia-Larsen V., Pattaro C.. Association of dietary proteins with serum creatinine and estimated glomerular filtration rate in a general population sample: The CHRIS study. J. Nephrol. (2022.0) 36 103-114. DOI: 10.1007/s40620-022-01409-7 43. Kim E., Chung S., Hwang J.-T., Park Y.J.. 2020 Korean Dietary Reference Intakes for Protein: Estimation of protein requirements and the status of dietary protein intake in the Korean population. J. Nutr. Health (2022.0) 55 10-20. DOI: 10.4163/jnh.2022.55.1.10 44. Ko G.J., Obi Y., Tortorici A.R., Kalantar-Zadeh K.. Dietary protein intake and chronic kidney disease. Curr. Opin. Clin. Nutr. Metab. Care (2017.0) 20 77-85. DOI: 10.1097/MCO.0000000000000342 45. Oba R., Kanzaki G., Sasaki T., Okabayashi Y., Haruhara K., Koike K., Kobayashi A., Yamamoto I., Tsuboi N., Yokoo T.. Dietary Protein Intake and Single-Nephron Glomerular Filtration Rate. Nutrients (2020.0) 12. DOI: 10.3390/nu12092549 46. Nunes S., Vieira P., Gomes P., Viana S.D., Reis F.. Blueberry as an Attractive Functional Fruit to Prevent (Pre)Diabetes Progression. Antioxidants (2021.0) 10. DOI: 10.3390/antiox10081162 47. El Morsy E.M., Ahmed M.A., Ahmed A.A.. Attenuation of renal ischemia/reperfusion injury by açaí extract preconditioning in a rat model. Life Sci. (2015.0) 123 35-42. DOI: 10.1016/j.lfs.2014.11.013 48. Nair A.R., Elks C.M., Vila J., Del Piero F., Paulsen D.B., Francis J.. A blueberry-enriched diet improves renal function and reduces oxidative stress in metabolic syndrome animals: Potential mechanism of TLR4-MAPK signaling pathway. PLoS ONE (2014.0) 9. DOI: 10.1371/journal.pone.0111976 49. Ma L., Sun Z., Zeng Y., Luo M., Yang J.. Molecular Mechanism and Health Role of Functional Ingredients in Blueberry for Chronic Disease in Human Beings. Int. J. Mol. Sci. (2018.0) 19. DOI: 10.3390/ijms19092785 50. Belmonte-Herrera B.H., Domínguez-Avila J.A., Wall-Medrano A., Ayala-Zavala J.F., Preciado-Saldaña A.M., Salazar-López N.J., López-Martínez L.X., Yahia E.M., Robles-Sánchez R.M., González-Aguilar G.A.. Lesser-Consumed Tropical Fruits and Their by-Products: Phytochemical Content and Their Antioxidant and Anti-Inflammatory Potential. Nutrients (2022.0) 14. DOI: 10.3390/nu14173663 51. Wan E.Y.F., Yu E.Y.T., Chan L., Mok A.H.Y., Wang Y., Chan E.W.Y., Wong I.C.K., Lam C.L.K.. Comparative Risks of Nonsteroidal Anti-Inflammatory Drugs on CKD. Clin. J. Am. Soc. Nephrol. CJASN (2021.0) 16 898-907. DOI: 10.2215/CJN.18501120 52. Gooch K., Culleton B.F., Manns B.J., Zhang J., Alfonso H., Tonelli M., Frank C., Klarenbach S., Hemmelgarn B.R.. NSAID use and progression of chronic kidney disease. Am. J. Med. (2007.0) 120 280.e281-e287. DOI: 10.1016/j.amjmed.2006.02.015 53. Are Berries High in Potassium? 54. Kovesdy C.P., Matsushita K., Sang Y., Brunskill N.J., Carrero J.J., Chodick G., Hasegawa T., Heerspink H.L., Hirayama A., Landman G.W.D.. Serum potassium and adverse outcomes across the range of kidney function: A CKD Prognosis Consortium meta-analysis. Eur. Heart J. (2018.0) 39 1535-1542. DOI: 10.1093/eurheartj/ehy100 55. Yamada S., Inaba M.. Potassium Metabolism and Management in Patients with CKD. Nutrients (2021.0) 13. DOI: 10.3390/nu13061751
--- title: Retinol and Pro-Vitamin A Carotenoid Nutritional Status during Pregnancy Is Associated with Newborn Hearing Screen Results authors: - Rebecca Slotkowski - Matthew Van Ormer - Anum Akbar - Olivia Paetz - Taija Hahka - Maranda Thompson - Alyssa Freeman - Alexandra Hergenrader - Sarah Sweeney - Zeljka Korade - Thiago Genaro-Mattos - Corrine Hanson - Ann Anderson-Berry - Melissa Thoene journal: Nutrients year: 2023 pmcid: PMC9967333 doi: 10.3390/nu15040800 license: CC BY 4.0 --- # Retinol and Pro-Vitamin A Carotenoid Nutritional Status during Pregnancy Is Associated with Newborn Hearing Screen Results ## Abstract The prenatal period is critical for auditory development; thus, prenatal influences on auditory development may significantly impact long-term hearing ability. While previous studies identified a protective effect of carotenoids on adult hearing, the impact of these nutrients on hearing outcomes in neonates is not well understood. The purpose of this study is to investigate the relationship between maternal and umbilical cord plasma retinol and carotenoid concentrations and abnormal newborn hearing screen (NHS) results. Mother–infant dyads ($$n = 546$$) were enrolled at delivery. Plasma samples were analyzed using HPLC and LC–MS/MS. NHS results were obtained from medical records. Statistical analysis utilized Mann–Whitney U tests and logistic regression models, with p ≤ 0.05 considered statistically significant. Abnormal NHS results were observed in $8.5\%$ of infants. Higher median cord retinol (187.4 vs. 162.2 μg/L, $$p \leq 0.01$$), maternal trans-β-carotene (206.1 vs. 149.4 μg/L, $$p \leq 0.02$$), maternal cis-β-carotene (15.9 vs. 11.2 μg/L, $$p \leq 0.02$$), and cord trans-β-carotene (15.5 vs. 8.0 μg/L, $$p \leq 0.04$$) were associated with abnormal NHS. Significant associations between natural log-transformed retinol and β-carotene concentrations and abnormal NHS results remained after adjustment for smoking status, maternal age, and corrected gestational age. Further studies should investigate if congenital metabolic deficiencies, pesticide contamination of carotenoid-rich foods, maternal hypothyroidism, or other variables mediate this relationship. ## 1. Introduction Pro-vitamin A carotenoids, including α-carotene, β-carotene, and β-cryptoxanthin, are anti-inflammatory nutrients found in vegetables, fruits, and fish which can be converted into vitamin A in the body [1,2,3]. Vitamin A plays an important role in pregnancy and fetal development by regulating normal organogenesis, tissue differentiation, and development of the immune system and the inner ear [4]. Furthermore, numerous studies have shown that the antioxidant and anti-inflammatory properties of carotenoids can ameliorate pregnancy-associated morbidities such as pre-eclampsia, intrauterine growth restriction, gestational diabetes, and pregnancy induced hypertension [5,6,7,8]. However, vitamin A deficiency is still prevalent and is considered a worldwide public health problem, affecting an estimated 19 million pregnancies every year [9]. This deficiency is associated with long-term health consequences for the infant, including hearing loss [10]. Vitamin A is indispensable in inner ear development. During in utero development, essential enzymes in the inner ear convert vitamin A (retinol) into a biologically active form, retinoic acid. Previous studies reported that retinoic acid regulates fibroblast growth factors (FGF) such as FGF3 and FGF10, which may in turn modulate several downstream target molecules that are necessary for normal inner ear development [11,12]. Surface ectoderm gives rise to the otic placode during the first trimester of gestation, which in turn forms the mature inner ear. Studies have shown that vitamin A induces development of the otic placode, which further corroborates the role of vitamin A in inner ear development. In fact, experiments in rats, zebrafish, and chicks demonstrate that deficiency of vitamin A leads to aberrant otic placode development [12,13]. Developmental defects of the inner ear, which is composed of the cochlear and vestibular systems, are a major cause of congenital sensorineural hearing loss [14,15] and approximately one-third of children with congenital sensorineural hearing loss have significant hearing impairment [16]. The American Academy of Pediatrics recommends that a newborn hearing screen (NHS) should be done within the first month of life to detect congenital sensorineural hearing loss [17]. There are several known causes of congenital hearing loss including genetic factors and congenital infections (such as congenital cytomegalovirus, toxoplasmosis, and rubella infection). Early detection of and intervention for congenital sensorineural hearing loss is important, as previous studies have shown that children diagnosed with congenital sensorineural hearing loss earlier have better language and socioemotional outcomes, likely resulting from early interventions [18]. The protocols for NHS vary in different countries; in the United States, NHS may include otoacoustic emissions (OAE) or automated auditory brainstem response (AABR) tests [19]. OAE is helpful in assessing the inner ear (cochlear function) whereas automated AABR assesses the auditory pathway in addition to assessing the function of the external and inner ear [20,21]. This comprehensive screening of newborns reduces the chance of missing auditory neuropathy spectrum disorders [22]. However, although previous studies have demonstrated that vitamin A is required for normal ear development, there is a gap in our knowledge regarding how maternal and infant nutritional status of pro-vitamin A carotenoids are associated with the results of these newborn hearing screens. To address this gap, we conducted a study to assess the relationship between NHS results and retinol and pro-vitamin A carotenoid maternal intake, maternal plasma levels, and umbilical cord plasma levels. As pro-vitamin A carotenoids can be converted into vitamin A and vitamin A plays a key role in inner ear development, we hypothesized that higher concentrations of retinol, α-carotene, β-carotene, and β-cryptoxanthin in maternal and umbilical cord plasma would correlate with better NHS outcomes. ## 2.1. Participant Enrollment Ethical approval for this study was obtained from the University of Nebraska Medical Center Institutional Review Board (#112-15-EP). Eligible mothers were enrolled at the time of delivery ($$n = 546$$). Mothers provided written consent for both themselves and their infant(s) prior to participation. Inclusion criteria included women ≥ 19 years old admitted to the Labor and Delivery Unit at Nebraska Medicine during the study period (June 2015 to August 2021) who delivered at least one live-born infant. Exlcusion criteria included gastrointestinal, liver, or kidney diseases which would affect normal nutrient metabolism, inborn errors of metabolism, congenital abnormalities in the infant, and infants who were deemed wards of the state. ## 2.2. Biological Sample Collection Maternal blood samples were collected in K2 EDTA tubes during routine labs upon admission for delivery. Umbilical cord blood samples were collected in K2 EDTA tubes during routine labs at the time of delivery. Whole blood samples were protected from heat and light to preserve nutrient integrity. Samples were seperated into plasma and frozen at −80 °C within a maximum of 12 h after collection. Blood samples were available for 404 maternal–infant dyads. ## 2.3. Carotenoid Laboratory Analysis Plasma samples ($$n = 404$$) were analyzed for retinol and pro-vitamin A carotenoids (α-carotene, β-carotene, and β-cryptoxanthin). The first 328 dyad plasma samples were analyzed at the Biomarker Research Institute at the Harvard T.H. Chan School of Public Health via HPLC as previously described by Thoene et al. [ 23]. The remaining 76 dyad plasma samples were analyzed at the University of Nebraska Medical Center via LC–MS/MS as previously described by McConnell et al. [ 24]. Quality control at both labs was achieved using NIST standards. ## 2.4. Dietary Intake and Socioeconomic Questionnaires The Harvard Food Frequency Questionnaire (FFQ) [25] was administered to maternal participants at the time of delivery by trained study personnel. The FFQ was completed for 472 participants. De-identified FFQs were analyzed by the Harvard T.H. Chan School of Public Health to quantify average daily nutrient intake from foods and supplements over the previous year. The FFQ is validated for use in pregnant women [26]. Women with estimated daily caloric intakes outside a probable range (800–8000 Cal/day) were excluded from analysis [27]. A socioeconomic status questionnaire collected information on annual household income and household size prior to delivery of the infant ($$n = 363$$). The income-to-poverty ratio was calculated by dividing the reported annual household income by the federal poverty level for the reported household size [28]. ## 2.5. NHS Results and Clinical Data Collection NHS were conducted using automated auditory brainstem response (AABR; NATUS ALGO® 5 Newborn Hearing Screener) prior to infant discharge from the Newborn Nursery or Neonatal Intensive Care Unit. To minimize the rate of falsely abnormal AABR results [29], NHS were conducted per standard of care guidelines. Initial NHS results were collected from the infant electronic medical record (EMR). Additional variables collected from the infant EMR included corrected gestational age at birth (CGA), sex, NICU admission, birthweight, days of antibiotic therapy, and days of oxygen therapy. Maternal age, race/ethnicity, and smoking status were collected from the maternal EMR. ## 2.6. Statistical Analysis Descriptive statistics were calculated, including medians and interquartile ranges for continuous variables and frequencies and percentages for categorical variables. Retinol intake was categorized as insufficient (≤770 μg RAE/day), adequate (771–2999 μg RAE/day), or excessive (≥3000 μg RAE/day) according to Institute of Medicine guidelines [30] and retinol plasma concentrations were categorized as deficient (≤200.5 μg/L), insufficient (200.5–300.8 μg/L), or adequate (>300.8 μg/L) according to the World Health Organization guidelines [9]. Chi-square tests were used to identify differences in categorical retinol nutritional status between normal vs. abnormal NHS groups. Mann–Whitney U tests were used to identify differences in median retinol and pro-vitamin A carotenoids intake/plasma concentrations between normal vs. abnormal NHS groups. Logistic regression models were used to evaluate the association between natural-log transformed plasma nutrient concentrations and NHS results, with adjustment for CGA, maternal age, and smoking status (Model 1) or CGA, maternal age, smoking status, and income-to-poverty ratio (Model 2). Odds ratios with $95\%$ confidence intervals ($95\%$ CI) are reported for a 1-unit increase in natural log-transformed nutrient concentrations (ORln = eB; $95\%$ CI = eB ± 1.96SE) and for a $10\%$ increase in nutrient plasma concentrations (not natural-log transformed; OR$10\%$ = 1.1B; $95\%$ CI = 1.1B ± 1.96SE) [31]. A 1-unit increase in natural log-transformed nutrient concentration is equivalent to a $172\%$ increase in nutrient plasma concentration (not natural-log transformed). Subjects with non-detectable plasma concentrations or missing maternal intake information were excluded from the corresponding analysis. A p-value ≤ 0.05 was considered statistically significant. For twin pregnancies, mothers and twin A were included in the analysis; twin B was excluded. ## 3.1. Demographic Characteristics A total of 546 mother–infant dyads participated in this study, including 47 ($8.6\%$) infants with abnormal NHS results and 499 ($91.4\%$) with normal NHS results. The demographic characteristics of the sample population by NHS result are shown in Table 1. Data on CGA and maternal age was available for all participants. Income-to-poverty ratio and energy intake was available for 363 (330 normal, 33 abnormal NHS) and 472 (433 normal, 39 abnormal NHS) maternal–infant dyads, respectively. Although very low birth weight (≤1500 g), administration of antibiotics, and hypoxia (measured as days of oxygen therapy) have previously been associated with hearing loss in infants [32], these factors were rarely present in our population of infants with abnormal NHS. ## 3.2. Retinol Nutritional Status and NHS Results Maternal retinol intake over the past year was categorized into insufficient, adequate, or excessive using the Institute of Medicine guidelines [30]. Maternal and infant retinol plasma concentrations were categorized into deficient, insufficient, or adequate using WHO guidelines [9]. There were no significant associations between either retinol intake groups or retinol plasma status categories and NHS result groups (Table 2). However, median cord retinol plasma concentrations were significantly higher among infants with abnormal NHS compared to those with normal NHS ($$p \leq 0.01$$; Table 3). After adjustment for CGA, maternal age, and smoking status, natural log-transformed cord plasma concentrations of retinol remained significantly associated with abnormal NHS (Table 4; Model 1). A $10\%$ increase in cord retinol plasma concentrations was associated with 1.15 ($95\%$ CI 1.05, 1.27; $$p \leq 0.01$$) times higher odds of abnormal NHS (Table 4; Model 1). Poverty-to-income ratio was available on a subset of maternal–infant dyads. After additional adjustment for income-to-poverty ratio, a $10\%$ increase in cord retinol plasma concentration was associated with 1.20 ($95\%$ CI 1.05, 1.36; $$p \leq 0.01$$) times higher odds of abnormal NHS results (Table 4; Model 2). There were no significant differences in maternal plasma retinol concentrations or maternal retinol intake between groups. ## 3.3. β-Carotene Intake, Maternal Plasma, and Cord Plasma Concentrations Median maternal plasma concentrations of trans-β-carotene ($$p \leq 0.02$$) and cis-β-carotene ($$p \leq 0.02$$) were significantly higher in mothers of infants with abnormal NHS results compared to mothers of infants with normal NHS results (Table 5). Cord plasma concentrations of trans-β-carotene ($$p \leq 0.02$$) and total β-carotene ($$p \leq 0.04$$) were significantly higher in infants with abnormal NHS results. Maternal β-carotene intake during pregnancy was not significantly different between mothers of infants with normal NHS versus abnormal NHS results. After adjustment for CGA, maternal age, and smoking status, natural log-transformed maternal trans-β-carotene ($95\%$ CI 1.00, 1.11; $$p \leq 0.04$$), maternal cis-β-carotene ($95\%$ CI 1.00, 1.11; $$p \leq 0.04$$), cord trans-β-carotene ($95\%$ CI 1.01, 1.12; $$p \leq 0.03$$), and cord total β-carotene ($95\%$ CI 1.02, 1.10; $$p \leq 0.003$$) remained significantly associated with abnormal NHS (Table 6; Model 1). After additional adjustment for poverty-to-income ratio, maternal trans-β-carotene ($95\%$ CI 1.02, 1.18; $$p \leq 0.02$$), maternal cis-β-carotene ($95\%$ CI 1.02, 1.18; $$p \leq 0.02$$), and cord total β-carotene ($95\%$ CI 1.00, 1.10; $$p \leq 0.05$$) remained significantly associated with abnormal NHS results (Table 6; Model 2). ## 3.4. Intake, Maternal Plasma, and Cord Plasma Concentration of α-carotene and β-cryptoxanthin No significant differences in α-carotene or β-cryptoxanthin intake, maternal plasma concentrations, or infant plasma concentrations were observed between NHS groups (Table 7). ## 4. Discussion Contrary to our original hypothesis, we detected that higher infant plasma concentrations of retinol, as well as higher maternal and infant plasma concentrations of β-carotene, were associated with increased odds of abnormal NHS. Other studies investigating the association between retinol and carotenoid nutritional status and hearing loss have focused on intake in adult populations, rather than plasma concentrations in infants. Although not statistically significant, median maternal intake of β-carotene was also higher for dyads in this study with an abnormal NHS result (5570.9 vs. 4881.8 μg/day, $$p \leq 0.15$$). These findings are in stark contrast to studies in adult populations that have demonstrated an opposite effect, with increased β-carotene intake being associated with a decreased likelihood of acquired hearing loss [33,34,35]. Differences in population age and mechanism of hearing loss (congenital vs. acquired) between our study cohort and previous studies in adult populations could explain our divergent findings. Studies examining the relationship between acquired hearing loss and retinol in children, rather than adults, have reported mixed results. One randomized placebo-controlled trial by Schmitz et al. reported that supplementation with vitamin A during preschool was associated with a decreased risk of hearing impairment during adolescence [36]. In contrast, a randomized placebo-controlled trial by Ambalavanan et al. reported that there was no difference in the risk of infant hearing impairment at 18 to 22 months adjusted age between extremely low birth weight infants who receive vitamin A supplementation compared to those who received placebo [37]. Similar to Ambalavanan et al., we observed no association between maternal retinol intake and odds of infant hearing impairment as measured by NHS results. However, infant plasma concentrations of both retinol and β-carotene were associated with increased odds of abnormal NHS results in our cohort. There are several potential explanations for the observed association between increased retinol and β-carotene plasma concentrations and increased odds of abnormal NHS. Excessive maternal vitamin A intake during gestation is associated with abnormal fetal inner ear development in both humans and animal models [12,38]. The Institute of Medicine Food and Nutrition Board recommends 770 μg RAE/day of vitamin A, with a tolerable upper limit of 3000 μg RAE/day for preformed vitamin A [30]; however, intake of up to 9000 μg RAE/day of preformed vitamin A, almost twelve times the recommended daily dose and three times the tolerable upper limit, has been shown to be safe during pregnancy [39]. Additionally, although β-carotene can be metabolized into vitamin A, excessive intake of β-carotene is not thought to result in vitamin A toxicity during pregnancy [39]. In this study, median maternal intake of retinol for the mothers of infants with abnormal NHS results over the past year was 1,755.9 μg RAE/day, well within recommended limits [30]. Among mothers of infants with abnormal NHS results, only three ($7.7\%$) consumed less than 770 μg RAE/day and only two ($5.1\%$) consumed more than 3000 μg RAE/day. These intakes are comparable to mothers of infants with normal NHS results, where $9.9\%$ consumed less than 770 μg RAE/day and $6.2\%$ consumed more than 3000 μg RAE/day. Likewise, although most infants in this study had insufficient or deficient retinol plasma concentrations of retinol, the median retinol plasma concentration was significantly higher among infants with an abnormal NHS. Alternatively, it is possible that impaired vitamin A metabolism or transport during inner ear development may be associated with both congenital hearing loss and increased plasma concentrations of retinol and β-carotene. During normal inner ear development, β-carotene is converted to retinol [1,2,3], and retinol is converted to the biologically active retinoic acid [11,12]. Retinoic acid is then utilized in signaling cascades to promote inner ear development [11,12]. Impaired metabolism of retinol into retinoic acid could result in accumulation of plasma retinol. A high concentration of retinol could, in turn, impair metabolism of β-carotene into retinol and result in accumulation of plasma β-carotene. Similarly, an imbalance in retinol-binding protein could result in impaired transport of retinol to target tissues, leading to accumulation of retinol, and ultimately β-carotene, in plasma. However, vitamin A plays an important role in organogenesis, tissue differentiation, and immune system development, in addition to inner ear development [4]. It is unlikely that impaired metabolism of vitamin A would result in congenital hearing loss without significantly affecting the development of other organ systems. In this study, $87\%$ of infants with an abnormal NHS result were otherwise healthy infants who were not admitted to the NICU at delivery, though long-term health of these infants has not been evaluated. Another potential explanation for the observed association between retinol and β-carotene and abnormal NHS is that infants with higher retinol and β-carotene plasma concentrations may have been exposed to higher levels of ototoxic environmental contaminants in vitamin A and carotenoid-rich foods, such as pesticides applied to fruits and vegetables. We observed a trend towards higher median maternal β-carotene intake in mothers of infants with an abnormal NHS result, which may support this theory, although the relationship was not significant and there were similarly no significant differences in maternal intake of other evaluated carotenoids. Multiple studies have reported associations between pesticide or other environmental contaminant exposure and hearing loss [40,41,42,43,44,45]. In a rat model, prenatal exposure to the environmental contaminants 2,3,7,8-tetrachorodibenzo-p-dioxin (TCDD) [40] or polychlorinated biphenyls (PCB) [41,42] were associated with hearing deficits. In humans, higher prenatal exposure to multiple organochlorine pesticides, as measured in cord plasma concentrations, were associated with significantly worse cochlear function, as measured by distortion product otoacoustic emission (DPOAE) at 45 months of age [43]. Alternative rationales may exist for these findings, including variables unknown or not evaluated. For example, metabolic and endocrine disorders such as hypothyroidism [46,47,48,49], diabetes [50,51], and polycystic ovarian syndrome (PCOS) [52,53] have been associated with both hearing loss and alterations in vitamin A nutritional status. Subclinical hypothyroidism is particularly common during pregnancy, affecting an estimated $15.5\%$ of pregnant women in the United States [54]. Interestingly, hypothyroidism is associated with increased concentrations of β-carotene [46,47]. Maternal hypothyroidism and congenital hypothyroidism are also linked to neurosensory hearing loss [48,49], although less is known about how subclinical hypothyroidism may affect infant hearing. One study by Radetti et al. failed to detected any significant association between maternal subclinical hypothyroidism and infant hearing outcomes [55], while a study by G et al. observed that infants born to mothers with subclinical hypothyroidism had minor alterations in hearing which self-corrected within 6–8 months [56]. As the American Thyroid Association does not currently recommend asymptomatic screening for subclinical hypothyroidism during pregnancy [57], thyroid function was not evaluated in this study. However, it is possible that maternal–infant dyads with subclinical hypothyroidism may jointly exhibit higher β-carotene concentrations and newborn hearing screen fails, though these results would not indicate a causal relationship between nutritional status of β-carotene and infant NHS results. Future research is therefore needed to evaluate the relationship between these variables. ## Limitations This study was conducted at a single academic medical center in the Midwest United States (University of Nebraska Medical Center/Nebraska Medicine) with a majority non-Hispanic White cohort of maternal–infant dyads, which may limit generalizability of our results. We were unable to collect information on some factors that have been previously associated with hearing loss in infants, such as hyperbilirubinemia, craniofacial abnormalities, administration of loop diuretics, or various environmental exposures (e.g., viral infections) [32]. However, our analysis did account for several other variables potentially associated with neonatal hearing loss, including gestational age of the infant, maternal age, maternal smoking status, and income-to-poverty ratio. Additionally, this analysis focused on associations between first NHS result and nutritional status at time of delivery. Neonates may have abnormal NHS results which resolve on repeat testing and nutritional status at time of delivery may differ from nutritional status during the critical period of inner ear development in the first trimester of pregnancy. Future studies should assess associations between diagnosed congenital hearing loss and retinol and β-carotene nutritional status across multiple timepoints in pregnancy. ## 5. Conclusions The observed relationship between higher retinol and β-carotene infant plasma concentrations and increased odds of abnormal NHS was unexpected. While other studies suggest both deficient and excessive levels of vitamin A can impact inner ear development, median retinol plasma levels in our study were within normal limits. One possible explanation for our observed results is that higher neonatal retinol and β-carotene plasma concentrations may be indicative of impaired retinol metabolism or transport to the inner ear. Alternatively, exposure to ototoxic environmental contaminants, maternal metabolic disorders, maternal endocrine disorders, perinatal infections, or other variables not assessed in this study could be associated with both increased retinol and β-carotene plasma concentrations and higher odds of abnormal NHS results. Future research should evaluate the association between plasma retinol and β-carotene concentrations and diagnosed congenital hearing loss in a larger cohort of maternal–infant dyads to further investigate this relationship. ## References 1. Milani A., Basirnejad M., Shahbazi S., Bolhassani A.. Carotenoids: Biochemistry, pharmacology and treatment. Br. J. Pharmacol. (2017.0) 174 1290-1324. DOI: 10.1111/bph.13625 2. Ngamwonglumlert L., Devahastin S., Chiewchan N., Raghavan V.. Plant carotenoids evolution during cultivation, postharvest storage, and food processing: A review. Compr. Rev. Food Sci. Food Saf. (2020.0) 19 1561-1604. DOI: 10.1111/1541-4337.12564 3. Toti E., Chen C.O., Palmery M., Villaño Valencia D., Peluso I.. Non-Provitamin A and Provitamin A Carotenoids as Immunomodulators: Recommended Dietary Allowance, Therapeutic Index, or Personalized Nutrition?. Oxidative Med. Cell. Longev. (2018.0) 2018 4637861. DOI: 10.1155/2018/4637861 4. Sommer A., Vyas K.S.. A global clinical view on vitamin A and carotenoids. Am. J. Clin. Nutr. (2012.0) 96 1204S-1206S. DOI: 10.3945/ajcn.112.034868 5. Zielińska M.A., Wesołowska A., Pawlus B., Hamułka J.. Health Effects of Carotenoids during Pregnancy and Lactation. Nutrients (2017.0) 9. DOI: 10.3390/nu9080838 6. Cohen J.M., Kramer M.S., Platt R.W., Basso O., Evans R.W., Kahn S.R.. The association between maternal antioxidant levels in midpregnancy and preeclampsia. Am. J. Obstet. Gynecol. (2015.0) 213 695.e1-695.e13. DOI: 10.1016/j.ajog.2015.07.027 7. Gao Q., Zhong C., Zhou X., Chen R., Xiong T., Hong M., Li Q., Kong M., Han W., Sun G.. The association between intake of dietary lycopene and other carotenoids and gestational diabetes mellitus risk during mid-trimester: A cross-sectional study. Br. J. Nutr. (2019.0) 121 1405-1412. DOI: 10.1017/S0007114519000606 8. Kumar A., Ranjan R., Basu S., Khanna H.D., Bhargava V.. Antioxidant levels in cord blood of low birth weight newborns. Indian Pediatr. (2008.0) 45 583-585. PMID: 18695279 9. 9. WHO Global Prevalence of Vitamin A Deficiency in Populations at Risk 1995–2005: WHO Global Database on Vitamin A DeficiencyWHOGeneva, Switzerland2009. Global Prevalence of Vitamin A Deficiency in Populations at Risk 1995–2005: WHO Global Database on Vitamin A Deficiency (2009.0) 10. Emmett S.D., West K.P.. Gestational vitamin A deficiency: A novel cause of sensorineural hearing loss in the developing world?. Med. Hypotheses (2014.0) 82 6-10. DOI: 10.1016/j.mehy.2013.09.028 11. Lahlou H., Lopez-Juarez A., Fontbonne A., Nivet E., Zine A.. Modeling human early otic sensory cell development with induced pluripotent stem cells. PloS ONE (2018.0) 13. DOI: 10.1371/journal.pone.0198954 12. Frenz D.A., Liu W., Cvekl A., Xie Q., Wassef L., Quadro L., Niederreither K., Maconochie M., Shanske A.. Retinoid signaling in inner ear development: A “Goldilocks” phenomenon. Am. J. Med. Genet. Part A (2010.0) 152A 2947-2961. DOI: 10.1002/ajmg.a.33670 13. Wu D.K., Kelley M.W.. Molecular mechanisms of inner ear development. Cold Spring Harb. Perspect. Biol. (2012.0) 4 a008409. DOI: 10.1101/cshperspect.a008409 14. Billings K.R., Kenna M.A.. Causes of pediatric sensorineural hearing loss: Yesterday and today. Arch. Otolaryngol. Head Neck Surg. (1999.0) 125 517-521. DOI: 10.1001/archotol.125.5.517 15. Germiller J.A., Wetmore R.F., Bell L.M.. Chapter 4- Hearing Loss in Children. Pediatric Otolaryngology (2007.0) 60 16. Karimi-Boroujeni M., Zahedi-Amiri A., Coombs K.M.. Embryonic Origins of Virus-Induced Hearing Loss: Overview of Molecular Etiology. Viruses (2021.0) 13. DOI: 10.3390/v13010071 17. American Academy of Pediatrics J.C.o.I.H.. Year 2007 position statement: Principles and guidelines for early hearing detection and intervention programs. Pediatrics (2007.0) 120 898-921. PMID: 17908777 18. Neumann K., Mathmann P., Chadha S., Euler H.A., White K.R.. Newborn Hearing Screening Benefits Children, but Global Disparities Persist. J. Clin. Med. (2022.0) 11. DOI: 10.3390/jcm11010271 19. Kanji A., Khoza-Shangase K., Moroe N.. Newborn hearing screening protocols and their outcomes: A systematic review. Int. J. Pediatr. Otorhinolaryngol. (2018.0) 115 104-109. DOI: 10.1016/j.ijporl.2018.09.026 20. Kemp D.T.. Otoacoustic emissions, their origin in cochlear function, and use. Br. Med. Bull. (2002.0) 63 223-241. DOI: 10.1093/bmb/63.1.223 21. van Straaten H.L.. Automated auditory brainstem response in neonatal hearing screening. Acta Paediatr. (1999.0) 88 76-79. DOI: 10.1111/j.1651-2227.1999.tb01165.x 22. Chen M.M., Oghalai J.S.. Diagnosis and Management of Congenital Sensorineural Hearing Loss. Curr. Treat. Options Pediatr. (2016.0) 2 256-265. DOI: 10.1007/s40746-016-0056-6 23. Thoene M., Anderson-Berry A., Van Ormer M., Furtado J., Soliman G.A., Goldner W., Hanson C.. Quantification of Lutein + Zeaxanthin Presence in Human Placenta and Correlations with Blood Levels and Maternal Dietary Intake. Nutrients (2019.0) 11. DOI: 10.3390/nu11010134 24. McConnell C., Thoene M., Van Ormer M., Furtado J.D., Korade Z., Genaro-Mattos T.C., Hanson C., Anderson-Berry A.. Plasma Concentrations and Maternal-Umbilical Cord Plasma Ratios of the Six Most Prevalent Carotenoids across Five Groups of Birth Gestational Age. Antioxidants (2021.0) 10. DOI: 10.3390/antiox10091409 25. T.H. Chan School of Public Health, D.o.N.H.U. Nutrition Questionnaire Service Center \| Harvard T.H. Chan School of Public Health. (2022.0) 26. Suitor C.J., Gardner J., Willett W.C.. A comparison of food frequency and diet recall methods in studies of nutrient intake of low-income pregnant women. J. Am. Diet. Assoc. (1989.0) 89 1786-1794. DOI: 10.1016/S0002-8223(21)02469-X 27. Cave C., Hein N., Smith L.M., Anderson-Berry A., Richter C.K., Bisselou K.S., Appiah A.K., Kris-Etherton P., Skulas-Ray A.C., Thompson M.. Omega-3 Long-Chain Polyunsaturated Fatty Acids Intake by Ethnicity, Income, and Education Level in the United States: NHANES 2003–2014. Nutrients (2020.0) 12. DOI: 10.3390/nu12072045 28. 28. HHS Annual Update of the HHS Poverty GuidelinesDepartment of Health and Human ServicesWashington, DC, USA2022. Annual Update of the HHS Poverty Guidelines (2022.0) 29. Johnson L.C., Toro M., Vishnja E., Berish A., Mills B., Lu Z., Lieberman E.. Age and Other Factors Affecting the Outcome of AABR Screening in Neonates. Hosp. Pediatr. (2018.0) 8 141-147. DOI: 10.1542/hpeds.2017-0060 30. 30. Institute of Medicine Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and ZincNational Academies Press (US)Washington, DC, USA2001. Dietary Reference Intakes for Vitamin A, Vitamin K, Arsenic, Boron, Chromium, Copper, Iodine, Iron, Manganese, Molybdenum, Nickel, Silicon, Vanadium, and Zinc (2001.0) 31. Group U.S.C.. How Do I Interpret a Regression Model When Some Variables Are Log Transformed? 32. Wroblewska-Seniuk K.E., Dabrowski P., Szyfter W., Mazela J.. Universal newborn hearing screening: Methods and results, obstacles, and benefits. Pediatr. Res. (2017.0) 81 415-422. DOI: 10.1038/pr.2016.250 33. Choi Y.H., Miller J.M., Tucker K.L., Hu H., Park S.K.. Antioxidant vitamins and magnesium and the risk of hearing loss in the US general population. Am. J. Clin. Nutr. (2014.0) 99 148-155. DOI: 10.3945/ajcn.113.068437 34. Curhan S.G., Stankovic K.M., Eavey R.D., Wang M., Stampfer M.J., Curhan G.C.. Carotenoids, vitamin A, vitamin C, vitamin E, and folate and risk of self-reported hearing loss in women. Am. J. Clin. Nutr. (2015.0) 102 1167-1175. DOI: 10.3945/ajcn.115.109314 35. Gopinath B., Flood V.M., McMahon C.M., Burlutsky G., Spankovich C., Hood L.J., Mitchell P.. Dietary antioxidant intake is associated with the prevalence but not incidence of age-related hearing loss. J. Nutr. Health Aging (2011.0) 15 896-900. DOI: 10.1007/s12603-011-0119-0 36. Schmitz J., West K.P., Khatry S.K., Wu L., Leclerq S.C., Karna S.L., Katz J., Sommer A., Pillion J.. Vitamin A supplementation in preschool children and risk of hearing loss as adolescents and young adults in rural Nepal: Randomised trial cohort follow-up study. BMJ (2012.0) 344 d7962. DOI: 10.1136/bmj.d7962 37. Ambalavanan N., Tyson J.E., Kennedy K.A., Hansen N.I., Vohr B.R., Wright L.L., Carlo W.A., National Institute of Child H., Human Development Neonatal Research N.. Vitamin A supplementation for extremely low birth weight infants: Outcome at 18 to 22 months. Pediatrics (2005.0) 115 249. DOI: 10.1542/peds.2004-1812 38. Lammer E.J., Chen D.T., Hoar R.M., Agnish N.D., Benke P.J., Braun J.T., Curry C.J., Fernhoff P.M., Grix A.W., Lott I.T.. Retinoic acid embryopathy. New Engl. J. Med. (1985.0) 313 837-841. DOI: 10.1056/NEJM198510033131401 39. Miller R.K., Hendrickx A.G., Mills J.L., Hummler H., Wiegand U.W.. Periconceptional vitamin A use: How much is teratogenic?. Reprod. Toxicol. (1998.0) 12 75-88. DOI: 10.1016/S0890-6238(97)00102-0 40. Safe T.M., Luebke A.E.. Prenatal low dosage dioxin (TCDD) exposure impairs cochlear function resulting in auditory neuropathy. Hear. Res. (2016.0) 331 7-12. DOI: 10.1016/j.heares.2015.09.015 41. Poon E., Powers B.E., McAlonan R.M., Ferguson D.C., Schantz S.L.. Effects of developmental exposure to polychlorinated biphenyls and/or polybrominated diphenyl ethers on cochlear function. Toxicol. Sci. Off. J. Soc. Toxicol. (2011.0) 124 161-168. DOI: 10.1093/toxsci/kfr214 42. Lasky R.E., Widholm J.J., Crofton K.M., Schantz S.L.. Perinatal exposure to Aroclor 1254 impairs distortion product otoacoustic emissions (DPOAEs) in rats. Toxicol. Sci. Off. J. Soc. Toxicol. (2002.0) 68 458-464. DOI: 10.1093/toxsci/68.2.458 43. Sisto R., Moleti A., Palkovičová Murínová Ľ., Wimmerová S., Lancz K., Tihányi J., Čonka K., Šovčíková E., Hertz-Picciotto I., Jusko T.A.. Environmental exposure to organochlorine pesticides and deficits in cochlear status in children. Environ. Sci. Pollut. Res. Int. (2015.0) 22 14570-14578. DOI: 10.1007/s11356-015-4690-5 44. Jusko T.A., Sisto R., Iosif A.M., Moleti A., Wimmerová S., Lancz K., Tihányi J., Sovčiková E., Drobná B., Palkovičová L.. Prenatal and postnatal serum PCB concentrations and cochlear function in children at 45 months of age. Environ. Health Perspect. (2014.0) 122 1246-1252. DOI: 10.1289/ehp.1307473 45. Silver M.K., Shao J., Ji C., Zhu B., Xu L., Li M., Chen M., Xia Y., Kaciroti N., Lozoff B.. Prenatal organophosphate insecticide exposure and infant sensory function. Int. J. Hyg. Environ. Health (2018.0) 221 469-478. DOI: 10.1016/j.ijheh.2018.01.010 46. Aktuna D., Buchinger W., Langsteger W., Meister E., Sternad H., Lorenz O., Eber O.. Beta-carotene, vitamin A and carrier proteins in thyroid diseases. Acta Med. Austriaca (1993.0) 20 17-20. PMID: 8475673 47. Goswami U.C., Choudhury S.. The status of retinoids in women suffering from hyper- and hypothyroidism: Interrelationship between vitamin A, beta-carotene and thyroid hormones. Int. J. Vitam. Nutr. Res. (1999.0) 69 132-135. DOI: 10.1024/0300-9831.69.2.132 48. Wasserman E.E., Nelson K., Rose N.R., Eaton W., Pillion J.P., Seaberg E., Talor M.V., Burek L., Duggan A., Yolken R.H.. Maternal thyroid autoantibodies during the third trimester and hearing deficits in children: An epidemiologic assessment. Am. J. Epidemiol. (2008.0) 167 701-710. DOI: 10.1093/aje/kwm342 49. Zhang K., Fried J., Nguyen S.A., Meyer T.A., White D.R.. Dysfunction of the Auditory System in Children With Hypothyroidism: A Systematic Review and Meta-Analysis. Ear. Hear. (2022.0) 43 23-31. DOI: 10.1097/AUD.0000000000001082 50. Horikawa C., Kodama S., Tanaka S., Fujihara K., Hirasawa R., Yachi Y., Shimano H., Yamada N., Saito K., Sone H.. Diabetes and risk of hearing impairment in adults: A meta-analysis. J. Clin. Endocrinol. Metab. (2013.0) 98 51-58. DOI: 10.1210/jc.2012-2119 51. Taneera J., Awadallah S., Khader Mohammed A., Unnikannan H., Sulaiman N.. Vitamin A levels are decreased but not influenced by glucose- or lipid-lowering medications in subjects with type 2 diabetes. Saudi. J. Biol. Sci. (2021.0) 28 572-577. DOI: 10.1016/j.sjbs.2020.10.044 52. Oghan F., Coksuer H.. Does hyperandrogenism have an effect on hearing loss in patients with polycystic ovary syndrome?. Auris Nasus Larynx (2012.0) 39 365-368. DOI: 10.1016/j.anl.2011.06.006 53. Jia J., Bai J., Liu Y., Yin J., Yang P., Yu S., Ye J., Wang D., Yuan G.. Association between retinol-binding protein 4 and polycystic ovary syndrome: A meta-analysis. Endocr. J. (2014.0) 61 995-1002. DOI: 10.1507/endocrj.EJ14-0186 54. Blatt A.J., Nakamoto J.M., Kaufman H.W.. National status of testing for hypothyroidism during pregnancy and postpartum. J. Clin. Endocrinol. Metab. (2012.0) 97 777-784. DOI: 10.1210/jc.2011-2038 55. Radetti G., Gentili L., Paganini C., Oberhofer R., Deluggi I., Delucca A.. Psychomotor and audiological assessment of infants born to mothers with subclinical thyroid dysfunction in early pregnancy. Minerva Pediatr. (2000.0) 52 691-698. PMID: 11268926 56. Mohiyuddin S.A., Apoorva H.M., Prasad K.N.V., Munikrishna N.. Hearing outcome in infants following correction of maternal hypothyroidism during pregnancy. Int. J. Pediatr. Otorhinolaryngol. (2021.0) 142 110597. DOI: 10.1016/j.ijporl.2020.110597 57. Alexander E.K., Pearce E.N., Brent G.A., Brown R.S., Chen H., Dosiou C., Grobman W.A., Laurberg P., Lazarus J.H., Mandel S.J.. 2017 Guidelines of the American Thyroid Association for the Diagnosis and Management of Thyroid Disease During Pregnancy and the Postpartum. Thyroid (2017.0) 27 315-389. DOI: 10.1089/thy.2016.0457
--- title: Renin–Angiotensin System Antagonism Protects the Diabetic Heart from Ischemia/Reperfusion Injury in Variable Hyperglycemia Duration Settings by a Glucose Transporter Type 4-Mediated Pathway authors: - Aisha Al-Kouh - Fawzi Babiker - Maie Al-Bader journal: Pharmaceuticals year: 2023 pmcid: PMC9967344 doi: 10.3390/ph16020238 license: CC BY 4.0 --- # Renin–Angiotensin System Antagonism Protects the Diabetic Heart from Ischemia/Reperfusion Injury in Variable Hyperglycemia Duration Settings by a Glucose Transporter Type 4-Mediated Pathway ## Abstract Background: Diabetes mellitus (DM) is a risk factor for cardiovascular diseases, specifically, the ischemic heart diseases (IHD). The renin–angiotensin system (RAS) affects the heart directly and indirectly. However, its role in the protection of the heart against I/R injury is not completely understood. The aim of the current study was to evaluate the efficacy of the angiotensin-converting enzyme (ACE) inhibitor and Angiotensin II receptor (AT1R) blocker or a combination thereof in protection of the heart from I/R injury. Methods: Hearts isolated from adult male Wistar rats ($$n = 8$$) were subjected to high glucose levels; acute hyperglycemia or streptozotocin (STZ)-induced diabetes were used in this study. Hearts were subjected to I/R injury, treated with Captopril, an ACE inhibitor; Losartan, an AT1R antagonist; or a combination thereof. Hemodynamics data were measured using a suitable software for that purpose. Additionally, infarct size was evaluated using 2,3,5-Triphenyltetrazolium chloride (TTC) staining. The levels of apoptosis markers (caspase-3 and -8), antioxidant enzymes, superoxide dismutase (SOD) and catalase (CAT), nitric oxide synthase (eNOS), and glucose transporter type 4 (GLUT-4) protein levels were evaluated by Western blotting. Pro-inflammatory and anti-inflammatory cytokines levels were evaluated by enzyme-linked immunosorbent assay (ELISA). Results: Captopril and Losartan alone or in combination abolished the effect of I/R injury in hearts subjected to acute hyperglycemia or STZ-induced diabetes. There was a significant ($p \leq 0.05$) recovery in hemodynamics, infarct size, and apoptosis markers following the treatment with Captopril, Losartan, or their combination. Treatment with Captopril, Losartan, or their combination significantly ($p \leq 0.05$) reduced pro-inflammatory cytokines and increased GLUT-4 protein levels. Conclusions: The blockade of the RAS system protected the diabetic heart from I/R injury. This protection followed a pathway that utilizes GLUT-4 to decrease the apoptosis markers, pro-inflammatory cytokines, and to increase the anti-inflammatory cytokines. This protection seems to employ a pathway which is not involving ERK$\frac{1}{2}$ and eNOS. ## 1. Introduction Cardiovascular diseases (CVDs) are among the leading causes of morbidity and mortality worldwide [1]. Recent studies have reported ischemic heart disease (IHD) as the worldwide leading cause of disability and death [2]. The ultimate effect of ischemia is myocardial infarction (MI), which results from either partial or complete lack of oxygen and nutrients supply to the myocardium [3]. The myocardial reperfusion procedure is the only contemporary standard treatment for the protection of the heart against ischemia/reperfusion (I/R) injury. Nevertheless, development of complications such as in-hospital deaths, reoccurrence of myocardial infarction, and left ventricular (LV) reconditioning leading to heart failure following infarction still exist [4]. Several treatment interventions to prevent myocardial I/R injury have been investigated. Regrettably, the translation of such cardioprotective procedures to the clinical trials did not produce the expected outcome [5,6]. However, these studies were well-planned postconditioning clinical studies, and they resulted in negative results which did not support the previous preclinical findings. Therefore, an urgent need for more effective regimens persists. Despite the availability of optimal therapy, IHD and its consequences remain high in diabetic patients [7]. Globally, the occurrence of cardiovascular disease in the diabetic population is far greater compared to the nondiabetic population [8]. Furthermore, it was reported that signaling mechanisms of protection from ischemic injury are impaired in chronic diabetes mellitus (DM) [9]; therefore, the threshold for protection of the diabetic heart is significantly increased to a limit that may not be achieved by the traditional procedures [10]. Alteration in signaling pathways, such as survivor activating factor enhancement (SAFE) [11] and reperfusion injury salvage kinase (RISK) [12] is well-established in diabetes, necessitating tailoring new potential therapies for this disease. The effect of hyperglycemia on the outcome of ischemic heart disease is unclear [13]. Coexistence of acute hyperglycemia at occurrence of ischemia may worsen the prognosis of diabetic and nondiabetic patients [14]. The alteration of signaling pathways by diabetes leads to the decrease of major glucose transporter in the heart, glucose transporter 4 (GLUT-4) [15], leading eventually to cardiomyocyte death [16]. Therefore, evaluation of GLUT-1 and GLUT-4 levels will be crucial for understanding the protection of the diabetic heart from I/R injury. The renin–angiotensin system (RAS) plays a pivotal role in combating myocardial diseases and might ultimately participate in the aggravation of reperfusion injury [17]. Although RASblockers are beneficial during myocardial ischemia, its effects on the diabetic myocardium remain controversial [18,19]. Losartan is a selective antagonist of type 1 angiotensin II receptors (AT1R) and has been used in medical treatments of a variety of cardiovascular diseases [20]. Losartan has been reported to prevent I/R-induced cardiac injury by inhibiting reactive oxygen species (ROS)-induced injury [21]. However, the studies of the effect of Losartan on the diabetic heart remain unresolved and require more clarification. On the other hand, using Captopril, a nonspecific ACE inhibitor, administration resulted in protection of myocardial tissue after I/R insult by preventing left ventricular hypertrophy [22]. We have previously reported the involvement of ACE inhibitors in the protection of the isolated rat heart [23] from the detrimental effect of the locally produced ACE [24,25]. Although dual therapy of RAS antagonism was proven to be protective to the normotensive heart [26,27] and four-week diabetic rats [28], its role in acute hyperglycemia and chronic diabetes was not investigated. Inhibiting angiotensin activity in diabetes by Captopril and Losartan showed great improvement of heart hemodynamics after I/R injury [28]. These regimens could be promising treatment procedures for the protection of the diabetic heart against I/R injury. We aimed in this study to investigate the potential protective effects of Losartan and Captopril and their combination in the protection of heart subjected to acute or chronic hyperglycemia which were known to cause variable pathological effects on the heart. Chronic inflammation [29,30] and oxidative stress [31] that persist after myocardial infarction were reported to be key mediators of cardiac remodeling. Chronic inflammation interferes metabolically with the heart, indirectly as it jeopardizes cardiac function [32], or directly by interference of the cytokines with calcium transport [33]. Indeed, ROS are crucial for normal cellular function at low levels; however, they were recognized to damage the mitochondria when they reach high concentration levels [18,34,35]. Understanding the effects of inflammation, oxidative stress, and the RAS system in cardiac remodeling in nondiabetic and diabetic hearts will help in developing effective procedures which may produce a better protection to the heart against deterioration from IHD to heart failure. ## 2. Results In this study, the role of RAS system in the protection of the diabetic heart from I/R injury was investigated. Left ventricular dynamics were assessed throughout the experiment by evaluating left ventricular pressure (LVEDP), LV maximum developed pressure (DPmax), and LV contractility index (±dP/dt). The coronary vascular dynamics were assessed by evaluating coronary flow (CF) and coronary vascular resistance (CVR). There were no significant differences detected in ventricular dynamics, contractility, and coronary vascular dynamics between the groups when at baseline levels. Body and LV weights were not significantly different between the experimental groups (data not shown). Ischemia resulted in a remarkable worsening in the heart functions. Captopril, Losartan, or a combination thereof were administered at reperfusion in hearts subjected to hyperglycemia, four weeks, or six weeks of diabetes. These treatments significantly ($p \leq 0.05$) improved cardiac hemodynamics and vascular dynamics in all three different conditions (hyperglycemia, four weeks, and six weeks diabetic hearts) compared to untreated controls (Figure 1A–C, Table 1). Interestingly, there were no differences in the effects of the three experimental conditions on the protection of the heart, and the combination of the drugs did not show additive effects in all treatments compared to untreated controls (Figure 1, Table 1). Hearts subjected to hyperglycemia, four weeks, or six weeks of diabetes were protected from I/R injury by Captopril, Losartan, or their combination. These drugs resulted in a significant decrease in the infarct size ($p \leq 0.001$) and cardiac TnT levels. There were no significant differences in the effects of the three experimental conditions on the protection of the heart, and the combination of the drugs did not show additive effects in all treatments compared to untreated controls ($p \leq 0.01$) (Figure 2 and Table 2). These results were confirmed by a decrease of apoptosis in the myocytes by Captopril and Losartan or their combination. Administration of these drugs at reperfusion resulted in a marked decrease in apoptosis markers caspase-3 and caspase-8 ($p \leq 0.01$) compared to untreated controls (Figure 3). We further evaluated the possible downstream signaling pathways that could be involved in the protection by Captopril, Losartan, or their combination, to the diabetic heart. We tested the effect of these treatments on the protein levels and basal ratio of phosphorylated to total extracellular signal-regulated protein kinase (ERK$\frac{1}{2}$) or endothelial nitric oxide synthase (eNOS) during I/R injury. Surprisingly, neither ERK$\frac{1}{2}$ nor eNOS were affected by these treatments. Both protein levels did not show a significant difference between the diabetic hearts and the untreated controls (Figure 4). Surprisingly, the protein levels of superoxide dismutase (SOD) and catalase (CAT) did not show a significant increase in the presence of Captopril, Losartan, or their combination in all the three treatment conditions. ( Figure 5). There were no differences in the effects of the three experimental conditions on the protection of the heart, and the combination of the drugs did not show additive effects in all treatments compared to untreated controls (Figure 4 and Figure 5). To identify a potential role for glucose transporters, glucose transporter 1 (GLUT-1) and glucose transporter 4 (GLUT-4) protein levels were evaluated using Western blotting. Interestingly, GLUT-4 protein levels were significantly ($p \leq 0.01$) increased with the treatment of Losartan, Captopril, and their combination, However, no changes in GLUT-1 protein levels were observed. There were no differences in the effects of the three experimental conditions on the protection of the heart, and the combination of the drugs did not show additive effects in all treatments compared to untreated controls (Figure 6). To study the effects of cytokines in the protection of the diabetic heart against I/R injury, we evaluated the levels of the pro-inflammatory cytokines tumor necrosis factor alpha (TNF-α), interleukin 1 beta (IL-1β), and interleukin 6 (IL-6), and anti-inflammatory cytokine interleukin 10 (IL-10) in the cardiomyocyte lysate by enzyme-linked immunosorbent assay (ELISA). Administration of Captopril, Losartan, or their combination resulted in a significant ($p \leq 0.01$) decrease in TNF-α, IL-1β, and IL-6 levels which was increased at the beginning by ischemia compared to untreated controls (Figure 7A–I). The same treatment resulted in a significant ($p \leq 0.05$) increase in the anti-inflammatory cytokine IL-10. There were no differences in the effects of the three experimental conditions on the protection of the heart, and the combination of the drugs did not show additive effects in all treatments compared to untreated controls (Figure 7J–L). ## 3. Discussion The studies of the effects of I/R and its treatments on the diabetic heart are inconsistent. The major area of controversy resides in the dependence of the efficiency of the heart protection on the duration of diabetes before the intervention. Some laboratories consider four weeks of diabetes satisfactory for the pathological effects of diabetes to appear in the heart, and thereafter the animals will be suitable models for the studies. Other laboratories consider this period rather short for the pathology of diabetes to occur and the animal models are similar to normoglycemic animals. Therefore, acute hyperglycemia, four-, and six-week diabetic (chronic hyperglycemia) rats were used in this study to assess the protection of the heart against I/R injury. Although RAS antagonism was reported to be protective to the heart by many investigators, its role in the protection of the diabatic heart is controversial [23,36,37]. The diabetic heart was reported by many researchers to be resistant to protection against I/R injury [10,38]. This alteration in the effect of postconditioning is a consequence of the remodeling caused by the diabetes which blocks or impairs activation of many kinases including, among others, PI3 kinase/Akt, ERK, p70S6 kinase, and/or GSK-3β [39]. However, the data available to date did not reveal significant variations in the injury of the diabetic heart relative to the normotensive heart [23,40,41]. The present study investigated the prospective protection of heart compromised by hyperglycemia and DM (diabetic heart) against I/R injury. The study used hearts subjected to high glucose levels to mimic acute hyperglycemia and diabetic hearts isolated from four- and six-weeks diabetic rats treated at reperfusion with RAS member antagonists (ACE and AT1R). The choice of these conditions was made mainly to understand the potential protection of the diabetic heart in different scenarios with different durations of hyperglycemia. The animal models used for the study of heart protection from I/R injury showed contradicting results when extrapolated to the trials of the protection of the diabetic heart. Some studies reported a notable protection to the diabetic heart [27,42]; however, some other studies reported lack of protection [40,43]. This controversy is evident even in the studies reported by the same laboratory [28,40]. However, substantial uncertainty persisted as two weeks or four weeks are considered the minimum required duration for the diabetic heart-healing disturbance to manifest [44]. Indeed, hyperglycemia even for a short period could raise the threshold for the protection of the diabetic heart [45]; however, some studies reported protection of the four-week diabetic heart against I/R injury [28]. Therefore, this study investigated protection of hearts subjected to different hyperglycemia durations including acute hyperglycemia, four-, and six-week diabetic hearts, to understand the effect of the hyperglycemia duration on the protection, the response of the diabetic heart to the available heart protection procedures, and to find a suitable regimen for the protection of the diabetic heart. This study demonstrates that the diabetic heart could be protected from ischemic injury by RAS antagonism. Captopril, Losartan, or a combination thereof protected the diabetic heart against I/R injury by a signaling pathway that is not utilizing ERK$\frac{1}{2}$ or eNOS. This protection targets mainly the apoptotic enzymes and pro-inflammatory and anti-inflammatory cytokines (Figure 8). However, although the combination of Captopril and Losartan was protective to the diabetic heart, it does not show additivity in its protection. The diabetic heart was reported previously to be resistant to protection from I/R injury [10]. That could be due to the permanent tissue and organ remodeling induced by diabetes which deteriorates the protective machinery of the classical protective pathways such as RISK and SAFE [9]. In contrast, a lack of blockade for the protection of the nondiabetic heart by the antagonism of AT1R was reported [43]. Nevertheless, controversy in the protection of the diabetic heart still exists. Protection by Losartan was reported in diabetic hearts; however the duration of diabetes was four weeks in these rats [42]. We and others also reported protection of four-week diabetic hearts by ACE and AT1R antagonism [28,42]. In addition, Shi-Ting et al. [ 46] also showed that four-week diabetic mice have a tolerance to I/R injury which supports the notion that four-week diabetic rats are not fully affected by diabetes. However, this tolerance to ischemia and ease of protection is lacking in six- and eight-week diabetic rats. Combined therapy using ACE inhibitors and AT1R blockers, though not additive, was also reported to be protective to the nondiabetic heart [47]. These notions support the finding of this study as Captopril, Losartan, and their combination protected the heart subjected to hyperglycemia, and four- and six-week diabetic hearts. To investigate the potentials of this regimen in the treatment of long duration hyperglycemia we used six-week diabetic hearts. Captopril, Losartan, or their combination protected the six-week diabetic hearts as well (Figure 8). In contrast to six- and eight-week diabetic rats, protection of four-week diabetic rats was reported [39]. This indicates the prevalence of permanent pathological changes in six- and eight-week diabetic rats [39]. The intolerance and lack of protection to the diabetic hearts of a long diabetes duration was confirmed by Drenger et al. [ 11]. These notions indicate an inverse relation between increased diabetic duration and the possibility of heart protection. This study proved the possibility of protection of six-week diabetic hearts from I/R injury. This denotes that although the protective signaling pathways for the protection of the six-week diabetic hearts are defective, some treatment regimen could be protective to these hearts. This protection might have followed a pathway different from the pathways which were previously proven to be effected. The protection reported in the present study seems to follow a pathway using the GLUT-4 receptor. This study showed that the protection given by Losartan, Captopril, or their combination is through a pathway; however, it is not employing antioxidant enzymes but targeting the pro-inflammatory cytokines (Figure 8). Our results are in line with reports from other laboratories that reported no increase in SOD or CAT activity in the protection of the nondiabetic heart from I/R injury [48,49]. However, although there is a lack of reporting in the literature regarding the effect of RAS antagonism in diabetic animals, we anticipate the lack of increase of the antioxidant enzymes levels in the protection of the diabetic heart is due to the impairment of the AKT/ERK$\frac{1}{2}$/Nrf2 pathway. The lack of increase in SOD and CAT protein levels in this study could be because of the nature of the RAS antagonists used. The ACE inhibitors and AT1 antagonist blocks the activation of oxidant enzymes and redox-sensitive genes, which may spare the antioxidant enzymes and blunt their increase [50,51]. A lack of synchronous increase in antioxidant enzyme levels with an increased ROS level was reported previously [52]. Furthermore, the protection reported in this study may be following a novel pathway using GLUT-4 in the protection of the diabetic heart. On the other hand, the regulation of the pro-inflammatory cytokines is crucial for the protection of the ischemic heart. Lowering IL-1 and IL-6 availability is associated with reduced heart injury and cardiomyocyte death [53]. Reduced protein levels of TNF-α, IL-1β, and IL-6 presented in this study are similar to those reported by Liu et al. [ 54]. These notions indicate that RAS antagonism protect the heart by a pathway targeting the pro-inflammatory cytokines. On the other hand, the increased expression of IL-10 in the protection of the diabetic hearts seen in this study was reported before [55]. This study showed an increase in GLUT-4 protein levels (Figure 6). This increase was reported before in the studies of heart protection from I/R injury [56]. The protection reported in the recent study could be due to increased expression of GLUT-4 and improved glucose uptake. GLUT-4 expression was also reported in obese rats with high glucose levels [57]. Similar results that showed increased GLUT-4 but not GLUT-1 levels were reported in the protection of the diabetic heart [58]. A potential limitation of this study is that we did not show the changes in the levels of the cleaved caspases which could give a better indication of apoptosis compared to the levels of the procaspases. ## 4.1. Ischemia Reperfusion Study Wistar rats of the same age weighing between 270–300 g ($$n = 8$$ per group) were used in this study. Animal use and treatment were done in agreement with the guidelines of Animal Resource Centre of Health Science Centre, Kuwait University, Kuwait. Rats were housed at 22 °C on a 12 h light/dark cycle (7 am–7 pm), water and food were unrestrictedly available. Anesthesia was performed with intraperitoneal injections of sodium pentobarbital (60 mg/kg) and the animals were anticoagulated by heparin (1000 U/kg). Heart mounting cannulation and perfusion were done as described previously [59,60]. Briefly, isolated heart was reversely perfused with a freshly prepared Krebs–Hensleit (KH) buffer, pH 7.35 to 7.45 at 37.0 ± 0.5 °C and was gassed with CO2 ($5\%$) and O2 ($95\%$). Heart was supplied by pacing electrodes connected to the right atrium (RA) appendage to maintain physiological heart rhythm. Local blood supply was induced by 30 min occlusion of the left anterior descending (LAD) coronary artery. The LAD was surrounded by a string at 0.5 cm below the atrioventricular groove, and a small plastic rod was positioned between the heart and the string to ensure the closure of the coronary artery. Thereafter, the heart was reperfused for 30 min. The flow of the perfusion buffer used a “Statham pressure transducer” (P23 Db). A constant preload was kept at 6 mmHg during acclimatization of the heart on the Langendorff system and the perfusion pressure (PP) was kept constant at 50 mmHg during the experimental procedure. Constant PP was secured digitally by means of the perfusion assembly, Module PPCM type 671 (Hugo Sachs Elektronik-Harvard Apparatus GmbH, March, Germany”)). ## 4.2. Induction of Diabetes Diabetes was induced by a single intraperitoneal injection of 55 mg/kg body weight STZ as described previously [40]. ## 4.3. Study Protocol and Study Groups A total of 96 rats were subdivided into 3 groups ($$n = 32$$ per group) and subjected to 4 different experimental protocols. The first group contained nondiabetic rats. All hearts isolated from these animals in this group were subdivided into 4 subgroups subjected to acute hyperglycemia which was created by adding 6 g/L to the perfusion buffer. One subgroup (Ctr) was subjected to only I/R and served as control. The second subgroup (Cap) was subjected to I/R and treated with Captopril (100 μM; cat. #: C4042 Sigma-Aldrich (St Louis, MI, USA)). The third subgroup (Los) was subjected to I/R and treated with Losartan (4.5 μM, Santa Cruz Biotechnology). The fourth subgroup (Cap + Los) was subjected to I/R and treated with Losartan (4.5 μM) and Captopril (100 μM). All drugs were injected 5 min before the end of ischemia and continued for 10 min thereafter (Figure 9). The second group was subjected to DM for four weeks, then divided into four subgroups, Ctr, Cap, Los, and Los + Cap (Figure 9). The third group was subjected to DM for six weeks, then like the previous groups, was subdivided into four similar subgroups (Ctr, Cap, Los, and Cap + Los (Figure 9). ## 4.4. Data Collection and Processing The function of LV was evaluated by the assessment of LV end diastolic pressure (LVEDP), LV maximum developed pressure (DPmax), and LV contractility index (±dP/dt). The coronary–vascular dynamics were assessed by the measurement of the coronary flow (CF) and coronary vascular resistance (CVR). Cardiac hemodynamics were evaluated as described in the previously [60]. Briefly, a latex balloon filled with water was fixed in the LV cavity and connected to a “DC-Bridge amplifier (DC-BA)” of the pressure module (DC-BA type 660, Hugo-Sachs Electronik, Germany) which was connected to a computer for online determination of DPmax. Left ventricular developed pressure was calculated from the acquired values of the “LVESP using Max-Min module Number MMM type 668” (Hugo Sachs Elektronik-Harvard ApparatusGmbH, Germany). Coronary flow was estimated by electromagnetic flow probe attached to the inflow of the aortic cannula and connected to a computer. This was done digitally and was confirmed manually. The CVR and hemodynamics data were acquired every 10 sec using an “online data acquisition program” (Isoheart software V 1.524-S, Hugo-Sachs Electronik, Germany). ## 4.5. Sample Collection and Storage Perfusion buffer sample was collected immediately below the cannulated heart from the coronary outflow a few minutes before the end of the reperfusion. The hearts were also collected at the end of the experiments. All samples were frozen in liquid nitrogen and stored at −80 °C for the required biochemical analysis. ## 4.6. Infarct Size Evaluation The infarct size was evaluated as described previously [60]. Briefly, at the end of experiments the hearts were stored at −20 °C to the next day. Thereafter, the hearts were sliced into thin sections. The sections were then incubated in $1\%$ 2,3,5-Triphenyl-2H-tetrazolium chloride (TTC) solution and then fixed in $4\%$ formaldehyde. The infarcted area and the rest of the LV are located manually on the image. The infarct size was calculated as a percentage of the LV area for every heart. The calculation of the LV area and the infarct size was done using ImageJ (Image J, Wayne Rasb and National Institute of Health, Bethesda, Maryland, USA). Cardiomyocyte injury was confirmed by measuring troponin T (TnT) release in the coronary effluent during the reperfusion period using TnT immunoassay as suggested by the manufacturer’s assay protocols. ## 4.7. Protein Extraction from the Hearts The hearts were snap-frozen in liquid nitrogen and thereafter stored at −80 °C for biochemical analysis. The left ventricle was homogenized in a lysis buffer (MOPS, 20 mM; KCl, 150 mM; Mg Acetate, 4.5 mM; Triton X, $1\%$). Protease inhibitor cocktail (Roche Applied Science, Mannheim, Germany, Cat#05892970001) was added to the buffer. The whole mixture was centrifuged at 12,000× g for 15 min at 4 °C. The supernatant was then aliquoted and stored for biochemical analysis. Protein levels were measured using bicinchoninic acid protein assay (Pierce Chemical Co., Rockford, IL, USA, Cat# 23227). ## 4.8. Western Blot Analysis The frozen protein samples were used in Western blot analysis to evaluate the protein expression using: anti-superoxide dismutase [Cu-Zn] (SOD) antibody (cat# 07-403-I EMD Millipore, Sigma, St. Louis, Missouri, USA), anti-catalase (CAT) antibody (Cat# 12980); anti-extracellular regulated kinases $\frac{1}{2}$ (ERK$\frac{1}{2}$) antibody (cat# 9102); anti-phosphorylated ERK$\frac{1}{2}$ antibody (cat #9743); anti-caspase-3 antibody (cat #9662); anti-caspase-8 antibody (cat #4927); anti-phosphorylated endothelial nitric oxide synthase (eNOS) (Ser1177) antibody (cat #9571); anti-eNOS (49G3) antibody (cat #Ab #9586) (all from Cell Signaling Technology, Danvers, Massachusetts, USA) and anti-glucose transporter 1 (GLUT-1) antibody (cat #ab115730, Abcam, Cambridge, UK), anti-glucose transporter 4 (GLUT-4) antibody (phosphor S488) (cat #ab188317 Abcam) by standard immunoblot procedures as previously described [61]. Protein extracts were normalized with the lysis buffer to the desired protein concentration. Equal concentrations of protein per lane were calculated for all the samples. The protein band density was corrected to the total loaded protein. After incubation with secondary antibody, the detection was done with the enhanced chemiluminescence technique. A computerized image-acquisition system was used for densitometry analysis. ## 4.9. Estimation of the Inflammatory Cytokines The aliquoted LV protein ($$n = 4$$) was used to evaluate the expression of the pro-inflammatory and anti-inflammatory cytokines: tumor necrosis factor alpha (TNF-α) (cat# MBS175904), interleukin 1 beta (IL-1β) (cat# MBS825017), interleukin 6 (IL-6) (cat# MBS355410), and IL-10 (cat# MBS355232) using enzyme-linked immunosorbent assay (ELISA). The protein content was according to the manufacturer’s assay protocols (Biosource International, Camarillo, California USA). ## 4.10. Statistical Analysis All data were represented as mean ± S.E.M. For analyzing the data, two-way analysis of variance (ANOVA), was performed on absolute values even when the data are presented as % of baseline. For further confirmation of a statistically significant difference, post-hoc analysis with Tukey’s correction was used. Student’s T-Test was used to assess the significance in molecular experiments, infarct size, pre-inflammatory, and anti-inflammatory cytokines (Microsoft Excel). In all cases, $p \leq 0.05$ was considered statistically significant. ## 5. Conclusions A blockade of the RAS system protected the diabetic heart from I/R injury. This protection followed a pathway that decreases the apoptosis markers, pro-inflammatory cytokines, and increases the anti-inflammatory cytokines. This protection seems to employ GLUT-4 and a pathway which is not involving ERK$\frac{1}{2}$ and eNOS. ## References 1. Go A.S., Mozaffarian D., Roger V.L., Benjamin E.J., Berry J.D., Borden W.B., Bravata D.M., Dai S., Ford E.S., Fox C.S.. Heart disease and stroke statistics—2013 update: A report from the American Heart Association. Circulation (2013) 127 e6-e245. DOI: 10.1161/CIR.0b013e31828124ad 2. Lecour S., Botker H.E., Condorelli G., Davidson S.M., Garcia-Dorado D., Engel F.B., Ferdinandy P., Heusch G., Madonna R., Ovize M.. ESC working group cellular biology of the heart: Position paper: Improving the preclinical assessment of novel cardioprotective therapies. Cardiovasc. Res. (2014) 104 399-411. DOI: 10.1093/cvr/cvu225 3. Neumar R.W., Nolan J.P., Adrie C., Aibiki M., Berg R.A., Bottiger B.W., Callaway C., Clark R.S., Geocadin R.G., Jauch E.C.. Post-cardiac arrest syndrome: Epidemiology, pathophysiology, treatment, and prognostication. A consensus statement from the International Liaison Committee on Resuscitation (American Heart Association, Australian and New Zealand Council on Resuscitation, European Resuscitation Council, Heart and Stroke Foundation of Canada, InterAmerican Heart Foundation, Resuscitation Council of Asia, and the Resuscitation Council of Southern Africa); the American Heart Association Emergency Cardiovascular Care Committee; the Council on Cardiovascular Surgery and Anesthesia; the Council on Cardiopulmonary, Perioperative, and Critical Care; the Council on Clinical Cardiology; and the Stroke Council. Circulation (2008) 118 2452-2483. DOI: 10.1161/CIRCULATIONAHA.108.190652 4. Cung T.T., Morel O., Cayla G., Rioufol G., Garcia-Dorado D., Angoulvant D., Bonnefoy-Cudraz E., Guerin P., Elbaz M., Delarche N.. Cyclosporine before PCI in Patients with Acute Myocardial Infarction. New Engl. J. Med. (2015) 373 1021-1031. DOI: 10.1056/NEJMoa1505489 5. Freixa X., Bellera N., Ortiz-Perez J.T., Jimenez M., Pare C., Bosch X., De Caralt T.M., Betriu A., Masotti M.. Ischaemic postconditioning revisited: Lack of effects on infarct size following primary percutaneous coronary intervention. Eur. Hear. J. (2012) 33 103-112. DOI: 10.1093/eurheartj/ehr297 6. Limalanathan S., Andersen G.O., Klow N.E., Abdelnoor M., Hoffmann P., Eritsland J.. Effect of ischemic postconditioning on infarct size in patients with ST-elevation myocardial infarction treated by primary PCI results of the POSTEMI (POstconditioning in ST-Elevation Myocardial Infarction) randomized trial. J. Am. Heart Assoc (2014) 3 e000679. DOI: 10.1161/JAHA.113.000679 7. Li H., Liu Z., Wang J., Wong G.T., Cheung C.W., Zhang L., Chen C., Xia Z., Irwin M.G.. Susceptibility to myocardial ischemia reperfusion injury at early stage of type 1 diabetes in rats. Cardiovasc. Diabetol. (2013) 12 133. DOI: 10.1186/1475-2840-12-133 8. Brown J.R., Edwards F.H., O’Connor G.T., Ross C.S., Furnary A.P.. The diabetic disadvantage: Historical outcomes measures in diabetic patients undergoing cardiac surgery—The pre-intravenous insulin era. Semin. Thorac. Cardiovasc. Surg. (2006) 18 281-288. DOI: 10.1053/j.semtcvs.2006.04.004 9. Miki T., Miura T., Hotta H., Tanno M., Yano T., Sato T., Terashima Y., Takada A., Ishikawa S., Shimamoto K.. Endoplasmic reticulum stress in diabetic hearts abolishes erythropoietin-induced myocardial protection by impairment of phospho-glycogen synthase kinase-3beta-mediated suppression of mitochondrial permeability transition. Diabetes (2009) 58 2863-2872. DOI: 10.2337/db09-0158 10. Badalzadeh R., Azimi A., Alihemmati A., Yousefi B.. Chronic type-I diabetes could not impede the anti-inflammatory and anti-apoptotic effects of combined postconditioning with ischemia and cyclosporine A in myocardial reperfusion injury. J. Physiol. Biochem. (2017) 73 111-120. DOI: 10.1007/s13105-016-0530-4 11. Drenger B., Ostrovsky I.A., Barak M., Nechemia-Arbely Y., Ziv E., Axelrod J.H.. Diabetes blockade of sevoflurane postconditioning is not restored by insulin in the rat heart: Phosphorylated signal transducer and activator of transcription 3- and phosphatidylinositol 3-kinase-mediated inhibition. Anesthesiology (2011) 114 1364-1372. DOI: 10.1097/ALN.0b013e31820efafd 12. Ghaboura N., Tamareille S., Ducluzeau P.H., Grimaud L., Loufrani L., Croue A., Tourmen Y., Henrion D., Furber A., Prunier F.. Diabetes mellitus abrogates erythropoietin-induced cardioprotection against ischemic-reperfusion injury by alteration of the RISK/GSK-3beta signaling. Basic Res. Cardiol. (2011) 106 147-162. DOI: 10.1007/s00395-010-0130-3 13. Blake D.R., Meigs J.B., Muller D.C., Najjar S.S., Andres R., Nathan D.M.. Impaired glucose tolerance, but not impaired fasting glucose, is associated with increased levels of coronary heart disease risk factors: Results from the Baltimore Longitudinal Study on Aging. Diabetes (2004) 53 2095-2100. DOI: 10.2337/diabetes.53.8.2095 14. Deedwania P., Kosiborod M., Barrett E., Ceriello A., Isley W., Mazzone T., Raskin P.. Hyperglycemia and acute coronary syndrome: A scientific statement from the American Heart Association Diabetes Committee of the Council on Nutrition, Physical Activity, and Metabolism. Circulation (2008) 117 1610-1619. DOI: 10.1161/CIRCULATIONAHA.107.188629 15. Tian R., Abel E.D.. Responses of GLUT4-deficient hearts to ischemia underscore the importance of glycolysis. Circulation (2001) 103 2961-2966. DOI: 10.1161/01.CIR.103.24.2961 16. Jia G., DeMarco V.G., Sowers J.R.. Insulin resistance and hyperinsulinaemia in diabetic cardiomyopathy. Nat. Rev. Endocrinol. (2016) 12 144-153. DOI: 10.1038/nrendo.2015.216 17. Taskin E., Tuncer K.A., Guven C., Kaya S.T., Dursun N.. Inhibition of Angiotensin-II Production Increases Susceptibility to Acute Ischemia/Reperfusion Arrhythmia. Med. Sci. Monit. (2016) 22 4587-4595. DOI: 10.12659/MSM.896350 18. Pain T., Yang X.M., Critz S.D., Yue Y., Nakano A., Liu G.S., Heusch G., Cohen M.V., Downey J.M.. Opening of mitochondrial K(ATP) channels triggers the preconditioned state by generating free radicals. Circ. Res. (2000) 87 460-466. DOI: 10.1161/01.RES.87.6.460 19. Yang H., Zeng X.J., Wang H.X., Zhang L.K., Dong X.L., Guo S., Du J., Li H.H., Tang C.S.. Angiotensin IV protects against angiotensin II-induced cardiac injury via AT4 receptor. Peptides (2011) 32 2108-2115. DOI: 10.1016/j.peptides.2011.09.015 20. Xu F., Mao C., Hu Y., Rui C., Xu Z., Zhang L.. Cardiovascular effects of losartan and its relevant clinical application. Curr. Med. Chem. (2009) 16 3841-3857. DOI: 10.2174/092986709789178046 21. Klishadi M.S., Zarei F., Hejazian S.H., Moradi A., Hemati M., Safari F.. Losartan protects the heart against ischemia reperfusion injury: Sirtuin3 involvement. J. Pharm. Pharm. Sci. (2015) 18 112-123. DOI: 10.18433/J3XG7T 22. Gurevitch J., Pevni D., Frolkis I., Matsa M., Paz Y., Mohr R., Yakirevich V.. Captopril in cardioplegia and reperfusion: Protective effects on the ischemic heart. Ann. Thorac. Surg. (1997) 63 627-633. DOI: 10.1016/S0003-4975(96)00932-0 23. Babiker F., Al-Jarallah A., Joseph S.. The Interplay between the Renin Angiotensin System and Pacing Postconditioning Induced Cardiac Protection. PLoS ONE (2016) 11. DOI: 10.1371/journal.pone.0165777 24. Nishimoto M., Takai S., Sawada Y., Yuda A., Kondo K., Yamada M., Jin D., Sakaguchi M., Asada K., Sasaki S.. Chymase-dependent angiotensin II formation in the saphenous vein versus the internal thoracic artery. J. Thorac. Cardiovasc. Surg. (2001) 121 729-734. DOI: 10.1067/mtc.2001.112467 25. Urata H.. Pathological involvement of chymase-dependent angiotensin II formation in the development of cardiovascular disease. J. Renin Angiotensin Aldosterone Syst. (2000) 1 S35-S37. DOI: 10.3317/JRAAS.2000.054 26. Ozhan O., Parlakpinar H., Acet A.. Comparison of the effects of losartan, captopril, angiotensin II type 2 receptor agonist compound 21, and MAS receptor agonist AVE 0991 on myocardial ischemia-reperfusion necrosis in rats. Fundam. Clin. Pharmacol. (2021) 35 669-680. DOI: 10.1111/fcp.12599 27. Shi X., Shan Z., Yuan H., Guo H., Wang Y.. The effect of captopril and losartan on the electrophysiology of myocardial cells of myocardial ischemia rats. Int. J. Clin. Exp. Med. (2014) 7 5310-5316. PMID: 25664036 28. Benter I.F., Babiker F., Al-Rashdan I., Yousif M., Akhtar S.. RU28318, an aldosterone antagonist, in combination with an ACE inhibitor and angiotensin receptor blocker attenuates cardiac dysfunction in diabetes. J. Diabetes Res. (2013) 2013 427693. DOI: 10.1155/2013/427693 29. Bujak M., Dobaczewski M., Chatila K., Mendoza L.H., Li N., Reddy A., Frangogiannis N.G.. Interleukin-1 receptor type I signaling critically regulates infarct healing and cardiac remodeling. Am. J. Pathol. (2008) 173 57-67. DOI: 10.2353/ajpath.2008.070974 30. Ferdinandy P., Danial H., Ambrus I., Rothery R.A., Schulz R.. Peroxynitrite is a major contributor to cytokine-induced myocardial contractile failure. Circ. Res. (2000) 87 241-247. DOI: 10.1161/01.RES.87.3.241 31. Suematsu N., Tsutsui H., Wen J., Kang D., Ikeuchi M., Ide T., Hayashidani S., Shiomi T., Kubota T., Hamasaki N.. Oxidative stress mediates tumor necrosis factor-alpha-induced mitochondrial DNA damage and dysfunction in cardiac myocytes. Circulation (2003) 107 1418-1423. DOI: 10.1161/01.CIR.0000055318.09997.1F 32. Kantor P.F., Lucien A., Kozak R., Lopaschuk G.D.. The antianginal drug trimetazidine shifts cardiac energy metabolism from fatty acid oxidation to glucose oxidation by inhibiting mitochondrial long-chain 3-ketoacyl coenzyme A thiolase. Circ. Res. (2000) 86 580-588. DOI: 10.1161/01.RES.86.5.580 33. Frati G., Schirone L., Chimenti I., Yee D., Biondi-Zoccai G., Volpe M., Sciarretta S.. An overview of the inflammatory signalling mechanisms in the myocardium underlying the development of diabetic cardiomyopathy. Cardiovasc. Res. (2017) 113 378-388. DOI: 10.1093/cvr/cvx011 34. Forbes R.A., Steenbergen C., Murphy E.. Diazoxide-induced cardioprotection requires signaling through a redox-sensitive mechanism. Circ. Res. (2001) 88 802-809. DOI: 10.1161/hh0801.089342 35. Vanden Hoek T., Becker L.B., Shao Z.H., Li C.Q., Schumacker P.T.. Preconditioning in cardiomyocytes protects by attenuating oxidant stress at reperfusion. Circ. Res. (2000) 86 541-548. DOI: 10.1161/01.RES.86.5.541 36. Li Y., Yao Y., Li J., Chen Q., Zhang L., Wang Q.K.. Losartan protects against myocardial ischemia and reperfusion injury via vascular integrity preservation. FASEB J. (2019) 33 8555-8564. DOI: 10.1096/fj.201900060R 37. Marques-Neto S.R., Ferraz E.B., Rodrigues D.C., Njaine B., Rondinelli E., Campos de Carvalho A.C., Nascimento J.H.. AT1 and aldosterone receptors blockade prevents the chronic effect of nandrolone on the exercise-induced cardioprotection in perfused rat heart subjected to ischemia and reperfusion. Cardiovasc. Drugs Ther. (2014) 28 125-135. DOI: 10.1007/s10557-013-6503-8 38. Badalzadeh R., Mohammadi M., Najafi M., Ahmadiasl N., Farajnia S., Ebrahimi H.. The additive effects of ischemic postconditioning and cyclosporine-A on nitric oxide activity and functions of diabetic myocardium injured by ischemia/reperfusion. J. Cardiovasc. Pharmacol. Ther. (2012) 17 181-189. DOI: 10.1177/1074248411416118 39. Przyklenk K., Maynard M., Greiner D.L., Whittaker P.. Cardioprotection with postconditioning: Loss of efficacy in murine models of type-2 and type-1 diabetes. Antioxid. Redox Signal. (2011) 14 781-790. DOI: 10.1089/ars.2010.3343 40. Babiker F., Al-Jarallah A., Al-Awadi M.. Effects of Cardiac Hypertrophy, Diabetes, Aging, and Pregnancy on the Cardioprotective Effects of Postconditioning in Male and Female Rats. Cardiol. Res. Pr. (2019) 2019 3403959. DOI: 10.1155/2019/3403959 41. Babiker F.A., van Golde J., Vanagt W.Y., Prinzen F.W.. Pacing postconditioning: Impact of pacing algorithm, gender, and diabetes on its myocardial protective effects. J. Cardiovasc. Transl. Res. (2012) 5 727-734. DOI: 10.1007/s12265-012-9390-7 42. Akhtar S., Yousif M.H., Chandrasekhar B., Benter I.F.. Activation of EGFR/ERBB2 via pathways involving ERK1/2, P38 MAPK, AKT and FOXO enhances recovery of diabetic hearts from ischemia-reperfusion injury. PLoS ONE (2012) 7. DOI: 10.1371/journal.pone.0039066 43. Babiker F.A., Lorenzen-Schmidt I., Mokelke E., Vanagt W.Y., Delhaas T., Waltenberger J., Cleutjens J.P., Prinzen F.W.. Long-term protection and mechanism of pacing-induced postconditioning in the heart. Basic Res. Cardiol. (2010) 105 523-533. DOI: 10.1007/s00395-010-0095-2 44. Kim H.J., Jung B.H., Yoo K.Y., Han J.W., Um H.S., Chang B.S., Lee J.K.. Determination of the critical diabetes duration in a streptozotocin-induced diabetic rat calvarial defect model for experimentation regarding bone regeneration. J. Periodontal Implant. Sci. (2017) 47 339-350. DOI: 10.5051/jpis.2017.47.5.339 45. Zhang Y., Yuan D., Yao W., Zhu Q., Liu Y., Huang F., Feng J., Chen X., Huang Y., Chi X.. Hyperglycemia Aggravates Hepatic Ischemia Reperfusion Injury by Inducing Chronic Oxidative Stress and Inflammation. Oxid. Med. Cell. Longev. (2016) 2016 3919627. DOI: 10.1155/2016/3919627 46. Shi-Ting W., Mang-Hua X., Wen-Ting C., Feng-Hou G., Zhu-Ying G.. WITHDRAWN: Study on tolerance to ischemia-reperfusion injury and protection of ischemic preconditioning of type 1 diabetes rat heart. Biomed. Pharmacother. (2010) 1 56-60. DOI: 10.1016/j.biopha.2010.09.004 47. Qiu Y., Shil P.K., Zhu P., Yang H., Verma A., Lei B., Li Q.. Angiotensin-converting enzyme 2 (ACE2) activator diminazene aceturate ameliorates endotoxin-induced uveitis in mice. Investig. Ophthalmol. Vis. Sci. (2014) 55 3809-3818. DOI: 10.1167/iovs.14-13883 48. Dianat M., Hamzavi G.R., Badavi M., Samarbafzadeh A.. Effects of losartan and vanillic Acid co-administration on ischemia-reperfusion-induced oxidative stress in isolated rat heart. Iran. Red Crescent Med. J. (2014) 16 e16664. DOI: 10.5812/ircmj.16664 49. Maulik S.K., Kumari R., Maulik M., Manchanda S.C., Gupta S.K.. Captopril and its time of administration in myocardial ischaemic-reperfusion injury. Pharmacol. Res. (2001) 44 123-128. DOI: 10.1006/phrs.2001.0832 50. Warnholtz A., Nickenig G., Schulz E., Macharzina R., Brasen J.H., Skatchkov M., Heitzer T., Stasch J.P., Griendling K.K., Harrison D.G.. Increased NADH-oxidase-mediated superoxide production in the early stages of atherosclerosis: Evidence for involvement of the renin-angiotensin system. Circulation (1999) 99 2027-2033. DOI: 10.1161/01.CIR.99.15.2027 51. Zalba G., Beaumont F.J., San Jose G., Fortuno A., Fortuno M.A., Etayo J.C., Diez J.. Vascular NADH/NADPH oxidase is involved in enhanced superoxide production in spontaneously hypertensive rats. Hypertension (2000) 35 1055-1061. DOI: 10.1161/01.HYP.35.5.1055 52. Real J.T., Martinez-Hervas S., Tormos M.C., Domenech E., Pallardo F.V., Saez-Tormo G., Redon J., Carmena R., Chaves F.J., Ascaso J.F.. Increased oxidative stress levels and normal antioxidant enzyme activity in circulating mononuclear cells from patients of familial hypercholesterolemia. Metabolism (2010) 59 293-298. DOI: 10.1016/j.metabol.2009.07.026 53. Chen X., Li X., Zhang W., He J., Xu B., Lei B., Wang Z., Cates C., Rousselle T., Li J.. Activation of AMPK inhibits inflammatory response during hypoxia and reoxygenation through modulating JNK-mediated NF-kappaB pathway. Metabolism (2018) 83 256-270. DOI: 10.1016/j.metabol.2018.03.004 54. Liu H., Li S., Jiang W., Li Y.. MiR-484 Protects Rat Myocardial Cells from Ischemia-Reperfusion Injury by Inhibiting Caspase-3 and Caspase-9 during Apoptosis. Korean Circ. J. (2020) 50 250-263. DOI: 10.4070/kcj.2019.0107 55. Zhang W., Xing B., Yang L., Shi J., Zhou X.. Icaritin Attenuates Myocardial Ischemia and Reperfusion Injury Via Anti-Inflammatory and Anti-Oxidative Stress Effects in Rats. Am. J. Chin. Med. (2015) 43 1083-1097. DOI: 10.1142/S0192415X15500627 56. Tian Y., Li H., Liu P., Xu J.M., Irwin M.G., Xia Z., Tian G.. Captopril Pretreatment Produces an Additive Cardioprotection to Isoflurane Preconditioning in Attenuating Myocardial Ischemia Reperfusion Injury in Rabbits and in Humans. Mediat. Inflamm. (2015) 2015 819232. DOI: 10.1155/2015/819232 57. Lekli I., Szabo G., Juhasz B., Das S., Das M., Varga E., Szendrei L., Gesztelyi R., Varadi J., Bak I.. Protective mechanisms of resveratrol against ischemia-reperfusion-induced damage in hearts obtained from Zucker obese rats: The role of GLUT-4 and endothelin. Am. J. Physiol. Heart Circ. Physiol. (2008) 294 H859-H866. DOI: 10.1152/ajpheart.01048.2007 58. Ramasamy R., Hwang Y.C., Whang J., Bergmann S.R.. Protection of ischemic hearts by high glucose is mediated, in part, by GLUT-4. Am. J. Physiol. Heart Circ. Physiol. (2001) 281 H290-H297. DOI: 10.1152/ajpheart.2001.281.1.H290 59. Ismaeil A., Babiker F., Al-Sabah S.. Discrepancy between the Actions of Glucagon-like Peptide-1 Receptor Ligands in the Protection of the Heart against Ischemia Reperfusion Injury. Pharmaceuticals (2022) 15. DOI: 10.3390/ph15060720 60. Khalaf A., Babiker F.. Discrepancy in calcium release from the sarcoplasmic reticulum and intracellular acidic stores for the protection of the heart against ischemia/reperfusion injury. J. Physiol. Biochem. (2016) 72 495-508. DOI: 10.1007/s13105-016-0498-0 61. Al-Herz W., Babiker F.. Acute Intravenous Infusion of Immunoglobulins Protects Against Myocardial Ischemia-Reperfusion Injury Through Inhibition of Caspase-3. Cell Physiol. Biochem. (2017) 42 2295-2306. DOI: 10.1159/000480002
--- title: Nomogram and Risk Calculator for Postoperative Tracheostomy after Heart Valve Surgery authors: - Xiangchao Ding - Bing Sun - Liang Liu - Yuan Lei - Yunshu Su journal: Journal of Cardiovascular Development and Disease year: 2023 pmcid: PMC9967351 doi: 10.3390/jcdd10020073 license: CC BY 4.0 --- # Nomogram and Risk Calculator for Postoperative Tracheostomy after Heart Valve Surgery ## Abstract Postoperative tracheostomy (POT) is an important indicator of critical illness, associated with poorer prognoses and increased medical burdens. However, studies on POTs after heart valve surgery (HVS) have not been reported. The objectives of this study were first to identify the risk factors and develop a risk prediction model for POTs after HVS, and second to clarify the relationship between POTs and clinical outcomes. Consecutive adults undergoing HVS from January 2016 to December 2019 in a single cardiovascular center were enrolled, and a POT was performed in $1.8\%$ of the included patients ($\frac{68}{3853}$). Compared to patients without POTs, the patients with POTs had higher rates of readmission to the ICU and in-hospital mortality, as well as longer ICU and hospital stays. Five factors were identified to be significantly associated with POTs after HVS by our multivariate analysis, including age, diabetes mellitus, pulmonary edema, intraoperative transfusion of red blood cells, and surgical types. A nomogram and a risk calculator were constructed based on the five factors, showing excellent discrimination, calibration, and clinical utility. Three risk intervals were defined as low-, medium-, and high-risk groups according to the nomogram and clinical practice. The findings of this study may be helpful for early risk assessment and perioperative management. ## 1. Introduction Postoperative complications are prevalent after cardiovascular surgery, and a postoperative tracheostomy (POT) is often required to strengthen airway management when respiratory failure, circulatory failure, multiple organ dysfunction, and other critical adverse events develop [1,2,3]. As an important indicator of poor prognoses, POT operations often indicate a higher risk of mortality, prolonged hospital stay, increased medical burden, and declined quality of life [4,5,6,7]. The incidence of POTs varies widely in the previous literature due to the different surgical populations in different studies [3,7,8,9,10,11]. Compared to other surgical types, patients undergoing cardiovascular surgery have been reported to have a relatively higher POT rate, mostly in the range of 1.4–$11.8\%$ [3,12,13]. For patients who are expected to be unable to escape from mechanical ventilation in the short term, a tracheostomy is a routine operation to effectively relieve airway obstruction, reduce airway resistance, reduce airway dead space, and increase effective ventilation volume [10,14]. Previous studies have reported that among critical patients requiring prolonged mechanical ventilation in intensive care unit (ICU) settings, compared with patients who did not receive a tracheostomy, patients who received a tracheostomy had significantly declined mortality and improved prognoses, despite a longer duration of mechanical ventilation and hospitalization [14,15,16,17,18]. In some major operations, such as liver transplantation and heart surgery, patients requiring a tracheostomy have a significantly longer hospital stay and lower overall survival rate compared with patients who do not require a tracheostomy [4,10]. Several studies focused on POTs have been conducted in patients undergoing cardiovascular surgery due to the high prevalence and significantly poorer outcomes [3,5,10,19,20]. Some significant risk factors for POTs have been reported and several predictive systems have been established in previous studies [4,5,21,22]. However, relevant clinical studies conducted in this field are still currently limited, and none of these previous studies were carried out specifically in patients undergoing heart valve surgery (HVS). Therefore, our understanding of the risk factors for POTs after HVS needs to be deepened and the construction of a credible and convenient risk prediction model is still urgently needed. The primary objective of this study was to identify independent risk factors, develop a risk prediction model for POTs after HVS, and perform risk stratification according to the established model. The second objective of this study was to clarify the relationship between POTs and in-hospital clinical outcomes, and thus to provide evidence-based support for clinical practice. ## 2.1. Ethical Statement This study was conducted based on the Declaration of Helsinki’s ethical principles. The Ethics Committee of Union Hospital, Tongji Medical College, Huazhong University of Science and Technology approved this study. Due to its observational and retrospective nature, patients’ signed informed consent was not needed. ## 2.2. Study Population This was a retrospective, observational single-center cohort study. From January 2016 to December 2019, consecutive adult patients undergoing HVS in our hospital were identified and analyzed. A number of conditions were excluded from the current study, including: [1] younger than 18 years; [2] immunodeficiency, immunosuppression, or organ transplant history; [3] intraoperative death, early postoperative death or discharge (within the first 48 h after surgery); and [4] incomplete data in medical records. ## 2.3. Data Collection Clinical data were extracted from the hospital’s electronic medical record system, including preoperative, intraoperative, and postoperative variables. Preoperative data included demographics (gender, age, body mass index, smoking and drinking history), underlying conditions (hypertension, diabetes mellitus, chronic obstructive pulmonary disease, cerebrovascular disease, peripheral vascular disease, renal insufficiency, gastrointestinal tract disease, atrial fibrillation, pulmonary edema, New York Heart Association (NYHA) class, cardiac surgery history, and general surgery history), ultrasound results (pulmonary artery hypertension, pericardial effusion, left ventricular ejection fraction, diameters of the left atrium, left ventricle, right atrium, and right ventricle), and laboratory values (white blood cell count, red blood cell count, hemoglobin, platelet count, serum creatinine, serum albumin, and serum globulin). Intraoperative data included cardiopulmonary bypass time, aortic cross clamp time, surgical types (isolated valve surgery, combined coronary artery bypass grafting (CABG), and combined aortic surgery), and transfusion of red blood cells (RBCs). Postoperative data included the rates of readmission to ICU and in-hospital mortality, as well as the lengths of patients’ ICU and hospital stays. ## 2.4. Endpoints The primary endpoint was tracheostomy operations after HVS in this study. All the operations were performed via a percutaneous route using disposable sterile percutaneous tracheostomy surgical instrument package at patients’ bedsides by experienced operators. In this study, tracheostomy was indicated for the following reasons: repeated intubation, predicted difficult reintubation, one or more failed trials of extubation, bypass of upper airway obstruction, prolonged mechanical ventilation, and tracheal access that was necessary for removing thick pulmonary secretions. ## 2.5. Statistical Analysis Statistical analysis of the data was performed with IBM SPSS (version 26.0) and R software (version 4.0.5). Differences were considered to be statistically significant if the two-tailed p values were less than 0.05. Categorical data were presented as numbers (proportions) and continuous variables were presented as means ± standard deviations or medians (interquartile ranges) according to whether they were normally distributed. For univariate analysis, categorical data were analyzed using chi-square test or Fisher’s exact test, and continuous variables were analyzed using Student’s t-test or Mann–Whitney U test, as appropriate. Factors initially screened by univariate analysis (p values less than 0.1) were then entered into a forward stepwise multivariate logistic regression procedure to identify significant risk factors for POTs after HVS. The results of multivariate analysis were presented as p values, coefficients, and odds ratios (ORs) with $95\%$ confidence intervals (CIs). A nomogram and a web-based risk calculator were then constructed based on the multivariate logistic regression model. Finally, risk stratification was performed to further facilitate the clinical application based on the nomogram model. The performance of the model was assessed with discrimination, calibration, and clinical utility. Bootstrap method with 1000 replicates was used for internal validation. The area under the receiver operating characteristic (ROC) curve (AUC) was used to assess the discrimination. The Hosmer–Lemeshow goodness-of-fit test and calibration plot were used to assess the calibration. The decision and clinical impact curves were used to assess the clinical utility. ## 3.1. Demographic Characteristics A total of 3853 adult patients undergoing HVS met the inclusion criteria and were analyzed in the current study. The average age of these patients was 51.3 ± 12.5 years. Female patients accounted for $46.2\%$ of the patients. The incidence rate for POTs in this population was $1.8\%$ ($\frac{68}{3853}$). The average body mass index of this study population was 23.0 ± 3.3 kg/m2. A total of $20.1\%$ of the patients had a history of drinking and $26.7\%$ had a history of smoking. A significant proportion of patients had at least one underlying disease, including pulmonary artery hypertension ($32.1\%$), general surgery history ($29.7\%$), hypertension ($24.2\%$), atrial fibrillation ($23.3\%$), pericardial effusion ($15.6\%$), chronic obstructive pulmonary disease ($12.9\%$), renal insufficiency ($8.2\%$), gastrointestinal tract disease ($8.2\%$), cardiac surgery history ($8.0\%$), pulmonary edema ($6.0\%$), and diabetes mellitus ($5.7\%$). Isolated valve surgeries were performed on $75.2\%$ of the patients, combined CABG on $12.5\%$, and combined aortic surgeries on $12.3\%$. The median cardiopulmonary bypass time was 108 [86, 139] minutes, the aortic cross clamp time was 72 [54, 95] minutes, and the transfusion of intraoperative RBC was 1 [1, 3] units, respectively. The incidence rate for POTs was $0.9\%$ in the patients undergoing isolated valve surgeries, $3.7\%$ in those undergoing concomitant CABG, and $5.1\%$ in those undergoing concomitant aortic surgeries. The types of HVS are as follows: isolated aortic valve surgeries accounted for $26.8\%$, isolated mitral valve surgeries $25.1\%$, isolated tricuspid valve surgeries $8.1\%$, combined aortic and mitral valve surgeries $15.8\%$, combined aortic and tricuspid valve surgeries $1.2\%$, combined mitral and tricuspid valve surgeryies $15.8\%$, and combined aortic, mitral, and tricuspid valve surgeries $7.2\%$. Their corresponding POT rates were, respectively, $2.4\%$, $1.2\%$, $1.0\%$, $1.6\%$, $2.2\%$, $2.1\%$, and $1.1\%$. ## 3.2. Development of the Nomogram and Risk Calculator A univariate analysis was first conducted to explore possible risk factors for POTs after HVS (Table 1). The factors screened (p values less than 0.1) by the univariate analysis were then entered into a forward stepwise multivariate logistic regression procedure to further identify independent risk factors, including gender, age, smoking history, hypertension, diabetes mellitus, chronic obstructive pulmonary disease, cerebrovascular disease, renal insufficiency, pulmonary edema, cardiac surgery history, NYHA class, white blood cell count, red blood cell count, platelet count, serum creatinine, serum albumin, surgical types, cardiopulmonary bypass, and intraoperative transfusion of RBCs. A multicollinearity test was conducted before the regression analysis in order to exclude confounded variables with potential multicollinearity. Finally, five independent risk factors for POT after HVS were identified in the multivariate logistic regression analysis, including being of an older age, diabetes mellitus, pulmonary edema, combined aortic surgeries, and more transfusions of intraoperative RBCs (Table 2). Based on the logistic regression model established using the above five risk factors, we constructed a graphical nomogram for convenience in clinical use (Figure 1). The nomogram can proportionally convert each regression coefficient in the multivariate analysis to a scale of 0–100 points, reflecting their relative importance. The individualized POT risk of each patient can be easily obtained by summing the points of all the five risk factors and then identifying the corresponding probability at the bottom of the nomogram. Older patients who have had diabetes mellitus, pulmonary edema, combined aortic surgeries, and more intraoperative transfusions of RBCs may obtain more points and thus are at a higher risk of POTs after HVS. An example showing the usage of the nomogram is given in Figure 1. To facilitate the usage of this system in modern clinical work, we further created an internet-based risk calculator (available at https://pothvs.shinyapps.io/dynnomapp/, accessed on 7 December 2022). When using this online predictive system, we only need to choose the information of the patient and then click the “Predict” button; the estimated risk of POTs after HVS is calculated in the “Graphical summary” area on the right (Figure 2). When calculating the risk of another patient, one can simply change the information on the left. This makes it possible to assess the risks of multiple patients simultaneously, as well as the risk comparison among different patients. The specific information of the patients and the model can also be obtained by clicking the “Numerical summary” and “Model summary” on the right. When a user cannot log in with a new device, we recommend logging out first by pressing the “Quit” button at the left bottom and then reloading the procedure. ## 3.3. Validation and Assessment of the Model The model was well validated internally by the bootstrap method with 1000 replicates. By plotting the ROC curves and calculating the AUC, the model demonstrated excellent discrimination, with an AUC of 0.938 ($95\%$ CI, (0.912–0.964), Figure 3A). By plotting the calibration curves and the goodness-of-fit test, the model demonstrated good consistency between the predicted and the actual probabilities, with a Hosmer–Lemershow chi-square value of 3.260 ($$p \leq 0.860$$, Figure 3B). By plotting the decision and clinical impact curves, the model demonstrated good clinical utility (Figure 3C,D), which may bring more clinical net benefits compared to the “treat-all/none” strategies. ## 3.4. Risk Stratification To facilitate clinical applications, we further propose a more concise risk stratification for POTs after HVS on the basis of the nomogram and clinical practice (Table 3). We stratified all the patients into three risk intervals: low-risk, medium-risk, and high-risk groups. The selected cutoffs of the estimated probabilities were, respectively, 0.01 and 0.05, corresponding to scores of 127 and152 points on the nomogram. In the current study, $81.5\%$ of the patients were stratified into the low-risk group, $10.9\%$ into the medium-risk group, and $7.6\%$ into the high-risk group. Both the estimated and observed probabilities demonstrated a significant difference across the three risk intervals and the estimated and observed probabilities showed good consistency within each risk interval, indicating the rationality of the risk stratification (Figure 4). ## 3.5. Clinical Outcomes The overall mortality rate was $2.9\%$ ($\frac{111}{3853}$) in the current study, with a significant increase in patients who experienced POTs ($57.4\%$ versus $1.9\%$, $p \leq 0.001$). In addition, the rate of readmission to ICU was significantly higher in patients with POTs, and the lengths of their ICU and hospital stays were significantly longer compared to those of patients who did not require POTs. The comparison details of these outcomes between patients with and without POTs are presented in Table 4. For the 68 patients who underwent POTs after HVS, five patients were operated on within the first postoperative week, with a mortality rate of $20.0\%$ ($\frac{1}{5}$); forty-one patients were operated on between the first and the second postoperative week, with a mortality rate of $58.5\%$ ($\frac{24}{41}$); and twenty-two patients were operated on after two postoperative weeks, with a mortality rate of $63.6\%$ ($\frac{14}{22}$). ## 4. Discussion Undergoing a tracheostomy is an important indicator of the increased risk of poor prognoses in patients undergoing cardiovascular surgeries [3,7,9,20], which was again confirmed in the current study. Due to the difference of surgical populations in different studies, the reported rates of POTs in the previous literature were quite different [3,7,8,9,10,11]. The overall incidence rate of POTs after HVS was $1.8\%$ in our analysis, falling within the range of incidence rates reported in the previous literature [3,12,13]. The overall mortality rate was $2.9\%$; however, patients with POTs had a significantly higher rate of mortality compared with patients without POTs. Moreover, a higher rate of readmission to the ICU, as well as prolonged ICU and hospital stays were also observed in patients with POTs. The increased risk of multiple poor clinical outcomes in patients with POTs stressed the importance of identifying significant risk factors for POTs after HVS and developing a compelling risk prediction model. In the current study, using clinical data of 3853 adult patients who underwent HVS at a single cardiovascular center, we analyzed the risk factors of POTs after HVS and developed a parsimonious risk prediction model. Through a univariate and multivariate analysis, we identified five independent risk factors for POT after HVS, including being of an older age, having diabetes mellitus, pulmonary edema, combined aortic surgery, and more transfusions of intraoperative RBCs. To facilitate the clinical application of the logistic regression model, we further constructed a visual nomogram and an internet-based risk calculator. The model demonstrated excellent discrimination, calibration, and clinical utility, and was well validated internally. On the basis of the nomogram and clinical practice, we defined three risk intervals: low-risk, medium-risk, and high-risk groups. To the best of our knowledge, this is the first report that has targeted the risk factors of POTs after HVS and the first attempt to construct a nomogram model and an internet-based risk calculator worldwide, which may have certain clinical guiding significance. Numerous studies have been conducted on POTs after various surgical procedures due to the adverse outcomes, and the risk factors identified in our analysis have also been reported in different reports [3,4,5,19,20,21,22]. Being of an older age was identified to associate with a higher risk of POTs in the current study; however, the results of whether the risk of POTs would increase with age were inconsistent in previous studies, which may be due to the difference in disease types and study populations [4,22,23]. Diabetes mellitus and pulmonary edema as risk factors for POTs have also been reported in previous studies, which may be mainly associated with higher risks of various pulmonary complications [7,24,25]. The relationship between combined aortic surgery and ventilation dependence was reported a long time ago, and patients undergoing aortic procedures have been identified to have a higher risk of respiratory failure [26,27]. Intraoperative transfusions of RBCs may significantly increase the risk of transfusion-related acute lung injury and systemic inflammatory response syndrome, which may lead to prolonged hospitalization, increased medical costs, and a higher risk of mortality [28,29,30]. Although RBC transfusions are routine in traditional cardiovascular surgery to deal with bleeding and improve tissue oxygen delivery, there is growing evidence that the restrictive RBC transfusion strategy is safe and effective, which has been recommended by practice guidelines [31,32,33,34]. In addition, previous studies have found that the risk factors identified in this study are related to the development of various postoperative respiratory complications, such as pneumonia to some extent, which may indirectly increase the risk of the need for POTs [35,36,37,38,39,40,41,42]. Several other factors have also been reported to be related to an increased risk of POTs in previous studies but were not identified in the current study, including renal insufficiency, chronic obstructive pulmonary disease, white blood cell count, smoking history, body mass index, and platelet transfusion [4,5,7,9,43]. Additionally, although some postoperative variables have been identified to be related to POTs in the literature [44], we did not include these variables in our analysis. This was because the inclusion of these variables would not achieve the purpose of early prediction as they were not available early. Nonetheless, the results of our analysis demonstrated that a model constructed using only the preoperative and intraoperative variables identified in this study could also perform well. Using the nomogram and risk calculator, we can accurately and easily estimate personalized POT risks, identify high-risk subsets and then take early and appropriate intervention measures. In the past few years, some measures have been proposed to be effective in reducing the risk of POTs, such as prophylactic administration of sivelestat at the beginning of cardiopulmonary bypasses. Taking appropriate measures targeting high-risk patients identified by our risk prediction model may significantly improve prognoses and achieve greater financial success. Tracheostomies have been proven to be an effective treatment for various critically ill patients in recent years [15,17,45]. For patients undergoing high-risk surgeries, such as cardiovascular procedures, performing tracheostomies at an optimal time point when needed may significantly improve their prognoses [3,8,19,20]. However, the optimal timing is still unclear and controversial even though a lot of effort has been made by scientists and clinicians [1,8,16]. The results of this study showed that the in-hospital mortality rate increased in patients undergoing late POTs compared to patients undergoing early POT, consistent with the majority of the published reports [3,8,11,16,20]. However, we cannot simply conclude that the earlier the POT is performed, the better the prognosis will be. We must realize that this result was only based on a small sample which only included 68 patients who underwent POTs, and we cannot guarantee that all these patients had the same basic conditions when the POTs were performed, which may also have a significant impact on patients’ outcomes. Therefore, a prospective large sample study is still needed to further determine the timing of POTs after cardiovascular surgery. There are several limitations in the current study that should be noted. First, this was a single-center study and was not validated externally in an independent dataset, which may limit the generalizability of our findings. Second, some possible risk factors, such as the N-terminal fragment of B-type natriuretic propeptide [46], were not collected and included in our analysis, even though the established model performed well. Third, the data we collected were limited to hospitalization, and long-term prognoses after discharge were not followed or analyzed, which needs to be strengthened in future studies. Fourth, due to the nature of the retrospective observational real-world study, we cannot accurately judge whether there were some patients who needed POTs theoretically but POTs were not performed actually among the dead patients, and therefore we did not know whether a POT would reduce the expected mortality in this subset of the patients, which may lead to some difference between the actual situation and the ideal judgment. ## 5. Conclusions A POT after HVS is not uncommon, and is associated with poorer outcomes. In a multivariate analysis, five independent risk factors for POTs after HVS were identified, including being of an older age, having diabetes mellitus, pulmonary edema, combined aortic surgery, and more transfusions of intraoperative RBCs. The prediction model constructed using the above five risk factors demonstrated excellent discrimination, calibration, and clinical utility, which may be helpful for early risk assessment and perioperative management. A graphical nomogram and an internet-based risk calculator were constructed to facilitate clinical applications on the basis of the multivariate model, and three risk groups were defined based on the nomogram and clinical practice. To our knowledge, this is the first report that has targeted the risk factors of POTs after HVS and the first attempt to construct a nomogram model and an internet-based risk calculator worldwide, which may have certain clinical guiding significance. ## References 1. Kohne J.G., MacLaren G., Rider E., Carr B.D., Mallory P., Gebremariam A., Friedman M.L., Barbaro R.P.. Tracheostomy Practices and Outcomes in Children during Respiratory Extracorporeal Membrane Oxygenation. Pediatr. Crit. Care Med. (2022). DOI: 10.1097/PCC.0000000000002902 2. Cotts T., Hirsch J., Thorne M., Gajarski R.. Tracheostomy after pediatric cardiac surgery: Frequency, indications, and outcomes. J. Thorac. Cardiovasc. Surg. (2011) 141 413-418. DOI: 10.1016/j.jtcvs.2010.06.027 3. Sareh S., Toppen W., Ugarte R., Sanaiha Y., Hadaya J., Seo Y.J., Aguayo E., Shemin R., Benharash P.. Impact of Early Tracheostomy on Outcomes after Cardiac Surgery: A National Analysis. Ann. Thorac. Surg. (2021) 111 1537-1544. DOI: 10.1016/j.athoracsur.2020.07.027 4. Graham R.C., Bush W.J., Mella J.S., Fridell J.A., Ekser B., Mihaylov P., Kubal C.A., Mangus R.S.. Tracheostomy Post Liver Transplant: Predictors, Complications, and Outcomes. Ann. Transpl. (2020) 25 e920630-1. DOI: 10.12659/AOT.920630 5. Songdechakraiwut T., Aftab M., Chatterjee S., Green S.Y., Price M.D., Preventza O., de la Cruz K.I., LeMaire S.A., Coselli J.S.. Tracheostomy after Thoracoabdominal Aortic Aneurysm Repair: Risk Factors and Outcomes. Ann. Thorac. Surg. (2019) 108 778-784. DOI: 10.1016/j.athoracsur.2019.02.063 6. Jarosz K., Kubisa B., Andrzejewska A., Mrówczyńska K., Hamerlak Z., Bartkowska-Sniatkowska A.. Adverse outcomes after percutaneous dilatational tracheostomy versus surgical tracheostomy in intensive care patients: Case series and literature review. Ther. Clin. Risk Manag. (2017) 13 975-981. DOI: 10.2147/TCRM.S135553 7. Wang D., Wang S., Du Y., Song Y., Le S., Wang H., Zhang A., Huang X., Wu L., Du X.. A Predictive Scoring Model for Postoperative Tracheostomy in Patients Who Underwent Cardiac Surgery. Front. Cardiovasc. Med. (2022) 8 799605. DOI: 10.3389/fcvm.2021.799605 8. Robba C., Galimberti S., Graziano F., Wiegers E.J.A., Lingsma H.F., Iaquaniello C., Stocchetti N., Menon D., Citerio G.. Tracheostomy practice and timing in traumatic brain-injured patients: A CENTER-TBI study. Intens. Care Med. (2020) 46 983-994. DOI: 10.1007/s00134-020-05935-5 9. Wang D., Wang S., Song Y., Wang H., Zhang A., Wu L., Huang X., Ye P., Du X.. Predictors and outcomes of postoperative tracheostomy in patients undergoing acute type A aortic dissection surgery. Bmc Cardiovasc. Disor. (2022) 22. DOI: 10.1186/s12872-022-02538-4 10. Mastropietro C.W., Benneyworth B.D., Turrentine M., Wallace A.S., Hornik C.P., Jacobs J.P., Jacobs M.L.. Tracheostomy after Operations for Congenital Heart Disease: An Analysis of the Society of Thoracic Surgeons Congenital Heart Surgery Database. Ann. Thorac. Surg. (2016) 101 2285-2292. DOI: 10.1016/j.athoracsur.2016.01.046 11. Miyoshi R., Chen-Yoshikawa T.F., Hamaji M., Kawaguchi A., Kayawake H., Hijiya K., Motoyama H., Aoyama A., Date H.. Effect of early tracheostomy on clinical outcomes in critically ill lung transplant recipients. Gen. Thorac. Cardiovasc. Surg. (2018) 66 529-536. DOI: 10.1007/s11748-018-0949-3 12. Jin M., Ma W., Liu S., Zhu J., Sun L., Lu J., Cheng W.. Predictors of Prolonged Mechanical Ventilation in Adults after Acute Type-A Aortic Dissection Repair. J. Cardiothor. Vasc. An. (2017) 31 1580-1587. DOI: 10.1053/j.jvca.2017.03.036 13. Comas G.M., Leshnower B.G., Halkos M.E., Thourani V.H., Puskas J.D., Guyton R.A., Kilgo P.D., Chen E.P.. Acute Type A Dissection: Impact of Antegrade Cerebral Perfusion under Moderate Hypothermia. Ann. Thorac. Surg. (2013) 96 2135-2141. DOI: 10.1016/j.athoracsur.2013.06.085 14. Salna M., Tipograf Y., Liou P., Chicotka S., Biscotti M., Agerstrand C., Abrams D., Brodie D., Bacchetta M.. Tracheostomy Is Safe During Extracorporeal Membrane Oxygenation Support. Asaio J. (2020) 66 652-656. DOI: 10.1097/MAT.0000000000001059 15. Angel L.F., Amoroso N.E., Rafeq S., Mitzman B., Goldenberg R., Shekar S.P., Troxel A.B., Zhang Y., Chang S.H., Kwak P.. Percutaneous Dilational Tracheostomy for Coronavirus Disease 2019 Patients Requiring Mechanical Ventilation. Crit. Care Med. (2021). DOI: 10.1097/CCM.0000000000004969 16. Tripathi S., Swayampakula A.K., Deshpande G.G., Astle M., Wang Y., Welke K.F.. Illustration of the current practice and outcome comparison of early versus late tracheostomy after pediatric ECMO. Int. J. Artif. Organs (2020) 43 726-734. DOI: 10.1177/0391398820913571 17. Smith M.C., Evans P.T., Prendergast K.M., Schneeberger S.J., Henson C.P., McGrane S., Kopp E.B., Collins N.E., Guillamondegui O.D., Dennis B.M.. Surgical outcomes and complications of bedside tracheostomy in the ICU for patients on ECMO. Perfusion (2022) 37 26-30. DOI: 10.1177/0267659120979564 18. Schneider H., Hertel F., Kuhn M., Ragaller M., Gottschlich B., Trabitzsch A., Dengl M., Neudert M., Reichmann H., Wöpking S.. Decannulation and Functional Outcome after Tracheostomy in Patients with Severe Stroke (DECAST): A Prospective Observational Study. Neurocrit. Care (2017) 27 26-34. DOI: 10.1007/s12028-017-0390-y 19. Xin M., Wang L., Li C., Hou D., Wang H., Wang J., Jia M., Hou X.. Percutaneous dilatation tracheotomy in patients on extracorporeal membrane oxygenation after cardiac surgery. Perfusion (2022) 1961970979. DOI: 10.1177/02676591221099811 20. Okada M., Watanuki H., Masato T., Sugiyama K., Futamura Y., Matsuyama K.. Impact of Tracheostomy Timing on Outcomes after Cardiovascular Surgery. J. Cardiothor. Vasc. An. (2022) 36 2335-2338. DOI: 10.1053/j.jvca.2021.10.001 21. Wang Y., Guo Z., Fan D., Lu H., Xie D., Zhang D., Jiang Y., Li P., Teng H.. A Meta-Analysis of the Influencing Factors for Tracheostomy after Cervical Spinal Cord Injury. Biomed Res. Int. (2018) 2018 5895830. DOI: 10.1155/2018/5895830 22. Humble S.S., Wilson L.D., McKenna J.W., Leath T.C., Song Y., Davidson M.A., Ehrenfeld J.M., Guillamondegui O.D., Pandharipande P.P., Patel M.B.. Tracheostomy risk factors and outcomes after severe traumatic brain injury. Brain Inj. (2016) 30 1642-1647. DOI: 10.1080/02699052.2016.1199915 23. Mu Z., Zhang Z.. Risk factors for tracheostomy after traumatic cervical spinal cord injury. J. Orthop. Surg. (2019) 27 920545668. DOI: 10.1177/2309499019861809 24. Szylinska A., Listewnik M., Ciosek Z., Ptak M., Mikolajczyk A., Pawlukowska W., Rotter I.. The Relationship between Diabetes Mellitus and Respiratory Function in Patients Eligible for Coronary Artery Bypass Grafting. Int. J. Environ. Res. Public Health (2018) 15. DOI: 10.3390/ijerph15050907 25. Lauruschkat A.H., Arnrich B., Albert A.A., Walter J.A., Amann B., Rosendahl U.P., Alexander T., Ennker J.. Diabetes mellitus as a risk factor for pulmonary complications after coronary bypass surgery. J. Thorac. Cardiovasc. Surg. (2008) 135 1047-1053. DOI: 10.1016/j.jtcvs.2007.07.066 26. Murthy S.C., Arroliga A.C., Walts P.A., Feng J., Yared J., Lytle B.W., Blackstone E.H.. Ventilatory dependency after cardiovascular surgery. J. Thorac. Cardiovasc. Surg. (2007) 134 484-490. DOI: 10.1016/j.jtcvs.2007.03.006 27. Filsoufi F., Rahmanian P.B., Castillo J.G., Chikwe J., Adams D.H.. Predictors and early and late outcomes of respiratory failure in contemporary cardiac surgery. Chest (2008) 133 713-721. DOI: 10.1378/chest.07-1028 28. Crawford T.C., Magruder J.T., Fraser C., Suarez-Pierre A., Alejo D., Bobbitt J., Fonner C.E., Canner J.K., Horvath K., Wehberg K.. Less Is More: Results of a Statewide Analysis of the Impact of Blood Transfusion on Coronary Artery Bypass Grafting Outcomes. Ann. Thorac. Surg. (2018) 105 129-136. DOI: 10.1016/j.athoracsur.2017.06.062 29. Likosky D.S., Paone G., Zhang M., Rogers M.A.M., Harrington S.D., Theurer P.F., DeLucia A., Fishstrom A., Camaj A., Prager R.L.. Red Blood Cell Transfusions Impact Pneumonia Rates after Coronary Artery Bypass Grafting. Ann. Thorac. Surg. (2015) 100 794-801. DOI: 10.1016/j.athoracsur.2015.03.089 30. LaPar D.J., Crosby I.K., Ailawadi G., Ad N., Choi E., Spiess B.D., Rich J.B., Kasirajan V., Fonner E., Kron I.L.. Blood product conservation is associated with improved outcomes and reduced costs after cardiac surgery. J. Thorac. Cardiovasc. Surg. (2013) 145 796-804. DOI: 10.1016/j.jtcvs.2012.12.041 31. Naeem S.S., Sodha N.R., Sellke F.W., Ehsan A.. Impact of Packed Red Blood Cell and Platelet Transfusions in Patients Undergoing Dissection Repair. J. Surg. Res. (2018) 232 338-345. DOI: 10.1016/j.jss.2018.06.048 32. Mazer C.D., Whitlock R.P., Fergusson D.A., Hall J., Belley-Cote E., Connolly K., Khanykin B., Gregory A.J., de Médicis É., McGuinness S.. Restrictive or Liberal Red-Cell Transfusion for Cardiac Surgery. N. Engl. J. Med. (2017) 377 2133-2144. DOI: 10.1056/NEJMoa1711818 33. Carson J.L., Guyatt G., Heddle N.M., Grossman B.J., Cohn C.S., Fung M.K., Gernsheimer T., Holcomb J.B., Kaplan L.J., Katz L.M.. Clinical Practice Guidelines from the AABB: Red Blood Cell Transfusion Thresholds and Storage. JAMA (2016) 316 2025-2035. DOI: 10.1001/jama.2016.9185 34. Ferraris V.A., Brown J.R., Despotis G.J., Hammon J.W., Reece T.B., Saha S.P., Song H.K., Clough E.R., Shore-Lesserson L.J., Goodnough L.T.. 2011 Update to the Society of Thoracic Surgeons and the Society of Cardiovascular Anesthesiologists Blood Conservation Clinical Practice Guidelines. Ann. Thorac. Surg. (2011) 91 944-982. DOI: 10.1016/j.athoracsur.2010.11.078 35. Yao X., Wang J., Lu Y., Huang X., Du X., Sun F., Zhao Y., Xie F., Wang D., Liu C.. Prediction and prognosis of reintubation after surgery for Stanford type A aortic dissection. Front. Cardiovasc. Med. (2022) 9 1004005. DOI: 10.3389/fcvm.2022.1004005 36. Wang D., Huang X., Wang H., Le S., Yang H., Wang F., Du X.. Risk factors for postoperative pneumonia after cardiac surgery: A prediction model. J. Thorac. Dis. (2021) 13 2351-2362. DOI: 10.21037/jtd-20-3586 37. Wang D., Lu Y., Sun M., Huang X., Du X., Jiao Z., Sun F., Xie F.. Pneumonia after Cardiovascular Surgery: Incidence, Risk Factors and Interventions. Front. Cardiovasc. Med. (2022) 9 1727. DOI: 10.3389/fcvm.2022.911878 38. Wang D., Ding X., Su Y., Yang P., Du X., Sun M., Huang X., Yue Z., Sun F., Xie F.. Incidence, Risk Factors, and Outcomes of Severe Hypoxemia after Cardiac Surgery. Front. Cardiovasc. Med. (2022) 9 934533. DOI: 10.3389/fcvm.2022.934533 39. Wang D., Chen X., Wu J., Le S., Xie F., Li X., Wang H., Huang X., Zhang A., Du X.. Development and Validation of Nomogram Models for Postoperative Pneumonia in Adult Patients Undergoing Elective Cardiac Surgery. Front. Cardiovasc. Med. (2021) 8 750828. DOI: 10.3389/fcvm.2021.750828 40. Wang D., Abuduaini X., Huang X., Wang H., Chen X., Le S., Chen M., Du X.. Development and validation of a risk prediction model for postoperative pneumonia in adult patients undergoing Stanford type A acute aortic dissection surgery: A case control study. J. Cardiothorac. Surg. (2022) 17 22. DOI: 10.1186/s13019-022-01769-y 41. Wang D., Li Y., Sheng W., Wang H., Le S., Huang X., Du X.. Development and validation of a nomogram model for pneumonia after redo cardiac surgery. J. Cardiovasc. Med. (2022) 23 325-334. DOI: 10.2459/JCM.0000000000001302 42. Wang D., Huang X., Wang H., Le S., Du X.. Clinical risk score for postoperative pneumonia following heart valve surgery. Chin. Med. J.-Peking (2021) 134 2447-2456. DOI: 10.1097/CM9.0000000000001715 43. Luc N.F., Rohner N., Girish A., Didar S.S.U., Neal M.D., Sen G.A.. Bioinspired artificial platelets: Past, present and future. Platelets (2022) 33 35-47. DOI: 10.1080/09537104.2021.1967916 44. Verma A., Hadaya J., Tran Z., Dobaria V., Madrigal J., Xia Y., Sanaiha Y., Mendelsohn A.H., Benharash P.. Incidence and Outcomes of Laryngeal Complications Following Adult Cardiac Surgery: A National Analysis. Dysphagia (2022) 37 1142-1150. DOI: 10.1007/s00455-021-10377-2 45. Jones A., Olverson G., Hwang J., Bhagat R., McGann K., Bradburn K., Miller M., Louis C.. The effect of tracheostomy on extracorporeal membrane oxygenation outcomes. J. Card Surg. (2022) 37 2543-2551. DOI: 10.1111/jocs.16666 46. Duchnowski P.. The Role of the N-Terminal of the Prohormone Brain Natriuretic Peptide in Predicting Postoperative Multiple Organ Dysfunction Syndrome. J. Clin. Med. (2022) 11. DOI: 10.3390/jcm11237217

Subsets and Splits

No community queries yet

The top public SQL queries from the community will appear here once available.