1. IntroductionDiabetic retinopathy (DR) is a major microvascular complication of diabetes mellitus (DM) that leads to sight impairment in one-third of diabetic patients [1]. In the retina, DR causes vascular endothelial dysfunction, pericyte loss, and neuronal abnormalities through multiple interlinked pathways [2,3]. Since retinal microvascular endothelial cells (RMECs) line the inner wall of vessels, its dysfunction is an early event in DR due to long-term contact with glucose-rich blood [4,5]. Impairment of the RMECs leads to an increase in endothelial permeability, which eventually fuels vascular complications and damage of the blood–retinal barrier (BRB) [6]. Thus, improving RMEC dysfunction is critical for remission of DR-induced vision loss.Mitochondria are central organelles for energy production and are closely related to cell fate. The mitochondria regulate cellular metabolism and redox homeostasis, which mediate pathological processes in many diseases, including DR [7]. Furthermore, the mitochondria play an important role in calcium ion (Ca2+) homeostasis. Imbalance of mitochondrial Ca2+ homeostasis has been shown to be associated with a variety of diseases including diabetes [8]. Under normal physiological conditions, the mitochondria do not harbor Ca2+ [9]. However, since the mitochondria have good Ca2+ buffering capacity, a large amount of Ca2+ enter mitochondria under cellular stress, resulting in the surge of mitochondrial Ca2+ [10]. The rate of energy metabolism regulates mitochondrial-free Ca2+ that stimulate mitochondrial Ca2+-sensitive enzymes [11]. In addition, the balance between formation and scavenging of free radicals was shown to be disrupted by excess mitochondrial Ca2+, which exacerbates oxidative stress and induces apoptosis [12]. The endoplasmic reticulum (ER) is the main intracellular organelle that stores Ca2+, which is a crucial factor for cellular homeostasis. In addition, the ER is a site of protein synthesis, folding, and modification. Previous data have demonstrated that a high concentration of Ca2+ in the ER lumen ensures proper functioning of the ER [13,14]. Dysregulation of ER Ca2+ can trigger an ER stress (ERS) response that leads to cell death [15], and sustained outflow of ER Ca2+ results in elevation of cytoplasmic Ca2+. The elevation of Ca2+ initiates a series of physiological reactions such as caspase activation, release of apoptotic factors from the mitochondria, and direct activation of Ca2+-dependent endonucleases that cleave cellular DNA [16,17]. Therefore, intracellular Ca2+ imbalance defines the initiation of DR, and thus adjustment of the mitochondrial and/or ER dysfunction may ameliorate the microvascular endothelial injury.Maintenance of cellular homeostasis is related to the biological functions of several organelles that play coordinated or antagonistic roles. Crosstalk between organelles, especially the interaction between the ER and mitochondria, has attracted a large amount of research interest in recent decades [18]. Mitochondria-associated membranes (MAMs) tether the mitochondria to the ER and regulate cell processes through a cascade of signaling events such as Ca2+ homeostasis, oxidative stress, apoptosis, and lipid biosynthesis [19,20]. Moreover, there are many core proteins in mammals that either directly regulate the physical connectivity of the MAMs or indirectly regulate the binding complexes in the MAMs. Bridging the inositol triphosphate receptor 1 (IP3R1)–glucose-regulated protein 75 (GRP75)–voltage-dependent anion channel 1 (VDAC1) complex, which is one of the main members of the core proteins, regulates the mitochondria Ca2+ homeostasis [21]. GRP75 interconnects two organelles by assembling the IP3R1–GRP75–VDAC1 complex, which enhances the ER–mitochondria interaction. Therefore, increased GRP75 expression accelerates Ca2+ transfer from the ER to the mitochondria, causing Ca2+ overload, which prompts mitochondrial membrane potential impairment and increased mitochondria ROS production [22]. In contrast, decreased expression of GRP75 suppresses mitochondrial Ca2+ levels [23]. Together, this evidence indicates that any abnormalities in the IP3R1–GRP75–VDAC1 axis induces mitochondrial Ca2+ overload and apoptosis.

In this study, we aimed to evaluate the role of the IP3R1–GRP75–VDAC1 axis in modulating Ca2+ homeostasis in RMECs. Our results revealed that ERS transports Ca2+ to the mitochondria via the IP3R1–GRP75–VDAC1 axis and promotes the formation of MAMs mainly by increasing the GRP75 expression, which ultimately induces mitochondrial Ca2+ overload leading to apoptosis. Thus, this pathway presents a potential treatment target for DR.

2. Materials and Methods 2.1. Materials and ReagentsGlucose, streptozocin (STZ), and Evans blue were purchased from Sigma-Aldrich (St. Louis, MO, USA), while advanced glycosylation end products (AGEs) were acquired from BioVision (Palo Alto, CA, USA). Tunicamycin (TUN) was obtained from Abcam (Cambridge, CA, USA), while 4-phenylbutyric acid (4-PBA), BAPTA-AM, and Protein A/G magnetic beads were purchased from MedChemExpress (Lowell, NJ, USA). Cell counting kit-8 (CCK8), JC-1 assay kit, Calcein/PI assay kit, MPTP assay kit, Mito-Tracker Green, ER-Tracker Red, and a mitochondria isolation kit were purchased from Beyotime (Nantong, China). On the other hand, the TUNEL assay kits were purchased from Roche (Basel, Switzerland). MitoSOX™ Red, DAPI, ECL kits, and goat anti-mouse/rabbit IgG (H+L, Alexa Fluor 555/488) were acquired from Thermo Fisher (Waltham, CA, USA). We used primary antibodies including GRP75, Cyt c, Bcl-xl, Bax, cleaved caspase-3, and CHOP (Cell Signaling Technology, Boston, MA, USA); 4-HNE and Brn3a (Abcam, Cambridge, CA, USA); IP3R1, VDAC1, VEGF, 8-OHDG, and 3-NT (Santa Cruz Biotechnology, Santa Cruz, CA, USA); and COX IV, Bcl-2, and β-actin (Proteintech, Wuhan, China). More detailed information about the materials and reagents is offered in Table S1. 2.2. Animals and Models

Male Sprague-Dawley rats (SD, 8 weeks old, 200–250 g) were purchased from Changzhou Cavans Experimental Animal Co., Ltd. (Changzhou, China). The animals were housed at 22–26 °C, while food and water were provided ad libitum during the entire experiment. All experimental procedures were performed according to the National Institutes of Health Guidelines for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee of Nanjing Medical University (SYXK(SU)2020-0010).

The rats were divided into 2 groups randomly: normal control rats (Ctrl, n = 15) and diabetes rats (n = 45). Diabetes was induced by intraperitoneal injection of STZ (1.5%, 60 mg/kg) dissolved in citrate buffer (pH 4.5, 0.1 mol/L), while the control group received an equivalent volume of citrate buffer. Fasting blood glucose level was assessed on the third day after STZ treatment. Only 32 rats with diabetes were modelled successfully, for which random blood glucose was higher than 16.7 mM. The 32 rats were once again randomized into 2 groups: the diabetes group (n = 16) and the diabetes with 4-PBA treatment group (STZ + 4-PBA, n = 16). 4-PBA (500 mg/kg) was administered daily by oral gavage from week 4 to week 14. The diabetes group was treated with equal volumes of the vehicle solution. The rats were sacrificed by overdose of anesthesia at the 14th week, and then the eyes and retinas were isolated for further analyses.

2.3. Cell Culture

Human retinal microvascular endothelial cells (RMECs) were acquired from the Beina Chuanglian Biotechnology Institute (Beijing, China). The cells were maintained in complete DMEM medium (Gibco, Grand Island, New York, NY, USA) with 10% FBS (Gibco Grand Island, New York, NY, USA) and 1% penicillin and streptomycin (Beyotime, Nantong, China) at 5% CO2, 37 °C.

2.4. siRNA TransfectionThe RMECs were transfected with GRP75 or negative control (NC) siRNA at 80% density with a riboFECT CP Transfection Kit (RiboBio, Guangzhou, China). The RMECs were cultivated in 6-well plates overnight. A total of 120 μL Opti-MEM medium including 12 μL riboFECT CP, 100 nM GRP75 siRNA, or 100 nM NC was blended in a 6-well plate. Detailed siRNA information is shown in Table S2. 2.5. Cell Counting Kit 8 (CCK-8)

We analyzed cell viability using the CCK8 kit (Beyotime, Nantong, China). RMECs (1000 per well) were planted in 96-well plates followed by the addition of CCK-8 solution (100 μL/each well). The cells were incubated for 1.5 h at 37 °C in a 5% CO2 incubator. Absorbance was detected at 450 nm by a microplate reader.

2.6. ER and Mitochondrion Staining

The ER was stained using the ER-Tracker Red kit (1 μM, Red, Beyotime, Nantong, China). The mitochondria were stained using MitoTracker-Green (200 nM, Green, Beyotime, Nantong, China) or MitoTracker-Red (200 nM, Red, Beyotime, Nantong, China). The cells were assessed by laser scanning confocal microscopy (40×, Leica, Weztlar, Germany).

2.7. Detection of Mitochondrial Ca2+

Mitochondrial Ca2+ was detected using Rhod-2 AM (2 μM, Red, YEASEN, Shanghai, China), which contains Pluronic F-127 (0.02%, Beyotime, Nantong, China). In addition, MitoTracker-Green (200 nM, Green, Beyotime, Nantong, China) was used to co-stain the mitochondria. The cells were evaluated under the laser scanning confocal microscope (40×, Leica, Weztlar, Germany).

2.8. Analysis of Mitochondrial ROS and Mitochondrial Membrane Potential (MMP)

The levels of mitochondrial ROS were assayed by MitoSOX™ Red (Red, Thermo Fisher, Waltham, MA, USA). Briefly, the RMECs were stained with the MitoSOX™ Red (5 μM, 10 min) and then co-stained with MitoTracker-Green (200 nM, Green, 30 min, Beyotime, Nantong, China). The mitochondrial ROS was analyzed by laser scanning confocal microscopy (40×, Leica, Weztlar, Germany).

On the other hand, MMP was assessed by a JC-1 assay kit (Beyotime, Nantong, China). Briefly, the RMEC cells were incubated with JC-1 staining solution. The cells were visualized under the laser scanning confocal microscope (40×, Leica, Weztlar, Germany). MMP was calculated as the ratio of red/green using ImageJ software.

2.9. Measurement of Mitochondrial Permeability Transition Pore (mPTP)

The mitochondrial permeability transition pore opening was analyzed by an mPTP assay kit (Beyotime, Nantong, China). Calcein-AM dye selectively gathers inside the mitochondria and stimulates green fluorescence, while cobalt chloride (CoCl2) quenches the green fluorescence of calcein in the cytoplasm. Once mPTP is opened, calcein is released from the mitochondria and then quenched by CoCl2. Therefore, the stronger the calcein green fluorescence, the less mPTP opening. Briefly, the RMECs were loaded with Calcein-AM dye (2 μM) containing CoCl2 (250 μM). The Mitotraker-red (200 nM, 30 min, Beyotime, Nantong, China) was then used to co-stain the mitochondria. The cells were analyzed by laser scanning confocal microscopy (40×, Leica, Weztlar, Germany).

2.10. Calcein-AM/Propidium Iodide (PI) Staining

Live or dead cells were analyzed by a Calcein-AM/PI double stain kit (Beyotime, Nantong, China). The RMECs were stained with Calcein-AM (2 μM) and PI (5 μM), and then the analysis of the stained cells was immediately performed by laser scanning confocal microscopy (20×, Leica, Weztlar, Germany).

2.11. Vascular Permeability of the Blood–Retinal Barrier (BRB)Vascular permeability was evaluated by assessing Evans blue dye leakage from retinal vessels as previously described [24]. The rats received intravenous injection (through the tail vein) of Evans blue (3%, dissolved in 0.9% NaCl). After 2 h, the animals turned blue. Thereafter, the eyes were isolated and the retinas were removed from the eyecup, and following this, we used the BX-51 light microscope (10×, Olympus, Tokyo, Japan) for observation. For quantification, retinas were weighed and then incubated in 150 μL formamide for 18 h at 78 °C to extract the Evans blue. The extract was centrifuged at 12,000× g, 30 min at 4 °C, and the absorbance was determined at 620 nm. The results were expressed in microgram per gram of retina. 2.12. Retinal Digestion and Periodic Acid–Schiff (PAS) Staining

Retinal PAS staining is a method used to detect the amounts of acellular capillaries in the retina. The retinas were fixed in 4% paraformaldehyde for 1.5 h and digested in 3% trypsin for 2 h at 37 °C. The samples were gently washed until the retinal leaves were completely spread. Visualized vessels were dried on carrier slides and then retinal capillaries were dyed with the PAS/haematoxylin stain kit (Solarbio, Beijing, China), following the manufacturer’s instructions. The samples were assessed under a light microscope (20×, Olympus, Tokyo, Japan).

2.13. HE Staining

To explore morphological alterations in the retina, we performed hematoxylin and eosin (H&E) staining. Briefly, eye paraffin-embedded tissues were cut into 5 μm sections at the periphery of the retinas. The slides were deparaffinized, rehydrated, and then stained using the H&E staining kit (BOSTER, Wuhan, China). The samples were analyzed using a light microscope (20×, Olympus, Tokyo, Japan).

2.14. TUNEL Assay

Apoptosis was evaluated using the TUNEL assay kit (Roche, Basel, Switzerland). Briefly, retinal frozen sections were permeabilized with 0.5% TritonX-100 and then incubated with TUNEL reaction mixture. The samples were then counterstained with DAPI (1:1000). The analysis of the samples was performed using laser scanning confocal microscopy (20×, Leica, Weztlar, Germany).

2.15. Immunohistochemistry and Immunofluorescence

Paraffin-embedded tissues were cut into 5 μM sections. The sections were stained with an immunohistochemistry kit (BOSTER, China). Briefly, the sections were incubated with 3% H2O2 for 15 min, and then antigen repair was performed in citrate buffer for 20 min. After blocking with 5% BSA for 1 h, the samples were incubated with 4-HNE (1:200, Abcam, Cambridge, CA, USA), Brn3a (1:500, Abcam, Cambridge, CA, USA), 8-OHDG (1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA), VEGF (1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA), or 3-NT (1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at 4 °C. Thereafter, the samples were incubated with polymerized HRP-labeled anti-rabbit/mouse IgG and then detected with diaminobenzidine (DAB, BOSTER, Wuhan, China). Finally, the sections were counterstained with hematoxylin, then observed with a light microscope (20×, Olympus, Tokyo, Japan).

Cells or frozen sections were fixed with 4% paraformaldehyde for 15 min, then permeabilized with 0.2% Triton X100 (Beyotime, Nantong, China) for 15 min and blocked with 5% BSA for 1 h. Thereafter, the samples were incubated with CHOP (1:500, CST, Boston, MA, USA), c-caspase-3 (1:400, CST, Boston, MA, USA), or Rhod (1:100, Santa Cruz Biotechnology, Santa Cruz, CA, USA) overnight at 4 °C. The samples were then incubated with goat anti-rabbit/mouse IgG (H+L) and Alexa Fluor 555/488 (Thermo Fisher, Waltham, MA, USA) while nuclei were stained with DAPI (1:1000). Fluorescence was detected using laser scanning confocal microscopy (20×/40×, Leica, Weztlar, Germany), and we quantified the fluorescence intensity using ImageJ software (ImageJ, Bethesda, AR, USA).

2.16. Co-ImmunoprecipitationCo-immunoprecipitations (Co-IP) was performed as previously with minor adaptations [25]. The RMECs were lysed for 30 min using Pierce™ IP lysis buffer (Thermo Fisher, Waltham, MA, USA) containing protease inhibitor (Roche, Basel, Switzerland). The lysed cells were centrifuged at 12,000× g for 15 min at 4 °C to obtain the supernatant, and then protein concentration was determined using the BCA, assay kit (Beyotime, Nantong, China). A total of 1/10 of the supernatant was kept as the input group and the remaining 500 μL was taken for immunoprecipitation. Here, the supernatant was incubated with 1 μg primary antibody, IP3R1 or VDAC1, or normal control IgG overnight at 4 °C. The samples were incubated with 50 μL protein A/G magnetic beads (MedChemExpress, South Brunswick, NJ, USA) at 4 °C for 3 h and then centrifuged at 4 °C for 5 min at 1000× g. The magnetic beads were washed five times with buffer, and then proteins were eluted from the beads. The samples were boiled for 5 min in 1 × loading buffer and then resolved in SDS-PAGE. The samples were further analyzed by Western blot. 2.17. Western Blot Analysis

RMECs and tissues were extracted using RIPA lysis buffer (Beyotime, Nantong, China). Protein concentrations were quantified using the BCA, kit (Beyotime, Nantong, China). The samples were resolved in an electrophoresis set up and then transferred to PVDF membranes. After blocking with 5% milk for 2 h, the membranes were incubated with the following primary antibodies: GRP75 (1:1000, CST, Boston, MA, USA), IP3R1 (1:200, Santa Cruz Biotechnology, Santa Cruz, CA, USA), VDAC1 (1:200, Santa Cruz Biotechnology, Santa Cruz, CA, USA), Cyt c (1:1000, CST, Boston, MA, USA), Bcl-2 (1:2000, Proteintech, Wuhan, China), Bcl-xl (1:1000, CST, Boston, MA, USA), Bax (1:1000, CST, Boston, MA, USA), c-caspase-3 (1:1000, CST, Boston, MA, USA), COX IV (1:5000, Proteintech, Wuhan, China), or β-actin (1:5000, Proteintech, Wuhan, China) at 4 °C overnight. Thereafter, the blots were incubated with horseradish peroxidase-conjugated secondary antibody for 2 h and then analyzed using the chemiluminescence kit (Thermo Fisher, Waltham, MA, USA).

2.18. Statistical Analysis

At least three rounds of each experiment were performed. All data are expressed as mean ± SD (standard deviation). Statistical analyses were performed using Student’s two-tailed t-test or one-way ANOVA with Tukey’s honest significant difference test using GraphPad Prism version 8.0 (GraphPad Software, San Diego, CA, USA). A p < 0.05 was considered to be statistically significant.

4. DiscussionThe pathogenesis of DR, in which retinal vascular endothelial cell (RMEC) dysfunction plays a crucial role, is a complex process. However, the mechanisms involved in RMEC dysfunction have not been fully established. Various studies have evaluated the roles of ERS or mitochondrial dysfunction in RMECs during DR progression; however, the molecular mechanisms involved in interactions between the two organelles are yet to be conclusively determined. We found that ER–mitochondrial coupling was increased under the DR state and was associated with the upregulated IP3R1–GRP75–VDAC1 axis. Mitochondrial Ca2+ overload was induced by persistent Ca2+ transport from ER to the mitochondria, leading to mitochondrial dysfunction and RMECs damage. The main mechanisms are shown in Figure 8.AGEs, H2O2, and HG are the frequently used agents for in vitro induction of diabetic-related complications [31,32]. The situation is more complex with exposure to HG, which closely resembles the human condition. AGEs, which are closely associated with DR, are products of excess glucose and protein formation in the body [33]. Therefore, HG and AGEs were used to simulate pathological changes in DR in vitro. Information communication between organelles is a universal physiological phenomenon. We established that HG and AGEs enhance ER–mitochondrial coupling in RMECs. Under physiological conditions, ER–mitochondrial coupling regulates Ca2+ and lipid transport between organelles, as well as the maintenance of mitochondrial morphology, dynamics, and energy metabolism [34,35]. However, in pathological conditions, increased or decreased ER–mitochondrial coupling induces various pathological changes, such as mitochondrial Ca2+ overload, mitochondrial dysfunction, lipid accumulation, and oxidative stress [36]. Therefore, studies have aimed at evaluating the roles of ER–mitochondrial coupling in disease progression.Mitochondria-associated membranes (MAMs) are synaptic-like subcellular structures formed between the ER membrane and the mitochondria. As sites for bidirectional communication between ER and the mitochondria, MAMs play an essential role in regulating basic cell physiological activities [37]. Furthermore, MAMs are a particular class of cellular structures with flexible and plastic features and are regulated by various regulatory factors [18]. MAM-associated Ca2+ homeostasis is a research focus in metabolism-related diseases [38,39]. Ca2+ is an essential intracellular second messenger, regulating multiple pathophysiological processes in cells [40]. Intracellular Ca2+ homeostasis depends on coordination between enzymes, ion pumps, and Ca2+ channels, as well as storage of ER, mitochondria, and lysosomes [41,42]. The mitochondria rapidly detect and respond to changes in cytosolic Ca2+, which is an important buffer for regulating Ca2+ homeostasis [39]. Mitochondrial Ca2+ uptake is primarily regulated by VDAC, which anchors VDAC to Ca2+ release channels in ER/sarcoplasmic reticulum (SR) via scaffolding proteins [43]. IP3R and ryanodine receptor (RyR) are the main Ca2+ release channels on the ER, while GRP75 is a bridge between IP3R and VDAC [42]. Therefore, the IP3R1/GRP75/VDAC1 axis is a key pathway in the regulation of mitochondrial Ca2+ homeostasis on MAMs. In skeletal muscles of insulin resistance mice models, increased MAM formation mediated Ca2+ transport to the mitochondria via the IP3R1–GRP75–VDAC1 complex, resulting in mitochondrial Ca2+ overload and dysfunction [44]. In this study, we found increased MAM formation accompanied by the upregulated IP3R1/GRP75/VDAC1 axis and mitochondrial Ca2+ overload under DR conditions. However, GRP75 knockdown significantly reduced mitochondrial Ca2+ overload and dysfunction. Therefore, GRP75 may be a bridge protein linking ER–mitochondrial coupling and mitochondrial Ca2+. As a member of the HSP70 protein family, GRP75 is involved in formation of MAMs and participates in cellular endocytosis, proliferation, and tumorigenesis [45,46]. Apolipoprotein-E4-induced ERS in neurons upregulates GRP75 expressions, accelerates MAM formation, and impairs neuronal mitochondrial functions through mitochondrial Ca2+ overload [23]. In palmitate-induced pancreatic β cells, GRP75 overexpression promoted MAM formation and caused mitochondrial dysfunction [22]. However, after axon dissection in primary neurons, GRP75 overexpression enhanced the tricarboxylic acid cycle and increased ATP production by upregulating mitochondrial Ca2+, accelerating the regrowth of damaged axons [47]. Under different disease progression states, the ER–mitochondrial coupling alterations play different roles. Therefore, the roles of increased ER–mitochondrial coupling in RMECs under DR states should be investigated.Hyperglycemia induces mitochondrial Ca2+ overload in RMECs [29]; however, the underlying mechanisms have yet to be fully elucidated. ER, as the principal site of Ca2+ storage, plays important roles in Ca2+ homeostasis. Increased ER–mitochondrial coupling is closely associated with mitochondrial Ca2+ overload. Various studies are evaluating the physiological significance of mitochondrial Ca2+ homeostasis [48,49]. Mitochondrial Ca2+ imbalance is involved in multiple disease processes, including diabetes, neurodegeneration, heart failure, and cancer [50,51,52]. However, the role of mitochondrial Ca2+ homeostasis in DR requires further studies. Mitochondria, the central site of energy metabolism, is closely correlated with Ca2+ homeostasis. Under physiological conditions, key enzymes of the tricarboxylic acid cycle are activated by Ca2+, thereby increasing ATP production to maintain higher energy requirements. However, under pathophysiological conditions, mitochondrial Ca2+ overload promotes mitochondrial ROS levels, mitochondrial membrane potential disruption, and mPTP opening in cells [53]. ER–mitochondrial coupling impairs mitochondrial membrane potential and increases mitochondrial Ca2+ in pancreatic β-cells [22]. We established that BAPT-AM (Ca2+ chelator) reverses hyperglycemia induced-mitochondrial dysfunction and apoptosis via the mPTP-mediated Cyt c/caspase-3 pathway, suggesting that mitochondrial Ca2+ overload accelerates the DR process via mitochondrial dysfunction.Dysfunction of the ER and/or the mitochondria plays a significant role in DR development [30,54]. The mechanisms of ER–mitochondrial coupling are still unclear. Disruption of ER homeostasis results in accumulation of unfolded and misfolded proteins in the ER, leading to ERS. ERS may cause cell death by inducing Ca2+ overload, inhibiting protein synthesis and promoting apoptosis [55]. Khaled Elmasry et al. found that upregulated Ca2+ levels in the retina promotes 12/15-lipoxygenase-induced ERS and microvascular dysfunction by activating the ERS/NADPH oxidase/VEGFR2 signaling pathway [29]. In this study, TUN (an inducer of ERS) increased mitochondrial and ER coupling in RMECs, accompanied by mitochondrial Ca2+ overload, mitochondrial dysfunction, and apoptosis. These effects were associated with upregulated IP3R1–GRP75–VDAC1 axis.

This study has some limitations. First, retinal mitochondrial Ca2+ levels cannot be detected in vivo. Second, GRP75 knockout mice were not used. However, despite these limitations, we found that DR promotes MAM formation via the IP3R1–GRP75–VDAC1 axis, facilitates Ca2+ transport from the ER to the mitochondria, and contributes to mitochondrial Ca2+ overload, leading to mitochondrial dysfunction and endothelial cell apoptosis.