Ginsenoside Rg1 ameliorates the senescence and neurite injury of retinal ganglion cells in DR via targeting VDR and promoting mitochondrial biogenesis

Abstract

Background

Injury of retinal ganglion cells (RGCs) is one of the earliest signs of diabetic retinopathy (DR), preceding retinal microvascular abnormalities. Driven by metabolic and biochemical cascades, diabetes-dependent senescence in the retinal neural cells is responsible for neurodegeneration and subsequent permanent visual loss. This study investigated the involvement of ginsenoside Rg1 (Rg1) in neuropathy associated with DR to identify a possible therapeutic target.

Methods

The anti-aging and synaptogenesis effects, and neuroprotective mechanism of Rg1 were investigated in high glucose-induced RGCs and STZ-induced DR mice.

Results

Rg1 effectively reduced the β-galactosidase activity, promoted the neurite outgrowth, and reversed the expression of senescence and synaptic development-related proteins. Mechanistically, the compromised mitochondrial biogenesis induced by hyperglycaemia manifested as a critical driver of functional and structural impairments in RGCs. Meanwhile, Rg1 interacts with VDR to potentiate transcription of PGC-1α via the VDR/cAMP/PKA/CREB pathway. Activation of PGC-1α by Rg1 revitalized hyperglycaemia-hampered mitochondrial biogenesis, and resultantly alleviated senescence and neurite outgrowth inhibition of RGCs both in vitro and in vivo models.

Conclusion

Rg1 ameliorates neuropathy of DR by activating VDR non-genomic pathway and facilitating mitochondrial biogenesis. These results suggest a therapeutic approach for mitigating neurodegeneration in early DR, and provide insights into the potential clinical application of VDR agonism with Rg1 in regulating mitochondrial quality control.

Graphical abstract

Highlights

• •Compromised mitochondrial biogenesis triggers RGCs senescence and neurite injury.
• •Rg1 targets VDR to increase PGC-1α level and accelerate mitochondrial biogenesis by VDR/cAMP/PKA/CREB pathway.
• •Rg1 ameliorates RGCs senescence and neurite injury of DR.

1. Introduction

Diabetic retinopathy (DR) continues to be a prevalent sight-threatening complication in individuals with chronic hyperglycaemia, characterized by a multifaceted pathogenesis and limited therapeutic options [1]. Traditionally categorized as a microvascular complication, recent advancements in understanding retinal pathology have highlighted the significance of neuropathy, prompting the American Diabetes Association to recognize DR as a neurovascular complication [2,3]. Retinal ganglion cells (RGCs) serve as the only projection neurons responsible for transmitting visual information from upstream retinal neurons to cortex [4]. As terminally differentiated neurons, the absence of endogenous regenerative capacity results in synaptic degeneration or loss of RGCs, contributing to the irreversible visual functional defects in DR [5]. Currently, the available clinical treatments for neuroprotection in DR are constrained. RGCs exhibit senescence prior to irreparable cell death in mouse models of ischemic retinopathy and DR, as well as in patients suffering from proliferative diabetic retinopathy (PDR) [6,7]. Notably, the senescence-associated secretory phenotype (SASP) can drive paracrine senescence and exacerbate aberrant angiogenesis. Therefore, the establishment of neuroprotective and anti-aging therapies for RGCs disorders remains an emergent task for preventing vision impairments in patients with DR.

Mitochondrial biogenesis is a critical process of mitochondrial quality control [8]. In the mitochondrial DNA (mtDNA) mutator-initiated mouse model of progeria, activation of mitochondrial biogenesis has been shown to mitigate the systemic aging phenotype [9]. The continuous and intensive energy demand renders RGCs reliant on well-coordinated bioenergetics, largely depending on mitochondrial biogenesis [10]. The compromised mitochondrial biogenesis and subsequent dysfunction have been implicated both in the animal models and clinical samples of DR [11]. Furthermore, inadequate mitochondrial repair caused by high glucose accelerates senescence and triggers the apoptosis of peripheral neurons in diabetic neuropathy [12,13]. Further investigations are warranted to identify novel therapeutic approaches aimed at treating DR by revitalizing mitochondrial biogenesis in RGCs.

Ginsenoside Rg1 (Rg1) is a tetracyclic triterpene, isolated from Panax notoginseng (Burk.) F. H. Chen, with abundant activities in neurological disorders [14]. Previous pharmacological studies have indicated that Rg1 attenuates synaptic neurodegeneration and apoptosis of RGCs, excessive proliferation of endothelial cells, and fibrosis of Müller cells in DR [[15], [16], [17], [18]]. Our prior study has revealed that Rg1 promotes neurite growth of RGCs through activating the cAMP/PKA/CREB pathways and strengthening glycolysis under physiological conditions [19], whereas little is known about the efficacy on neural aging and the molecular targets for treating DR. Consequently, we elucidated the protective effects of Rg1 on RGCs from the perspective of mitochondrial biogenesis, with the goal of identifying potential therapeutic targets and developing pharmacological intervention strategies against DR.

2. Materials and methods

2.1. Animals

8-weeks-old male C57BL/6J mice weighing 20–22 g were purchased from Shanghai Sippr-BK Laboratory Animal Co. Ltd. (Shanghai, China). With access to drinking water and standard chow ad libitum, the mice were housed under standard conditions (12:12h light-dark cycle at 22–24 °C). A review and approval process of the animal studies was conducted by the Institutional Animal Care and Use Committee, and Institutional Ethics Committee of China Pharmaceutical University (approval number: 2023-08-015). All animal studies were conducted in accordance with the guidelines set by the National Institutes of Health.

Mouse model with DR was induced by intraperitoneal injection of STZ. After fasting for 10 h, STZ (55 mg/kg) was dissolved in sterile citrate buffer (0.01 M, pH 4.5) and injected intraperitoneally into C57BL/6J mice for 5 consecutive days. The concentration of serum glucose was measured with a glucometer (Roche Diagnostics, Mannheim, Germany) 7 days after injection. Mice with fasting blood glucose levels above 16.7 mM and metabolic syndrome were considered diabetic.

In the group (A), mice receiving equal solvent, with blood glucose levels below 6 mM, served as controls. The diabetic mice were randomly divided into four groups: (B) DR model, (C) DR+1 % Rg1, (D) DR+1 % Rg1+1β, 25(OH)₂D₃, (E) DR+1α, 25(OH)₂D₃ respectively. Rg1 (0.1 g) was dissolved adequately in 10 mL of normal saline under aseptic conditions as Rg1 eye drops. One week after the last STZ injection, groups C and D were topically treated with 10 μl 1 % (w/v) ginsenoside Rg1 eye drops every day consecutively for three months. After anesthesia, 1 μl of 1β, 25(OH)₂D₃ (100 μM) and 1 μl of 1α, 25(OH)₂D₃ (100 μM) were injected into vitreous of mice in groups D and E using 33-gauge needle twice a week, separately. In both the normal and DR groups, mice received an equal volume of vehicle eye drops. Three months after the first administration, samples of retinal tissue were harvested. The duration and dosages of eye drops and other drugs were conducted according to our previous report [19].

2.2. Isolation and culture of primary RGCs

The retinas of day 2–4 mice were dissected under aseptic conditions, and the retinas were ground on an 80-mesh stainless steel sieve after dissection. The fragments were digested with trypsin and collagenase II for approximately 25 min at 37 °C. After neutralization, the cells were passed through a 40 μm cell strainer. The cells were separated by centrifugation and resuspended in neuron-conditioned medium (Gibco, CA, USA) with B27 (1x, Invitrogen, CA, USA), 1 % (v/v) penicillin/streptomycin and 5 % (v/v) fetal bovine serum, and plated on a poly-D-lysine (0.1 mg/mL) coated cell plate. After 24 h, 10 μM arabinoside (Sigma) was added to inhibit non-neuronal growth. The cell purity was determined by immunofluorescence staining. The neuronal characteristics of primary RGCs were confirmed by positive signals of Thy1.1 (Fig. S1A). The cells were negative for markers of Müller cells (GFAP), endothelial cells (CD31) and microglia (Iba1) (Figure S1A). The duration of high glucose or mannitol treatment for primary RGCs was 48 h.

2.3. Western blotting

The retinal fragments, RGC-5s and primary RGCs were homogenized in RIPA lysis buffer with protease inhibitor complex (1x, Beyotime) on ice. Cytosolic and nuclear proteins were isolated as described in the nucleoprotein extraction kit (Beyotime). Protein content was measured using BCA assay. Proteins were subjected to SDS-PAGE gel electrophoresis and then transferred to polyvinylidene difluoride (PVDF) membranes (Millipore, Billerica, MA, USA). The membranes were blocked with 5 % bovine serum albumin in TBST buffer, and incubated with indicated primary antibodies at 4 °C overnight. The chemiluminescent substrate system (Tanon, Shanghai, China) was used to acquire immunoreactive signals after incubation with species-specific secondary antibodies conjugated with HRP.

2.4. PCR

The total RNA from cells were extracted using Trizol reagent (Takara, Otsu, Japan) according to the manufacturer's instructions. RNA samples were reverse-transcribed to cDNA using HiScript II Q RT SuperMix (Vazyme, Nanjing, Jiangsu, China). Real-time qPCR was performed using Ace qPCR SYBR Green Master Mix (Vazyme, Nanjing, Jiangsu, China) with a StepOne Real-Time PCR System (Applied Biosystems). The primer sequences are as follows: ND4 (mtDNA): for 5’ -ATTATTATTACCCGATGAGGGAACC - 3′, rev 5’ - ATTAAGATGAGGGCAATTAGCAGT - 3’; GAPDH: for 5’ - GTGAGGGAGATGCTCAGTGT - 3′, rev 5’ - CTGGCATTGCTCTCAATGAC - 3’.

2.5. Affinity ultrafiltration mass spectrometry assay

Affinity ultrafiltration mass spectrometry analyses were performed as described [20]. 30 μM Rg1 was incubated with 100 mg/mL recombinant human VDR protein (Cusabio, Wuhan, China) in PBS buffer at 37 °C for 1 h. Meanwhile, the inactivated protein (boiled for 5 min) was served as the negative control. The binding complex was fractionated using a 10 kDa cut-off ultrafiltration membrane (Millipore). After washing with ice-cold PBS, Rg1 with specific binding to VDR was separated by 80 % ice acetonitrile and analyzed using a triple quadrupole mass spectrometry tandem ultra high liquid chromatography (Waters, MC, USA).

2.6. Cellular thermal shift assay

For the cellular thermal shift assay (CETSA), RGC-5s were lysed by a mammalian active protein extraction reagent (Beyotime). The cell lysates were incubated with 30 μM Rg1 or DMSO for 30 min at 37 °C, and then heated at each temperature point (37–70 °C) in metal bath for 3 min. After centrifugation, the lysates were analyzed by Western blotting.

2.7. Drug affinity responsive target stability assay

For the drug affinity responsive target stability assay (DARTS), the RGC-5s were lysed by a mammalian active protein extraction reagent (Beyotime), and mixed with the gradient concentrations of Rg1 for 1 h at 37 °C. Then, pronase (5 mg/mL) was added and incubated for 15 min, and the lysates were analyzed by Western blotting.

2.8. Microscale thermophoresis assay

The binding affinity of Rg1 for VDR was determined using microscale thermophoresis assay (MST). HEK293T cells were transfected with the plasmids encoding EGFP-tagged VDR (GenePharma) and lysed by a mammalian active protein extraction reagent (Beyotime). The lysate cell supernatant was collected for MST experiments, and analyzed by NT.115 NanoTemper Monolith Instrument (NanoTemper, Germany).

2.9. Statistical analysis

The data were presented as the mean ± standard error of the mean (SEM). All statistical analyses and plots were generated using GraphPad Prism software (San Diego, CA, USA). Statistical variation between two groups was analyzed the two-tailed Student's t-test. Multi-group comparisons were evaluated utilizing one-way ANOVA with post-hoc test. p-value <0.05 was considered statistically significant.

3. Results

3.1. Ginsenoside Rg1 alleviated high glucose-induced cellular senescence and impaired neurite outgrowth of RGCs

High glucose treatment, as opposed to mannitol, increased the number of senescence-associated β-galactosidase (SA-β-gal) staining positive cells in RGC-5s and primary RGCs (Figure S1 B-C and F-G). Both RGC-5s and primary RGCs exhibited impaired neurite outgrowth and complexity when exposed to high glucose concentrations (Figure S1 D-E and H-I). Following a screening of the primary saponins of Panax notoginseng, Rg1 was selected for further comprehensive investigation, due to its superior apparent potency (Fig. 1A–D). Rg1 reduced SA-β-gal staining and enhanced neurite growth of RGCs in a concentration-dependent manner (Fig. 1E–H and Figure S1 J-M). Rg1 reversed high glucose-driven overexpression of p16, p21, and p53 as well as the loss of Lamin B in RGC-5s and primary RGCs (Fig. 1I and J, Figure S2 A-D and G-J). The synaptic growth-related regulatory proteins (Pax 6 and GAP-43) were downregulated by high glucose and subsequently normalized by Rg1 (Fig. 1I and J, Figure S2 E-F and K-L). Overall, Rg1 effectively prevented high glucose-mediated senescence and neurite growth damage of RGCs in vitro.

Fig. 1.

Ginsenoside Rg1 alleviated high glucose-induced cellular senescence and impaired neurite outgrowth of RGCs. (A–B) SA-β-gal staining in RGC-5s treated with high glucose and different ginsenosides (n = 5). Scale bar = 120 μm. (C–D) Neurite growth in RGC-5s treated with high glucose and a variety of different ginsenosides. The concentrations of saponins used in Fig. 1A–D were 30 μM (n = 5). Scale bar = 60 μm. (E–F) SA-β-gal activity in RGC-5s treated with high glucose and Rg1 (3, 10, and 30 μM) (n = 5). Scale bar = 120 μm. (G–H) RGC-5s were observed under microscope after treatment with high glucose and Rg1 (n = 5). Scale bar = 60 μm. (I) Representative blots of p16, p21, p53, Lamin B, Pax 6, and GAP-43 expressions in RGC-5s (n = 5). (J) The proteins levels of p16, p21, p53, Lamin B, Pax 6, and GAP-43 in primary RGCs (n = 5). ##p < 0.01 vs. the control, ∗∗p < 0.01 vs. high glucose group.

3.2. Ginsenoside Rg1 reinforced high glucose-impeded mitochondrial biogenesis of RGCs

High glucose induced the opening of mitochondrial permeability transition pore (MPTP), leading to a reduction in mitochondrial membrane potential (MMP) and ATP synthesis (Fig. 2A–E). Rg1 concentration-dependently suppressed MPTP opening, rescued MPP collapse, and augmented ATP levels in RGC-5s (Fig. 2A–E). High glucose contributed considerably to diminishing mitochondrial biomass and mtDNA copies in RGC-5s, which was subsequently relieved by Rg1 in a concentration-dependent manner (Fig. 2F and G). In high glucose-treated RGC-5, the therapeutic impact of Rg1 on sustaining mitochondrial biogenesis was also recapitulated through the detection of mitochondrial biogenesis-related proteins (COX Ⅳ, TFAM, NRF1, and PGC-1α) in a concentration-dependent manner (Fig. 2H and I). Collectively, Rg1 could potentiate mitochondrial biogenesis hampered by high glucose, ultimately mitigating mitochondrial dysfunction in RGCs.

Fig. 2.

Ginsenoside Rg1 reinforced high glucose-impeded mitochondrial biogenesis of RGCs. (A–B) TMRE staining in RGC-5s treated with high glucose and Rg1 (n = 5). Scale bar = 120 μm. (C–D) MPTP opening in RGC-5s (n = 5). Scale bar = 120 μm. (E) The ATP content was measured by ATP assay kit (n = 5). (F) Mitochondrial biomass was evaluated by Mito Track (n = 5). (G) qPCR analysis of mtochondrial DNA copy number over nuclear DNA (mtDNA/GAPDH) (n = 5). (H–L) Representative blots of COX Ⅳ, TFAM, NRF1 and PGC-1α, and statistical graphs of relative protein expression in RGC-5s (n = 5). ##p < 0.01 vs. the control, ∗p < 0.05, ∗∗p < 0.01 vs. high glucose group.

3.3. PGC-1α-inducted mitochondrial biogenesis was required for the neuroprotective efficacy of ginsenoside Rg1 in RGCs

In RGC-5s, Rg1 and PGC-1α activator, ZLN-005, reinforced high glucose-damaged mitochondrial biogenesis to preserve mitochondrial function, whereas the PGC-1α inhibitor SR-18292 limited these effects of Rg1 (Fig. 3A–H and Figure S3 A-C). Rg1 reduced the population of senescent cells stained with SA-β-gal and increased neurite length and branch number, which was largely abrogated by SR-18292 (Fig. 3I–L). Concordantly, Rg1 and ZLN-005 counteracted the changes in expression of cellular senescence-related proteins and neurite growth-related proteins after high glucose challenge (Fig. 3H and Figure S3 D-I). Conversely, the effects of Rd were partially abolished by SR-18292 (Fig. 3H and Figure S3 D-I). In summary, Rg1 might strengthen mitochondrial biogenesis, and resultantly protected RGCs from high glucose-induced cellular senescence and neurite degeneration by boosting PGC-1α expression.

Fig. 3.

PGC-1α-inducted mitochondrial biogenesis was required for the neuroprotective efficacy of ginsenoside Rg1 in RGCs. (A–D) MMP and MPTP opening in RGC-5s treated with high glucose, 30 μM Rg1, 5 μM ZLN-005, 10 μM SR-18292 (n = 5). Scale bar = 120 μm. (E–G) Mitochondrial function was expressed as ATP level, mitochondrial biomass and mitochondrial copy number (n = 5). (H) Expression levels of COX Ⅳ, TFAM, NRF1, p16, p21, p53, Lamin B, Pax 6, and GAP-43 (n = 5). (I–L) SA-β-gal staining and neurite growth in RGC-5s (n = 5). Scale bar on picture I = 120 μm, Scale bar on picture K = 60 μm ##p < 0.01 vs. the control, ∗p < 0.05, ∗∗p < 0.01 vs. high glucose group or indicated group.

3.4. Ginsenoside Rg1 promoted PGC-1α transcription through activating the VDR/cAMP/PKA/CREB pathway in RGCs

Based on the findings of our previous study, the upregulation of Rac1, Pax 6, and GAP-43 in Rg1-treated RGC-5s was observed to be mediated by the cAMP/PKA/CREB pathway [19]. Given the role of phosphorylated CREB in initiating the transcription of PGC-1α [10], it was plausible that Rg1 activated the cAMP/PKA/CREB pathway to enhance PGC-1α expression. The ability of Rg1 to magnify PGC-1α levels was attenuated by adenylate cyclase inhibitor (SQ22536), cAMP competitive antagonist (Rp-cAMPs), PKA inhibitor (H-89), and CREB inhibitor (666-15) (Fig. 4A–C and Figure S4 A-C). There was a notable inhibition of Rg1-induced CREB and PKA phosphorylation in RGC-5s following treatment with the corresponding inhibitors (Fig. 4A–C and Figure S4 A-C). As the positive control, cAMP analogue 8-Br-cAMP and adenylate cyclase activator forscolin exhibited a similar effect on PGC-1α, CREB and PKA compared to Rg1 (Fig. 4A–C and Figure S4 A-C). The phosphorylation of CREB was accompanied by nuclear accumulation of CREB, and CHIP assay results indicated that CREB bound to the CRE region of the PGC-1α promoter upon Rg1 stimulation (Fig. 4D and E and Figure S4 D-F).

Fig. 4.

Ginsenoside Rg1 promoted PGC-1α transcription through activating the VDR/cAMP/PKA/CREB pathway in RGCs. (A–C) The PGC-1α, p-CREB, CREB, p-PKA and PKA level in RGC-5s treated with high glucose, 30 μM Rg1, 100 μM SQ22536, 10 μM Rp-cAMPs, 10 μM H-89, 1 μM 666-15, 100 μM 8-Br-cAMP and 10 μM forscolin (n = 5). (D) Immunoblotting analysis of nuclear and cytosolic CREB in RGC-5s stimulated with high glucose, 30 μM Rg1, 500 nM 1β, 25(OH)₂D₃ and 100 nM 1α, 25(OH)₂D₃ (n = 5). (E) Analysis of the effect of CREB on PGC-1α promoter by ChIP-qPCR (n = 5). (F–G) PGC-1α, p-CREB, CREB, p-PKA and PKA expressions were detected by Western blotting (n = 5). (H–I) The cAMP levels were analyzed in RGC-5s (n = 5). (J) The affinity between Rg1 and VDR was tested by HPLC-QQQ-MS/MS (n = 5). (K–L) The thermal stability of VDR in RGC-5s treated with vehicle and 30 μM Rg1 (n = 5). (M) The stability of VDR against enzymatic hydrolysis (n = 5). (N) The binding affinity of Rg1 with VDR was determined by the MST (n = 5). #p < 0.05, ##p < 0.01 vs. the control, ∗p < 0.05, ∗∗p < 0.01 vs. Rg1 treatment group or indicated group.

Deficiencies of vitamin D was related to the development of diabetic complications and neurodegenerative diseases [21]. In addition to the genomic pathway, vitamin D receptor (VDR) in membrane elicited non-genomic actions manifested as the activation of signaling molecules [22]. Based on these related literature, we speculated that the neuroprotective activity of Rg1 might be dependent on VDR. Rg1 augmented the levels of PGC-1α, p-CREB, p-PKA and intracellular cAMP, which was regressed by the VDR inhibitor 1β, 25(OH)₂D₃ or knockdown of VDR (Fig. 4F–I and Figure S4 G-L). Furthermore, 1β, 25(OH)₂D₃ was able to preclude the action of Rd on the nuclear translocation and enrichment at PGC-1α promoter of CREB (Fig. 4D and E and Figure S4 D-F). Concomitantly, the induction of PGC-1α via VDR-triggered non-genomic pathway was observed in RGC-5s treated with the VDR agonist (1α, 25(OH)₂D₃) (Fig. 4 D–E, G, I, Figure S4 D-F, and J-L). Validations of siRNA targeting VDR were seen in Fig. S4 M. Together, these findings suggested that PGC-1α was transcriptionally upregulated by Rg1 through the VDR/cAMP/PKA/CREB pathway in RGCs.

3.5. VDR was identified as a possible target protein binding with ginsenoside Rg1

The analysis using affinity ultrafiltration mass spectrometry analysis revealed a higher concentration of Rg1 intercepted by ultrafiltration tube in the active VDR group, indirectly confirming the binding between VDR and Rg1 (Fig. 4 J). Rg1 enhanced the heat degradation and enzymatic hydrolysis resistance of VDR, as demonstrated by the cellular thermal shift assay (CETSA) and drug affinity responsive target stability assay (DARTS) (Fig. 4K–M). The microscale thermophoresis assay (MST) showed that Rg1 bound to VDR with an equilibrium dissociation constant (Kd) of 9.25 μM (Fig. 4 N). Collectively, the investigation situated VDR as a potential target of Rg1 through a series of thermodynamic and kinetic experiments.

3.6. Ginsenoside Rg1 restored high glucose-induced cellular senescence and defective neurite outgrowth of RGCs via VDR/PGC-1α pathway

The VDR inhibitor 1β, 25(OH)₂D₃ and PKA inhibitor H-89 attenuated the protective effects of Rg1 on the loss of MMP, MPTP opening, ATP content, mitochondrial biomass and mtDNA copy number in RGC-5s (Fig. 5A–G). Additionally, Rg1 and 1α, 25(OH)₂D₃ potentiated the expressions of COX Ⅳ, TFAM, NRF1, and PGC-1α in RGC-5s and primary RGCs, which was restrained by 1β, 25(OH)₂D₃ (Fig. 5L and M and Figure S5 A-H). The beneficial influence of Rg1 on high glucose-induced cellular senescence and defective neurite outgrowth was compromised by 1β, 25(OH)₂D₃ in RGC-5s and primary RGCs (Fig. 5H–M and Figure S6 A-P). As a positive control, 1α, 25(OH)₂D₃ recapitulated the neuroprotection of Rg1 in RGC-5s and primary RGCs after high glucose challenge (Fig. 5H–M and Figure S6 A-P). These results suggested the function for VDR/PKA/CREB/PGC-1α pathway in the efficacy of Rg1 in reversing mitochondrial biogenesis-related senescence and neurite degeneration of RGCs.

Fig. 5.

Ginsenoside Rg1 restored high glucose-induced cellular senescence and defective neurite outgrowth of RGCs via VDR/PGC-1α pathway. (A–G) Mitochondrial function of RGC-5s was measured via MMP, MPTP opening, ATP level, mitochondrial biomass and mitochondrial copy number (n = 5). Scale bar = 120 μm. (H–K) The SA-β-gal activity and neurite morphology in RGC-5s (n = 5). Scale bar on picture H = 120 μm, Scale bar on picture J = 60 μm. (L–M) The COX Ⅳ, TFAM, NRF1, PGC-1α, p16, p21, p53, Lamin B, Pax 6, and GAP-43 levels in RGC-5s and primary RGCs (n = 5). ##p < 0.01 vs. the control, ∗p < 0.05, ∗∗p < 0.01 vs. high glucose group or indicated group.

3.7. Ginsenoside Rg1 ameliorated RGCs injury by targeting VDR and enhancing mitochondrial biogenesis in the retina of diabetic mice

Hyperglycaemia initiated the loss of RGCs and the decreased thickness of retina (Fig. 6A and B). The senescence of RGCs in DR retina was confirmed by the retinal SA-β-gal staining and senescence-related proteins (Fig. 6C and F and Figure S7 A-D). Hyperglycemia diminished the density of dendritic spine and nerve fibre, as well as the expression of synaptic growth-related proteins (Fig. 6D–F and Figure S7 E-F). The above detrimental alterations in retinal morphological structure, neural aging and neurite growth inhibition were eliminated by Rg1 (Fig. 6A–F and Figure S7 A-F). For in vivo mechanism validation, Rg1 expectedly increased the levels of COX Ⅳ, TFAM, NRF1, PGC-1α, p-PKA, and p-CREB under prolonged hyperglycemic conditions (Fig. 6G and H and Figure S7 G-L). However, the therapeutic effects of Rg1 was absent when administered concurrently with 1β, 25(OH)₂D₃ in the retinas with DR (Fig. 6A–H and Figure S7 A-L). The positive control 1α, 25(OH)₂D₃ revealed a comparable effect in revitalizing mitochondrial biogenesis and repairing neurologic damage of RGCs in DR (Fig. 6A–H and Figure S7 A-L). In summary, the results ascertained the possible role of Rg1 in ameliorating hampered mitochondrial biogenesis and subsequent RGCs injury by modulation of the VDR-activated PGC-1α pathway.

Fig. 6.

Ginsenoside Rg1 ameliorated RGCs injury by targeting VDR and enhancing mitochondrial biogenesis in the retina of diabetic mice. (A–B) Retinal section was stained with HE (n = 5). Scale bars: 20 μm. (C) Retinal section was stained with β-galactosidase (n = 5). Scale bar, 20 μm. (D) Retinal section was stained with Golgi-Cox (n = 5). Scale bar, 20 μm. (E) Retinal section was observed by SEM (n = 5). Up scale bar: 8 μm, down scale bar: 4 μm. (F–H) The levels of p16, p21, p53, Lamin B, Pax 6, GAP-43, COX IV, TFAM, NRF1, PGC-1α, PKA, p-PKA, CREB and p-CREB in retina (n = 5). ##p < 0.01 vs. the control, ∗∗p < 0.01 vs. DM group or indicated group.

4. Discussion

Retinal neurologic injury precedes any noticeable vascular lesions in DR, with degeneration of RGCs being identified as one of the earliest pathological events [23]. Regrettably, existing available treatments primarily target advanced vascular abnormalities, which has the drawbacks of destructiveness, drug resistance and lack of visual recovery [24]. The study not only unearths a previously unrecognized efficacy of Rg1 in attenuating DR, but also articulates that reversing the senescent phenotype of RGCs, which might serve as a potential early therapeutic strategy to prevent irreversible neurodegeneration and vision deficits.

There is growing recognition that the impact of diabetes mellitus broadly affects the retinal neural cells. The thicknesses of NFL and GCL decreases by 0.54 μm annually in diabetic patients without evidence of vascular pathology [25]. The neurologic impairment in DR consists of the delayed dark adaptation, contrast sensitivity, diminished hue discrimination and abnormal visual fields, ultimately manifesting a decreased vision-related quality of life [26]. Above retinal electrical dysfunction may be linked to the compromised synaptic function and plasticity of RGCs, amacrine cells, bipolar cells, rods and cones [2]. Despite the crucial role of senescence in numerous diabetic complications, there is a lack of investigations in DR. The endothelial senescence was caused mainly by the cGAS/STING activation, which in turn exacerbates retinal inflammation and capillary angiopathy in DR [27]. Additionally, Pink1-dependent mitophagy deterioration uncouples mitochondrial quality control, and leads to Müller cells senescence in DR [11]. Only one article reported the hyperglycemia-driven senescence of RGCs, which primarily examined the correlation between SASP-associated inflammatory cytokines and downstream neovascularization, while ignored the upstream regulatory pathways in the aging process of RGCs [6]. In this study, we address for the first time the detailed mechanisms leading to RGCs senescence from the perspective of mitochondrial quality control, and lay a foundation for pharmacological interventions.

Mitochondrial biogenesis is currently regarded as a potential target for the management of chronic metabolic disorders and aging-related diseases. The mitochondrion functions as a semi-self-replicating organelle in response to the environmental stressors, which is orchestrated by PGC-1α. Downstream of PGC-1α, NRF1 and NRF2 execute a synergistic effect with PGC-1α to magnify the transcription of TFAM and are responsible for facilitating the expression of nuclear genes encoding mitochondrial proteins [28,29]. Metformin has been shown to prevent methylglyoxal-induced retinal pigment epithelial cell death through PGC-1α-dependent mitochondrial biogenesis [30]. Activation of PGC-1α attenuates neuronal apoptosis and angiogenesis in both db/db mice and STZ-induced diabetic rats [31]. Similar advantageous impacts of targeting PGC-1α are evident in age-related macular degeneration, neurodegenerative disorders and cardiovascular senescence [32,33]. Simultaneously, Rg1 has been shown to enhance cardiac function by elevation of mitochondrial biogenesis in sepsis and diabetic cardiopathy [34,35]. Our findings suggest for the first time that PGC-1α-triggered mitochondrial biogenesis is a possible process required for restoring the high glucose-induced progressive functional decline of RGCs, and accordingly, with Rg1 effectively inhibiting senescence and revitalizing neurite growth. Investigating the imbalance and uncoupling of various mitochondrial quality control process in RGCs will provide new insights on neuropathy in DR and better comprehension with regards to the clinical applications of Rg1.

Apart from transcriptional regulation, post-translational modification pathways could affect mitochondrial biogenesis in neurons via AMPK/SIRT1-PGC-1α axis. Specifically, PGC-1α is directly phosphorylated by AMPK and deacetylated by SIRT1, thereby intensifying the mitochondrial biogenesis [10,36]. In the previous study, we observed that ginsenoside Rd, a structural analogue of Rg1, strengthened AMPK/SIRT1 interaction by elevating NAD⁺/NADH levels and LKB1 deacetylation in endothelial cells exposed to high glucose [37]. Ginsenoside Rd targeted GPR30, activated AMPK-driven fatty acid oxidation, and maintained mitochondrial redox homeostasis to restore vascular barrier in DR [38]. Additionally, several ginsenosides, including Rg1, are reported with the superior pharmacokinetic properties when administered topically [39]. Despite the valuable insights gained from the pharmacology and pharmacokinetics of Rg1, further study remains to be executed to investigate the intricate network involved in the ginsenosides-driven mitochondrial biogenesis in DR.

Numerous epidemiological and clinical studies have investigated the relationship between the incidence of DR and vitamin D deficiency. The widespread distribution of vitamin D-related metabolic enzymes and VDR throughout the retina provides a foundation for the regulation of VDR in ocular diseases [40]. Most clinical observations demonstrate a negative correlation between the development of DR and serum vitamin D levels in both type 1 and type 2 diabetes [41]. Some contrasting studies report that DR patients with vitamin D deficiency may experience increased severity of neuropathy and microangiopathy [40,42]. In high glucose-induced human retinal endothelial cells, vitamin D reduces the release of inflammatory cytokines and maintains the integrity of paracellular barrier [43]. Although vitamin D supplementation may have neuroprotective effects, the basic research regarding vitamin D and DR predominantly concentrates on microvascular complications [40,42]. However, the current mechanisms underlying the critical position of VDR pathway in DR remains too limited to tackle the unmet clinical demands.

Preliminary studies have been conducted to determine the effectiveness and mechanism of Rg1 in DR. Rg1 inhibits TLR4/NF-κB signaling pathway, upregulates SIRT3 expression, and then reduces high glucose-induced vascular growth factors and inflammatory cytokines in human retinal microvascular endothelial cells [17,44]. For retinal macroglial cells, high glucose initiates mesenchymal activation and fibrosis in Müller cells, a process attenuated by Rg1 through modulation of the miR-2113/RP11-982M15.8/Zeb1 pathway [18]. Intriguingly, hyperphosphorylated tau appears as a key contributor to synaptic neurodegeneration of RGCs in DR, whereas Rg1 can prevent synaptic injury by activating IRS-1/Akt/GSK3β signaling before progressive neural cell death [15]. The report fails to address the role of RGCs senescence in DR, focusing solely on synaptic dysfunction, which is incongruent with our research priorities. The point is that there are no target exploration and confirmation of Rg1 in the study. We firstly identify VDR as a potential target of Rg1 and decipher that the agonism at VDR may be indispensable for the anti-senescence and neurite repair of Rg1 mediated by mitochondrial biogenesis. Rg1, a pleiotropic compound, could interact with other nuclear receptors, including estrogen receptor and glucocorticoid receptor [45,46]. Consequently, future research on Rg1 should investigate the involvement of alternative receptors or signaling pathways beyond VDR, through comparative binding or functional assays with alternative receptors.

In conclusion, Rg1 significantly improved RGCs senescence and impaired neurite outgrowth both in vitro and in vivo models. The presence of hyperglycemia results in a decrease in PGC-1α expression, leading to dampen mitochondrial biogenesis, disrupted mitochondrial function, premature senescence and synaptic dysfunction in neurons. Furthermore, Rg1 exhibits potential as a VDR activator that increases transcription of PGC-1α through the VDR/cAMP/PKA/CREB pathway, and ameliorates mitochondrial biogenesis-related retinal neurodegeneration. Collectively, our findings provide an engaging dimension for understanding previously unappreciated aspects of how hyperglycemia alters retinal neuronal structure and function, and suggest that the VDR activated by Rg1 might be a novel therapeutic strategy with translational potential against early DR.

Author contribution

Ninghua Tan: Conception and design of study, Revising the manuscript critically for important intellectual content, Approval of the version of the manuscript to be published (the names of all authors must be listed), Zhen Wang: Conception and design of study, Approval of the version of the manuscript to be published (the names of all authors must be listed), Kai Tang: Conception and design of study, Acquisition of data, Formal analysis, Drafting the manuscript, Approval of the version of the manuscript to be published (the names of all authors must be listed), Congcong Huang: Conception and design of study, Acquisition of data, Formal analysis, Drafting the manuscript, Approval of the version of the manuscript to be published (the names of all authors must be listed), Zhengjie Huang: Formal analysis, Approval of the version of the manuscript to be published (the names of all authors must be listed), Zhilei Lin: Formal analysis, Approval of the version of the manuscript to be published (the names of all authors must be listed)

Data availability

All data supporting the findings of this study are available from the corresponding author on reasonable request.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: