Targeting Bcl3/NF-κB p50 Pathway for Neuroinflammation Attenuation and RGCs Protection in Retinal Ischemia/Reperfusion Injury

Abstract

Purpose

Retinal ischemia/reperfusion (IR) injury caused by pathologically high intraocular pressure (ph-IOP) induces excessive inflammation, contributing to retinal ganglion cell (RGC) death in glaucoma. Lowering IOP alone is insufficient, highlighting the need for neuroprotective strategies. Resveratrol (RSV) exhibits anti-inflammatory and neuroprotective effects, but its molecular mechanisms remain unclear. This study aims to evaluate RSV’s neuroprotective role and underlying mechanisms in retinal IR injury.

Methods

Retinal morphology and RGC survival were assessed via immunofluorescence and hematoxylin and eosin (H&E) staining. Retinal function was evaluated using flash visual evoked potential (F-VEP) and flash electroretinogram (F-ERG). Inflammation and microglial activation were analyzed by quantitative real-time PCR (qRT-PCR) and immunohistochemistry. Pyroptosis and apoptosis were examined using Western blotting, TUNEL staining, and electron microscopy. RNA sequencing, qRT-PCR, and Western blotting identified molecular pathways.

Results

RSV significantly protected RGCs and preserved retinal function. It reduced inflammation by inhibiting microglial activation and redistribution. Electron microscopy confirmed its protective effects against apoptosis and pyroptosis. Most importantly, we identified the Bcl3/NF-κB p50 pathway as a key target of RSV. Using the Bcl3-NF-κB p50-specific inhibitor JS-6, we validated this pathway's role in reducing neuroinflammation, pyroptosis, and apoptosis.

Conclusions

This study provides insights into RSV’s molecular mechanisms and identifies new therapeutic targets for glaucoma.

Retinal ischemia/reperfusion (IR) injury, referring to retinal tissue damage that occurs when the blood supply returns to the tissue after a period of ischemia, is linked to many ocular diseases, including diabetic retinopathy, retinal vascular occlusion, and especially acute glaucoma.^1,2 A self-reinforcing destructive cascade involving neuronal depolarization, calcium influx, and blood-retina barrier breach is triggered in acute glaucoma when transient intraocular hypertension quickly obstructs retinal blood flow. Following the end of the ischemia, blood reperfusion causes an extreme release of free radicals and an overabundance of inflammatory reactions in addition to bringing oxygen and glucose.³ This overwhelms normal cellular antioxidant defenses, ultimately leading to the death of retinal ganglion cells (RGCs), morphological degeneration of the retina, and the impairment of retinal function.⁴ The mechanisms driving retinal damage from IR injury have been extensively studied, revealing that oxidative stress and inflammation play crucial roles in the initiation and progression of RGC damage.⁴ Therefore, combating inflammation might be a promising strategy for rescuing IR-injured RGCs in glaucoma.

Emerging evidence suggests that phytochemicals, due to their antioxidant and anti-inflammatory properties, may play a significant role in preventing and treating RGC damage.⁵ Resveratrol (RSV), a biologically active polyphenol found in grapes, has been shown in clinical studies to possess potent antioxidant and anti-inflammatory effects, demonstrating cardioprotective, chemotherapeutic, neuroprotective, and anti-aging properties.^6,7 The anti-inflammatory and neuroprotective effects of RSV have been observed in in vitro and in vivo studies of glaucoma. However, there is a relative scarcity of studies investigating the underlying molecular mechanisms that may support this promising therapeutic agent.

NF-κB signaling is a pivotal pathway that impacts various cell functions, such as proliferation and differentiation, induction of apoptosis, and immune response.⁸ One of its best known roles is to serve as a key mediator of inflammatory responses, that leads to the expression of inflammatory cytokines such as IL-1β, IL-6, TNF-α, and inflammasomes like NLRP3, one of the most important inflammasomes regulating neuroinflammation and pyroptosis.⁹ The most abundant form of NF-κB activated via the canonical pathway is the p50/RelA heterodimer.¹⁰ Canonical NF-κB activation results in the release of p50/RelA, which acts as a transcription factor to activate the transcription of target genes.¹⁰ Bcl3 is an atypical member of the IκB protein family that plays a crucial role in regulating the activity of nuclear factor NF-κB.¹¹ Depending on the cell type and the nature of stimulation, Bcl3 can either promote or inhibit the NF-κB signaling pathway, thereby modulating downstream gene transcription,¹¹ ultimately exhibiting a dual role in inflammation, which is exerting either pro-inflammatory or anti-inflammatory functions under different circumstances.¹² However, its involvement in retinal diseases remains unexplored, and no studies have investigated the Bcl3/NF-κB/NLRP3 pathway and its regulatory role in this context.

JS-6 is a novel small molecule showing potent intracellular Bcl3-inhibitory activity.¹³ JS-6 is designed to inhibit the protein-protein interaction between the ankyrin repeat domain of Bcl3 and its regulatory protein partner p50.¹³ JS-6 has been shown to prevent tumor growth and metastasis,¹³ although Zhao et al. have questioned its antitumor efficacy, as JS-6 does not suppress all Bcl3 functions in tumor cells.¹⁴ To our knowledge, JS-6 has not been studied in diseases other than tumors, including retinal diseases. Using JS-6 as a specific inhibitor may provide valuable insights into Bcl3-related signaling.

Historically, apoptosis has been considered the primary mechanism of RGC death in glaucoma.¹⁵ Nevertheless, researchers have discovered that ocular cells involved in glaucoma progression not only typically undergo apoptosis but also exhibit characteristics of other forms of cell death, including pyroptosis.¹⁶ Pyroptosis is a form of programmed cell death involving membrane rupture and the release of IL-1β and IL-18.¹⁶ Mechanistically, pyroptosis can occur through the canonical pathway (caspase-1 [CASP1] activation) or the non-canonical pathway (caspase-4/5/11 activation) via different inflammasomes.¹⁷ Recent studies indicate that pyroptosis can also be triggered in an inflammasome-independent manner via caspase-3/8 activation or other stimuli.¹⁶ Caspases will then cleave the gasdermin family protein to form pores on the plasma membrane, amplifying inflammatory responses.¹⁸ Various morphological features of pyroptosis, such as cell swelling, membrane blebbing, membrane rupture, bubble-like protrusions, and pore formation, can be observed over time using electron microscopy.¹⁶ Despite the known implications of pyroptosis in glaucoma,^16,19 little research has focused on the effect of RSV on pyroptosis in glaucoma. Our study aims to address this gap by investigating the relationship between pyroptosis and RGC cell death in glaucoma and exploring how RSV may exert its anti-inflammatory effects through the Bcl3/NF-κB pathway. This research could provide new insights into the pathogenesis of glaucoma and contribute to a better understanding of therapeutic strategies targeting inflammation in RGCs.

Methods

Animals

C57BL6 mice aged 8 weeks were obtained from Hunan SJA Laboratory Animal Co. Ltd. (Hunan, China), and maintained under specific pathogen-free conditions with a 12-hour light-dark cycle, and fed with food and water ad libitum in the Center of Laboratory Animals, Central South University Xiangya Hospital (Hunan, China). All procedures used during animal experiments were approved by the Institutional Animal Care and Use Committee of Central South University, and were conducted in strict accordance with the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, Bethesda, MD, USA).

Animal Model of Retinal IR Injury and Drug Administration

To explore the protective effect of RSV on retinal nerve injury, the animals were randomly grouped into four sets: a control group, a control treated group (RSV), an IR injury group, and an IR + RSV group. To verify the effect of inhibition of Bcl3 on retinal nerve injury, the animals were randomly grouped into four sets: a control group, a control treated group (JS-6), an IR injury group, and an IR + JS-6 group.

RSV powder (Resveratrol, Cat. HY-16561; MedChemExpress, Monmouth Junction, NJ, USA) and JS-6 powder (Bcl3 inhibitor JS-6, Cat. PC-38199, ProbeChem, Shanghai, China) were stored at −20°C protected from light. An appropriate volume of RSV was freshly diluted into a concentration of 1 millimolar (mM) solution, and JS-6 was freshly diluted into a concentration of 1, 2.5, and 5 mM, respectively, with phosphate-buffered saline (PBS; 0.01 M; pH 7.4) for each use.

Animals were anesthetized with 1% sodium pentobarbital solution at a dose of 100 mg/kg. Then, 0.5% tropicamide phenylephrine (Santen Pharmaceutical Co., Ltd., Shiga Plant) followed by 0.4% obtained hydrochloride (Benoxil; Santen Pharmaceutical Co., Ltd.) were then applied to the eyes for mydriasis and cornel topical anesthesia. Then, 2 µL volume of intravitreal drug administration was performed using a 5-µL Hamilton syringe assisted with a tunneled injection made with a 32G needle. The other cohort of mice received an equal amount of PBS injection into the vitreous cavity as a control.

The following day, IR injury was induced after anesthetization by inserting a micro-glass needle linked with a reservoir of normal saline into the anterior chamber of the eye, maintaining an IOP of 120 millimeters of mercury (mm Hg) for 60 minutes. Successful establishment of the model was verified by observation of anterior segment whitening and blanching of episcleral veins under surgical microscopy (BELONA, China). A sham operation was performed without raising the pressure in the eyes of the control group. After the procedure, 0.3% tobramycin and dexamethasone eye ointment (s.a. Alcon-Couvreur n.v.) was applied to the eyes. Animal body temperature was kept at 37°C with a heating pad.

Quantification of RGC Soma Density on Flat-Mounted Retinas

The mice were euthanized and the eyes were immediately enucleated and fixed in a 4% paraformaldehyde solution. A retinal flat-mount was prepared under surgical microscopy, and incubated in 0.3% Triton X-100 for 30 minutes. After removing Triton X-100 drops, 5% bovine serum albumin (BSA) was added dropwise to block antigen at room temperature for 1 hour. After aspiration of the blocking solution, RGCs were labeled and counted using an antibody against the RBPMS (1:200, Cat. ab152101; Abcam, Cambridge, UK) at 4°C overnight. After being washed with PBS 5 times the next day, the retina tissues were then incubated with fluorescent secondary antibody (Anti-rabbit IgG [H+L], F(ab')2 Fragment; Alexa Fluor 488 Conjugate; Cat. 4412, Cell Signaling Technology, Danvers, MA, USA) at room temperature for 1 hour. Images were acquired using a fluorescence microscope (N2-DM4B; Nikon, Tokyo, Japan). The number of RBPMS-positive RGCs was quantified in 3 nonoverlapping areas along the median line of each quadrant, with 12 regions per retina starting from the optic disc to the border at 400 µm intervals.²⁰ The resulting data were shown as percentage loss of RGCs compared to the control group.

Hematoxylin and Eosin Staining

Histological evaluations were performed on retinal cross-sections. Hematoxylin and eosin (H&E)-stained retinal morphology was acquired post IR injury, to visualize and obtain a count of cells in the ganglion cell layer (GCL). Eyes were immediately enucleated post-execution and fixed in FAS eye fixation fluid (Cat. G1109; Servicebio, Wuhan, China) for 24 hours at 4°C. The eyes were embedded in paraffin and were cut into sections of 5-micrometre (µm) thickness through the optic nerve, prepared in the standard way, then stained with H&E. Retinal morphology was observed, and photomicrographs were taken using a light microscope. Sections containing the optic nerve stump were chosen for consistency, and at least three discontinuous sections (per animal) were analyzed using CaseViewer software.

Flash Visual Evoked Potentials and Flash Electroretinogram

Flash visual evoked potential (F-VEP) recording in mice has been described previously.²¹ Briefly, the measurement was conducted 24 hours after IR injury. Mice were anesthetized by intraperitoneal injection of 1% sodium pentobarbital. Body temperature of the animals was maintained at 37°C. Pupils were dilated with 0.5% tropicamide phenylephrine.

The recording and reference electrodes were inserted into the subcutaneous space, touching the occipital bone and the frontal bone surface, respectively; the ground needle electrode was inserted into the subcutaneous space near the tail. The unstimulated eye was occluded with a dark patch during recording. F-VEP was consecutively recorded at different light intensities, consecutive 100 traces in response to flash stimulation were averaged to yield one waveform for each intensity. The first positive peak in the F-VEP waveform was designated as P1, and the first negative peak was N1. The N1-P1 amplitudes and P1 wave latency were measured and analyzed.

For flash electroretinogram (F-ERG) recordings, preparation steps were the same as for F-VEP, whereas mice were adapted to darkness for more than 12 hours before the examination, and a dark testing environment was maintained using a red light for illumination. Gold-wire ring electrodes were placed on the corneal surface, along with the application of carboxymethyl cellulose eye drops to increase current conduction and cornea moisturization. The placement of the ground needle electrode is identical to that of the F-VEP recording, as described above, whereas the reference electrodes were inserted subcutaneously on both sides of the nose. Stimulation and detection refer to International Society for Clinical Electrophysiology of Vision (ISCEV) standard. The amplitude of the a-wave and b-wave in each group were analyzed after detection completion.

Bulk RNA Sequencing

The whole retina was collected 24 hours after IR injury and immediately frozen in liquid nitrogen for subsequent experiments. The total RNA of mouse retina tissue was extracted using Trizol reagent (Cat. 15596018; Thermo Scientific, Waltham, MA, USA). The concentration, quality, and integrity were determined using a NanoDrop spectrophotometer (Thermo Scientific). Three micrograms of RNA were used as input material for the sequencing library preparation. Sequencing libraries were generated according to the following steps. Briefly, the mRNA was first purified from the total RNA using poly-T oligo-attached magnetic beads. Fragmentation was performed using divalent cations at an elevated temperature in Illumina's proprietary fragmentation buffer. First-strand cDNA was synthesized using random oligonucleotides and SuperScript II reverse transcriptase. Second-strand cDNA synthesis was subsequently performed using DNA polymerase I and RNase H. The remaining overhangs were converted to blunt ends via exonuclease/polymerase activities, and the enzymes were removed. After the adenylation of the 3′ ends of the DNA fragments, Illumina PE adapter oligonucleotides were ligated to prepare for hybridization. To select the cDNA fragments of the preferred 400 to 500 bp length, the libraries were further purified using the AMPure XP system (Beckman Coulter, Beverly, CA, USA). DNA fragments with ligated adaptor molecules on both ends were selectively enriched using the Illumina PCR primer cocktail in a 15-cycle PCR reaction. Products were purified (AMPure XP system) and quantified using the Agilent high-sensitivity DNA assay on a Bioanalyzer 2100 system (Agilent). The sequencing library was then sequenced on a NovaSeq 6000 platform (Illumina) by Shanghai Personal Biotechnology Co. Ltd.

Transcriptome Analysis

The original data in FASTQ format (Raw Data) was filtered using fastp (0.22.0) software to get a high-quality sequence (Clean Data) for further analysis. The reference genome (Mus_musculus. GRCm39) and gene annotation files were downloaded from the genome website. The filtered reads were mapped to the reference genome using HISAT2 (version 2.1.0). We used HTSeq (version 0.9.1) statistics to compare the Read Count values on each gene as the original expression of the gene. RSEM (version 1.2.31) was used for the transcript abundance estimation. The following bioinformatic analysis is conducted in R software (4.2.3). The RSEM-expected counts were then subjected to DESeq2 (1.38.3)²² for the identification of differentially expressed genes (DEGs), with a criterion of |log2 fold change| >1 and adjusted P value of < 0.05. Venn diagrams showing the number of common DEGs across the experimental comparisons were drawn by using Jvenn.²³ At the same time, we used ComplexHeatmap (version 2.16.0)²⁴ to perform a bidirectional clustering analysis of all different genes of samples. We get a heatmap according to the expression level of the same gene in different samples and the expression patterns of different genes in the same sample with the Euclidean method to calculate the distance and the Complete Linkage method to cluster. ClusterProfiler (version 4.6.0)²⁵ software was used to carry out the enrichment analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway of differential genes, focusing on the significant enrichment pathway with a P value < 0.05. Other R packages used to declaratively creating graphics includes ggplot2 (3.4.4),²⁶ cowplot (1.1.3),²⁷ RColorBrewer (1.1.3),²⁸ and ggrepel (0.9.5).²⁹ Gene set enrichment analysis (GSEA) was performed using the clusterProfiler (version 4.6.2) package in R software. The input was a ranked gene list ordered by the log2 fold change from differential expression analysis. Gene sets were obtained from the Molecular Signatures Database (MSigDB, https://www.gsea-msigdb.org/), including curated (C2) and hallmark (H) gene sets. Enrichment scores and normalized enrichment scores (NES) were calculated based on 1000 permutations. Gene sets with adjusted P values < 0.05 were considered statistically significant. KEGG pathway enrichment analysis was conducted using the clusterProfiler (version 4.6.2) package in R software. DEGs were mapped to KEGG pathways using the enrichKEGG function, with the organism parameter set to “mmu” for a mouse or “as” for a human. Pathways with adjusted P values (Benjamini-Hochberg correction) less than 0.05 were considered significantly enriched. Visualization of enriched pathways was performed using dot plots and bar plots generated by the ggplot2 and enrichplot packages. All original data tables from the bioinformatics analyses are available in Supplementary Data S1.

RNA Extraction and Quantitative Real-Time PCR

Retinal samples were harvested 24 hours after IR injury. Total RNA was extracted using TRIzol reagent. A First-Strand cDNA Synthesis Kit (Cat. 11141; Yeasenbio, Shanghai, China), and 1 µg of total RNA were used to perform first-strand cDNA quantitative real-time PCR (qRT-PCR). To detect gene expression, qRT-PCR was using SYBR Green qPCR Master Mix (Cat. 11202; Yeasenbio) according to the manufacturer's instructions. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as an endogenous control. The primers (Sangon Biotech, Shanghai, China) used are listed in the Table.

Table.

Primer Sequences Used for qRT-PCR

Gene	Forward	Reverse
TNF-α	ACTCCAGGCGGTGCCTATGT	GTGAGGGTCTGGGCCATAGAA
IL-6	CCACTTCACAAGTCGGAGGCTTA	CCAGTTTGGTAGCATCCATCATTTC
IL-1β	CAACCAACAAGTGATATTCTCCATG	GATCCACACTCTCCAGCTGCA
Bcl3	GGAGCCGCGAAGTAGACGT	TGTGGTGATGACAGCCAGGT
CCL24	CTCCTTCTCCTGGTAGCCTG	ATGGCCCTTCTTGGTGATGA
Lrg1	CCATGTCAGTGTGCAGATTC	AAGAGTGAGAGGTGGAAGAG
LTB4R1	TACTAAGGCCTTTGCCCGAT	TACTAAGGCCTTTGCCCGAT
GAPDH	GGCAAATTCAACGGCACAGTCAAG	TCGCTCCTGGAAGATGGTGATGG

Western Blotting

The level of NF-κB p105/p50, Bcl3, TNF-α, IL-6, NLRP3, CASP1/cleaved-CASP1 (C-CASP1), N-GSDMD, IL-1β, Bax, and Bcl2, in the retina was measured using Western blotting. Mouse retinas were harvested 24 hours after IR injury and homogenized in RIPA lysis buffer (25 mM Tris pH 7.6, 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% sodium dodecyl sulfate; Cat. WB3100; New Cell & Molecular Biotech, Suzhou, China) with 1% protease inhibitor cocktail and 1% phosphatase inhibitor cocktails (Cat. P002; New Cell & Molecular Biotech) in a ratio of 100:1, and then centrifuged (12,000g, 10 minutes, 4°C). The supernatant was drawn and measured via a BCA protein assay kit (Cat. 23225; Thermo Fisher). A 20 µg aliquot of proteins from each individual animal was subjected to SDS-polyacrylamide gel electrophoresis at a constant voltage of 120 V was applied for 60 minutes, and transferred into a polyvinylidene difluoride (PVDF) membrane. The membrane was then incubated with 5% non-fat milk at room temperature for 1 hour and washed with PBST. The blot was incubated with antibody against NF-κB p105/p50 (1:1000, Cat. ab32360; Abcam), Bcl3 (1:500, Cat. sc-32741; Santa Cruz), NLRP3 (1:1000, Cat. 15101; Cell Signaling Technology), IL-1β (1:1000, Cat. ab254360; Abcam), CASP1 (1:2000, Cat. 22915; Proteintech, Wuhan, China), N-GSDMD (1:1000, Cat. DF13758; Affinity), Bax (1:2000, Cat. A19684; Abclonal), Bcl2 (1:2000, Cat. ET1702-53; HUABIO), TNF-α (1:2000, Cat. 60291; Proteintech), IL-6 (1:2000, Cat. 21865; Proteintech), and GAPDH (1:10000, Cat. 60004-1; Proteintech) or β- actin (1:10000, Cat. 811115; Proteintech) at 4°C overnight, and washed with PBST 5 times, for 5 minutes per time. The secondary antibody (1:10000, Cat. SA00001-1; SA00001-2; Proteintech,) was then added and the membrane was incubated at room temperature for 1 hour. After the secondary antibody incubation, the membrane was washed the same as after the primary antibody incubation described above. The bands were visualized with ECL (NCM Biotech) Western blotting detection reagents and quantified using Image J software. The relative expression was determined after normalizing to the individual GAPDH or β-actin levels.

Scanning Electron Microscopy and Transmission Electron Microscopy

Ultrastructural retinal morphology was observed via scanning electron microscopy (SEM) and transmission electron microscopy (TEM). Eyes were immediately enucleated after painless euthanization (24 hours after IR injury), fixed in 2.5% glutaraldehyde in 0.1 M phosphate buffer for 24 hours at 4°C, and sent to the Department of Pathology of Xiangya Hospital for sample preparation. The upper temporal retina was dissected and post-fixed in 1% osmium tetroxide for 2 hours. Following serial dehydration, the tissue was stained with 3% uranyl acetate and lead nitrate, and then embedded in Eponate 12. Sections of 1-µm thickness were stained with 1% toluidine blue and examined with a light microscope. Sections of 0.1-µm thickness were analyzed with a transmission electron microscope. The prepared tissue was freeze-dried under a vacuum, and a small amount of the sample was adhered directly to the conductive adhesive. The sample was sputter-coated with gold using an Oxford Quorum SC7620 sputter coater and observed and photographed using SEM. The SEM imaging was focused on the GCL, with careful sample orientation to avoid deeper layers. Cells that were large, rounded, and localized to the inner retinal surface, were defined as RGCs.

Immunohistochemistry Staining

All samples for immunohistochemistry staining were harvested 24 hours after IR injury. Frozen retina tissue sections were placed on a glass slide, fixed in 4% paraformaldehyde for 30 minutes, and incubated with 0.3% Triton X-100 solution for 10 minutes, then washed with PBS 5 times, blocked with 5% FBS for 30 minutes, and incubated with primary antibody ionized calcium-binding adaptor molecule 1 (Iba1; 1:200, Cat. 019-19741; WAKO, Osaka, Japan); CASP1 (1:100, Cat. 22915; Proteintech) at 4°C overnight. After being washed with PBS 3 times the next day, the retina tissues were then incubated with fluorescent Alexa Fluor 488 Conjugate (1:1000, Cat. 4412; Cell Signaling Technology) at room temperature for 1 hour. Finally, nuclei were counterstained with DAPI, and the sections were sealed. Images were acquired using a fluorescence microscope (N2-DM4B, Nikon).

Multiplex immunofluorescence staining was conducted using a tyramide signal amplification (TSA) kit (Servicebio Technology Co., Ltd., Wuhan, China), following the manufacturer's protocol. Briefly, paraffin-embedded retinal tissue sections were deparaffinized, rehydrated, and subjected to antigen retrieval. Sections were then blocked in 5% BSA in PBS and incubated overnight at 4°C with the primary antibody GSDMD (1:200, Cat. DF13758; Affinity, Cincinnati, OH, USA) The next day, the sections were incubated with the appropriate HRP-conjugated secondary antibody for 1 hour at room temperature, followed by a 10-minute incubation with TSA-iF555 (1:400) at room temperature. Following amplification, sections underwent additional antigen retrieval and blocking before being incubated with the second primary antibody, RBPMS (1:200, Cat. ab152101; Abcam). This was followed by incubation with the corresponding HRP-labeled secondary antibody and a 10-minute incubation with TSA-iF488 (1:500) to achieve signal amplification for the second target. Finally, nuclei were counterstained with DAPI, and the sections were sealed. Images were acquired using a fluorescence microscope (N2-DM4B, Nikon).

Flat-mounted retinas were prepared as described above and incubated with primary Anti-Iba1 (1:200, Cat. 019-19741; WAKO) and Anti-MHC-II (1:200; Abcam, ab233990) at 4°C overnight. After being rinsed with PBS 3 times the next day, the retina tissues were then incubated with secondary antibody Alexa Fluor 488 conjugate (1:1000, Cat. 4412; Cell Signaling Technology) and Alexa Fluor 555 conjugate (1:1000, Cat. 4409; Cell Signaling Technology), respectively.

Cell apoptosis was detected via a TUNEL kit (Cat. T6014; UElandy, Suzhou, China) following the manufacturer's instructions. Tissue sections were deparaffinized and permeabilized with proteinase K for 30 minutes, then rinsed with 0.1% Triton X-100 in PBS 3 times dropwise. Tissue slices were incubated with TUNEL test solution at 37°C, protected from light for 1 hour, and then washed with PBS 5 times.

All slides were counterstained with DAPI (Cat. EK-5103; Ecotop, Shanghai, China). Images were acquired using a confocal microscope (AirScan, Zeiss) or fluorescence microscope (N2-DM4B, Nikon). The staining intensity of the target protein was quantified using Image J software.

Statistical Analysis

All data are represented as mean ± standard deviations (SDs), with n representing the number of independent experiments conducted. Statistical analysis was performed using GraphPad Prism software. The P values were calculated by ordinary 1-way ANOVA with Tukey's multiple comparison test to analyze differences among multiple groups. Student’s t-test was used to analyze differences between two independent samples. *P < 0.05, **P < 0.01, and ***P < 0.001 were considered statistically significant.

Results

IR Injury-Induced RGC Death

Initially, we established the retinal IR injury model in mice to mimic retinal damage caused by pathologically high IOP (ph-IOP).³⁰ We collected the whole retina for flat mount and H&E staining at 12 hours, 24 hours, and 48 hours after IR injury. Results of flat mount (Fig. 1A) showed that the number of RGCs, which are stained with RBPMS, decreased over time. By 24 hours after IR injury, RGCs have already reduced significantly in the peripheral area and slightly decreased after 48 hours (Fig. 1B). H&E staining also confirmed this trend, showing reduced RGC density over time in the IR group (Fig. 1C), and serves as a valuable supplement to flat-mount quantification by enabling RGC counting from cross sections. The GCL appeared discontinuous, along with the presence of some vacuoles. Other retinal structures, including the inner nuclear layer (INL) and outer nuclear layer (ONL), also became loose and disordered over time, especially in the 24-hour group and the 48-hour group. The retina damage mainly occurred in the peripheral area in both the 24-hour and 48-hour groups. At the same time, 48 hours also showed a more severe loss of RGCs in the middle and central location and thinner retina in the peripheral area (Fig. 1D). Taken together, 24 hours after IR injury has already shown significant RGC loss and morphological changes in the model, so we chose 24 hours as the observation period for the following experiments.

Figure 1.

IR injury induced RGC death. (A) RGCs identified using RBPMS antibodies by immunofluorescence with flat-mounted retinas. The lower images are the enlarged representations of the boxed regions of the upper pictures. Scale bar = 200 µm. (B) Quantitative analysis of the RBPMS⁺ cell density at different timepoint after IR injury (Con, 12 hours, 24 hours, and 48 hours). (C) H&E staining showed retinal structure changes at different timepoint after IR injury (Con, 12 hours, 24 hours, and 48 hours) (left). Scale bar = 20 µm; representative images in the central, middle, and peripheral areas of the retina were selected. (D) Quantitative analysis of cell counts in GCL in the central, middle, and peripheral areas of the retina. RGCs, retinal ganglion cells; GCL, ganglion cell layer; Con: control; IR, ischemia reperfusion; RSV, resveratrol. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus control. ns, not significant.

Intravitreal Injection of Resveratrol Rescued RGC Death and Retinal Function

To evaluate the protective effect of RSV on the retina, we injected RSV into the vitreous body of the mice. Flat-mounted retinas from the IR + RSV group exhibited a higher density of RBPMS-positive cells compared to the IR group (Figs. 2A, 2B). A similar trend was confirmed via H&E staining that the density of RGCs was higher in the IR + RSV group compared with the IR group (Figs. 2C, 2D). In addition to morphological changes, we also evaluated the visual function of each group via F-VEP and F-ERG. The combined application of F-ERG and F-VEP provides a more comprehensive evaluation of retinal function. F-VEP was used to assess the function of the retinogeniculocortical pathway objectively. F-VEP recordings demonstrated a significant reduction in N1-P1 amplitudes and a delay in P1 latency in the IR group compared with controls, indicating impaired functional output of RGCs and slowed signal conduction along the visual pathway. Following RSV treatment, both N1-P1 amplitudes and P1 latency showed partial recovery, suggesting that RSV exerts a protective effect on RGC function and helps preserve visual signal transmission (Figs. 2E, 2F). The results of F-ERG, which were used to examine retinal function, showed that both a-waves and b-waves were prominently decreased after IR injury, but were able to be saved by RSV intravitreal treatment (Figs. 2G, 2H), which indicates that the widespread impact of IR injury on the outer and inner retinal layers was partially alleviated by RSV treatment, highlighting the neuroprotective effect of RSV on the whole retina. The above results suggested that intravitreal injection of RSV could significantly prevent RGC death and preserve retinal function.

Figure 2.

RSV rescued RGC death and retinal function. (A) The protective effects of RSV on RGCs in a 24-hour IR model were detected by RBPMS staining. Scale bar = 200 µm. (B) Quantitative analysis of the RBPMS⁺ cell density. (C) H&E staining was used to show the protective effect of RSV in the 24-hour IR model. Scale bar = 200 µm (upper) and 20 µm (lower). (D) Quantitative analysis of cell counts in GCL. (E) F-VEP results of Con, RSV, IR, and IR + RSV-treated groups. (F) Quantification of N1-P1 amplitudes and latency of P1-wave in the above four groups. (G) F-ERG results of Con, RSV, IR, and IR + RSV-treated groups. (H) Quantification of a-wave and b-wave amplitudes of the above four groups. RGCs, retinal ganglion cells; GCL, ganglion cell layer; Con, control; IR, ischemia reperfusion; RSV, resveratrol; F-VEP, flash visual evoked potentials; F-ERG, flash electroretinogram. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus Con.

The Bcl3/NF-κB p50 Signaling Pathway is Involved in the Protective Effect of Resveratrol

We used bulk RNA sequencing to elucidate the molecular mechanisms by which RSV alleviates retinal IR injury. As expected, the retinal tissues of mice in the three groups (Con versus IR versus IR + RSV, n = 3) exhibited distinct transcriptomic patterns (Fig. 3A). DEGs were identified using the R software package DESeq2, with the criteria of |log2FoldChange| >1 and adjusted P value (Padj) < 0.05. Overall, we identified 1250 upregulated and 1799 downregulated genes in the IR group compared to the control group, and 1435 upregulated and 174 downregulated genes in the IR + RSV group compared to the IR group (Figs. 3B, 3C). Among these, 84 genes were found to be upregulated in the IR group and downregulated in the IR + RSV group (Fig. 3D), suggesting potential molecular targets of RSV. Among the 84 genes, we focused on the top 10 highly expressed genes (Fig. 3E, left) and assessed the IR biological functions. Based on published studies, we selected four genes (Bcl3, Ccl24, Lrg1, and Ltb4r1) known to be associated with inflammation for further validation (see Fig. 3E, right). The mRNA expression levels of these genes (Bcl3, Ccl24, Lrg1, and Ltb4r1) were confirmed using qPCR (Figs. 3F–I). Among these four genes, Bcl3 exhibited the most significant differential expression between the IR group versus the IR + RSV group. Bcl3 is identified as a non-canonical regulator of NF-κB transcription factor, which is important in assisting NF-κB p50 or p52 to induce gene transcription.¹¹ NF-κB p50 is one of the most abundant subunits participating in the NF-κB pathway, which primarily triggers inflammatory response.³¹ Additionally, the KEGG pathway enrichment indicated that the NF-κB signaling pathway is highly involved following IR injury (Supplementary Fig. S1B). Transcriptomic comparisons among the Con, IR, and IR + RSV groups revealed that the non-canonical NF-κB pathway was upregulated after IR injury and downregulated after RSV treatment, making it one of the most prominently affected pathways (Supplementary Figs. S1C, S1D). Thus, we hypothesized that Bcl3-modulated NF-κB p50 signaling may serve as a downstream target of RSV. To test this hypothesis, we assessed the protein expression levels of Bcl3 and NF-κB p50/p105. As shown in Figure 3J, IR increased the expression of Bcl3 and NF-κB p50/p105 in the retinas of mice, whereas RSV treatment reversed the effects induced by IR.

Figure 3.

Bcl3/NF-κB p50 signaling pathway is involved in the molecular mechanism of RSV. (A) Heatmap of bulk-RNA sequencing data of the three groups: Con, IR, and IR + RSV. (B) Volcano plots displaying DEGs between the Con group and the IR group (upper), and between IR group and IR + RSV group (lower). (C) Bar chart showing the number of DEGs for all pairwise comparison among three groups. (D) Venn diagram illustrating the number of common genes which are upregulated in the IR group and downregulated in the IR + RSV group. (E) Heatmap of mRNA expression levels for the top 10 DEGs among common genes which are upregulated in the IR group, whereas downregulated in the IR + RSV group (left); mRNA expression level of four genes known to be associated with inflammation from the top 10 genes (right). (F–I) The qRT-PCR results of the relative mRNA expressions of Bcl3, Ccl24, Lrg1, and Ltb4r1 (from left to right). (J) Representative Western blot images of Bcl3, NF-κB p105, and NF-κB p50, and GAPDH. (K–M) Quantitative analysis of the relative expressions of Bcl3, NF-κB p105, and NF-κB p50, compared with GAPDH. Samples were harvested 24 hours after IR injury. DEGs, differentially expressed genes; Con, control; IR, ischemia reperfusion; RSV, resveratrol; qRT-PCR, quantitative real-time PCR. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus Con.

Resveratrol Alleviated IR-Induced Retinal Neuroinflammation

Given that the NF-κB signaling pathway is well-known for regulating inflammation, we identified whether RSV can fight against retinal neuroinflammation caused by IR injury. The qRT-PCR analysis identified a significantly elevated expression of inflammatory factors, including TNF-α, IL-1β, and IL-6, in the retina of the IR group, which were prominently downregulated after giving RSV treatment (Figs. 4A–C). This trend was further validated at the protein level by Western blot analysis, which showed consistent changes in the expression of these inflammatory cytokines (Figs. 4D, 4E, 5B, 5E). Iba1 is a cytoplasmic calcium-binding protein expressed in microglia, peripheral macrophages, or even infiltrating myeloid cells.^32,33 Although it is not an exclusive marker in microglia, it is still regarded as a reliable microglial marker, because Iba1 expresses uniformly throughout the cell body and process of microglia, thereby presenting a strong signal on different types of sections.^34,35 Elevated number of cells combining with changes in microglia's morphology and location shown by Iba1 staining can be used to evaluate microglia activation.^34,35 We used immunostaining for Iba1 (Fig. 4F) and collected images spanning all retina layers containing Iba1-positive cells, then compared cell morphology, number, and distribution. Microglia of the IR group presented less ramified and increased soma size, indicating microglia were activated after IR injury. Besides, an increase in the number of microglia localized in GCL was observed (see Fig. 4F), suggesting that IR-induced inflammation affect involves the GCL and might be a vital culprit for RGCs’ death. However, RSV treatment can significantly inhibit microglia’s activity and redistribution. Therefore, we supposed that RSV protects the retina via its anti-inflammation property.

Figure 4.

RSV alleviated IR-induced retinal neuroinflammation. (A–C) The qRT-PCR results showed that mRNA expression of TNF-α, IL-6, and IL-1β was increased by IR induction and downregulated by RSV. (D) Representative Western blot images of TNF-α, IL-6, and GAPDH. (E) Quantitative analysis of the relative expressions of TNF-α, IL-6, compared with β-actin. (F) Iba1 staining showed that the number of microglia was significantly increased in IR group and decreased in the IR + RSV group. Scale bar = 50 µm. (G) Quantification of cell counts of Iba1+ cells in the whole retina. Samples were harvested 24 hours after IR injury. qRT-PCR, quantitative real-time PCR; Con, control; IR, ischemia reperfusion; RSV, resveratrol. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus Con.

Figure 5.

RSV attenuated IR-induced RGC pyroptosis. (A) Western blot images of the relative expressions of NLRP3, CASP1 with C-CASP1, N-GSDMD, and IL-1β, compared with GAPDH. (B–E) Quantitative analysis of the relative expressions of NLRP3/GAPDH, C-CASP1/ CASP1, N-GSDMD/GAPDH, and IL-1β/GAPDH. (F) Representative immunofluorescence images of retinal sections limited to GCL. Sections were stained for GSDMD (red), RBPMS (green), and DAPI (blue). Scale bar = 20 µm (upper) and 5 µm (lower). (G) Quantitative analysis of fluorescence intensity of GSDMD and RBPMS-positive cells. (H) SEM images at × 1000 (upper), × 4000 (middle), × 10,000 (lower) magnification. Scale bar = 50 µm (upper) 10 µm (middle), and 5 µm (lower). (asterisk, *) Vesicles adhered to cell surface, (well number, #) rupture of cell membrane. CASP1, caspase 1; C-CASP1, cleaved-caspase 1; Con, control; IR, ischemia reperfusion; RSV, resveratrol; GCL, ganglion cell layer. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus Con.

Resveratrol Attenuated IR-Induced RGC Pyroptosis and Apoptosis

As shown above, we have demonstrated that the number of RGCs reduced in the IR group was significantly saved by RSV intravitreal treatment. To investigate whether this protective effect is associated with programmed cell death, we performed pathway analyses on the RNA sequencing data. GSEA revealed significant upregulation of gene sets related to apoptosis, pyroptosis, and necroptosis in the IR group compared to the control group (see Supplementary Fig. S1A), with changes in the pyroptosis and apoptosis pathways being particularly prominent. The KEGG pathway enrichment analysis of DEGs in the IR group further highlighted significant activation of immune- and inflammation-related pathways. Notably, the NF-κB signaling pathway (highlighted in red) was significantly enriched among these pathways (see Supplementary Fig. S1B). Given that NF-kB signaling pathway has been widely reported to be associate with pyroptosis and apoptosis, we decided to focus our subsequent analyses on these two cell death forms.

Western blot analysis revealed that NLRP3/C-CASP1/N-GSDMD/IL-1β protein level was significantly elevated in the IR group, and RSV treatment attenuated these elevations (see Figs. 5A–E). Co-immunostaining of N-GSDMD and RBPMS (Fig. 5F), as well as CASP1 staining (Supplementary Figure S2A), further confirmed that RSV reduced the fluorescence intensity of N-GSDMD and CASP1 in RGCs compared to the IR group, supporting the inhibitory effect of RSV on pyroptosis in this model. In addition, the Bax/Bcl-2 expression ratio was elevated after IR injury, indicating activation of the apoptotic pathway, and was downregulated following RSV administration (Figs. 6A, 6B). Similarly, the number of RBPMS and TUNEL double-positive cells significantly increased after IR injury but was markedly reduced following RSV treatment (Fig. 6C), suggesting attenuation of RGC apoptosis.

Figure 6.

RSV attenuated IR-induced RGC apoptosis. (A) Western blot images of the relative expressions of Bax and Bcl2, compared with β-actin. (B) Quantitative analysis of the relative expression of Bax/Bcl2, compared with β-actin. (C) Representative immunofluorescence images of retinal sections limited to GCL. Sections were stained for TUNEL (red), RBPMS (green), and DAPI (blue). Scale bar = 20 µm (upper) and 6 µm (lower). (D) Bar chart showing the number of TUNEL and RBPMS co-stained cells in the GCL. (E) TEM images at × 5000 (upper) and × 20,000 (lower) magnification. Scale bar = 5 µm (upper) and 1 µm (lower). Red arrow = cytoplasmic condensation; yellow arrow = nuclear chromatin condensation, along with apoptotic bodies; and blue arrow = mitochondrial vacuolization. Con, control; IR, ischemia reperfusion; RSV, resveratrol. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus Con. ns, not significant.

Then, we observed the cell morphology under SEM and TEM to determine the mode of cell death regulated by RSV. We found that some RGCs exhibited swelling and rounding, accompanied by disrupted cell membrane integrity showing as forming membrane pores under SEM (Fig. 5H). TEM confirmed pore-like membrane disruptions and membrane blebbing in damaged RGCs (Supplementary Fig. S2C). If treated with RSV before IR injury, the cell surface of RGCs remained intact and looked more similar to the normal RGCs in the control and RSV-treated group, suggesting that RSV helps preserve membrane integrity and suppress pyroptotic cell death. Moreover, TEM revealed classic apoptotic features in RGCs from the IR group, including chromatin condensation, mitochondrial vacuolization, and the presence of histiocytes phagocytosing karyorrhectic debris (apoptotic bodies; Fig. 6E). These apoptotic changes were significantly reduced in the IR + RSV group, indicating that RSV also mitigates IR-induced RGC apoptosis. Collectively, these findings suggest that RSV treatment exerts neuroprotective effects by attenuating both pyroptosis and apoptosis of RGCs following retinal IR injury.

Intravitreal Injection of JS-6 Rescued RGC Death and Retinal Function

Our RNA sequencing results identified Bcl3/ NF-κB p50 as a potential signaling pathway regulated by RSV. To investigate the role of the Bcl3/ NF-κB p50 signaling pathway in IR injury and the mechanism of RSV treatment, JS-6, known as a Bcl3 inhibitor, was thereby introduced to our experiments. By utilizing JS-6, the number of RBPMS-positive cells in the flat-mounted retina was increased in a dose-dependent manner compared with the IR group (Fig. 7A). We sectioned each flat-mounted retina into four leaves and chose three same-sized areas on the central, middle, and peripheral sites of each leaf for counting (Figs. 7B–E). After IR injury, RBPMS-positive cells were significantly decreased. However, JS-6 could alleviate this impact, and as the concentration of JS-6 elevated, the protective effect seemed to be increased. The morphology of the retina shown by H&E staining also verified the same trends. Cell counts in the GCL gradually increased as the dose of JS-6 increased (Figs. 7F, 7G). The above results suggested that JS-6 can significantly rescue RGC death and attenuate retinal morphological changes as its dose increased. We also confirmed the JS-6 protective effect by using F-VEP and F-ERG. The decreased amplitudes and latency of P1 waves in F-VEP (Figs. 7H, 7I) and the decline of a- and b-waves in F-ERG (Figs. 7J, 7K) were also saved by JS-6. In addition, the protective effect increased in a concentration-dependent manner with increasing levels of JS-6. As shown by the above results, we selected 5 mM as the best concentration of JS-6 in our following experiments. In addition, we suggested that using JS-6 to inhibit the Bcl3 signaling pathway alleviates RGC cell death and restores retinal function. The similar protective effect between RSV and JS-6 also indicated the importance of the Bcl3/NF-κB p50 pathway in RGC survival.

Figure 7.

Intravitreal injection of JS-6 rescued RGC death and retinal function. (A) RBPMS staining of retinal flat mounts. The lower images are the enlarged representations of the boxed regions of the upper pictures. Scale bar = 200 µm. (B) Diagrammatic sketch of RBPMS⁺ cell quantitation. (C–E) Quantitative analysis of the RBPMS⁺ cell density in the central, middle, and peripheral area of retina. (F) H&E staining images at × 4 (upper) and × 63 magnification (lower). Scale bar = 200 µm (upper) and 20 µm (lower). (G) Quantitative analysis of cell counts in GCL. (H) Images of merged F-VEP waves. (I) Quantification of N1-P1 amplitudes and latency of p1-waves. (J) Images of merged F-ERG waves. (K) Quantification of a-wave and b-wave amplitudes. GCL, ganglion cell layer; Con, control; IR, ischemia reperfusion; F-VEP, flash visual evoked potentials; F-ERG, flash electroretinogram. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus Con.

JS-6 Alleviated IR-Induced Retinal Neuroinflammation

Given the consistent protective effects of RSV and JS-6, and the role of the NF-κB signaling pathway in regulating inflammation, we investigated whether JS-6 administration affected microglial activation. Thus, we were able to further clarify the ability of JS-6 to fight retinal neuroinflammation caused by IR injury. We co-stained Iba1 (green) with MHC-II (red) on flat-mounted retinas and observed cell morphology, number, and distribution (Fig. 8). Microglia in retinas after IR injury exhibited an increase in quantity and superficiality in location (see Fig. 8A), along with a reduction of ramification and a merge of Iba1-positive cells with MHC-II perinuclear positive cells (see Figs. 8B–D), indicating a quantitative increase and activation of microglia in the GCL after IR injury. In comparison, JS-6 treatment inhibited the activation and redistribution of microglia; therefore, it leads to the inference that JS-6 has anti-inflammatory properties and supports the assumption of JS-6 being a protective factor against retinal IR injury.

Figure 8.

JS-6 alleviated IR-induced retinal neuroinflammation. (A) (i) Local fluorescence Images of Iba1 stained retinal flat mounts under 40 × objective lens. Scale bar = 50 µm. (ii) Microglia cell morphology extraction of previous images. (B) Local fluorescence Images of Iba1 (green, upper), MHC-II (red, middle), co-stained (merged, lower) retinal flat mounts under 20 × objective lens. Scale bar = 20 µm. (C) Quantification and morphological classification of IBA1⁺ cells. (D) Mean fluorescence intensity of MHC-II stained cells and cell counts of IBA1⁺/ MHC-II⁺ cells. ****P < 0.0001. Con, control; IR, ischemia reperfusion.

JS-6 Attenuated IR-Induced RGC Pyroptosis and Apoptosis

To assess whether the Bcl3 pathway regulates RGC pyroptosis and apoptosis, we intravitreally injected 5 mM JS-6 24 hours before IR injury. We tested the protein expression level of pyroptosis, including NLRP3/N-GSDMD/C-CASP1/IL-1β, and apoptotic markers, including Bax and Bcl2. We found that the elevated protein level of NLRP3/N-GSDMD/C-CASP1/IL-1β induced by IR injury was significantly downregulated if treated with JS-6 (Fig. 9A–E). The increased expression ratio of Bax/Bcl2 was also restored using JS-6 (Figs. 9F, 9G). These data indicated that pyroptosis and apoptosis were significantly inhibited by blocking the Bcl3 signaling pathway. The similar trends between using RSV and JS-6 also suggested that RSV's protective effect against pyroptosis and apoptosis might be achieved by inhibiting the Bcl3/NF-κB signaling pathway.

Figure 9.

JS-6 attenuated IR-induced RGC pyroptosis and apoptosis. (A) Western blot images of the relative expressions of NLRP3, CASP1 with C-CASP1, N-GSDMD, IL-1β, and β-actin. (B–E) Quantitative analysis of the relative expressions of NLRP3, C- CASP1/ CASP1, N-GSDMD, and IL-1β, compared with β-actin. (F) Western blot images of Bax, Bcl2, and β-actin. (G) Quantitation of Bax/Bcl2. Con, control; IR, ischemia reperfusion. One-way ANOVA was used for the comparison. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001 versus Con.

Discussion

Glaucoma is a multifactorial neurodegenerative disease of the optic nerve, and one of the leading causes of blindness in the world.¹ IR injury is a well-established experimental model to mimic the acute phase of acute angle-closure glaucoma, which activates retinal glial cells, leading to the release of pro-inflammatory cytokines and exacerbation of retinal inflammation, ultimately contributing to RGC death.³⁶ The fate of RGCs is highly associated with the responses of surrounding glial cells.³⁷ The uncontrolled response of glial cells occurs in glaucoma would have a harmful impact on the retina.³⁸ Glial cells in the retina, which are typically classified as macroglia (Müller cells and astrocytes) and microglia, have an important role in mediating inflammatory responses.³⁹ Retinal microglia cells are relatively inactive under normal conditions, and constantly monitor their surrounding milieu.⁴⁰ They have been shown to become reactive and redistributed in the retina, producing inflammatory cytokines and causing further neuronal damage in patients with glaucoma and experimental glaucoma models.⁴¹ Because neuroinflammation is a vital component in RGC survival, seeking agents that can control retinal inflammation and inhibit microglia activation and redistribution may provide a new strategy for treating glaucoma.

RSV is one of the potential candidates for ocular disease prevention and treatment, due to its multifunctional power, including its anti-inflammatory and antioxidant properties.⁴² In this study, we demonstrated that RSV could relieve retinal inflammation caused by IR injury by reducing the activation of microglia and their migration to GCL. We also evaluated the condition of retinas between the groups with or without pretreatment of RSV not only by counting the number of remaining RGCs in GCL through flat-mount and H&E staining, which showed a trend that is consistent with other studies conducted in different animal glaucoma models,^43–45 but also through evaluating the retinal function through F-VEP and F-ERG. Our research indicated that RSV could significantly prevent RGC death and preserve retinal function by controlling microglia-mediated neuroinflammation. Emerging studies in recent years focused on the molecular mechanisms of RSV on RGCs. For instance, Ji et al. investigated that RSV protected RGCs by decreasing the HIF-1α/VEGF and p38/p53 signaling axes and increasing the PI3K/Akt pathway.⁴⁶ RSV protected RGC axons from injuries by suppressing the phosphorylation of JNK proteins via SIRT1.⁴⁷ Studies have shown that RSV administration can remarkably decrease NF-κB, IL-6 expression.^48,49 Our study identified the Bcl3/NF-κB p50 pathway as a key molecular mechanism underlying the neuroprotective effects of RSV on RGCs, based on RNA sequencing analysis. Western blotting further confirmed that the expression levels of Bcl3 and NF-κB p50 were significantly upregulated following IR injury and were downregulated when RSV was administered 24 hours prior to IR injury. Besides, the results of molecular docking experiments showed that hydrogen bonds were formed between RSV and the Bcl3 protein structure (Supplementary Fig. S3). A hydrogen bond is an electrostatic interaction between a hydrogen bonding donor (a hydrogen atom in a polar bond) and a hydrogen bonding acceptor (a strongly electronegative atom with a lone pair available for bonding, typically O, N, or F).⁵⁰ Small-molecule drug interactions are impacted by hydrogen bonding at several degrees of complexity, ranging from interactions with other small molecules to the most complex supramolecular assemblies, such as proteins and membranes.⁵⁰ The biological activity, pharmacokinetics, and physicochemical characteristics of medications may all be greatly impacted by these interactions.⁵¹ The formation of hydrogen bonds may indicate a promotion of protein-ligand's affinity occurring between drug and its target protein.^52,53 As hydrogen bonds in the protein also help stabilize the secondary and tertiary structures of the protein and help drive the protein to the folded state,⁵⁴ it is possible that the interaction between RSV and Bcl3 facilitating by hydrogen bonds may have affected the protein structure conformation of Bcl3 and thus affected the following translation process, which was inconsistent with Western blot results that intravitreal injection of RSV reduced the protein expression of Bcl3.

Bcl3 is an atypical member of the ikappa B inhibitor family.⁵⁵ Unlike classical members, Bcl3 is a nuclear transcription cofactor that is mainly expressed in the nucleus.⁸ Interestingly, Bcl3 seemed to play a dual role in regulating NF-κB transcription, showing either promoting or inhibiting according to the promoter, cell type or received stimulation.⁵⁵ The NF-κB family of transcription factors, containing several proteins, including p65 (RelA), c-Rel, RelB, p105/p50, and p100/p52, acts as a master regulator of diverse physiological processes, of which a well-acknowledged function is regulation of inflammatory responses.⁸ Bcl3 is necessary for NF-κB signals, especially subunits p50 and p52, because p50 and p52 homodimers rely on the presence of Bcl-3 to initiate gene transcription.⁵⁶ Although more and more studies focus on the ability of Bcl3 to regulate the function of various cells through the NF-κB signaling pathway, most of the studies are conducted in tumors or inflammatory and autoimmune diseases.^11,55 In some diseases, such as acute liver injury, Bcl3 leads to increased cell apoptosis.⁵⁷ However, Bcl3 is usually considered an anti-apoptotic gene in many other diseases,^58–60 suggesting that whether Bcl3 promotes or inhibits the transcription of NF-κB-dependent antiapoptotic genes still needs further investigation in different disease models. Our investigation first identified the importance of Bcl3 in retinal degenerative disease, hoping to give new insight into glaucoma treatment and replenish the pharmaceutical mechanism studies of RSV. In addition, to the best of our knowledge, no studies have investigated the relationship between RSV and Bcl3 together with its downstream pathway. Studies showed that NF-κB pathways are linked to inflammatory pathway components in multiple retinal degenerative diseases, including glaucoma, diabetic retinopathy, etc.^61,62 Therefore, controlling NF-κB activation might be a potential measurement to fix the inflammatory chaos and slow down RGC death in retinal diseases, especially glaucoma. Previous studies have demonstrated that NF-κB upregulates the transcription of NLRP3, pro-IL-18, and pro-IL-1β, leading to the activation of pyroptosis.^63,64 Besides, NF-κB has a dual role in the regulation of apoptosis, depending on the balance between genes that controls cell survival and apoptosis.⁸ Given that our pathway analyses indicated the involvement of pyroptosis and apoptosis following IR injury, we propose that Bcl3/NF-κB may serve as a key signaling axis through which RSV mitigates RGC pyroptosis and apoptosis.

Apoptosis used to be considered as the primary mechanism of RGC death in glaucomatous neurodegeneration.⁶⁵ Previous studies of RSV reported its effect on preventing apoptosis of RGCs under elevated IOP conditions. Seong et al. showed that intraperitoneal administration of RSV resulted in a decreased loss of retinal cells and downregulation of the expression of caspase-3 and caspase-8, thus inhibiting apoptosis.⁶⁶ RSV inhibited RGC death by blocking the apoptotic-related genes, including Bax, caspase-3, and upregulating Bcl2, indicating its potential therapeutic effectiveness against glaucoma-induced IR injury.^49,67 A study conducted by Pang's team showed that RSV enhanced the survival of rat-originated RGCs and partially alleviated apoptosis.⁶⁸ Nevertheless, emerging studies have shown pyroptosis as another mode of RGC death during the past few years.^19,69 Pyroptosis is a lytic and inflammatory type of programmed cell death that is usually triggered by inflammasomes, such as NLRP3, and executed by gasdermin proteins.⁷⁰ The activation of microglia leads to the release of cytokines, which can be a stimulus triggering the RGC pyroptosis procedure. Plasma membrane disruption of the RGCs undergoing pyroptosis will lead to a subsequent release of cellular contents and pro-inflammatory mediators, including IL‐1β and IL‐18, which amplifies inflammatory response in the retina, thereby resulting in more severe retinal damage.¹⁶ Few studies focused on whether RSV could inhibit pyroptosis in retinal diseases. Feng et al. identified that the RSV treatment exerts a protective effect by inhibiting the NLRP3 inflammasome in the retinal IR injury model.⁴³ Xie et al. observed that RSV alleviates retinal IR injury by inhibiting the NLRP3-mediated pyroptosis pathway.⁷¹ However, none of them provide convincing direct evidence, such as SEM and TEM, to show the direct morphological features of either apoptosis or pyroptosis. In this research, we observed pyroptotic and apoptotic morphological features as stated above in RGCs of the IR group via SEM and TEM, respectively. After pretreating with RSV, we noticed that the morphological changes of RGCs are milder. Our research provided a more direct and convincing evidence to confirm the actual protective effect of RSV against pyroptosis and apoptosis of RGCs.

To further verify the importance of the Bcl3/NF-κB p50 pathway in RGC survival, for the first time, we pretreated mice with JS-6, a newly discovered small compound that specifically inhibits the protein-protein interaction between Bcl3 and NF-κB p50,¹³ before IR injury. JS-6 pretreatment before IR injury recapitulated the neuroprotective effects of RSV, including reduced pyroptosis and apoptosis of RGCs, suppressed microglial activation and retinal inflammation, and preservation of retinal function. Unlike nonselective inhibitors, JS-6 targets only the Bcl3/NF-κB p50 interaction, allowing for a more precise modulation of NF-κB activity. This selective inhibition may preserve therapeutic efficacy while minimizing potential systemic risks associated with global suppression of the NF-κB pathway.⁷² This provided more vital evidence that the Bcl3/NF-κB p50 signaling pathway is not only a crucial pathway regulated by RSV, but also a potential therapeutic target that is newly discovered for glaucoma treatment.

Although our results indicate that the Bcl3/NF-κB p50 signaling pathway plays a role in the protective mechanism of RSV, we were unable to confirm whether Bcl3 overexpression can reverse the anti-inflammatory and neuroprotective effects of RSV due to a lack of Bcl3-specific activators and technical constraints prevented us from achieving reliable in vivo Bcl3 overexpression. In addition, for future investigation of Bcl3 in retinal diseases and RSV mechanism, it is worth noting that the dual functional activity of Bcl3 is controlled by its induced expression and regulated by post-translational means, including phosphorylation and ubiquitination.¹¹ Research has shown that phosphorylation of specific serine residues of Bcl3 leads to its proteasomal degradation in cancer cells,⁷³ indicating that RSV might affect the functional activity of Bcl3 by means of regulating posttranslational modification, therefore reducing NF-κB-related inflammatory gene transcription. Besides, although most current studies focus on Bcl3 and its interacting mechanism with p50, evidence also showed that Bcl3 can directly interact with p52 homodimers, the mechanism of Bcl3 regulating p52 homodimer activation remains unclear.⁷⁴ Therefore, whether Bcl3/p52 is involved in the RSV-mediated neuroprotective process is still to be explored.

In summary, we comprehensively assessed the protective effect of RSV on the retina in the IR model through basic morphological evidence and by evaluating retinal function. Moreover, we demonstrated that RSV could reduce the activation and redistribution of microglia, thereby mitigating retinal inflammation. We also provide direct evidence via electron microscopy to confirm the occurrence of pyroptosis and apoptosis in the IR injured retina, which was mitigated by RSV. In addition, we identified Bcl3/ NF-κB p50 as a novel signaling pathway that is regulated by RSV and might become a potent target for the treatment of glaucoma (Fig. 10). More broadly, our work provides insights into the molecular mechanism of RSV and new targets for glaucoma diagnosis and treatment.

Figure 10.

RSV and JS-6 mitigates retinal inflammation, as well as pyroptosis and apoptosis of retinal ganglion cells in the ischemia-reperfusion-injured retina via inhibiting Bcl3/ NF-κB p50 signaling pathway.