PRMT5 Regulates Senescence in Retinal Ganglion Cells by Targeting the Wnt/β-Catenin Signaling Cascade

Abstract

Purpose

In patients with glaucoma, progressive degeneration of retinal ganglion cells (RGCs) leads to irreversible visual impairments. Despite recent studies indicating that senescence is associated with RGC death, the underlying molecular mechanisms remain unclear.

Methods

The chronic ocular hypertension (COH) mouse model was established by infusing a crosslinking hydrogel into the anterior chamber. Cellular senescence was evaluated using Western blot analysis, cell cycle, senescence-associated β-galactosidase (SA-β-gal) staining, enzyme-linked immunosorbent assay, and immunofluorescence. Functional experiments were conducted in retinal precursor (R28) cells through small interfering RNA–mediated knockdown and plasmid-mediated overexpression. Additionally, the role of the protein arginine methyltransferase 5 (PRMT5)–regulated Wnt/β-catenin pathway in RGC senescence was investigated via intravitreal injection of GSK3326595 and CHIR99021 in mice.

Results

We demonstrate that PRMT5 is markedly downregulated in RGC in a COH mouse model, correlating with increased RGC senescence induced by elevated intraocular pressure. Silencing PRMT5 significantly accelerated senescence, as evidenced by increased SA-β-gal activity, cell cycle arrest, and senescence marker upregulation. Cotreatment with GSK3β inhibitor CHIR99021 alleviated hypoxia-induced senescence and reactivated the Wnt/β-catenin pathway, while the antagonist FH535 negated the neuroprotective effects of PRMT5 overexpression. In vivo, the PRMT5 inhibitor GSK3326595 reduced RGC survival and heightened senescence markers, whereas CHIR99021 mitigated RGC loss and restored Wnt/β-catenin signaling.

Conclusions

Taken together, these findings highlight the critical role of the PRMT5-regulated Wnt/β-catenin pathway in RGC senescence and neurodegeneration. Targeting this pathway represents a promising therapeutic strategy for glaucoma.

Glaucoma is one of the leading causes of irreversible blindness and permanent visual impairment, primarily attributed to accelerated retinal ganglion cell (RGC) death, axonal degeneration, and optic nerve dysfunction.¹ With the elderly population increasing, the number of individuals affected by glaucoma is expected to reach around 111.8 million by 2040.² Numerous risk factors contribute to glaucoma onset, including elevated intraocular pressure (IOP), advancing age, and genetic predisposition; in particular, aging has been identified as one of the most significant, consistent risk factors for glaucoma.^3,4

Emerging evidence links glaucoma closely to the aging process. Recent studies have indicated that elevated oxidative stress, cellular senescence, and inflammation in aging retinas are significant risk factors for glaucoma and have thus become the focus of ongoing research.^5,6 Recent studies have demonstrated that aging retinas are more sensitive to mild IOP elevation than their younger counterparts, with aging RGCs expressing a larger number of aging-related markers, such as uPAR, p16^INK4a, and p19^ARF, in response to IOP-related stress.⁷ Notably, repeated stimulation of ocular hypertension accelerates the onset of senescence feature emergence in young retinas, accompanied by increased senescence-associated secretory phenotype (SASP) levels and age-related changes in DNA methylation.⁷ Therefore, understanding the mechanisms underlying RGC senescence in depth is critical for developing targeted therapeutic approaches for glaucoma.

Protein arginine methyltransferases (PRMTs) constitute a family of enzymes that covalently modify histone and nonhistone proteins, thereby playing critical roles in various physiological processes.⁸ PRMT5 is a type II PRMT, which specifically mediates the symmetrical dimethylation of arginine residues on various proteins.⁹ PRMT5 has been implicated in cell cycle regulation and aging processes in different cell types.^10,11 Banasavadi-Siddegowda et al.¹² and Yan et al.¹³ have reported that PRMT5 downregulation induces senescence and cell cycle arrest through the modulation of PTEN/AKT and EGFR/AKT/GSK3β signaling cascades, respectively. Moreover, PRMT5 depletion has been shown to promote cellular senescence by inducing DNA damage through the TXNIP/p21 axis.¹⁴ Despite these findings, the specific role of PRMT5 in RGC senescence and its potential contribution to glaucoma pathophysiology remain unclear.

In the present study, we report the first evidence of RGC senescence in a chronic ocular hypertension (COH) mouse model focusing on PRMT5. Our findings demonstrate that chronic IOP elevation leads to PRMT5 deficiency, which subsequently triggers RGC senescence and cell cycle arrest. These results suggest that PRMT5 may serve as a promising therapeutic target for glaucoma, providing novel strategies to halt RGC senescence and alleviate IOP-induced damage.

Materials and Methods

Mouse Model Establishment and IOP Measurement

C57BL/6 mice, aged 6 to 8 weeks, were sourced from Vital River (Shanghai, China) and housed in sterile conditions. They were allocated to groups to establish a COH model with elevated IOP, according to the methodology described by Chen et al.¹⁵ All animal experiments were approved by the Animal Research Committee at Shanghai Jiao Tong University and conducted with strict adherence to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. For sedation before surgical intervention, we administered xylazine (10 mg/kg; Sigma-Aldrich, St. Louis, MO, USA) combined with ketamine hydrochloride (25 mg/kg; Sigma-Aldrich) intraperitoneally. The topical anesthetic 0.5% proparacaine hydrochloride (Bausch and Lomb, Bridgewater, NJ, USA) was applied to the ocular surface.

To establish our COH mouse model, we dissolved HyStem and Extralink from the HyStem Cell Culture Scaffold kit (HYS020-1KT; Sigma-Aldrich) in deionized water at a 4:1 ratio. This solution was gently infused into the anterior chamber using a 31-gauge insulin syringe (BD Ultra-Fine, Franklin Lakes, NJ, USA) over the course of more than 2 minutes until the solution solidified. In the control group, we injected phosphate-buffered saline (PBS) into the anterior chamber of the right eyes, followed by the application of 0.3% ofloxacin eye ointment (Santen Pharmaceutical, Osaka, Japan) to prevent infection. IOP was measured noninvasively using a calibrated TonoLab Rebound Tonometer (Icare Finland, Vantaa, Finland) under mild anesthesia to minimize stress-induced fluctuations. Measurements were taken at baseline (premodel induction) and at days 3, 7, 14, 28, and 56 to track temporal changes. All measurements were conducted between 10 AM and 2 PM to eliminate bias. Averages of three readings for each measurement were recorded.

Intravitreal Injection

CHIR99021 (a potent GSK3β inhibitor and Wnt/β-catenin signaling pathway activator; HY-10182; MedChemExpress, China) or GSK3326595 (a selective PRMT5 inhibitor; HY-101563; MedChemExpress, Monmouth Junction, NJ, USA) (2 µL of a 10-µM solution) was intravitreally injected into the test eye of mice using a 34-gauge needle attached to a 10-µL microsyringe (Hamilton, Reno, NV, USA). Moreover, 2 µL of the vehicle control (0.1% dimethyl sulfoxide in PBS) was administered into the contralateral eye. All these intravitreal injections were administered to mice under deep anesthesia 3 days before COH induction and then weekly until postsurgical week 4.

R28 Cell Culture and Treatment

The retinal precursor (R28) cell line, kindly provided by Dr. Guotong Xu from the Tongji Eye Institute at Tongji University School of Medicine, was cultured in low-glucose Dulbecco's modified Eagle medium (DMEM; Sigma-Aldrich), supplemented with 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Waltham, MA, USA) and 1% penicillin/streptomycin (PS; Thermo Fisher Scientific) at 37°C in a humidified incubator with 5% CO₂. Hypoxic conditions were induced by treating the cells with 200 µM cobalt chloride (CoCl₂; Sigma-Aldrich) for 6, 12, or 24 hours. Cells were treated with tert-butyl hydroperoxide (TBHP; Sigma-Aldrich) at 12.5 µM, 25 µM, or 50 µM for 2 hours to induce oxidative stress, followed by a 46-hour incubation in complete medium. Both cells and supernatants were harvested posttreatment for subsequent analyses.

To study the effects of Wnt signaling on cellular senescence, FH535 (a novel Wnt/β-catenin signaling antagonist; HY-15721; MedChemExpress) was administered at concentrations of 0.5, 1, and 2 µM, whereas CHIR99021 was administered at concentrations of 1, 2.5, and 5 µM. R28 cells were seeded into 6-, 24-, or 96-well plates; allowed to adhere for 24 hours; and then treated with CoCl₂ or the aforementioned small molecules in fresh serum-free medium for an additional 12 hours. We used this experimental setup to examine the effects of Wnt signaling on cellular senescence.

Primary RGC Culture and Treatment

Primary RGCs were isolated from postnatal days 1 to 3 (P1–3) in C57BL/6 mice.¹⁶ Retinas dissected from enucleated eyes underwent enzymatic digestion with 15 U/mL papain (Worthington Biochemical, Lakewood, NJ, USA) and 200 U/mL DNase I (Sigma-Aldrich) at 37°C for 15 minutes, followed by inactivation with PBS/10% FBS. Triturated lysates were centrifuged and resuspended in PBS and incubated with Thy1.2-conjugated microbeads (Invitrogen, Carlsbad, CA, USA). Magnetic-activated cell sorting (MACS)-purified cells were cultured in DMEM/F12 containing 2% B27 (Sigma-Aldrich), 1% L-glutamine (Sigma-Aldrich), and 1% PS. RGC purity was confirmed via β-Ⅲ tubulin and Brn3a immunostaining (Supplementary Fig. S3A). Cells were treated with TBHP at 200 µM for 6, 12, or 24 hours to induce oxidative stress. Both cells and supernatants were harvested posttreatment for subsequent analyses. Table 1 lists the antibodies used here.

Table 1.

Primary and Secondary Antibodies Used in This Study

Antibody	Dilution	Manufacturer	Catalog No.
Rabbit anti-PRMT5	1:1000 (WB)/1:100 (IF)	Abcam	ab109451
Rabbit anti-active β-catenin	1:1000 (WB)/1:200 (IF)	Abcam	#8814
Rabbit anti-GSK3β	1000 (WB)	Proteintech	22104-1-AP
Rabbit anti–p-GSK3β	1:1000 (WB)	Cell Signaling Technology	#5558
Mouse anti–β-actin	1:1000 (WB)	Proteintech	66009-1-IG
Rabbit anti–γ-H2AX	1:1000 (WB)/1:250 (IF)	Cell Signaling Technology	#9718
Rabbit anti-p21CIP1	1:1000 (WB)	Proteintech	28248-1-AP
Rabbit anti-p16INK4a	1:1000 (WB)/1:100 (IF)	Abcam	ab211542
Rabbit anti-Brn3a	1:200 (IF)	Abcam	ab245230
Rabbit anti–β-Ⅲ tubulin	1:100 (IF)	Abcam	ab18207
HRP-conjugated Affinipure Goat Anti-Mouse IgG(H+L)	1:5000	Proteintech	SA00001-1
HRP-conjugated Affinipure Goat Anti-Rabbit IgG(H+L)	1:5000	Proteintech	SA00001-2
Alexa Fluor 488 donkey anti-rabbit secondary antibody	1:2000	Abcam	ab150117

HRP, horseradish peroxidase; WB, western blotting.

Immunofluorescence Staining Assay

Frozen tissue sections or cell slides were initially fixed with 4% paraformaldehyde (Servicebio, Wuhan, China) for 20 minutes, followed by permeabilization with 0.1% Triton X-100 for 10 minutes. To prevent nonspecific binding, tissues were treated with goat serum for 1 hour. Next, the samples were incubated with primary antibodies at 4°C overnight. The samples were rinsed with PBS and then exposed to secondary antibodies. The nuclei of the cells on the confocal dishes were counterstained using 4′,6-diamidino-2-phenylindole, and images were acquired under an LSM510 microscope (Zeiss, Oberkochen, Germany). Table 1 lists the antibodies.

Western Blotting

We prepared total cell lysates and mouse retinas by using RIPA buffer (Epizyme, Shanghai, China) with the addition of a protease inhibitor cocktail and phenylmethylsulfonyl fluoride (Beyotime, Shanghai, China), per the manufacturer's instructions. β-Actin was employed as the loading control. The proteins were subjected to denaturation and separation through 10% to 12.5% sodium dodecyl sulfate polyacrylamide gel electrophoresis (Epizyme, Shanghai, China). The separated proteins were transferred onto polyvinylidene fluoride (PVDF) membranes (Merck Millipore, USA). These membranes were blocked with 5% bovine serum albumin (Epizyme, Shanghai, China) at room temperature for 1 hour and then incubated with specific primary antibodies against PRMT5, p16^INK4a, p21^CIP1, γ-H2AX, non-phospho (active) β-catenin (Ser33/37/Thr41), GSK3β, phospho-GSK3β (Ser9), and β-actin at 4°C overnight. Subsequently, the membranes were incubated with horseradish peroxidase–conjugated IgG for 1 hour. The immunoreactivity was then visualized using an enhanced chemiluminescence kit (Thermo Fisher Scientific) and captured using an ImageQuant LAS 4000 Mini system (GE Healthcare Bio Sciences, Chicago, IL, USA). Table 1 lists the antibodies used here.

RNA Isolation and Quantitative PCR

Total RNA was extracted using TRIzol (Invitrogen), and RNA concentration was determined using a NanoDrop spectrophotometer (Thermo Fisher Scientific). The extracted RNA was then subjected to cDNA synthesis by using a reverse transcription Master Mix (Takara, Kusatsu, Japan), according to the manufacturer's instructions. Quantitative PCR (qPCR) was performed using a QuantiTect Reverse Transcription Kit (Qiagen, Venlo, Netherlands) on an ABI7500 (Thermo Fisher Scientific). β-Actin was used as the internal control for normalizing the qPCR results. Table 2 lists the qPCR primers used here.

Table 2.

Primer, siRNA, and Short Hairpin RNA Sequences Used in This Study

Gene	Sequence (5′-3′)
β-actin forward	GCTGTGCTATGTTGCCCTAGACTTC
β-actin reverse	GGAACCGCTCATTGCCGATAGTG
Prmt5 forward	GGAGTACAGTGGAGAGGAGAAGAC
Prmt5 reverse	GGAGAATGGCGGCTTTGATGG
siPrmt5-1 forward	CCUCAAGUUGGAAGUGCAAUUdTdT
siPrmt5-1 reverse	AAUUGCACUUCCAACUUGAGGdTdT
siPrmt5-2 forward	AGACGUACGAAGUGUUCGAAAdTdT
siPrmt5-2 reverse	UUUCGAACACUUCGUACGUCUdTdT
Prmt5-OE forward	GCCGCGAATTCGAAGTATACCTCGAGGCCACCATGGCGGCGATGGCAGTC
Prmt5-OE reverse	CCGTCATGGTCTTTGTAGTCGGATCCGAGGCCAATGGTATAGGAGCGACCA
siRNA NC forward	UUCUCCGAACGUGUCACGUdTdT
siRNA NC reverse	ACGUGACACGUUCGGAGAAdTdT

RNA Interference and Plasmid Transfection

Small interfering RNA (siRNA) oligonucleotides specifically targeting PRMT5 and a negative control siRNA (siNC) were procured from Genomeditech (Shanghai, China). A Prmt5-overexpression (OE) vector was also purchased from Genomeditech and verified via DNA sequencing. SiPrmt5, siNC, or Pmrt5-OE transfection was performed using Lipofectamine 3000 (Thermo Fisher Scientific), according to the manufacturer's instructions. At 48 hours after transfection, cells were collected for further experimentation. Table 2 lists the siRNA and Prmt5-OE sequences used here.

Flow Cytometry Analysis

To analyze the cell cycle in R28 cells, flow cytometry was performed on a flow cytometer from Beckman Coulter (Brea, CA, USA), and data were analyzed using FlowJo (version 10.8; Ashland, OR, USA). We used a propidium iodide (PI) flow cytometry kit per the manufacturer's protocol. In brief, 1 × 10⁵ cells were harvested, centrifuged, and resuspended in 200 µL of PI working solution containing RNase A. The mixture was incubated at 37°C for 30 minutes in the dark and then subjected to flow cytometry.

Senescence-Associated β-galactosidase Staining

Intracellular senescence-associated β-galactosidase (SA-β-gal) activity was evaluated using an SA-β-gal Staining Kit (Beyotime), according to the manufacturer's instructions. Five random images of SA-β-gal–positive cells from different groups were taken at a 20× or 40× magnification.

ELISA

SASP (IL-6, MCP-1, TNF-α, CXCL10) levels in retinas and cell culture supernatants were detected using ELISA kits (Weiao Biotech, Nanjing, China). Detection procedures were conducted according to the manufacturer's instructions. SASP concentrations were determined using the standard curve provided by the kits.

Statistical Analysis

All statistical analyses and scatterplots were generated using GraphPad Prism (version 9; GraphPad Software, La Jolla, CA, USA). All data were presented as mean ± standard deviation (SD). Differences between two groups were assessed using the unpaired Student's t-test, whereas one-way analysis of variance was employed to evaluate significant differences across multiple groups. Pairwise comparisons were performed using the Tukey test, and the Dunnett test was used to compare each experimental group's mean against that of the control group. A P value of <0.05 was considered to indicate statistical significance.

Results

PRMT5 Downregulation During RGC Senescence in our COH Mouse Model

We established a COH mouse model to simulate the pathological features of glaucoma. We observed that the IOP of the control (Ctrl) group remained stable at baseline levels, while in the COH group, IOP increased and remained stable for at least 8 weeks (Supplementary Fig. S1A). Compared with those in PBS-treated mice, the number of RGCs in COH mice progressively decreased in all retinal positions in a time-dependent manner from weeks 2 to 8 (Figs. 1A, 1B, Supplementary Figs. S1B, S1C). To assess RGC senescence, we performed SA-β-gal staining 2, 4, and 8 weeks after chronic IOP elevation and found that the number of senescent retinal cells significantly increased in the 4- and 8-week COH groups (Figs. 1C, 1D). Immunofluorescence (IF) demonstrated that the expression of p16^INK4a, a major cellular senescence marker, was significantly elevated in the ganglion cell layer (GCL) but that of PRMT5 was markedly lower in the 4-week COH group than in the Ctrl group (Figs. 1E, 1F). To quantify the relative abundance of PRMT5, we performed Western blotting on whole retina lysates. Compared with the Ctrl group, the 4-week COH group demonstrated significantly lower PRMT5 expression, along with an increase in the expression of the cellular senescence markers, p16^INK4a and p21^CIP1, and the DNA damage marker, γ-H2AX (Figs. 1G, 1H). This might be because DNA damage is a common mediator of cellular senescence. ELISA revealed significantly elevated levels of IL-6 (9.77 ± 1.15 vs. 25.78 ± 3.25 pg/mg), MCP-1 (11.32 ± 2.16 vs. 22.97 ± 3.01 pg/mg), and CXCL10 (30.03 ± 9.13 vs. 87.25 ± 8.53 pg/mg) in retinas at 4 weeks postinduction versus controls (n = 5; P < 0.0001 for all analytes). TNF-α levels showed earlier elevation at 2 weeks (33.94 ± 15.04 vs. 75.92 ± 12.18 pg/mg; n = 5, P < 0.0001) (Fig. 1I).

Figure 1.

Effects of COH on PRMT5 expression, senescence-related proteins, and RGC loss. (A) Immunofluorescence labeling of RGCs with Brn3a in retinal sections from Ctrl and COH-treated groups at 2, 4, and 8 weeks postinduction. Scale bars: 50 µm. (B) Quantitative analysis of RGC counts in the retinas of Ctrl and COH groups (n = 4). (C, D) Representative images and quantification of SA-β-gal staining in whole retinal preparations at various time points following COH. Scale bars: 50 µm. (E, F) Immunofluorescence staining and quantification of p16^INK4a (green) and PRMT5 (green) in COH 4-week retinas compared to the Ctrl group. Scale bars: 50 µm. (G) Western blot analysis of p16^INK4a, p21^CIP1, γ-H2AX, and PRMT5 levels in retinas collected at 2, 4, and 8 weeks after COH induction, along with Ctrl. β-Actin was used as an internal control. (H) Quantitation of p16^INK4a, p21^CIP1, γ-H2AX, and PRMT5 protein from panel E (n = 3). (I) IL-6, MCP-1, TNF-α, and CXCL10 levels detected by ELISA in mouse retina 0, 2, 4, and 8 weeks after COH induction (n = 5). Data are expressed as the mean ± SD. ns, not statistically significant. *P < 0.05, **P < 0.01, and ****P < 0.0001.

PRMT5 Knockdown–Induced DNA Damage and Senescence in Hypoxic R28 Cells In Vitro

To investigate the association of RGC senescence with PRMT5 expression, we mimicked a hypoxic retinal environment in R28 cells through stimulation with 200 µM CoCl₂. Accumulating evidence suggests that hypoxia typically initiates cellular senescence via oxidative stress induction.^17,18 SA-β-gal staining for cellular senescence detection revealed that CoCl₂ treatment considerably increased the number of senescent R28 cells compared with the control group (Fig. 2A). In the cell cycle assay, a marked reduction in the percentage of S phase cells (43.73% in control vs. 27.35% in the hypoxia group, P < 0.001) and an increase in the percentage of G0/G1 phase cells (40.48% in control vs. 59.03% in the hypoxia group, P < 0.001) were observed (Figs. 2B, 2C). IF showed an elevated cytosolic p16^INK4a and nuclear γ-H2AX expression in hypoxia-treated R28 cells (Figs. 2D, 2E), indicating cellular senescence and DNA damage. On the other hand, IF and Western blot analyses confirmed that hypoxia decreased PRMT5 expression while increasing p16^INK4a, p21^CIP1, and γ-H2AX levels (Figs. 2F–I). ELISA analysis of R28 cell supernatants revealed hypoxia-induced upregulation of IL-6 (31.47 ± 4.35 vs. 59.93 ± 13.39 pg/mg) and TNF-α (34.32 ± 3.85 vs. 68.00 ± 15.21 pg/mg) alongside elevated MCP-1 (17.40 ± 1.74 vs. 32.77 ± 7.45 pg/mg) and CXCL10 (18.78 ± 1.95 vs. 37.86 ± 8.96 pg/mg) compared to controls (n = 10; P < 0.0001 for all analytes) (Fig. 2J).

Figure 2.

Hypoxia-induced expression changes in PRMT5 and senescence-related proteins in R28 cells. (A) Representative images of SA-β-gal staining in R28 cells subjected to hypoxia. Scale bars: 50 µm. (B, C) Flow cytometry analysis showing an increased G0/G1 fraction in hypoxia-treated R28 cells. (D) Immunofluorescence staining of p16^INK4a (green) and γ-H2AX (green) in R28 cells exposed to hypoxia for 24 hours, with 4′,6-diamidino-2-phenylindole (blue) as a nuclear counterstain. Scale bars: 50 µm. (E) Quantitative analysis of p16^INK4a fluorescence intensity and γ-H2AX foci per cell. (F, G) Immunofluorescence staining and quantification of PRMT5 (green) in the Hypo group compared to the Ctrl group. Scale bars: 50 µm. (H) Western blot analysis of p16^INK4a, p21^CIP1, γ-H2AX, and PRMT5 levels in R28 cells treated with hypoxia for varying durations. β-Actin was used as an internal control. (I) Quantitation of p16^INK4a, p21^CIP1, γ-H2AX, and PRMT5 protein levels from panel H (n = 3). (J) IL-6, MCP-1, TNF-α, and CXCL10 levels detected by ELISA in R28 cells treated with hypoxia (n = 10). Data are expressed as the mean ± SD. ns, not statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Parallel TBHP experiments in R28 cells recapitulated G0/G1 phase arrest (49.20% in control vs. 62.69% in TBHP group, P < 0.01) with increased SA-β-gal⁺ cells; p16^INK4a, p21^CIP1, and γ-H2AX elevation; and SASP release (Supplementary Fig. S2), validating the universality of oxidative stress–induced senescence mechanisms. However, in postmitotic primary RGCs, TBHP treatment induced p16^INK4a elevation and γ-H2AX foci formation without a significant cell cycle shift (G0/G1 phase: 92.58 in control vs. 92.22% in TBHP group, P = 0.9898; Supplementary Fig. S3), consistent with their terminal differentiation status.

Next, we investigated the possible effects of PRMT5 on R28 cell senescence through siRNA knockdown of PRMT5. The results showed that both siPrmt5-1 and siPrmt5-2 knocked down PRMT5 mRNA and protein expression efficiently; therefore, we selected siPrmt5-1 and siPrmt5-2 for subsequent experimentation (Figs. 3A–C). SA-β-gal staining indicated that PRMT5 knockdown significantly increased the percentage of senescent cells (Fig. 3D). Cell cycle analysis revealed that PRMT5 knockdown induced an increase in G0/G1 phase cells (51.27% in control vs. 65.56% in siPrmt5-1, P < 0.001; 64.66% in siPrmt5-2, P < 0.001), with a concurrent reduction in S phase cells (32.60% in control vs. 15.62% in siPrmt5-1, P < 0.001; 19.53% in siPrmt5-2, P < 0.001) (Figs. 3E, 3F), indicating that PRMT5 knockdown led to cell cycle arrest. PRMT5 knockdown also increased γ-H2AX and p21^CIP1 levels (Figs. 3G–J), suggesting that PRMT5 knockdown causes persistent nuclear DNA damage and the consequent cellular senescence. However, p16^INK4a levels remained unchanged—as corroborated by our Western blotting results (Figs. 3G–J).

Figure 3.

Silencing PRMT5 leads to cell cycle arrest and senescence in R28 cells. (A) RT-PCR analysis of Prmt5 expression in R28 cells treated with various Prmt5-siRNAs, with β-actin serving as an internal control (n = 3). (B) Western blot analysis of PRMT5 protein levels in R28 cells following treatment with different Prmt5-siRNAs. (C) Quantification of PRMT5 protein levels from panel B (n = 3). (D) Representative images of SA-β-gal staining in cells transfected with NC, siPrmt5-1, and siPrmt5-2. Scale bars: 50 µm. (E, F) Flow cytometry analysis showing an increased G0/G1 fraction in siPrmt5-1– or siPrmt5-2–transfected cells. (G) Immunofluorescence staining for p16^INK4a (green) and γ-H2AX (green) in R28 cells transfected with siPrmt5s for 48 hours, with 4′,6-diamidino-2-phenylindole (blue) counterstaining. Scale bars: 50 µm. (H) Quantitative analysis of p16^INK4a fluorescence intensity and γ-H2AX foci per cell. (I) Western blot analysis of p16^INK4a, p21^CIP1, and γ-H2AX protein levels in siPrmt5-transfected cells, with β-actin as an internal control. (J) Quantification of p16^INK4a, p21^CIP1, and γ-H2AX protein levels from panel H (n = 3). Data are presented as mean ± SD. NC, negative control; ns, not statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Roles of PRMT5 in Wnt/β-catenin Axis Regulation In Vitro and In Vivo

Emerging evidence indicates that the Wnt/β-catenin axis is critical for suppressing cellular senescence and that it can be implicated in retinal development and glaucoma.^19,20 In the COH model, the protein levels of active β-catenin and p-GSK3β were initially upregulated at 2 weeks, followed by a decline at 4 and 8 weeks (Figs. 4A, 4B). This suggests an early compensatory activation of the Wnt/β-catenin signaling in RGC in vivo in response to elevated IOP, oxidative stress, and hypoxia, which is subsequently inhibited during prolonged COH. In vitro, CoCl₂ treatment for 12 hours induced nuclear translocation of active β-catenin, which was reversed after 24 hours, as shown by IF (Figs. 4C, 4D). Western blot analysis confirmed that Wnt pathway activation peaked at 12 hours and declined by 24 hours under hypoxia (Figs. 4E, 4F). Overexpression of PRMT5 in R28 cells markedly enhanced nuclear translocation of active β-catenin, while PRMT5 depletion attenuated the active β-catenin translocation in nuclei (Figs. 4G, 4H). Consistently, PRMT5 overexpression increased active β-catenin and p-GSK3β levels, indicating activation of the Wnt/β-catenin pathway (Figs. 4I, 4J).

Figure 4.

PRMT5 is essential for regulating the Wnt/β-catenin axis. (A) Western blot analysis of active β-catenin, GSK3β, and p-GSK3β levels in the retinas from Ctrl and COH mice at 2, 4, and 8 weeks postinduction. β-Actin was used as an internal control. (B) Quantitation of active β-catenin, GSK3β, and p-GSK3β protein levels from panel A (n = 3). (C, D) Immunofluorescence staining and quantification of active β-catenin (green) in R28 cells treated with CoCl₂ for varying durations, with 4′,6-diamidino-2-phenylindole (blue) counterstaining. Scale bars: 50 µm. (E) Western blot analysis of active β-catenin, GSK3β, and p-GSK3β protein levels in R28 cells treated with CoCl₂ for different time periods, with β-actin as an internal control. (F) Quantitation of active β-catenin, GSK3β, and p-GSK3β protein levels from panel D (n = 3). (G, H) Immunofluorescence staining and quantification of active β-catenin (green) in R28 cells pretreated with NC or oePrmt5 for 24 hours and then transfected with siNC or Prmt5-siRNAs. Scale bars: 50 µm. (I) Western blot analysis of PRMT5, active β-catenin, GSK3β, and p-GSK3β protein levels in oePrmt5-transfected cells, with β-actin serving as an internal control. (J) Quantitation of PRMT5, active β-catenin, GSK3β, and p-GSK3β protein levels from panel G (n = 3). Data are presented as mean ± SD. NC, negative control; ns, not statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

PRMT5-Induced Senescence Regulation in R28 Cells via the Wnt/β-catenin Pathway In Vitro

Considering the aforementioned results, we assessed whether PRMT5 is involved in RGC senescence via the Wnt/β-catenin pathway. PRMT5 inhibition with siRNA (siPrmt5-1 or siPrmt5-2) led to further activation of senescence in hypoxia-exposed R28 cells, as indicated by the SA-β-gal staining results, whereas CHIR99021 counteracted this prosenescence effect (Fig. 5A). Cell cycle analysis revealed that PRMT5 knockdown under hypoxia exacerbated G0/G1 phase arrest (53.79% in control vs. 64.78% in siPrmt5-1, P < 0.001; 63.97% in siPrmt5-2, P < 0.001), whereas cotreatment with CHIR99021 significantly mitigated G0/G1 arrest (64.78% in siPrmt5-1 vs. 44.25% with CHIR99021, P < 0.001 and 63.97% in siPrmt5-2 vs. 46.57% with CHIR99021, P < 0.001) (Figs. 5B, 5C). Given that PRMT5 knockdown led to persistent nuclear DNA damage, we next assessed whether Wnt/β-catenin signaling pathway activation reverses DNA damage induced by PRMT5 knockdown in R28 cells. As illustrated in Fig. 5D, CHIR99021 significantly attenuated the increased γ-H2AX levels induced by PRMT5 downregulation under hypoxia. Consistently, Western blotting revealed that PRMT5 knockdown considerably increased senescence-associated markers, while concurrently suppressing the Wnt/β-catenin signaling pathway (Figs. 5E, 5F). Notably, these molecular alterations were reversed by cotreatment with CHIR99021 (Figs. 5E, 5F).

Figure 5.

CHIR99021 mitigates cellular senescence in siPrmt5-transfected cells. (A) Representative images of SA-β-gal staining of R28 cells among the following groups: siPrmt5-1+Hypo, siPrmt5-2+Hypo, NC+Hypo, siPrmt5-1+Hypo+CHIR99021, siPrmt5-2+Hypo+CHIR99021, and NC+Hypo+CHIR99021. Scale bars: 50 µm. (B, C) Flow cytometry analysis showing that siPrmt5-1 or siPrmt5-2 transfection increased the G0/G1 fraction, while cotreatment with CHIR99021 reversed these changes. (D) Immunofluorescence staining of γ-H2AX (green) in the indicated groups. Scale bars: 20 or 50 µm. (E) Western blot analysis of PRMT5, active β-catenin, GSK3β, p21^CIP1, and γ-H2AX protein levels in the indicated groups, with β-actin as an internal control. (F) Quantitation of PRMT5, active β-catenin, GSK3β, p21^CIP1, and γ-H2AX protein levels from panel E (n = 3). Data are presented as mean ± SD. NC, negative control; ns, not statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

In contrast, PRMT5 overexpression led to considerable reductions in the percentages of hypoxia-induced senescent cells, highlighting the pivotal role of PRMT5 in cellular senescence modulation (Fig. 6A). Notably, treatment with FH535 effectively ameliorated these alterations (Fig. 6A). In the flow cytometric assay, PRMT5 overexpression was noted to significantly reduce the percentage of G0/G1 phase cells compared with the hypoxia group (62.44% vs. 53.99%, P = 0.043) (Figs. 6B, 6C), further substantiating the regulatory effects of PRMT5 on cell cycle dynamics. However, cotreatment with FH535 significantly mitigated the protective effects conferred by PRMT5 overexpression on cell cycle arrest, suggesting that PRMT5 inhibits cellular senescence via Wnt signaling cascade activation (Figs. 6B, 6C). Furthermore, PRMT5 overexpression was noted to counteract the hypoxia-induced activation of γ-H2AX in the nuclei, whereas cotreatment with FH535 significantly increased γ-H2AX expression in a dose-dependent manner (Fig. 6D). Moreover, R28 cells transfected with a Prmt5-OE vector exhibited a significant downregulation in the expression of the senescence-associated markers p21^CIP1 and γ-H2AX (Figs. 6E, 6F). Furthermore, the Wnt/β-catenin pathway was strongly activated, as indicated by reduced GSK3β expression and increased active β-catenin expression (Figs. 6E, 6F).

Figure 6.

FH535 abrogated the neuroprotective effects against cellular senescence in oePrmt5-transfected cells. (A) Representative images of SA-β-gal staining of R28 cells from the following groups: NC, NC+Hypo, oePrmt5+Hypo, oePrmt5+Hypo+FH535, oePrmt5+Hypo+FH535, and oePrmt5+Hypo+FH535. Scale bars: 50 µm. (B, C) Flow cytometry analysis showing that oePrmt5 transfection alleviated hypoxia-induced G0/G1 phase arrest, while cotreatment with FH535 reversed these protective effects. (D) Immunofluorescence staining of γ-H2AX (green) in the indicated groups, with cell nuclei counterstained using 4′,6-diamidino-2-phenylindole (blue). Scale bars: 20 or 50 µm. (E) Western blot analysis of PRMT5, active β-catenin, GSK3β, p21^CIP1, and γ-H2AX protein levels in the indicated groups, with β-actin as an internal control. (F) Quantitation of PRMT5, active β-catenin, GSK3β, p21^CIP1, and γ-H2AX protein levels from panel E (n = 3). Data are presented as mean ± SD. NC, negative control; ns, not statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

PRMT5-Induced Senescence Regulation in RGC via the Wnt/β-catenin Pathway In our COH Mouse Model

To assess the impact of the PRMT5-regulated Wnt/β-catenin pathway in vivo, we intravitreally administered the PRMT5 inhibitor GSK3326595, the GSK3β inhibitor CHIR99021, or a combination of the two before the surgical induction of COH and then weekly over 4 postsurgical weeks (Fig. 7A). IF revealed that RGC survival decreased in GSK3326595-treated COH mice at all retinal locations, especially in the 5/6 retinal radius (P = 0.015) (Figs. 7B, 7C, Supplementary Figs. S4A, S4B). In contrast, the administration of CHIR99021 significantly increased RGC survival, even when combined with GSK3326595, suggesting that activation of the Wnt/β-catenin signaling pathway effectively mitigates RGC loss under experimental glaucoma conditions (Figs. 7B, 7C). Western blotting analysis further revealed that treatment with GSK3326595 elevated the expression levels of senescence markers p16^INK4a and p21^CIP1, as well as GSK3β, while markedly reducing the levels of active β-catenin (Figs. 7D, 7E). Importantly, CHIR99021 administration significantly reversed these molecular changes, restoring active β-catenin levels and reducing the expression of senescence-associated proteins (Figs. 7D, 7E). Notably, even in the presence of GSK3326595, cotreatment with CHIR99021 markedly attenuated RGC senescence and mitigated neurodegeneration (Figs. 7D, 7E).

Figure 7.

The impact of GSK3326595, CHIR99021, or both on RGC loss, Wnt/β-catenin signaling, and cellular senescence in COH-induced mice. (A) Schematic representation of the experimental timeline, including surgical induction of COH, intravitreal administration of GSK3326595, CHIR99021, or both, and subsequent analysis. (B) Immunofluorescence labeling of RGCs with Brn3a in retinal sections from the dimethyl sulfoxide, GSK3326595, CHIR99021, and GSK3326595+CHIR99021 treatment groups following surgical induction of COH. Scale bars: 50 µm. (C) Quantitative analysis of RGC counts in retinas from the different treatment groups (n = 4). (D) Western blot analysis of PRMT5, active β-catenin, GSK3β, p21^CIP1, and γ-H2AX protein levels in retinas from the indicated groups. β-Actin was used as an internal control. (E) Quantitation of PRMT5, active β-catenin, GSK3β, p21^CIP1, and γ-H2AX protein levels from panel D (n = 3). (F) This schematic illustrates the proposed mechanism in which PRMT5 downregulation, driven by cellular stressors, disrupts β-catenin/GSK3β homeostasis, leading to DNA damage response (γ-H2AX), senescence marker (p21^CIP1) expression, and ultimately RGC senescence. Arrows (→) indicate activation; T-bars (⊣) indicate inhibition. Red lines: prosenescence effects; green lines: antisenescence effects. Created in BioRender. Data are presented as mean ± SD. ns, not statistically significant. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001.

Discussion

Here, we found that diminished PRMT5 expression promotes senescence in RGC via Wnt/β-catenin pathway suppression through GSK3β transactivation, thereby facilitating glaucoma progression. These findings expand current knowledge regarding molecular mechanisms underlying RGC degeneration involved in glaucoma progression and underscore the potential role of PRMT5 as a therapeutic target.

Recent studies have found that high IOP damage significantly alters pro-aging molecules such as p16^INK4a and anti-aging molecules such as SIRT6, leading to RGC senescence and degeneration, thus contributing to glaucoma pathogenesis.^21–23 While most previous studies have focused on acute ocular hypertension models, clinical glaucoma is predominantly caused by the long-term effects of chronic ocular hypertension. To address this, we established a COH mouse model using an in situ crosslinking hydrogel and validated RGC loss at all three retinal positions (Figs. 1A, 1B, Supplementary Figs. S1B, S1C). Furthermore, the COH model demonstrated significant upregulation of β-galactosidase–positive cells and senescence-associated proteins (Figs. 1C–I). Notably, RGC demonstrated increased expression of p16^INK4a, a widely recognized cellular senescence marker (Figs. 1E, 1F)—corroborating previous findings.^21,24 These results underscore the significant role of cellular senescence in RGC loss and glaucoma progression. Consistent with these findings, CoCl₂ and TBHP treatment in vitro, mimicking COH-induced oxidative stress, triggered cellular senescence (Fig. 2, Supplementary Figs. S2–S3). Intriguingly, in postmitotic primary RGCs, p16^INK4a and p21^CIP1 sustain cell cycle quiescence while executing multifaceted roles in DNA damage response, SASP modulation, and prevention of pathological dedifferentiation.^25,26 These findings redefine their canonical senescence-associated functions, establishing them as essential regulators of both proliferative arrest and senescence surveillance in terminally differentiated neural cells.

In our present study, we observed significant downregulation of PRMT5 expression in both the COH mouse model and the hypoxia-induced in vitro model, with reduced PRMT5 levels primarily localized to the RGC layer (Figs. 1E–H, Figs. 2F–I). The current results indicated that decreased PRMT5 expression may trigger a cascade of events leading to cellular senescence (Fig. 3). Suppression of PRMT5 arrested the cell cycle in the G0/G1 phase, consistent with findings in other diseases and cancers.^10,11,27 Studies have reported that PRMT5 regulates p21^CIP1 expression through an epigenetic silencing mechanism.^14,28,29 Similarly, here, we noted that PRMT5 inhibition sequentially activates a DNA damage response and p21^CIP1 expression in RGCs, inducing senescence and significantly increasing the number of senescent cells.

Our work adds to evidence linking Wnt/β-catenin signaling with RGC senescence. Several studies have demonstrated the presence of active Wnt signaling in developing and adult retinas, particularly in their GCLs.^30,31 However, the role and mechanism of the Wnt/β-catenin pathway in COH remain unclear. The current results revealed that in mice, the Wnt/β-catenin pathway undergoes early activation after COH induction, followed by significant attenuation in the later stages (Figs. 4A, 4B). This result suggests that the Wnt/β-catenin pathway is crucial for early neuroprotective responses in RGCs. In vitro, hypoxia-induced activation of β-catenin was observed in the early phase of treatment, followed by downregulation at later stages (Figs. 4C–F). Under acute hypoxia, β-catenin is activated through HIF-1α–mediated stabilization and GSK3β inactivation, enhancing cellular survival; conversely, chronic hypoxia triggers β-catenin suppression via induction of endogenous inhibitors (e.g., AXIN2/DKK1), disruption of the ER stress–LRP6 (Wnt coreceptor) axis, and GSK3β reactivation, restricting proliferation through apoptotic or senescent pathways.^32–34 This temporally regulated mechanism orchestrates adaptive survival while maintaining proliferative control under hypoxic stress. Overexpression of PRMT5 significantly enhanced β-catenin activation; however, PRMT5 knockdown partially inhibited this effect (Figs. 4G–J). These results suggest that PRMT5 may exert neuroprotective effects on RGCs by regulating the Wnt/β-catenin signaling cascade.

We subsequently explored the potential relationship between PRMT5-mediated modulation of the Wnt/β-catenin pathway and cellular senescence. Our findings suggested that PRMT5 depletion is correlated with enhanced GSK3β activity, resulting in decreased β-catenin levels and fostering a senescent state in RGCs. In vivo, intravitreal administration of CHIR99021 counteracted the prosenescence effects of GSK3326595, further confirming the critical role of the Wnt/β-catenin signaling in PRMT5-mediated neuroprotection on RGCs in glaucoma (Fig. 7). This finding highlights the critical role of the Wnt/β-catenin pathway in counteracting the effects of PRMT5 inhibition and suggests that PRMT5 acts as an upstream modulator of this signaling cascade.

Despite these findings, our study has limitations. While CoCl₂ serves as a practical HIF-1α–targeted hypoxia mimic, it cannot fully replicate the multifactorial complexity of physiological hypoxia. We will further validate results with a 1% O₂ hypoxic chamber and investigate hypoxia-specific metabolic changes. Furthermore, while we used a PRMT5 inhibitor to elucidate its role of PRMT5 in COH mice, conditional knockout models specifically targeting RGC will be necessary to fully delineate PRMT5’s in vivo function. Finally, the precise mechanism by which PRMT5 regulates the Wnt/β-catenin signaling pathway remains unclear. Further studies are warranted to investigate whether PRMT5 forms complexes with or methylates Wnt-related molecules to modulate this pathway.

Conclusions

In conclusion, our findings demonstrate that PRMT5 regulates RGC senescence through the Wnt/β-catenin signaling pathway, playing a critical role in glaucoma progression (Fig. 7F). By elucidating these molecular mechanisms, we provide a foundation for the development of novel therapeutic strategies aimed at mitigating RGC loss in glaucoma. Future research should further explore the intricacies of PRMT5-mediated regulation of Wnt/β-catenin signaling and its potential clinical applications in glaucoma management.