Ginsenoside Rg1 Inhibits High Glucose-Induced Proliferation, Migration, and Angiogenesis in Retinal Endothelial Cells by Regulating the lncRNA SNHG7/miR-2116-5p/SIRT3 Axis

Background Diabetic retinopathy (DR), including retinal angiogenesis and endothelial cell proliferation and migration, is a serious complication in diabetic patients. It has been reported that ginsenoside Rg1 can prevent retinal damage. However, the mechanism by which Rg1 prevents retinal damage is unknown. Therefore, the aim of the present study was to investigate the mechanism by which Rg1 inhibits high glucose-induced complications through the regulation of the lncRNA SNHG7/miR-2116-5p/SIRT3 axis. Methods Under high glucose (HG) conditions, human retinal endothelial cells (HRECs) were cultured to simulate a DR environment, and Rg1 was added after 48 h. Negative control (NC), miR-2116-5p mimic, si-SNHG7, pc-DNA SIRT3, and miR-2116-5p inhibitor were transfected into HRECs, and CCK-8 assay was used to detect the cell viability. Angiogenesis and transwell assays were used to evaluate angiogenesis and cell migration, respectively. qRT–PCR and Western blot were used to detect the expression of related genes and proteins. Luciferase reporter assays and bioinformatics were used to analyze the target binding sites of miR-2116-5p to lncRNA SNHG7 and SIRT3. Results The proliferation, migration and angiogenesis of HRECs were induced by HG. As expected, HG upregulated miR-2116-5p and VEGF expression but downregulated lncRNA SNHG7 and SIRT3 expression. Importantly, Rg1 inhibited HG-induced HREC proliferation, migration, and angiogenesis by upregulating the lncRNA SNHG7, and miR-2116-5p had a target regulatory relationship with both lncRNA SNHG7 and SIRT3. Conclusion Rg1 inhibits HG-induced proliferation, migration, angiogenesis, and VEGF expression in retinal endothelial cells through the lncRNA SNG7/miR-2116-5p/SIRT3 axis. This finding provides theoretical evidence for the clinical application of Rg1 in DR.


Introduction
Numerous studies have shown that diabetes causes various complications (diabetic nephropathy, diabetic retinopathy (DR) and cardiovascular disease), which have become the main cause of morbidity and mortality of diabetes [1]. Type 2 diabetes can lead to serious neurovascular complications, leading to visual impairment and blindness, and DR is one of the main causes [2]. Te basic pathological changes in DR include the selective loss of pericytes, capillary basement membrane thickening, microangioma formation, endothelial cell proliferation, and retinal detachment due to neovascularization [3]. Te frst barrier to monitoring blood glucose changes is the retinal endothelium. Te existing evidence suggests that a high concentration of glucose can lead to increasing numbers and migration of retinal endothelial cells, which is a key step in the occurrence of DR [4]. Although some new drugs and vitreoretinal microsurgery have been used in clinical DR treatment, the incidence of DR has dramatically increased in recent decades [5,6].
Terefore, further revealing the etiology of DR is important for improving the available treatment methods [7,8]. Ginsenoside Rg1 (Rg1) is a component of ginsenoside, which mainly exists in ginseng medicinal materials. Rg1 can quickly relieve fatigue, delay aging, stimulate the central nervous system, inhibit platelet aggregation, and improve learning and memory [9]. Rg1 has also been shown to be useful in the treatment of myocardial infarction [10], diabetic limb infarction [11], and ischemic necrosis of the skin [12]. Rg1 promotes neovascularization after myocardial infarction, diabetic limb infarction, skin ischemic necrosis, and neonatal hypoxic encephalopathy [13]. Experimental studies have shown that Rg1 has strong antioxidant and blood glucose-lowering activities [14,15]. Rg1 can promote angiogenesis and enhance endothelial progenitor cell angiogenesis. Moreover, Rg1 can improve endoplasmic reticulum stress-induced apoptosis in diabetic cardiomyopathy induced by streptozotocin (STZ) [16]. Rg1 prevents retinal damage by inhibiting retinal cell apoptosis [17]. Moreover, experimental studies have shown that ginsenoside Rg1 plays a role in promoting vascular regeneration and enhancing endothelial progenitor cell angiogenesis [18].
LncRNAs have been shown to afect the progression mechanisms of DR through various methods [19]. In human retinal endothelial cells, HG-induced angiogenesis can be inhibited by the lncRNA SNHG7 through the miR-543/ SIRT1 cascade [20]. Whether SNHG7 participates in the regulation of vascular growth in DR and whether it promotes angiogenesis in human retinal endothelial cells (HRECs) remain unclear. Notably, in the pathogenesis of DR, miR-NAs also play an indispensable role. Tere have been reports of the abnormal expression of miRNAs in the retina of diabetic rats induced by STZ [21,22]. In addition, it has been shown that miR-3197 and miR-2116-5p are immensely upregulated in DR patients and are efective diagnostic markers of DR [23].
As a conservative nicotinic adenine dinucleotidedependent (NAD-dependent) deacetylase, sirtuins consist of seven isomers [24]. In addition, sirtuin-3 (SIRT3), a core member of the sirtuin family, is located on the mitochondrial membrane. SIRT3 can deacetylate the target protein, which plays an important role in antioxidation, biosynthesis, and energy metabolism of mitochondria [25]. For example, under the mediation of SIRT3, autophagy-related proteins can be acetylated, thus afecting autophagy [26]. SIRT3 is necessary for coronary angiogenesis and glycolysis [27]. In type 2 diabetic mice, retinal dysfunction may be related to the loss of SIRT3 and SIRT5 [28] because SIRT3 may promote autophagy by downregulating the expression of angiogenesis-related genes in retinal endothelial cells [29]. In addition, in a rat model of diabetes and retinopathy, the expression of autophagy-related proteins was promoted by the overexpression of SIRT3, while VEGF was inhibited [30]. Tese fndings suggest that SIRT3 is a key therapeutic target for DR.
In the present study, StarBase website prediction suggested that miR-2116-5p has target binding sites for both the lncRNA SNHG7 and SIRT3, implying that miR-2116-5p, lncRNA SNHG7, and SIRT3 may act as an axis. Terefore, we investigated the mechanism by which ginsenoside Rg1 inhibits retinal endothelial cell lesions induced by high glucose by regulating the lncRNA SNHG7/miR-2116-5p/ SIRT3 axis.

Animal Breeding and
Modeling. In total, 120 healthy male SD rats of SPF class (Animal Experiment Center of Kunming Medical University), weighing 200 ± 25 g, were utilized, and all rats had no pathological changes in the anterior and anterior segments of the eyes after the examination. Te blood glucose levels were within the normal range as detected by a blood glucose meter after tail vein collection. Te rats were randomly divided into 3 groups (40/group): normal rats group (NC); DR rats (Model); Rg1treated DR rats group (Rg1). Te rats were acclimatized and housed for 1 week before the experiment, and they were fed and watered ad libitum. Te rats in the diabetic group were fasted for 12 h before modeling and weighed before the experiment. In the diabetic group, freshly prepared STZ in bufer (55 mg/kg; Sigma-Aldrich, Germany) was injected once into the left lower abdominal cavity, and the rats ate and drank normally after the injection. After the rats were injected with STZ for 48 h, the blood glucose and body weight were measured. Additionally, in the diabetic rat models, blood glucose >16.7 mmol/L, polyuria, and polyphagia were considered. Te blood glucose and body weight of the rats were observed once every 2 weeks. In the Rg1 group, gavage was started on the day of modeling, and 0.5 mL (12-5 g/ml) of Rg1 solution (Solarbio, Beijing, China) was given by gavage every day, while the same dose of saline was given to the model and NC groups. Te rats were sacrifced 8 weeks after modeling, and fresh retinal specimens were removed and preserved for the relevant assays. Te experimental scheme of this study was approved by the Animal Ethics Committee of Kunming Medical University and fully met the requirements of the National Institutes of Health Laboratory Animal Care Guide.

Cell Culture and Transfection. HRECs (American Type
Culture Collection, Inc.) were cultured at 37°C with saturated humidity for two days before passaging. Before adding 1 mL of trypsin, the cells were washed with PBS 3 times, which covered the entire cell layer. Te cells were observed under an inverted microscope until they shrank into a round shape, and then, 10% fetal bovine serum was added to neutralize trypsin. Te samples were centrifuged at 1000 rpm for 5 min to collect the cells. Ten percent fetal bovine serum was added, and the cells were inoculated in 75 cm 2 culture fasks at 10 mL/bottle (10 4 cells/ml). A concentration of 5 mM glucose is a normal glucose condition, and 25 mM is a high glucose condition. Te cells were incubated with 25

Cell Viability Assay.
In total, 5 × 10 4 cells/well were inoculated into a 96-well plate, 10 μL CCK-8 solution (Sangon, Shanghai, China) was added, and the cells were incubated for 4 h at 25°C. Te OD value was measured at 450 nm.

Angiogenesis Experiments.
Cells were inoculated into a 24-well plate at 37°C, and Matrigel (Sigma-Aldrich, Germany) was added to each well and allowed to harden for 30 min. HRECs were inoculated at a density of 1.2 × 10 5 cells/ well on top of the Matrigel-coated wells and cultured in a sterile incubator at 100% humidity, 37°C and 5% CO 2 for 6 h. An inverted microscope was used to observe the tube lumen and acquire the images. Image-Pro Plus software was used to calculate the number of Matrigel tubule formations in the feld of view and the tube formation capacity.

Transwell Experiment.
Cells were collected from each group 48 h after transfection and washed, and serum-free DMEM was used to adjust the cell concentration to 1 × 10 5 cells/mL. In 24-well Transwell plates (Corning, USA), 200 μL of cell suspension was added to the upper chamber, and in the lower chamber, 500 μL of DMEM containing 10% fetal bovine serum was added. After culturing for 24 h, the unstained cells were wiped of, while the stained cells were stained with crystal violet for 20 min. An inverted microscope was used to observe the cells, and fve randomly selected visual felds were imaged and counted.

qRT-PCR Experiments.
Te total RNA was extracted from tissues and cells using a Total RNA Extractor (Sangon Biotech). A cDNA synthesis kit (Vazyme, Nanjing, China) was used to reverse transcribe 2 μg mRNA into cDNA, which was then diluted 10 times. One microliter of the prepared cDNA was used for qPCR, and the U6 or GAPDH gene was used as the reference gene. All primers (Table 1) used in this study were designed with Premier 5.0. Te two-step reaction conditions for PCR were as follows: predenaturation (maintained at 95°C for 5 min), maintenance at 95°C for 10 s, annealing (30 s) and extension (30 s). Both annealing and extension were cycled 40 times. Te confdence of the PCR results was assessed by the dissociation curve and cycle threshold (CT) values. Te results were calculated by the 2 −ΔΔCt method after repetition at least 3 times.

Western Blot Assay.
Proteins were extracted from retinal tissue utilizing RIPA lysis bufer (Sangon Biotech, Shanghai), and a BCA assay (Sangon Biotech, Shanghai) was used to determine the total protein content. 10% SDS-PAGE gel was used to separate the total proteins, which were then transferred to PVDF membranes by a constant current fow at 200 mA. Subsequently, the PVDF membranes were incubated with antibodies (Abcam, USA) for 12 h at 4°C. Te PVDF membranes were washed with TBS bufer and incubated with secondary antibodies at 25°C for 1 h. After washing the membranes three times, chemiluminescent reagents were added, and the bands were analyzed for grayscale values using ImageJ software. Each experiment was repeated 3 times independently.

Bioinformatics and Dual Luciferase Gene Reporter
Analysis. In this study, StarBase (http://starbase.sysu.edu. cn/) was used to predict the binding sites of miRNAs and lncRNAs. Te dual-luciferase reporter vectors containing WT and mutant-type binding sites for SNHG7 or SIRT3 sequences were constructed by a rapid cloning kit (Vazyme, Nanjing, China) and named WT-SNHG7 or WT-SIRT3 and MUT-SNHG7 or SIRT3, respectively. Subsequently, WT-SNHG7 or WT-SIRT3 and MUT-SNHG7 or SIRT3 vectors were transfected into HRECs (Chinese Academy of Sciences Culture Collection) with NC mimic or miR-29b-3p mimic.
After transfection for 48 h, a dual luciferase reporter assay was used to detect luciferase activity.
2.11. Statistical Analysis. GraphPad Prism 8 software was used to analyze and prepare graphs of the experimental data. In this study, the results are shown as the mean ± standard deviation (SD). As expected, two groups and multiple groups of data were analyzed by unpaired Student's t-test and oneway analysis of variance, followed by Tukey's post-hoc test. Te P value representing statistical signifcance was 0.05. Compared with the control group, the detection of blood glucose values in the diferent treatment groups revealed that the model group rats had signifcantly higher blood glucose after the STZ injection, and the blood glucose level was higher than 16.7 mmol/L, demonstrating a successful diabetic model. In contrast, the treatment group had signifcantly lower blood glucose (Figure 1(a)). Compared with the control group, SNHG7 and SIRT3 were signifcantly lower in the model group, and the expression of both SNHG7 and SIRT3 increased after the Rg1 treatment as shown by qRT-PCR (Figures 1(b) and 1(d)). As expected, compared with the control group, the expression of miR-2116-5p was signifcantly higher in the model group, and the expression of miR-2116-5p was signifcantly lower after the Rg1 treatment (Figure 1(c)). Te HE staining results showed that the control rats had a clear and continuous inner boundary membrane and only a few vascular endothelial cells in the vitreous near the inner retinal boundary membrane. Te model rats showed edema on the retinal surface, and the number of vascular endothelial cells was considerably increased. Moreover, the rats in the Rg1treated group had a clear and continuous inner boundary membrane, reduced edema and decreased vascular endothelial cells (Figure 1(e)). Compared to the control group, SIRT3 was signifcantly reduced in the retinal tissues of the rats in the model group in the immunohistochemical assay.

Efect of
In contrast, SIRT3 in the Rg1-treated rats was signifcantly higher than that in the model group (Figure 1(f )). Te VEGF-immunopositive product was indicated by brownish-yellow granular staining, and immunopositive cells were mainly distributed in the retinal ganglion cell layer, which was opposite to that observed with SIRT3 (Figure 1(g)), and in the inner nuclear layer. Te results of Western blot detection also showed that compared with the control group, the expression of SIRT3 was down-regulated and VEGF was up-regulated in the model group, and Rg1 treatment reversed this phenomenon (Figure 1(h)). In summary, these fndings show that Rg1 downregulates miR-2116-5p and VEGF but upregulates the lncRNA SNHG7 and SIRT3 in the retinas of diabetic rats.

Efect of Rg1 on the Proliferation, Migration, and Angiogenesis of HG-Treated HRECs.
Cell viability was assessed using a CCK-8 assay to investigate the efect of Rg1 on HGinduced pathological phenomena (HREC proliferation, migration and angiogenesis). Te results showed that the HG treatment signifcantly increased the viability of HRECs, while the Rg1 treatment signifcantly inhibited cell viability (Figure 2(a)). As expected, in the HG group, the qRT-PCR analysis showed that the lncRNA SNHG7 and miR-2116-5p were signifcantly lower and higher, respectively, and they were signifcantly reversed after the addition of Rg1 (Figures 2(b) and 2(c)). Te number of migrating cells and angiogenesis were signifcantly higher in the HG group in the Transwell and angiogenesis assays, and the number of migrating cells and angiogenesis were decreased after the addition of Rg1 (Figures 2(d) and 2(e)). Similarly, the protein expression of SIRT3 was signifcantly lower and VEGF was elevated in the HG group as shown by the Western blot analysis. Te treatment with Rg1 signifcantly increased the protein level of SIRT3 but signifcantly decreased VEGF (Figure 2(f )). Tus, these fndings demonstrate that high glucose induces pathological phenomena in HRECs, but Rg1 signifcantly inhibits these changes.

Te Targeting Relationship between the lncRNA SNHG7
and miR-2116-5p. StarBase online software was used to predict the binding sites of lncRNA SNHG7 in miR-2116-5p (Figure 3(a)). As verifed by the dual luciferase assays, the luciferase activity of wild-type SNHG7 could be reduced by miR-2116-5p but had almost no efect on mutant SNHG7 (Figure 3(b)). Te transfection of diferent siRNAs, including siRNA NC (si-NC) and siRNA-SNHG7 (si-S1/2/3), was used to knockdown SNHG7. Because the transfection of si-S2 showed the best knockdown of SNHG7, it was used in the subsequent experiments (Figure 3(c)). Te knockdown or overexpression of SNHG7 was verifed by a qRT-PCR analysis. Te results showed that miR-2116-5p was signifcantly decreased after the overexpression of SNHG7, while miR-2116-5p was signifcantly increased after the knockdown of SNHG7 (Figure 3(d)). Tus, these fndings  Journal of Oncology demonstrate that the lncRNA SNHG7 negatively regulates miR-2116-5p by targeting the modulation of miR-2116-5p.

Rg1 Inhibits HG-Induced Cell Proliferation, Migration, and Angiogenesis by Upregulating the lncRNA SNHG7 in
HRECs. Next, we investigated the efects of Rg1 in HGinduced HRECs via the lncRNA SNHG7. Compared with the HG group, the cell viability was reduced in the Rg1 group, however, si-SNHG7 reversed the inhibitory efect of Rg1 on cell proliferation. Furthermore, compared with the Rg1+si-SNHG7 group, the cell viability was signifcantly reduced in the Rg1+si-SNHG7+miR-2116-5p inhibitor group (Figure 4(a)). Te results of qRT-PCR assay showed that compared with the HG group, the expression of SNHG7 was signifcantly increased and the expression of miR-2116-5p was signifcantly down-regulated in the Rg1 group, which was reversed by si-SNHG7. At the same time, compared with the Rg1+si-SNHG7 group, in the Rg1+si-SNHG7+miR-2116-5p inhibitor group, the expression of SNHG7 was upregulated and the expression of miR-2116-5p was downregulated (Figures 4(b) and 4(c)). Transwell and angiogenesis experiments showed that Rg1 treatment could effectively inhibit HG-induced cell proliferation and angiogenesis, while knockdown of SNHG7 could signifcantly attenuate the efect of Rg1. In addition, cotransfection of si-SNHG7+miR-2116-5p inhibitor could maintain the inhibitory efect of Rg1 on cell proliferation and angiogenesis to a certain extent (Figures 4(d) and 4(e)). Tese results suggest that Rg1 inhibits HG-induced HREC pathological phenomena through the upregulation of the lncRNA SNHG7.

Validation of the Targeting Relationship between miR-
2116-5p and SIRT3. StarBase online software was used to predict the miR-2116-5p-binding sites in SIRT3, and the results are shown in Figure 5(a). As verifed by the dual luciferase assays, miR-2116-5p reduced the activity of wild-type SIRT3 but had almost no efect on mutant SIRT3 ( Figure 5(b)). SIRT3 was decreased after the transfection of the miR-2116-5p mimic, and the transfection of the miR-2116-5p inhibitor increased the SIRT3 expression levels ( Figure 5(c)). At the expression level, SIRT3 was reduced under high glucose conditions and after the transfection of miR-2116-5p under normal glucose and HG conditions ( Figure 5(d)). Tus, these data illustrate that miR-2116-5p acts by targeting the negative regulation of SIRT3.

Rg1 Afects the Proliferation, Migration and Angiogenesis of HG-Induced HRECs via miR-2116-5p/SIRT3
. We further explored the efects of Rg1 via miR-2116-5p/SIRT3. Te results of CCK-8 assay showed that compared with the HG group, the cell viability of the Rg1 group was reduced, but the  transfection of miR-2116-5p mimic reversed the inhibitory efect of Rg1 on cell proliferation to a certain extent. In addition, compared with the Rg1 + miR-2116-5p mimic group, the Rg1 + miR-2116-5p mimic + pc-DNA SIRT3 group had lower cell proliferation activity (Figure 6(a)). Te results of qRT-PCR assay showed that compared with the HG group, Rg1 could inhibit the expression of miR-2116-5p and promote the expression of SIRT3, but this phenomenon was reversed by transfection of miR-2116-5p mimic.
Meanwhile, compared with the Rg1+miR-2116-5p mimic group, the expression of miR-2116-5p was down-regulated and the expression of SIRT3 was up-regulated in the Rg1+miR-2116-5p mimic + pc-DNA SIRT3 group (Figures 6(b) and 6(c)). Te results of Transwell and angiogenesis assays showed that the inhibitory efect of Rg1 on cell proliferation and angiogenesis could be reversed by transfection of miR-2116-5p mimic, but co-transfection of miR-2116-5p mimic + pc-DNA can maintain the inhibitory efect of Rg1 on cell proliferation and angiogenesis to a certain extent (Figures 6(d) and 6(e)). Similarly, Western blot detection results showed that the promoting efect of Rg1 on SIRT3 expression and the inhibitory efect of VEGF expression were reversed by the transfection of miR-2116-5p mimic, but the transfection of miR-2116-5p mimic + PC-DNA SIRT3 maintained this efect of Rg1 to a certain extent ( Figure 6(f )). Tus, these fndings demonstrate that Rg1 afects the proliferation, migration, and angiogenesis of HGinduced HRECs via miR-2116-5p/SIRT3. Verifcation of the relationship between miR-2116-5p and SIRT3 by a dual luciferase reporter assay, * P < 0.05 compared to the NC mimic group. (c) Analysis of the transfection efciency of SIRT3 by qRT-PCR, * * * P < 0.001 compared to the NC mimic group; ## P < 0.01 compared to the NC inhibitor group. (d) Te expression of SIRT3 was detected by qRT-PCR. * * P < 0.01 compared to the NC group; ### P < 0.001 compared to the NC-HG group. 12 Journal of Oncology

Discussion
Retinopathy caused by diabetes is a serious ocular complication that mainly manifests as retinal endocrine and hematological damage [31]. Hyperglycemia and hyperlipidemia are direct factors in the development of DR [32]. Endothelial cell damage caused by HG is one of the main clinical features of DR; therefore, endothelial cell activity regulation-related molecules are considered to play a key role in the pathogenesis of DR [33]. In diabetes modeling, higher blood glucose concentrations are an important marker of success [34]. During the construction of the diabetes model in this study, we successfully constructed a diabetic rat model because the blood glucose concentration of the rats was higher than 16.7 mmol/L. One month after the onset of diabetes, peripapillary cell degeneration, retinal thickness, and retinal apoptosis were reduced in the diabetic rats [35]. In this study, the pathological features of the retinal tissue in the rats with diabetes mellitus were described, and new blood vessels were observed in the diabetic retina [36]. Moreover, there was a signifcant increase in cellular angiogenesis in HRECs under HG induction and a signifcant increase in cell viability and migration. As a key factor in diabetes, Rg1 can protect molecules from damage. In diabetic rats treated with Rg1, cardiomyocyte apoptosis is inhibited, and caspase 3 expression is downregulated [37]. In the present study, Rg1 had a protective efect on the retina of DR rats and HRECs under HG induction. Regarding the gene expression level, Rg1 increased SIRT3 but decreased VEGF in rat retinal tissue and inhibited HRECs proliferation, migration and angiogenesis. It is consistent with the fnding by Gao et al. [17] that Rg1 can prevent DR by reducing apoptosis.
Te lncRNA SNHG7 was reduced in HRMECs under HG stimulation, and lncRNA SNHG7 overexpression inhibited HG-induced pathological phenomena (cell migration, proliferation and angiogenesis) by regulating the miR-543/SIRT1 axis [20]. Our study also demonstrated that HG conditions downregulated SNHG7 and its inhibitory efect on HG-induced pathological phenomena. Tere is a targeted binding site between SNHG7 and miR-2116-5p, and the inhibition of miR-2116-5p can efectively attenuate the efect of knockdown of SNHG7 on the proliferation and angiogenesis of RG1 cells. Furthermore, we found that the target of miR-2116-5p is SIRT3. As expected, as a downstream pathway of SNHG7, miR-2116-5p/SIRT3 mediated its protective efect on HRECs, while Rg1 functioned by upregulating SNHG7 to regulate the miR-2116-5p/SIRT3 axis. As a result, these fndings show that SIRT3 may play a role in regulating neovascularization [29]. Te overexpression of SIRT3 has been shown to inhibit retinal neovascularization under HG and insulin-induced conditions [29]. Our study found that overexpression of SIRT3 could reverse the promoting efect of miR-2116-5p on angiogenesis, which also indicated that SIRT3 could inhibit angiogenesis. In this study, SIRT3 was signifcantly reduced after the development of DR. VEGF can maintain ocular vascular integrity, and its expression is low and necessary in normal healthy eyes [38]. However, in DR, the levels of VEGF are higher than normal in cells and body fuids. Elevated VEGF levels alter capillary permeability, leading to retinal neovascularization, retinal vascular hemorrhage, exudation and increased angiogenesis and visual impairment. Importantly, inhibiting the expression of VEGF can inhibit the formation of retinal neovascularization [39]. Tis study shows that in the DR model, VEGF expression was (f ) Analysis of the SIRT3 and VEGF protein levels by a western blot analysis. * * * P < 0.001 and * * * * P < 0.0001 compared to the HG group; ## P < 0.01 and #### P < 0.0001 compared Rg1 group; △ P < 0.05, △△ P < 0.01 and △△△△ P < 0.0001 compared to the Rg1+miR-2116-5p mimic group.
increased. However, VEGF can be inhibited by SIRT3 overexpression, which may afect the formation of new blood vessels in the retina by regulating VEGF expression to protect against retinal injury.
In summary, the present study investigated the molecular mechanisms related to the alleviation of DR by Rg1. We demonstrated that Rg1 inhibits HG-induced cell proliferation, migration and angiogenesis and VEGF expression in retinal endothelial cells through the lncRNA SNHG7/ miR-2116-5p/SIRT3 axis. Tese fndings provide a theoretical basis for the clinical use of Rg1 for the treatment of DR. In addition, our study has the limitation of not verifying our molecular mechanism in vivo experiments. In the next study, we will verify that Rg1 alleviates DR through the lncRNA SNHG7/miR-2116-5p/SIRT3 axis in animal experiments.

Data Availability
Te datasets used and/or analyzed during the current study are available from the corresponding author upon reasonable request.

Disclosure
Liping Xue and Min Hu are the co-frst authors.

Conflicts of Interest
Te authors declare that they have no competing interests.