Regulation of AMPK and GAPDH by Transglutaminase 2 Plays a Pivotal Role in Microvascular Leakage in Diabetic Retinas Free

Diabetic retinopathy (DR) is the most common microvascular complication caused by chronic hyperglycemia and is the leading cause of blindness (1). DR is clinically divided into two pathogenic stages: early nonproliferative DR and late proliferative DR (2). Nonproliferative DR is associated with microvascular leakage and capillary occlusion as well as microaneurysms, retinal pericyte loss, and basement membrane thickening (2,3). Proliferative DR is characterized by neovascularization, leading to severe vision loss via vitreous hemorrhage and retinal detachment (4). Vascular leakage, the clinical hallmark of DR, is caused by vascular endothelial growth factor (VEGF)–mediated disassembly of adherens junctions and blood-retinal barrier (BRB) breakdown, resulting in macular edema and subsequent retinal detachment (5,6). Therefore, to treat DR, it is essential to prevent retinal vascular leakage in the early stages.

Transglutaminase 2 (TGase2) is a member of the transglutaminase family that catalyzes protein cross-linking reactions through the transamidation of glutamine residues to lysine residues in a Ca²⁺-dependent manner (7). TGase2 is expressed ubiquitously and is a multifunctional enzyme that functions as transamidase, serine/threonine kinase, protein disulfide isomerase, and GTPase (7,8). TGase2 is implicated in various diseases, including neurodegenerative disorders, cardiovascular diseases, autoimmune diseases, inflammation, and fibrosis (9,10). TGase2 has emerged as a key enzyme in the pathogenesis of diabetic vascular complications (11–13). In the diabetic retina, microvascular leakage is predominantly caused by hyperglycemia-induced overexpression of VEGF (5,11), and TGase2 plays an important role in VEGF-induced microvascular leakage through stress fiber formation and disassembly of adherens junctions in the retinas of diabetic mice (11,14,15). However, the molecular mechanism by which TGase2 contributes to hyperglycemia-induced microvascular leakage in the diabetic retina remains unclear.

AMPK is a serine-threonine kinase that acts as an energy sensor and maintains cellular energy homeostasis (16). AMPK is an emerging target for preventing complications of diabetes, as exhibited by the most common antihyperglycemic drugs, thiazolidinediones and metformin (17,18). AMPK is involved in enhanced glucose uptake and metabolism by cells, potentially through activation of sirtuin 1 and peroxisome proliferator‐activated receptor γ coactivator 1 α (17). Hyperglycemia-induced inhibition of AMPK is associated with the pathogenesis of diabetic microvascular complications, including neuropathy, nephropathy, and retinopathy (17,19). In the diabetic retina, AMPK activation is beneficial in protecting against hyperglycemia-induced retinal dysfunctions, including BRB breakdown, vascular leakage, inflammation, and photoreceptor cell degeneration (19–21). Therefore, it is important in the treatment of DR to understand the mechanism of hyperglycemia-driven inhibition of AMPK and its downstream events in the diabetic retina; however, its molecular mechanism is not fully understood.

In the current study, we aimed to elucidate the molecular mechanism by which TGase2 plays an essential role in hyperglycemia-induced microvascular permeability in the diabetic retina. We hypothesized that regulation of AMPK and GAPDH by TGase2 would play a pivotal role in hyperglycemia-induced microvascular leakage in DR. GAPDH is a key enzyme in glycolysis, converting glyceraldehyde-3-phosphate to 1,3-biphosphoglycerate (22). We found that TGase2, which is activated through sequential elevation of intracellular Ca²⁺ and reactive oxygen species (ROS) levels, inhibits APMK and GAPDH activities in the retinas of diabetic mice, leading to vascular endothelial (VE)–cadherin disassembly and microvascular leakage. Our findings suggest that regulation of AMPK and GAPDH by TGase2 is essential for diabetes-induced microvascular permeability in the retina.

Human retinal endothelial cells (HRECs) were purchased from the Applied Cell Biology Research Institute (Cell Systems, Kirkland, WA) and maintained as previously described (23). Cells in passages 7 to 9 were used for experiments and were grown on 2% gelatin-coated plates in M199 medium (Gibco, Grand Island, NY) supplemented with 20% FBS (Hyclone, South Logan, UT), 3 ng/mL basic fibroblast growth factor (MilliporeSigma, Burlington, MA), 100 units/mL penicillin (Gibco), and 100 mg/mL streptomycin (Gibco) in a humidified 5% CO₂ incubator. The cells were authenticated by short tandem repeat profiling. For the studies, cells were incubated for 12 h in low-serum medium supplemented with 2% FBS, 0.1 ng/mL basic fibroblast growth factor, and antibiotics and then treated for 3 days with 5.5 mmol/L d-glucose (normal glucose) or 30 mmol/L d-glucose (high glucose).

GAPDH activity was quantitatively analyzed based on the irreversible reduction of nonfluorescent resazurin to fluorescent resorufin. Aliquots (1 μL) of reaction mixtures containing 2 mmol/L glyceraldehyde-3-phosphate, 1 mmol/L NAD⁺, 0.5 units/mL diaphorase, 5 μmol/L resazurin, 0.2 mmol/L EDTA, 5 mmol/L sodium arsenate, 0.01% Tween-20, and various concentrations of GAPDH (ranging from 0.1 to 5 μg/mL) or cell lysate samples in 50 mmol/L Tris-HCl buffer (pH 7.5) were applied for 60 min to well-type arrays, which were fabricated by mounting polydimethylsiloxane gaskets onto clean glass slides as previously described (24). Droplet arrays were then analyzed using a fluorescence scanner (InnoScan 300; Innopsys, Carbonne, France). GAPDH activity (n = 4) was calculated using a standard curve, which was plotted as the logistic regression fit using OriginPro 2015 software (Origin Lab, Northampton, MA).

Six-week-old male C57BL/6 mice were obtained from DBL (EumSeong, Korea), and TGase2-null (Tgm2^−/−) mice (C57BL/6), generated by disrupting exons 5 and 6 of the TGase2 gene (25), were provided by Dr Soo-Youl Kim (National Cancer Center, Goyang, Republic of Korea). Mice were maintained in filtered-top cages under pathogen-free conditions in a temperature-controlled room (22°C) with a 12-h light/dark cycle. Diabetic mice were generated with a single intraperitoneal injection of freshly prepared streptozotocin (150 mg/kg body weight; MilliporeSigma) in 100 mmol/L citrate buffer (pH 4.5) as previously described (11). Mice with fasting blood glucose concentrations ≥19 mmol/L, polyuria, and glucosuria were considered diabetic. Blood glucose levels and body weights of mice were monitored weekly (n = 6). Four weeks after streptozotocin injection, nondiabetic or diabetic mice were anesthetized with 3% isoflurane and intravitreally injected with 2 μL of 50 mmol/L cystamine, 20 mmol/L monodansylcadaverine (MDC), 2 mmol/L Trolox, 50 mmol/L 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR), or 1 mmol/L compound C. An equal volume of PBS was injected into the contralateral eye. All animal experiments conformed to the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, Bethesda, MD) and were approved by the Institutional Animal Care and Use Ethics Committee of Kangwon National University (approval nos. KW-210709-2 and KW-230227-2).

ROS generation in mouse retinal sections was measured using CellROX green reagent as previously described (26). Briefly, tissues were dissected and embedded in Optimal Cutting Temperature compound (Sakura Finetek, Torrance, CA). Then, retinal cryosections (10 μm) were prepared using a microtome cryostat (Leica Biosystems, Wetzlar, Germany). Unfixed retinal sections were incubated with inhibitors and 5 μmol/L CellROX green reagent at 37°C for 30 min. ROS generation in stained retinas was visualized by confocal microscopy (K1-fluo) and quantitatively analyzed by measuring the fluorescence intensities of three sections per eye (n = 6).

In vivo TGase transamidation activity was determined in retinal sections as previously described (12). Briefly, unfixed retinal cryosections (10 μm) were incubated for 1 h at 37°C with 1 mmol/L BAPA and inhibitors and then fixed with 3.7% formaldehyde for 20 min. The retinal sections were permeabilized with 0.2% Triton X-100 and incubated with FITC-conjugated streptavidin (1:200; v/v) for 1 h and with 1 μg/mL DAPI for 10 min. Stained samples were observed by confocal microscopy (K1-fluo), and TGase activity was quantified by measuring the fluorescence intensities of three sections per eye (n = 6).

VE-cadherin was visualized in whole-mounted retinas as previously described (27). Briefly, mice were deeply anesthetized using 3% isoflurane, and enucleated eyes were immediately fixed with 4% paraformaldehyde for 45 min at room temperature and acetone for 3 min at −20°C. The fixed retinas were permeabilized for 4 h with 1.0% Triton X-100 in PBS, incubated with a monoclonal VE-cadherin antibody (1:200; BD Pharmingen, San Diego, CA), and probed with Alexa 647–conjugated goat anti-rat immunoglobulin G (1:300; Invitrogen). VE-cadherin in the superficial and deep plexus vessels was visualized by confocal microscopy (n = 4).

Vascular leakage in mouse whole-mounted retinas was visualized using fluorescein angiography as previously described (12). Briefly, mice were deeply anesthetized, and 1.25 mg 500-kDa FITC-dextran (MilliporeSigma) was injected into the left ventricle and circulated for 5 min. The enucleated eyes were immediately fixed using 4% paraformaldehyde for 2 h and then dissected in the Maltese cross configuration and flat mounted onto glass slides. Vascular permeability in the superficial layer of the retina was visualized by confocal microscopy (K1-fluo) and quantitatively analyzed by measuring the fluorescence intensities of extravasated FITC-dextran in three fields per eye (n = 6).

Data were analyzed using OriginPro 2015 software. Data are expressed as the mean ± SD of four or six independent experiments. Statistical significance was determined using one-way ANOVA with Holm-Sidak’s multiple comparisons test. P values <0.05 were considered statistically significant.

The data sets generated during the current study are available from the corresponding author on reasonable request. Data on the diabetic mice generated during the current study are available from the corresponding author on reasonable request.

For a detailed description of the Research Design and Methods used, please see the Supplementary Material.

To investigate the roles of ROS generation and TGase activity in hyperglycemia-induced vascular leakage in diabetic mouse retinas, we injected the ROS scavenger Trolox or the TGase inhibitor cystamine into the vitreous chamber, and then, retinal vascular leakage was studied using fluorescence angiography (Fig. 1A and D). Diabetic mice showed elevated food and water consumption, loss of body weight, and hyperglycemia compared with nondiabetic mice (Fig. 1B and C). Hyperglycemia-induced vascular leakage in the retinas of diabetic mice and this retinal microvascular leakage were suppressed by Trolox or cystamine (Fig. 1D), suggesting ROS generation and TGase activation play roles in hyperglycemia-induced vascular leakage in diabetic retinas.

To understand the underlying mechanism of hyperglycemia-induced TGase activation, we explored the roles of intracellular Ca²⁺ and ROS in high glucose–induced TGase activation in HRECs. High-glucose conditions elevated the intracellular Ca²⁺ level, which was inhibited by the Ca²⁺ chelator BAPTA-AM (Fig. 1E). High-glucose conditions elevated intracellular ROS levels, and this ROS generation was inhibited by BAPTA-AM or the ROS scavengers NAC and Trolox, but not by the TGase inhibitors MDC and cystamine, indicating the role of intracellular Ca²⁺, but not TGase, in high glucose–induced ROS generation (Fig. 1F). High-glucose conditions also increased in situ TGase activity, which was inhibited by BAPTA-AM, the ROS scavengers, or the TGase inhibitors (Fig. 1G). These results demonstrate that high-glucose conditions activate TGase through sequential elevation of intracellular Ca²⁺ and ROS levels in HRECs.

We next explored the contribution of TGase2 to high glucose–induced TGase activation by transfecting HRECs with human TGase2–specific siRNA. TGase2 siRNA, which suppressed protein expression of TGase2, inhibited high glucose–induced TGase activation, but control siRNA had no effect (Fig. 1H and I), suggesting that TGase2 mostly contributed to the high glucose–induced elevation of TGase activity. We then studied the roles of ROS generation and TGase2 in high glucose–induced endothelial permeability in HRECs. Consistent with the in vivo study (Fig. 1D), high-glucose conditions induced endothelial permeability, which was inhibited by Trolox or cystamine (Fig. 1J). High glucose–induced endothelial permeability was also suppressed by TGase2 siRNA, but not by control siRNA. These results suggest that TGase2 plays a key role in high glucose–induced vascular permeability in HRECs.

To verify our in vitro findings, we examined the role of TGase2 in adherens junction disassembly and microvascular leakage in the retinas of TGase2-null (Tgm2^−/−) mice. Diabetic Tgm2^−/− mice showed loss of body weight and hyperglycemia compared with nondiabetic Tgm2^−/− mice (Fig. 2A and B). TGase activity was barely detectable in the retinal sections of nondiabetic Tgm2^−/− mice, unlike in nondiabetic wild-type (C57BL/6) mice. This dramatically decreased TGase activity was not altered by hyperglycemia in diabetic Tgm2^−/− mice (Fig. 2C), demonstrating that TGase2 accounted for most of the TGase activity in the diabetic retina. Hyperglycemia induced VE-cadherin disassembly in the retinas of diabetic C57BL/6 mice, but this retinal VE-cadherin disassembly was not observed in diabetic Tgm2^−/− mice (Fig. 2D and E). Microvascular leakage was also undetectable in the retinas of diabetic Tgm2^−/− mice (Fig. 2F and G). Taken together, these results demonstrate that TGase2, activated by sequential elevation of intracellular Ca²⁺ and ROS levels, plays an essential role in hyperglycemia-induced microvascular leakage in the retinas of diabetic mice.

To investigate the possible role of AMPK in TGase2-mediated vascular leakage in the diabetic mouse retina, we initially determined whether hyperglycemia could inhibit AMPK in diabetic retinas, which was demonstrated by AMPK phosphorylation. Hyperglycemia decreased AMPK phosphorylation, which was determined by Western blotting, in diabetic compared with nondiabetic retinas (Fig. 3A and B). This hyperglycemia-induced dephosphorylation of AMPK was reversed by intravitreal injection of Trolox, cystamine, or the AMPK activator AICAR, indicating the roles of ROS generation and TGase activation in hyperglycemia-induced inhibition of AMPK in diabetic retinas. Hyperglycemia induced ROS generation in the retinas of diabetic mice, which was suppressed by Trolox, but not by cystamine or AICAR (Fig. 3C and D). Hyperglycemia also elevated TGase activity in diabetic retinas, and this TGase activation was inhibited by Trolox or cystamine, but not by AICAR (Fig. 3E and F). These results suggest that ROS-mediated activation of TGase2 is involved in hyperglycemia-induced inhibition of AMPK in the retinas of diabetic mice.

To support our in vivo findings, we investigated the role of TGase2 in high glucose–induced AMPK dephosphorylation in HRECs. AMPK phosphorylation, which was determined by Western blotting and immunofluorescence, was decreased by high-glucose conditions, and this AMPK dephosphorylation was restored by BAPTA-AM, the ROS scavengers, or the TGase inhibitors (Fig. 4A–C). However, AMPK expression was not altered by high-glucose conditions or the inhibitors. We next transfected HRECs with TGase2 siRNA to study the contribution of TGase2 to high glucose–induced dephosphorylation of AMPK. TGase2 siRNA restored the reduced AMPK phosphorylation induced by high-glucose conditions, but control siRNA did not (Fig. 4D and E), suggesting the role of TGase2 in high glucose–induced AMPK dephosphorylation. Taken together, these results suggest that ROS generation and TGase2 activation are involved in hyperglycemia-induced inhibition of AMPK in the retinas of diabetic mice.

We investigated the possible role of AMPK inhibition by TGase2 in hyperglycemia-induced retinal vascular leakage by analyzing VE-cadherin disruption and endothelial permeability in HRECs. High-glucose conditions induced VE-cadherin disassembly, which was recovered by treatment with cystamine or AICAR (Fig. 5B and C). High glucose–induced in vitro cell permeability was also inhibited by cystamine or AICAR (Fig. 5D), indicating the roles of TGase2 activation and AMPK inhibition in high glucose–induced endothelial permeability. The role of AMPK in high glucose–induced endothelial permeability was further studied using human AMPK–specific siRNA in HRECs. AMPK siRNA suppressed AMPK protein expression compared with control siRNA treatment (Fig. 5A) and induced VE-cadherin disassembly and endothelial permeability in endothelial cells under normal glucose conditions (Fig. 5B–D). These results indicate that AMPK inhibition through TGase2 is involved in high glucose–induced VE-cadherin disruption and endothelial permeability in HRECs.

To support our in vitro findings, we explored the role of AMPK in hyperglycemia-induced vascular leakage by analyzing VE-cadherin disassembly and FITC-dextran extravasation in the retinas of diabetic mice. Disruption of VE-cadherin was observed in both the superficial and deep plexus layers in the retinas of diabetic mice, in contrast with those of nondiabetic mice, and was restored by intravitreal injection of cystamine, MDC, or AICAR (Fig. 5E). Intravitreal injection of the AMPK inhibitor compound C also induced VE-cadherin disassembly in the retinas of nondiabetic mice. Increased FITC-dextran extravasation was observed in diabetic mouse retinas, and this retinal microvascular leakage was prevented by intravitreal administration of cystamine, MDC, or AICAR (Fig. 5F and G), demonstrating the roles of TGase2 activation and AMPK inhibition in hyperglycemia-mediated microvascular permeability in the diabetic retina. The role of AMPK inhibition in hyperglycemia-induced microvascular permeability was further demonstrated by intravitreal injection of compound C in nondiabetic mice (Fig. 5F and G). Therefore, our results suggest that AMPK, which is inhibited by TGase2, plays an important role in hyperglycemia-induced microvascular permeability in the diabetic retina.

Because GAPDH was reported to be associated with DR (22), we examined the possible role of GAPDH in TGase2-mediated microvascular leakage in HRECs. For this study, we designed an on-chip GAPDH activity assay using liquid droplet arrays based on irreversible reduction of resazurin to resorufin as illustrated in Fig. 6A. NADH generated by GAPDH reduces nonfluorescent resazurin to fluorescent resorufin, and the resulting fluorescence intensity represents GAPDH activity. For quantitative analysis of GAPDH activity, we created a standard curve, in which the fluorescence intensity of resorufin was increased by GAPDH in a concentration-dependent manner (Fig. 6B). In HRECs, GAPDH activity was decreased by high-glucose conditions, and this activity reduction was reversed by MDC, cystamine, or TGM2 siRNA, but not by control siRNA, suggesting the role of TGase2 in high glucose–induced inhibition of GAPDH (Fig. 6C). We next studied the role of GAPDH in high glucose–induced endothelial permeability by transfecting HRECs with human GAPDH–specific siRNA. GAPDH siRNA, which suppressed GAPDH protein expression (Fig. 6D), induced endothelial permeability, which was assessed by VE-cadherin disruption and endothelial cell permeability assays in HRECs under normal glucose conditions (Fig. 6E–G). Our findings suggest that GAPDH, which is inhibited by TGase2, is involved in endothelial permeability in HRECs.

We next examined the possible role of AMPK in high glucose–induced inhibition of GAPDH in HRECs. The high glucose–induced decrease in GAPDH activity was recovered by metformin or AICAR, and this recovery of GAPDH activity by the AMPK activators was reversed by the GAPDH inhibitor heptelidic acid (Fig. 7A). Treatment of HRECs with compound C produced a reduction in GAPDH activity. The role of AMPK as an upstream regulator of GAPDH was further studied using human AMPK siRNA. Transfection of HRECs with AMPK siRNA significantly suppressed GAPDH activity but did not affect GAPDH protein expression (Fig. 7B and C), demonstrating that high glucose–induced endothelial permeability is mediated by sequential inhibition of AMPK and GAPDH in HRECs. Taken together, our results suggest that regulation of AMPK and GAPDH by TGase2 plays a pivotal role in hyperglycemia-induced microvascular leakage in the retinas of diabetic mice (Fig. 7D).

DR is the most common microvascular complication of diabetes and the leading cause of blindness in working-age populations (1). Nevertheless, effective therapy for DR has not been achieved, because the mechanism involved in the development of DR is complex and not yet fully understood. In this study, we elucidated a novel molecular mechanism of DR by which TGase2 plays an essential role in hyperglycemia-induced microvascular leakage in the diabetic retina. TGase2, which is activated by sequential elevation of intracellular Ca²⁺ and ROS levels, is important for hyperglycemia-induced microvascular leakage in the diabetic retina. ROS generation–mediated activation of TGase2 inhibited AMPK and GAPDH, which were involved in microvascular leakage in the diabetic retina (Fig. 7D). Our results provide three key findings regarding the pathogenesis of DR: 1) TGase2 acts as a negative regulator of AMPK and GAPDH, 2) AMPK regulates GAPDH activity, and 3) GAPDH is involved in hyperglycemia-induced vascular permeability. Taken together, our results suggest that regulation of AMPK and GAPDH by TGase2 plays a key role in diabetes-induced microvascular permeability, leading to BRB disruption in the retina.

TGase2 has multiple enzymatic functions as transamidase, kinase, isomerase, and GTPase as well as nonenzymatic functions via interactions with extracellular proteins (7,28). TGase2 is involved in several pathophysiologic processes, such as apoptosis, inflammation, epithelial-mesenchymal transition, fibrogenic reactions, and mitochondrial dysfunction (7,8). Because of these multifunctional activities, TGase2 is involved in the pathogenesis of various diseases, including celiac disease, neurodegenerative diseases, cancers, inflammatory diseases, cardiovascular disease, and fibrosis (9,10,29,30). Numerous reports have suggested TGase2 is a key enzyme in the pathogenesis of diabetic microvascular and macrovascular complications, including diabetic aortic dysfunction, retinal vascular leakage, pulmonary diseases, and glomerular endothelial dysfunction (11–13,31–33). However, the underlying mechanism of TGase2 in these complications of diabetes is unknown. In the current study, we demonstrated that TGase2 plays a pivotal role in hyperglycemia-induced microvascular leakage through inhibition of AMPK and GAPDH in the retinas of diabetic mice. Considering the contribution of vascular dysfunction to diabetic cardiovascular disease, pulmonopathy, and nephropathy, the TGase2/AMPK/GAPDH pathway may function as an important molecular axis in the pathogenesis of these diabetic vascular complications.

Interestingly, inhibition of AMPK by TGase2 played a key role in microvascular leakage in diabetic mouse retinas. AMPK is an emerging preventive and therapeutic target in the treatment of diabetes complications (17). AMPK is regulated by phosphorylation at threonine residue 172 within the catalytic α subunit and AMP binding to the γ subunit, whereas ATP promotes dephosphorylation of AMPK (16). AMPK provides beneficial effects against diabetes complications by enhancing glucose uptake by muscle cells through expression of glucose transporter 4, protecting β-cells from apoptosis, and improving insulin sensitivity (17). In the diabetic retina, AMPK activation is associated with stabilization of the BRB and prevention of hyperglycemia-induced inflammation and pericyte apoptosis (19,20,34). In contrast, AMPK inhibition leads to inflammation and retinal vascular leakage through sirtuin 1 inactivation and subsequent activation of nuclear factor-κB and inflammation (35). Therefore, regulation of AMPK by hyperglycemia is important in the pathogenesis of DR; however, the molecular mechanism by which hyperglycemia inhibits AMPK remains unclear. In the current study, we found that TGase2, which is activated by hyperglycemia-induced ROS generation, reduced AMPK phosphorylation in the retinas of diabetic mice. The inhibition of AMPK by TGase2 was implicated in hyperglycemia-induced adherens junction disassembly and increased FITC-dextran extravasation in the diabetic retina. Therefore, TGase2-mediated inhibition of AMPK could be an important molecular event for hyperglycemia-induced microvascular leakage in the diabetic retina, although it would be necessary to elucidate the association of TGase2 with the upstream activators of AMPK, including liver kinase B1.

GAPDH is a key enzyme in glycolysis that catalyzes the conversion of glyceraldehyde-3-phosphate to 1,3-biphosphoglycerate in the presence of inorganic phosphate and NAD⁺ (22). GAPDH, which is inhibited by hyperglycemia-induced overproduction of mitochondrial ROS, is associated with the pathogenesis of diabetic vascular complications through activation of several pathways, including polyol pathway flux, elevated hexosamine biosynthesis, activation of protein kinase C, and increased formation of advanced glycation end products and expression of their receptors (36,37). GAPDH has been reported to be involved in the development of DR and its progression in diabetic rats (22,38). Hyperglycemia stimulates nuclear translocation of GAPDH with its ADP-ribosylation and nitration, leading to the formation of acellular capillaries in the rodent retina (22). GAPDH translocation to the nucleus is also implicated in high glucose–induced apoptosis of retinal Müller cells (38). However, the role of GAPDH in diabetic retinal vascular permeability and its underlying mechanism are unclear. In the current study, we showed that high-glucose conditions inhibited GAPDH, resulting in adherens junction disassembly and endothelial permeability, and this GAPDH inhibition was regulated by TGase2 and AMPK in HRECs. In this regard, downregulation of GAPDH by TGase2 and AMPK may play an important role in hyperglycemia-induced vascular dysfunction in the diabetic retina.

In conclusion, we found that regulation of AMPK and GAPDH by TGase2 plays a pivotal role in hyperglycemia-induced vascular leakage in the retinas of diabetic mice. TGase, which is activated by elevation of intracellular Ca²⁺ and ROS, plays an essential role in hyperglycemia-induced microvascular leakage in the diabetic retina. ROS generation–mediated TGase2 activation inhibited AMPK and GAPDH, leading to adherens junction disassembly and microvascular leakage in the retina. Therefore, the TGase2/AMPK/GAPDH pathway may function as an important molecular axis in the pathogenesis of DR and thus may be a potential therapeutic target in DR treatment.

Funding. This work was supported by the National Research Foundation of Korea (2021R1A2C2091794).

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Author Contributions. H.-Y.J. and A.-J.L. designed and performed the experiments, analyzed the data, and wrote the manuscript. C.-H.M. designed the experiments and analyzed the data. K.-S.H. conceptualized the study, designed the experiments, analyzed and interpreted the data, and wrote the manuscript. All authors approved the final version of the manuscript. K.-S.H. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.