Ginsenoside Rg1 alleviates HG-induced autophagy of retinal microvascular endothelial cells by activating ACE2/Ang1-7/Mas in vitro

Abstract

Diabetic retinopathy (DR) is highly related to blindness. Fortunately, Ginsenoside Rg1 (Rg1) can ameliorate retinal endothelial dysfunction, and this study thus aims to investigate its underlying mechanism under high glucose (HG) conditions. Human retinal microvascular endothelial cells (HRMECs) were exposed to HG, followed by 10 μM Rg1 treatment and cell viability detection. Protein expressions of LC3II/I and P62 as well as autophagosome formation were examined using Western blot and transmission electron microscopy, respectively, to unveil the effect of Rg1 on the autophagy of HG-exposed cells. Whether the ACE2/ Ang1-7/Mas axis participated in the Rg1-caused autophagy abatement as well as its upstream regulators were verified through treatment with Ang1-7, A779 or LY294002. Rg1 or Ang1-7 treatment enhanced cell viability, activated the ACE2/Ang1-7/Mas axis, downregulated LC3II/I, upregulated P62, and repressed autophagosome formation in HG-exposed HRMECs. However, the regulatory effects of Rg1 on cell viability, ACE2/Ang1-7/Mas axis and autophagy were reversed in the presence of either A779 or LY294002. Rg1 alleviates HG-induced autophagy of HRMECs through the ACE2/Ang1-7/Mas axis, mediated by the PI3K/Akt pathway, supporting the protective effect of Rg1 on retinal endothelium.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-025-29344-0.

Introduction

Diabetic retinopathy (DR) is a common microvasculature complication of diabetes mellitus, and a leading cause of diabetes-related vision damage and preventable blindness in adults¹. The global prevalence of DR is on the rise, with estimated 161 million cases of DR by 2045². It has been shown that the progression of DR occurs in parallel with hyperglycemia, blood-retinal barrier (BRB) leakage and neovascularization³. The key pathways damaging retina and glomerulus microvascular endothelial barrier in the development of DR mainly involve: hyperglycemia that induces retinal ischemia-hypoxia, oxidative stress, accumulation of inflammatory cells/cytokines, and activation of multiple metabolic pathways, thereby precipitating structural disruption of the endothelial barrier, increased microvascular permeability, and increased incidence of DR⁴. Although current strategies including laser-induced photocoagulation, vitreous surgery and anti-vascular endothelial growth factor (VEGF) therapy confer health benefits to DR patients, their therapeutic efficacy is limited by factors such as adverse effects and high economic costs^5,6. Therefore, more effective and safer drugs are urgently needed for DR patients.

Traditional Chinese medicine (TCM) has accumulated substantial experience in treating DR, and various herbs and TCM prescriptions have demonstrated efficacy^7,8. Ginsenosides, a class of triterpene saponins, are classified into three major categories, namely protopanaxatriols (PPTs; ginsenoside Rg1, Re, Rg2, Rh1 and Rf), protopanaxadiol (PPD; ginsenoside Rb1, Rb2, Rd, Rg3 and Rh2) and oleanolic acid derivates (ginsenoside Ro)⁹. Ro has already been shown to inhibit advanced glycation endproducts (AGE)-induced autophagy in retinal endothelial cells¹⁰. Ginsenoside Rg1 (Rg1), a famous traditional Chinese herbs derived from the root or stem of Panax ginseng¹¹, possesses strong anti-inflammatory and anti-oxidant properties, and has the potential to treat neurological disorders, liver diseases and diabetes mellitus^12–14. Recently, Rg1 has been evidenced to protect retinal permeability and attenuate retinal injury in mice with early DR¹⁵, and to exert protective effects on the function of human retinal endothelial cells^16,17. Given that hyperglycemia-induced retinal microvasculopathy, which can lead to lipid exudates, hemorrhage and retinal detachment, is central to the pathogenesis of DR¹⁸, it is of great significance to thoroughly elucidate the effect and regulatory mechanism of Rg1 on the dysfunction of retinal microvascular endothelial cells (RMECs) in DR.

In recent years, studies have shown that autophagy is closely associated with endothelial cell dysfunction in the context of high glucose (HG)^19,20. As a fundamental cellular catabolic process, autophagy is usually dysregulated in endothelial injury largely due to sustained inflammatory response and oxidative stress²¹. A previous study revealed that HG treatment increases Bax expression, decreases Bcl-2 expression, activates autophagy, elevates LC3B fluorescent spots, upregulates LC3B and Atg5 protein expressions, and downregulates p62 protein expression²². Further, phosphatidylinositol-3-kinase (PI3K)/protein kinase B (Akt)/mammalian rapamycin target protein (mTOR) pathway is inhibited after HG treatment in the retinal endothelial cells²³. PI3K/Akt/mTOR pathway is a canonical negative regulator of autophagy^24,25. Noteworthily, ginsenoside Rg1 activates PI3K/Akt phosphorylation to protect mesangial cells in diabetic nephropathy models²⁶. Also, Rg1 inhibits angiotensin II (Ang II)-induced renal injury via AMPK/mTOR/PI3K pathway²⁷, indicating that Rg1 has the potential to exert protective effects by regulating autophagy-related pathways. The activation of the protective arm of the renin-angiotensin system (RAS)-the angiotensin-converting enzyme 2 (ACE2)/angiotensin 1–7 (Ang1-7)/Mas axis has emerged as a novel strategy for the prevention or treatment of DR^28,29. The axis plays an important role in regulating autophagy: Ang-(1–7) prevents LPS-induced autophagy, and reduces α-synuclein aggregation by alleviating autophagy dysfunction in Parkinson’s disease^30,31. ACE2 has been reported to alleviate autophagy through the AMPK/mTOR pathway³². The ACE2/Ang(1-7)/MasR axis suppresses excessive inflammatory responses by metabolizing AngII, promoting autophagy, and accelerating the clearance of the NLRP3 inflammasome³³, highlighting its complex role in balancing inflammation and autophagy.

Accordingly, the present study hypothesized that Rg1 may attenuate retinal endothelial cell dysfunction in DR by reducing ACE2/Ang1-7/Mas axis-mediated autophagy.

Materials and methods

Reagent preparation

Rg1 (C₄₂H₇₂O₁₄, CAS No. 22427-39-0, HY-N0045) and Ang1-7 (HY-12403) were obtained from MedChemExpress (Monmouth Junction, NJ, USA). A Mas receptor antagonist, A779 (M9091), was provided by AbMole (Houston, TX, USA). A PI3K inhibitor, LY294002 (S1105), was purchased from Selleck Chemicals (Houston, TX, USA). Dimethyl sulfoxide (DMSO; D8418, Sigma-Aldrich, St Louis, MO, USA) was utilized to prepare Rg1 or LY294002 stock solution.

Cell culture and HG induction

Human RMECs (HRMECs; CL0157) were provided by Hunan Fenghui Biotechnology Co., Ltd. (Changsha, China). Cell culture was carried out using a complete medium (CM-H130, Procell Life Science& Technology Co., Ltd., Wuhan, China) at 37 °C with 5% CO₂. To mimic DR in vitro, HRMECs at the logarithmic growth phase were treated with 30 mM HG (D20903, Acmec Biochemical, Shanghai, China) at 37 °C for 48 h³⁴. HRMECs cultured in complete medium, treated with 5.5 mM normal glucose (NG), and exposed to 5.5 mM NG + 24.5 mannitol (MA, D93792, Acmec Biochemical, Shanghai, China) separately served as negative control and osmotic control.

Cell treatment and viability analysis

Following HG induction, HRMECs were incubated with 100 nmol/L Ang1-7, 1 μmol/L A779 or 50 μM LY294002 in the culture medium at 37 °C for 48 h, as previously described^35,36. To investigate the mechanism of Rg1 on DR, 48-h incubation with 10 μM Rg1 was performed on the HG-induced cells at 37 °C as per guidance³⁷.

After indicated cell treatments, methylthiazolyldiphenyl-tetrazolium bromide (MTT) assay was used for cell viability analysis. Briefly, HRMECs (3 × 10³ cells/well) were seeded in plates and cultured at 37 °C overnight before reaction with 10 μL MTT solution (AC12966, Acmec Biochemical, China) for 4 h. Next, the formazan product was dissolved with DMSO, and cell absorbance (570 nm) was detected using a Spark microplate reader (Tecan, Männedorf, Switzerland).

Flow cytometry

The collected cells were washed twice with PBS, then stained with 5 μL Annexin V/FITC and propyl iodide (PI) at room temperature, and re-suspended with 0.5 mL binding buffer. Apoptosis was detected on FACScan flow cytometry using CellQuest software (version 5.1, BD Biosciences, New Jersey, NJ, USA). The apoptotic rate was calculated as follows: Apoptotic rate (%) = early + late apoptosis.

Enzyme-linked immunosorbent assay (ELISA)

The human ELISA kits Ang1-7 (CSB-E14242h, CUSABIO, Wuhan, China), TNF-α (CSB-E04736h-IS, CUSABIO, Wuhan, China) and IL-6 (CSB-E04638h, CUSABIO, Wuhan, China) were employed to detect the levels of Ang1-7, TNF-α, and IL-6 in HRMECs undergoing indicated treatments. According to the manufacturer’s protocol, the cells were subjected to centrifugation (1,000 × g) at 4 °C for 15 min to obtain supernatant. Subsequently, 100 μL samples were maintained in each well of pre-coated ELISA plates at 37 °C for 2 h, followed by 1-h incubation with Biotin antibody. After washing with Wash Buffer, the samples were incubated with horseradish peroxidase (HRP) avidin at 37 °C for 1 h and TMB substrate without light for 30 min, in sequence. The absorbance in each well was measured at 450 nm by a microplate reader.

Western blot

HRMECs were homogenized in RIPA Lysis Buffer (C500005, Sangon Biotech, Shanghai, China) with protease and phosphatase inhibitors on ice for 10 min to extract total protein, followed by centrifugation (12,000 × g) at 4 °C for 5 min. Protein concentration was determined by BCA Protein Assay Kit (P0012S, Beyotime, Shanghai, China). 20 μg of total protein per sample was loaded in 10% SDS-PAGE gel under denaturation conditions, electrophoretically transferred to polyvinylidene fluoride membranes (FFP20, Beyotime, China), and blocked with 5% nonfat milk (6250, Sigma-Aldrich, USA) at room temperature for 1 h. Afterwards, primary antibodies against ACE2 (ab272500, 92 kDa, Abcam, Cambridge, UK), Mas (NBP1-78,444, 42 kDa, Novus Biologicals, Littleton, CO, USA), Akt (ab8805, 60 kDa, Abcam, UK), p-Akt (ab38449, 56 kDa, Abcam, UK), LC3 (ab192890, 14/16 kDa, Abcam, UK), P62 (ab109012, 62 kDa, Abcam, UK), eNOS (#9572, 140 kDa, CST, USA), p-eNOS (#9571, 140 kDa, CST, USA) and loading control GAPDH (ab8245, 37 kDa, Abcam, UK) were accordingly diluted with 1 × Tris buffered saline solution (T1085, Solarbio, Beijing, China). Then, the blots were incubated on the membranes at 4 °C overnight. The following day, the blots were probed with corresponding HRP-coupled secondary antibodies (31460/G-21040, Thermo Fisher, Waltham, MA, USA) at room temperature for 2 h, and then were visualized with chemiluminescence solution (P1000, Applygen, Beijing, China) in a Quantity One System image analyzer (Bio-Rad, Hercules, CA, USA).

Transmission electron microscopy (TEM)

HRMECs after indicated treatments were fixed with 2.5% glutaraldehyde (AP1126, Acmec Biochemical, China) at 4 °C overnight, and treated with 1% osmium tetroxide (1.24505, Sigma-Aldrich, USA) in phosphate-buffered saline (C0221A, Beyotime, China) at 4 °C for 2 h. After being rinsed with ddH₂O (B541017, Sangon Biotech, China), the cells were stained with 2% aqueous uranyl acetate (22,400–2, Avantor, Shanghai, China) for 2 h, followed by gradient dehydration and resin embedding. Next, the samples were cut into sections (70–90 nm thickness) using an ultramicrotome (UC6, Leica, Wetzlar, Germany). A transmission electron microscope (HT7800, Hitachi, Tokyo, Japan) was utilized to observe the microstructure of autophagosomes.

Statistical analysis

The data were presented as mean ± standard deviation (SD) from independent experiments repeated at least 3 times. GraphPad Prism v8.0 (GraphPad Software Inc., San Diego, CA, USA) was utilized for statistical analysis. The Shapiro–Wilk test was applied to examine normal distribution, and the Brown-Forsythe’s test was employed to check the homogeneity of variances. Multi-group comparisons were conducted using one-way analysis of variance, followed by Tukey’s post hoc test. P < 0.05 was indicative of statistical significance.

Results

Ang1-7 treatment enhanced viability, decreased apoptosis, activated the ACE2/Ang1-7/Mas axis and inhibited autophagy in HG-induced HRMECs

We first exposed HRMECs to HG for 48 h to mimic retinal endothelial dysfunction during the progression of DR, and investigated whether Ang1-7 is involved in the regulation of retinal endothelial function. As demonstrated in Fig. 1A–G, the HG-treated cells showed poor viability, high apoptosis, and lower expressions of ACE2, Mas and Ang1-7, compared to the NG-treated cells (P < 0.001). However, there was no significant difference in the results between the MA group and the NG group, indicating that no effects were generated by the osmotic pressure (Fig. 1A–G). Of note, it was observed that 100 nmol/L Ang1-7 treatment enhanced viability, decreased apoptosis and increased expressions of ACE2, Mas and Ang1-7 in the HG-treated cells (Fig. 1A–G, P < 0.001). Additionally, the results of Western blot and TEM revealed that HG induction increased LC3II/LC3I expression, decreased P62 expression and promoted the formation of autophagosomes in HRMECs (figure F–L, P < 0.001), suggesting the enhancement of autophagy. Ang1-7 treatment reversed the effects of HG by downregulating LC3II/LC3I expression, upregulating P62 level and reducing autophagosomes (Fig. 1F–L, P < 0.001).

Fig. 1.

Effects of Ang1-7 on viability, the ACE2/Ang1-7/Mas axis and autophagy in HG-induced HRMECs. (A) HRMECs were subjected to 48-h incubation with 5.5, 5.5 mM NG + 24.5 mannitol and 30 mM glucose (NG, MA and HG groups), respectively, followed by 100 nmol/L Ang1-7 treatment (HG + Ang1-7 group). MTT assay was used to detect cell viability in the three groups. (B and C) Cell apoptosis was assessed with the flow cytometry experiment. (D–F) Western blot was performed to measure protein expressions of ACE2 and Mas in the three groups. (G) ELISA was utilized to detect Ang1-7 level in the three groups. (H–J) Western blot was applied to measure protein expressions of LC3II/LC3I and P62 in the three groups. (K) and (L) Microstructural detection of autophagosomes by transmission electron microscopy (scale bar = 200 µm) in the three groups, arrow: autophagosomes. Data are shown as mean ± standard deviation (SD) (n = 3 biological replicates). GAPDH was used as the loading control. ^***P < 0.001, versus NG; ⁺⁺⁺P < 0.001, versus HG. HRMECs, human retinal microvascular endothelial cells; NG, normal glucose; HG, high glucose; Ang1-7, angiotensin 1–7; ACE2, angiotensin-converting enzyme 2; MTT, methylthiazolyldiphenyl-tetrazolium bromide; ELISA, Enzyme-linked immunosorbent assay.

Rg1 protected HRMECs against HG-induced autophagy by activating the ACE2/Ang1-7/Mas axis

Next, we examined the underlying role of Rg1 in DR. In HG-induced HRMECs, Rg1 (10 μM) treatment enhanced viability (Fig. 2A, P < 0.001), decreased apoptosis (Fig. 2B and C, P < 0.001), increased ACE2, Mas and Ang1-7 expressions (Fig. 2D–G, P < 0.001), increased p-eNOS/eNOS level (Fig. 2H–I, P < 0.001), decreased TNF-α and IL-6 level (Fig. 2J–K, P < 0.001), as well as suppressed autophagy via upregulating P62 level, downregulating LC3II/LC3I ratio and blocking autophagosome formation (Fig. 2L–P, P < 0.001). Moreover, 1 μmol/L A779 (a Mas receptor antagonist) treatment enhanced the effect of HG (Fig. 2A–P, P < 0.001), which was determined to further impair viability, increase apoptosis, decrease expressions of ACE2, Mas, Ang1-7 and P62, promote LC3II/LC3I expression, and elevate autophagosomes (Fig. 2A–P, P < 0.001). However, these effects of A779 on HG-induced HRMECs were reversed in the presence of Rg1 (Fig. 2A–P, P < 0.05).

Fig. 2.

The protective role of Rg1 in HG-induced HRMECs through the ACE2/Ang1-7/Mas axis. (A) HRMECs were subjected to 48-h incubation with 5.5 and 30 mM glucose (NG and HG groups), respectively, followed by treatments with 1 μmol/L A779 (a Mas receptor antagonist) and/or 10 μM Rg1 (HG + A779, HG + Rg1 and HG + Rg1 + A779 groups). MTT assay was used to detect cell viability in the five groups. (B) and (C) Cell apoptosis was assessed with the flow cytometry experiment. (D–F) Western blot was performed to measure protein expressions of ACE2 and Mas in the five groups. (G) ELISA was utilized to detect Ang1-7 level in the five groups. (H) and (I) Western blot was conducted to measure protein expressions of p-eNOS and eNOS in the five groups. (J) and (K) ELISA was used to detect the levels of TNF-α, and IL-6. (L–N) Western blot was conducted to measure protein expressions of LC3II/LC3I and P62 in the five groups. (O) and (P) Microstructural detection of autophagosomes by transmission electron microscopy (scale bar = 200 µm) in the five groups, arrow: autophagosomes. Data are shown as mean ± standard deviation (n = 3 biological replicates). GAPDH was used as the loading control. ^***P < 0.001, versus NG; ⁺P < 0.05, ⁺⁺P < 0.01, ⁺⁺⁺P < 0.001, versus HG; ^#P < 0.05, ^###P < 0.001, versus HG + Rg1; ^{^^^}P < 0.001, versus HG + A779. Rg1, ginsenoside Rg1.

PI3K inhibition reversed the mitigating effect of Rg1 on HG-caused autophagy in HRMECs by blocking the ACE2/Ang1-7/Mas axis

To explore whether the PI3K/Akt pathway regulates the ACE2/Ang1-7/Mas axis in the process of Rg1 alleviating HG-caused dysfunction of HRMECs, we treated the cells with 50 μM LY294002 (a PI3K inhibitor) in the presence or absence of Rg1. Compared with HG group, LY294002 brought about diminished viability, increased apoptosis and decreased expression of p-Akt/Akt (Fig. 3A–E, P < 0.05), which, however, were offset by Rg1 (Fig. 3A–E, P < 0.01). As shown in Fig. 3F–L, LY294002 decreased expressions of ACE2, Mas, Ang1-7 and P62, while increasing the expression of LC3II/LC3I in the HG-treated cells (P < 0.001), which were all counteracted by Rg1 (P < 0.001). Besides, LY294002-induced increase of autophagosome formation in the HG-treated cells was abrogated by Rg1 (Fig. 3N, P < 0.001). Rescue experiments data unraveled that the effects of Rg1 on promoting viability and expressions of p-Akt/Akt, ACE2, Mas, Ang1-7 and P62, as well as on inhibiting the expression of LC3II/LC3I and the formation of autophagosomes were neutralized in the presence of LY294002 (Fig. 3A–N, P < 0.05).

Fig. 3.

Involvement of PI3K/Akt pathway in the mitigating effect of Rg1 on HG-caused dysfunctional HRMECs. (A) HRMECs were subjected to 48-h incubation with 30 mM glucose (HG group), followed by treatments with 10 μM Rg1 and/or 50 μM LY294002 (HG + Rg1, HG + LY294002 and HG + Rg1 + LY294002 groups). MTT assay was used to detect cell viability in the four groups. (B and C) Cell apoptosis was assessed with the flow cytometry experiment. (D–H) Western blot was performed to measure protein expressions of p-Akt/Akt, ACE2 and Mas in the four groups. (I) ELISA was utilized to detect Ang1-7 level in the four groups. (J–L) Western blot was conducted to measure protein expressions of LC3II/LC3I and P62 in the four groups. (M) and (N) Microstructural detection of autophagosomes by transmission electron microscopy (scale bar = 200 µm) in the five groups, arrow: autophagosomes. Data are shown as mean ± standard deviation (n = 3 biological replicates). GAPDH was used as the loading control. ⁺P < 0.05, ⁺⁺⁺P < 0.001, versus HG; ^#P < 0.05, ^###P < 0.001, versus HG + LY294002; ^{^^}P < 0.01, ^{^^^}P < 0.001, versus HG + Rg1. Akt, protein kinase B; p-Akt, phosphorylated Akt.

Discussion

DR is non-proliferative at the early stage; however, if left unchecked, it will progress to an advanced proliferative stage, leading to blindness³⁸. Given the widespread prevalence of diabetes mellitus, prevention and treatment of DR are of great significance to improve the quality of life for diabetics. Progressive vascular endothelial dysfunction is an important pathological feature in the development of DR¹⁸. Thus, we aimed to identify medicinal herb-derived compounds that can restore retinal endothelial function. Herein, we demonstrated that Rg1 alleviated HG-induced autophagy of HRMECs by activating the PI3K/Akt pathway-mediated ACE2/Ang1-7/Mas axis.

Ang1-7 is a biologically active product of ACE2 metabolizing Ang II in the renin-angiotensin system (RAS), a key regulator of blood pressure and hydroelectrolyte balance³⁹, and Ang1-7 can activate its downstream receptor Mas to protect epithelial cells from multifactor-induced oxidative stress, apoptosis and senescence^40,41. Reportedly, the activation of ACE2/Ang1-7 can suppress diabetes-induced retinal pathophysiologic changes, including thickening of basement membrane, vascular damage and oxidative damage, by balancing the RAS in the eyes of rats, which supports the therapeutic potential of Ang1-7 in DR management²⁸. In this study, we observed that HG exposure impaired viability of HRMECs accompanied by downregulation of ACE2, Ang1-7 and Mas, indicating the involvement of the ACE2/Ang1-7/Mas axis in HG-induced retinal endothelial injury. However, how this axis regulates vascular endothelial dysfunction in DR remains unclear. It is accepted that microvascular endothelial cells in the eyeballs are the primary target of hyperglycemic insult, and the regulatory role of autophagy in the permeability, migration and angiogenesis of vascular endothelial cells in DR has received increasing attention⁴². Reportedly, HG upregulates the expressions of LC3-II/I and Beclin-1 proteins, while reducing p62 protein levels⁴³; consistently, this study found the same tends in HRMECs, together with increased autophagosomes. LC3II is a well-established autophagosome marker highly expressed during the activation of autophagy, whereas P62 aggregation typically reflects impaired autophagy⁴⁴. Intriguingly, the Ang1-7/Mas axis, which acts as a novel player in neuro- and cardio-protection, is frequently associated with the regulation of autophagy^31,45. Through Ang1-7 treatment and A779-induced Mas inhibition, this study found that the activation of Ang1-7 enhanced viability and repressed autophagy of HG-exposed HRMECs, whereas A779 treatment yielded opposite results. Collectively, these findings suggested that the ACE2/Ang1-7/Mas axis protected HRMECs against hyperglycemia-induced dysfunction in DR through the inhibition of autophagy. However, reports regarding the specific role of autophagy in DR appear contradictory. Some studies showed decreased autophagy of HRMECs in the HG environment as reflected by upregulated P62 and downregulated LC3II/LC3I, Beclin-1 and LAMP1⁴⁶. Others have highlighted that DR is associated with hyperglycemia, but short- versus long-term hyperglycaemia differentially modulate autophagy and apoptosis—neither process is simply “activated” or “inhibited”, as both are characterized by effects that depend on the duration of exposure⁴⁷. These observations underscore the complexity of “rationally utilizing” the protective effects of apoptosis and autophagy (rather than simply activating or inhibiting them), warranting further investigation into context-specific regulatory strategies.

Accumulating evidence indicated that Rg1 relieves oxidative stress and inflammation in various tissue injuries by inhibiting 5′-monophosphate-activated protein kinase (AMPK)/mTOR-dependent autophagy^48,49. Although Rg1 has been recently documented to promote the functional restoration of retinal endothelial cells in DR via suppressing proliferation, migration and angiogenesis, little is known about whether the therapeutic effect of Rg1 on retinal endothelial dysfunction is realized by regulating autophagy. In HG-treated HRMECs, this study found that Rg1 induced the activation of ACE2/Ang1-7/Mas axis and repressed autophagy. Importantly, such effects were reversed by A779-induced Mas inhibition. These results demonstrated that Rg1 can restore retinal endothelial function to mitigate DR by suppressing autophagy via ACE2/Ang1-7/Mas axis. Autophagy is a complex process strictly modulated by multiple signaling pathways, initiation of which is inseparable from mTOR. The PI3K/Akt pathway, one of the up-stream regulators of mTOR⁵⁰, has previously been reported to participate in attenuating Alzheimer’s disease-related cognitive deficits through activating the ACE2/Ang1-7/Mas axis⁵¹. Yang et al. has underlined that inhibiting PI3K/Akt/mTOR pathway-mediated autophagy by trimetazidine can normalize the function of retinal endothelial cells and thus hamper the development of DR in vitro²³. In diabetic nephropathy, it is worth noting that Rg1 activates the PI3K/Akt pathway and promotes FOXO3 expression, thereby preventing HG-triggered damage of glomerular mesangial cells²⁶. To confirm whether Rg1 activates the ACE2/Ang1-7/Mas axis in HG-exposed HRMECs by the PI3K/Akt pathway, we inhibited the expression of PI3K in the cells using LY294002. According to our findings in this study, PI3K inhibition reversed the mitigating effect of Rg1 on HG-caused autophagy of HRMECs by blocking the ACE2/Ang1-7/Mas axis. However, a previous study showed that the transcription factor EB protects HRMECs from HG-induced endothelial injury by regulating the autophagy process and decreasing apoptosis, while the autophagy inhibitor chloroquine attenuates the protective effect of transcription factor EB overexpression⁴⁶, implying the dual role of autophagy. Interestingly, short-term HG impairs mitochondrial function, whereas long-term HG improves mitochondrial function by enhancing “mitophagy” (the clearance of damaged mitochondria), a process that relies on dynamic mitochondrial network regulation⁵². These findings hinted that the effect of autophagy on HREC in the HG environment is complex, and its specific mechanism of promoting apoptosis needs further clarification.

Notably, there are some limitations in this study. Firstly, although we included a mannitol osmotic control to confirm that the observed effects were specifically due to glucose metabolism rather than osmotic pressure changes, this control was not extended to all pharmacological intervention stages. Future studies should simultaneously set up osmotic pressure controls in all inhibitor-treated groups to provide a more irrefutable chain of evidence. Secondly, the current findings are derived from in vitro cell models and require further validation in diabetic animal models.

Conclusion

In summary, our in vitro findings demonstrate that the ACE2/Ang1-7/Mas axis plays a critical role in regulating HG-induced autophagy and dysfunction in HRMECs. We provide evidence that ginsenoside Rg1 protects HRMECs by suppressing excessive autophagy through activation of the ACE2/Ang1-7/Mas axis, a process modulated by the PI3K/Akt signaling pathway. These data illuminate a previously unexplored mechanism underlying the protective effect of Rg1 in a model of DR. However, it must be acknowledged that this study is limited to an in vitro setting. The physiological relevance and therapeutic potential of these findings require further validation in animal models and more complex systems.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

Yuting Chen: Conceptualization, writing—original draft preparation. Yamei Jin: Data curation, methodology, visualization, investigation, writing—reviewing and editing.

Funding

This research was supported by the Zhejiang Province Traditional Chinese Medicine Science and Technology Project (Grant number: 2023ZL412).

Data availability

The original data sustaining the conclusions of this study will be provided by the corresponding author, without undue reservation.

Declarations

Competing interests

The authors declare no competing interests.