Trimethylamine-N-Oxide Promotes High-Glucose-Induced Dysfunction and NLRP3 Inflammasome Activation in Retinal Microvascular Endothelial Cells

Introduction Along with blood glucose levels, diabetic retinopathy (DR) development also involves endogenous risk factors, such as trimethylamine-N-oxide (TMAO), a product of intestinal flora metabolic disorder, which exacerbates diabetic microvascular complications. However, the effect of TMAO on retinal cells under high-glucose conditions remains unclear. Therefore, this study examined the effects of TMAO on high-glucose-induced retinal dysfunction in the context of NLRP3 inflammasome activation, which is involved in DR. Materials and Methods TMAO was assessed in the serum and aqueous humor of patients using ELISA. Human retinal microvascular endothelial cells (HRMECs) were treated for 72 h as follows: NG (normal glucose, D-glucose 5.5 mM), NG + TMAO (5 μM), HG (high glucose, D-glucose 30 mM), and HG + TMAO (5 μM). The CCK8 assay was then used to assess cell proliferation; wound healing, cell migration, and tube formation assays were used to verify changes in cell phenotype. ZO-1 expression was determined using immunofluorescence and western blotting. Reactive oxygen species (ROS) formation was assessed using DCFH-DA. NLRP3 inflammasome complex activation was determined using a western blot. Results The serum and aqueous humor from patients with PDR contained higher levels of TMAO compared to patients with nontype 2 diabetes (Control), non-DR (NDR), and non-PDR (NPDR). TMAO showed significant acceleration of high-glucose-induced cell proliferation, wound healing, cell migration, and tube formation. ZO-1 expression decreased remarkably with the combined action of TMAO and a high glucose compared to either treatment alone. TMAO also promoted high-glucose-activated NLRP3 inflammasome complex. Conclusion The combination of TMAO and high-glucose results in increased levels of ROS and NLRP3 inflammasome complex activation in HRMECs, leading to exacerbated retinal dysfunction and barrier failure. Thus, TMAO can accelerate PDR occurrence and development, thus indicating the need for early fundus monitoring in diabetic patients with intestinal flora disorders.


Introduction
Diabetic retinopathy (DR) is a common microvascular disorder complicated by diabetes and is now the leading cause of blindness and visual impairment among the working-age population [1]. With DR progression, it can be divided into nonproliferative DR (NPDR) and proliferative DR (PDR). NPDR usually demonstrates microaneurysms, intraretinal microvascular abnormalities (IRMA), increased vessel permeability, and hard and soft exudations. Blood fow changes, loss of pericytes, broken tight junctions of endothelial cells, and pachynsis of capillaries in the endothelial basement membrane form the pathological basis of NPDR [2]. Te core process of PDR is neovascularization, with gradual ischemia and hypoxia [3]. Oxidative stress and infammatory reactions play signifcant roles in pathological processes throughout the development of diabetic retinal vasculopathy [4]. Anti-VEGF targeting has been widely used as an efective intervention in patients with PDR and diabetic macular edema. Studies on the Joslin 50-year Medalist cohort, comprising patients sufering from insulindependent diabetes for 50 years or more, have indicated no signifcant correlation between the severity of diabetic retinopathy and the level of blood glucose control [5]. Tis suggests that endogenous risk factors and protective factors other than glucose levels may be involved in the occurrence and development of DR.
Trimethylamine N-oxide (TMAO) is an amine oxide that can be induced by gut dysbiosis [6]. Choline, betaine, and carnitine, which originate from animal dietary components, are converted into trimethylamine (TMA) under the infuence of the gut microbiota [7]. TMA is transported into the liver through the portal circulation, where it is oxidized to TMAO by favin-containing monooxygenase 3 (FMO3) [8]. Several studies have shown that most cardiovascular diseases (CVD) and renal diseases are closely related to TMAO [9]. A recent review indicated that gut dysbiosis and the development of type 2 diabetes (T2DM) as well as the related retinal, neurological, and renal microvascular complications went hand in hand [10]. TMAO might play a key role in diabetic cardiomyopathy (DCM) through the pathways of infammation, connexin remodeling, the increase of calcium ions (Ca 2+ ), myocardial fbrosis, and so on [11,12]. In addition, TMAO could activate renal infammation, oxidative stress, fbrosis, and endothelial dysfunction in the pathogenesis of diabetic kidney disease (DKD) [13,14]. A clinical study demonstrated that elevated TMAO concentrations in plasma were associated with increased incidence and severity of DR in T2DM [15]. However, the specifc mechanism of TMAO as a risk factor for DR and its efect on retinal cells under the action of high glucose remain unclear.
Te NLRP3 infammasome has been found to be upregulated in the retinal proliferative membranes of PDR patients as well as in in vitro and in vivo DR models [16]. Once infammatory stimuli are sensed, activated NLRP3 binds the adaptor protein ASC, which contains a pyrin domain (PYD) and caspase recruitment domain (CARD), to recruit and cleave caspase-1. Tis triggers the downstream reaction, resulting in interleukin-1β (IL-1β) release [17]. Considering the role of TMAO as a risk element in DR, it is important to examine whether TMAO can enhance NLRP3 infammasome activation in DR. In hyperglycemia, dysfunction and activation of the NLRP3 infammasome are found in retinal microvascular endothelial cells (RMECs) [18].
Terefore, in this study, we examined the mechanism by which TMAO is involved in the dysfunction and activation of NLRP3 infammasomes in RMECs under a high-glucose environment.

Serum and Aqueous Humor Sampling and Quantitative
TMAO Analyses. Participants with or without type 2 diabetes who had cataract extraction and participants with PDR and type 2 diabetes who underwent vitrectomy at the Second Afliated Hospital of Nantong University were recruited. Tis study followed the guidelines of the Declaration of Helsinki. Sampling was carried out with the informed consent of patients and approval from the hospital ethics committee. Te samples were stored at −80°C. Te concentrations of TMAO in serum and aqueous humor samples were determined using a human TMAO ELISA kit (JINGME, JiangSu, China), according to the manufacturer's protocol; the results were determined at 450 nm using a microplate reader (BIOTEK, USA). Te TMAO concentrations were calculated based on standard concentrations and expressed as ppm (mg/L).

ROS Assessment.
HRMECs were seeded into a 6-well plate and treated as needed. Cells were then loaded with DCFH-DA (10 μM, Beyotime Biotech, Shanghai, China) in serum-free medium and incubated for 20 minutes at 37°C. After washing 3 times with phosphate bufered saline, the images were captured under an inverted fuorescence microscope (ECLIPSE Ti2, NIKON, Japan). Fluorescent intensity was analyzed using Image J software.

Cell Migration and Healing
Assays. For migration assays, migration chambers were placed in a 24-well plate and pretreated HRMECs were seeded into the upper chamber. Ten, 4% paraformaldehyde was used to fx the cells, and they were incubated for 12 h. Cells were then stained with crystal violet and counted under a microscope as follows: For wound healing assays, HRMECs were incubated on a 6-well plate until confuence and then scratched with sterile 200 μL pipette tips. Te wells were then viewed under an inverted microscope (TS2, NIKON, Japan) and the wound healing percentage was calculated after treatment for 24 h, 48 h, and 72 h. Te scratch areas were calculated by Image J software, and then the wound healing areas were obtained by calculating the D-value between the scratch areas of 24, 48, and 72 hours and the initial scratch area of this group. Te wound healing percentages were the ratios of the wound healing areas to the initial scratch area of the group.
2.6. Tube Formation Assay. Matrigel (356234, Corning, NY, USA) was used to coat a 96-well plate (50 μL/L), which was subsequently incubated at 37°C for 45 minutes to solidify. Pretreated cell suspensions were then seeded on the solidifed Matrigel (20,000 cells/well). After incubation at 37°C for 6 h, the tube formation of HRMECs was observed using an inverted microscope. Te tube length was analyzed using Image J software.

Immunofuorescence Analysis for ZO-1.
After treatment, HRMECs were fxed using 4% paraformaldehyde for 15 minutes and then blocked with 1% BSA in PBS for 2 h at 37°C. Te cells were then incubated overnight at 4°C with anti-ZO-1 (1 : 400; Cell Signaling), followed by incubation with red-labeled antirabbit secondary antibody(1 : 500, Invitrogen, CA, USA) for 2 h at room temperature. Te cells were sealed with antifade mounting medium containing DAPI (Beyotime Biotech, Shanghai, China) and scanned under a fuorescence microscope (ECLIPSE NI-E, NIKON, Japan). Fluorescent intensity was analyzed using Image J software.

Statistical
Analysis. Statistical analysis was performed using GraphPad Prism version 8.0 and SPSS 23. Te data are presented as mean ± standard deviation (SD). A one-way ANOVA followed by Tukey's post hoc test was used to examine the signifcant diferences between and within diferent groups. Te Kruskal−Wallis test was used for nonnormally distributed data. To compare the diferences in clinical characteristics between the four groups, we used the Wilcoxon and McNemar tests as appropriate. Te level of statistical signifcance was set at P < 0.05.

TMAO Showed Higher Levels in Proliferative Diabetic
Retinopathy (PDR). We assessed the concentration of TMAO in the samples from four groups of participating patients: control group (CT): cataract without diabetes; non-DR group (NDR): cataract with diabetes but not diabetic retinopathy; non-PDR group (NPDR): cataract with nonproliferative diabetic retinopathy; and PDR group (PDR): patients with proliferative diabetic retinopathy who underwent vitrectomy. Each group contained 5 patients, and the clinical characteristics of the participants were shown in Table 1. No signifcant diferences in age, gender, or diabetes duration among the four groups are displayed in Table 1. Compared with the other three groups, the PDR group showed signifcantly higher expression of TMAO in both serum (Figure 1(a)) and aqueous humor samples (Figure 1(b)). However, there was no statistically signifcant diference among the other three groups. Tis suggests that TMAO may accelerate the progression of diabetic retinopathy and aggravate its condition.

TMAO Promotes Cell Proliferation Induced by High Glucose in Human Retinal Microvascular Endothelial Cells.
To investigate the efect of TMAO in HRMECs treated with high glucose (HG, D-glucose 30 mM), the cells were treated with NG (normal glucose, D-glucose 5.5 mM) + TMAO (0 μM) or HG + TMAO (0, 1.25, 2.5, 5, 10, 20, 50, or 100 μM) for 72 hours. Te CCK8 assay was then used to determine the cell proliferation based on the OD 450 value. As shown in Figure 2(a), TMAO at 2.5, 5, 10, and 20 μM could promote the cell proliferation induced by HG, and only HG + TMAO (5 or 10 μM) showed statistically signifcant diferences compared with NG + TMAO (0 μM) at 72 h; cell proliferation was also better increased with TMAO 5 μM than 10 μM. Terefore, 5 μM TMAO was used in the subsequent experiments. We continued to compare the cell proliferation in the four diferent groups; as shown in Figure 2(b), the cell proliferation in NG + TMAO (5 μM), HG, and HG + TMAO (5 μM) was signifcantly increased compared with that in NG at 72 h. Moreover, the combination of TMAO and HG could accelerate cell proliferation compared with that induced by HG or TMAO alone. Overall, high-glucose-induced proliferation of HRMECs can be increased by TMAO.

TMAO Enhances High-Glucose-Induced Wound Healing, Migration, and Vascular Tube Formation of Human Retinal
Microvascular Endothelial Cells. In our previous research, we detected that osmotic pressure had no efect on the migration and tube formation of HRMECs [19]. Tus, we did not conduct osmotic controls in this experiment. Both TMAO and high glucose can individually promote the wound healing of HRMECs, but when combined, the healing process could be accelerated further (Figures 3(a)  and 3(b)). To assess cell migration, we used the Transwell chamber and found that the number of migrating cells was signifcantly increased when HRMECs were cultured with HG and TMAO together, compared with the individual HG and TMAO treatments (Figures 3(c) and 3(d)). Tube formation was also more striking after costimulation with TMAO and high glucose (Figures 3(e) and 3(f )). Tese results indicate that overexpression of TMAO accelerates neovascularization under high glucose conditions.

TMAO Enhances High-Glucose-Induced Degradation of the Tight Junction Protein ZO-1 in HRMECs.
We then examined ZO-1 expression by HRMECs after diferent treatments for 72 h. As results illustrated, OSM (mannitol 24.5 mM and D-glucose 5.5 mM) exposure produced no efect on the fuorescence and protein expression of ZO-1 compared to NG (D-glucose 5.5 mM) and HG (D-glucose 30 mM) (Figures 4(a)-4(d)). Red fuorescence intensity of ZO-1 decreased signifcantly upon combined stimulation with TMAO and high glucose compared to that with either treatment alone (Figures 4(e) and 4(g)). Western blot analysis further confrmed a more serious decrease in ZO-1 protein levels with the joint action of TMAO and high glucose (Figures 4(f ) and 4(h)). ZO-1 protein is a component of the intercellular tight junctions. When tight junctions are disintegrated, ZO-1 expression decreases. Our results thus indicate that TMAO may accelerate the destruction of tight junctions induced by high glucose in HRMECs. Tis may suggest that endothelial integrity could be decreased and high glucose-induced vascular leakage could be aggravated by TMAO.

TMAO Can Increase High Glucose-Induced ROS.
Te intracellular ROS levels after 72 h of diferent stimulations were found to be signifcantly upregulated by TMAO and high glucose individually, but were further increased by the combination of TMAO and high glucose (Figures 5(a) and  5(b)). Previous studies have demonstrated that upregulation of oxidative stress can lead to infammation in the DR. Tese results indicate that TMAO can enhance high glucoseinduced ROS, suggesting that the infammation induced by high glucose can also be exacerbated by exposure to TMAO.

TMAO Can Enhance High-Glucose-Induced NLRP3
Infammasome Signaling Activation. High glucose can induce infammasome activation, which plays an important role in the occurrence and development of DR [20]. In this study, we investigated whether TMAO could enhance the NLRP3 infammasome activation induced by high-glucose  ( Figure 5(c)). Te protein expression levels of NLRP3 ( Figure 5(d)), ASC ( Figure 5(e)), Caspase-1 ( Figure 5(f)), cleaved-Caspase-1 ( Figure 5(g)), IL-1 ( Figure 5(h)), and Pro-IL-1β ( Figure 5(i)) were found to be increased by the combined treatment with TMAO and high glucose compared to that induced by either TMAO or high glucose alone. Tus, TMAO may accelerate the occurrence of DR by increasing the infammasome activation induced by high glucose.

Discussion
We found that the TMAO content in serum and aqueous humor were both much higher in the PDR group than in the CT, NDR, and NPDR groups. Our study aimed to confrm whether TMAO can accelerate high-glucose-induced dysfunction and NLRP3 infammasome activation in HRMECs to accelerate the pathogenesis of DR. We found that TMAO can enhance the proliferation, wound healing, migration, tube formation, and degradation of intercellular tight junctions induced by high glucose in HRMECs. Tus, TMAO could accelerate high glucose-caused neovascularization and vascular leakage. Interestingly, we found that a 5 μM concentration of TMAO combined with highglucose guided greater cell proliferation than 10 μM as well as higher concentrations. On the one hand, it was reported that TMAO could activate oxidative stress to induce pyroptosis in the vascular endothelial cells [21] or apoptosis in renal cells [22]. On the other hand, apoptosis in the pancreatic acinar cells has been investigated and could occur after treatment with TMAO via ER stress [23]. So, with the increase in TMAO concentration, we guess that pyroptosis or apoptosis may occur in the HRMECs. In that case, 5 μM concentration of TMAO could induce greater cell proliferation than higher concentration of TMAO in HRMECs. We demonstrated that TMAO can enhance the high-glucose-induced occurrence of oxidative stress and NLRP3 infammasome complex activation. Pathological neovascularization is known to play a key role in the development of DR [24]. Oxidative stress and NLRP3 infammasome complex activation, including the subsequent chronic infammation, are also crucial pathogenic processes in DR [25,26]. Gut dysbiosis has been indicated to be associated with the progression of diabetic microvascular complications, including DR, and TMAO overexpression can be induced by gut dysbiosis [10]. Based on these fndings, TMAO overexpression in the retina can accelerate the occurrence and development of DR. Te plasma levels of TMAO are reported to be associated with the incidence rate and severity of DR [15]. Our experimental results were consistent with these clinical fndings. However, we did not further investigate additional signaling pathways for the combined action of TMAO and high glucose in HRMECs; therefore, we attempted to conjecture the related mechanisms based on previous studies on DR and another diabetic vascular complication associated with TMAO. Te PKC (protein kinase C) pathway plays a key role in the oxidative stress of DR [27]. Hyperglycemia increases the synthesis of diacylglycerol (DAG) through glycolysis, which is an agonist of PKC [28]. PKC can enhance active NADPH oxidase and regulate the assembly and activation of NOX 2 and NOX 4 isoforms, which can further increase the level of oxidative stress in retinal cells [29]. PKC activity can also  increase NF-κB phosphorylation, which can increase infammation, activation of the NLRP3 infammasome, and loss of ZO-1 and Claudin-1, thus damaging the blood -retinal barrier [30][31][32]. TMAO has also been found to activate PKC in human umbilical vein endothelial cells (HUVECs) to induce monocyte adhesion [33]. TMAO may thus enhance high-glucose-induced oxidative stress through the PKC pathway.
Te NLRP3 infammasome can also be activated by several other pathways, including toll-like-receptor-4 (TLR4) [34], p38 [35], Nrf2 [36,37], ROS [38], PKR [39], and prostaglandin E [40,41], in retinal cells. In cardiovascular disease, TMAO has been indicated to activate the NLRP3 infammasome through ROS induction [41]. Terefore, we consider that TMAO can promote ROS to further accelerate high-glucose-induced activation of the NLRP3 infammasome.  Based on our fndings, TMAO can accelerate the highglucose-induced destruction of tight junctions; however, TMAO itself can activate the NLRP3 infammasome and induce dysfunction in HRMECs. It is widely reported that TMAO can increase the risk of cardiovascular disease [42,43]. Disorders of the gut microbial ecosystem are reported to increase the risk of cardiovascular disease (CVD), atherosclerosis, diabetes, and stroke [44,45]. As products of intestinal dysbacteriosis, TMAO may thus be a risk factor for ocular fundus diseases; thus, regular examination of the ocular fundus is necessary, especially in patients with diabetes.

Conclusion
In conclusion, our study confrmed higher levels of TMAO in patients with PDR. We further demonstrated that TMAO can promote NLRP3 infammasome activation and HRMEC dysfunction induced by high glucose. TMAO overexpression may thus accelerate the course of DR and loss of vision. Terefore, fundus monitoring needs to be conducted as early as possible in diabetic patients with intestinal fora disorders. But this study also has some limitations, such as the fact that no animal experiments were conducted to further verify this study and the number of clinical participants was small. In the future, we will recruit more participants in our research and conduct more experiments to explore the mechanism of TMAO with DR.

Data Availability
All data generated or analyzed during this study are included in this article. Further enquiries can be directed to the corresponding author.

Ethical Approval
All experiments in this study were approved by the Ethics Committee of the Second Afliated Hospital of Nantong University (2021KYG045). Tis study followed the guidelines of the Declaration of Helsinki. Sampling was carried out after written informed consents were signed by patients.

Conflicts of Interest
Te authors declare that they have no conficts of interest.

Authors' Contributions
Conception and study design were done by LDX; data acquisition was performed by XC; data analysis was done by YJT; manuscript drafting was done by LLH; manuscript revising was done by YS. All authors have read and approved the fnal version of this manuscript to be published. Lidan Xue and Lili Huang contributed equally to this work.