GSK840 Alleviates Retinal Neuronal Injury by Inhibiting RIPK3/MLKL-Mediated RGC Necroptosis After Ischemia/Reperfusion

Abstract

Purpose: This study aimed to explore the impact of GSK840 on retinal neuronal injury after retinal ischemia/reperfusion (IR) and its associated mechanism.

Methods: We established an in vivo mouse model of IR and an in vitro model of oxygen and glucose deprivation/reoxygenation (OGDR) in primary mouse retinal ganglion cells (RGCs). GSK840, a small-molecule compound, was used to specifically inhibit RIPK3/MLKL-dependent necroptosis. Retinal structure and function evaluation was performed by using hematoxylin and eosin staining, optical coherence tomography, and electroretinography. Propidium Iodide (PI) staining was used for detection of necroptotic cell death, whereas Western blot analysis and immunofluorescence were used to assess necroptosis-related proteins and inner retinal neurons.

Results: RIPK3/MLKL-dependent necroptosis was rapidly activated in RGCs following retinal IR or OGDR. GSK840 helped maintain relatively normal inner retinal structure and thickness by preserving inner retinal neurons, particularly RGCs. Meanwhile, GSK840 ameliorated IR-induced visual dysfunction, as evidenced by the improved amplitudes of photopic negative response, a-wave, b-wave, and oscillatory potentials. And GSK840 treatment significantly reduced the population of PI⁺ RGCs after injury. Mechanistically, GSK840 ameliorated RGC necroptosis by inhibiting the RIPK3/MLKL pathway.

Conclusions: GSK840 exerts protective effects against retinal neuronal injury after IR by inhibiting RIPK3/MLKL-mediated RGC necroptosis. GSK840 may represent a protective strategy for RGC degeneration in ischemic retinopathy.

Retinal ischemia is a prevalent pathological state resulting in irreversible visual impairment in a variety of ischemic retinopathies, such as diabetic retinopathy (DR), retinal vessel occlusion, and glaucoma.^1,2 Neural cells in the retina, especially retinal ganglion cells (RGCs), exhibit a high susceptibility to ischemia and therefore undergo cell death in these diseases.^3,4 Because there are no clinically approved therapies to rescue RGCs in ischemic retinopathy, new effective therapies are desperately needed. 

Necroptosis, a genetically programmed cell death pathway, exhibits characteristics of both programmed apoptosis and nonregulated necrosis.⁵ This process is regulated by two pivotal proteins, receptor-interacting protein kinase 3 (RIPK3) and mixed-lineage kinase domain-like protein (MLKL), which are successively phosphorylated and activated.⁶ The active MLKL then disrupts cellular integrity, causing cellular distension and membrane breakdown.⁶ Necroptosis has been connected to impaired neural function and cellular demise across different types of neurodegenerative diseases, such as Alzheimer's disease,⁷ Parkinson's disease,⁸ and multiple sclerosis.⁹ Moreover, RGCs have been shown to undergo necroptosis in models of glaucoma^4,10 and DR.¹¹ 

Fortunately, inhibition of neuronal necroptosis, either via pharmacological blockade^12–14 or genetic elimination of RIPK,^15,16 has shown strong neuroprotection within the central nervous system, as well as the retina. GSK840, a specific small-molecule inhibitor of RIPK3, is widely used to inhibit the necroptosis pathway in cell and animal studies.^17–22 Research has shown that therapeutic application of GSK840 can inhibit RIPK3-mediated necroptosis in microglia using an oxygen-induced retinopathy model.²¹ However, the therapeutic effect of GSK840 in retinal ischemia/reperfusion (IR) model remains to be investigated. In this study, by using a mouse model of IR along with an in vitro model of oxygen and glucose deprivation/reoxygenation (OGDR), we evaluated the efficacy of GSK840 in providing retinal neuroprotection against IR. 

Material and Methods

Reagents

Tetracaine, tropicamide phenylephrine, and carboxymethylcellulose sodium were from the Zhongshan Ophthalmic Center (Guangzhou, Guangdong, China). Neurobasal medium, B-27 supplement, GlutaMAX supplement, glucose-free DMEM, trypsin–EDTA, penicillin-streptomycin, and poly-D-lysine were obtained from Gibco Thermo Fisher Scientific Inc. (Waltham, MA, USA). Triton X-100, dimethyl sulfoxide, 4% paraformaldehyde, and normal goat serum were purchased from Solarbio Life Sciences (Beijing, China). Optimal cutting temperature (OCT) compound, 4',6-diamidino-2-phenylindole (DAPI), and RIPA lysis buffer were from Servicebio Technology (Wuhan, Hubei, China). BCA protein assay kit and skim milk powder were purchased from Biosharp Life Sciences (Hefei, Anhui, China). Pentobarbital sodium was from Sigma Chemical Co. (St. Louis, MO, USA). Propidium Iodide (PI) was from Invitrogen Thermo Fisher Scientific Inc. (Waltham, MA, USA). Anti-macrophage antiserum was from Accurate Chemical (Westbury, NY, USA). GSK840 was from MedChemExpress (Shanghai, China). Primary antibodies are listed in Supplementary Table S1. 

Animals

Male and female C57BL/6J mice (Animal Laboratory of Zhongshan Ophthalmic Center, Guangzhou, China) were kept in a 12-hour light/dark cycle, with the dark phase starting at 8:00 PM and ending at 8:00 AM. The experimental protocols (O2022030) were granted ethical approval by the Institutional Animal Care and Use Committee of Zhongshan Ophthalmic Center. Our animal procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. 

IR Model and Drug Administration

C57BL/6J mice were induced with the IR model according to previous reports.^10,23 Briefly, mice anesthetization was performed with intraperitoneal injection of a solution of 1% pentobarbital sodium (50 mg/kg), and corneas were given anesthesia via 0.5% tetracaine. Tropicamide phenylephrine was used to dilate pupils. Sterile saline solution was infused into the anterior chamber of the right eye using a 32-gauge needle to sustain the IOP at 110 mm Hg for 60 minutes. After 60 minutes, the needle was pulled out to restore a normal IOP. The left eye without IOP elevation was used as a non-IR control group. 

To investigate the dose-dependent effects of GSK840, the drug was dissolved in dimethyl sulfoxide and diluted to concentrations of 0.5, 1, and 5 mM. Immediately after reperfusion, 2 µL of GSK840 or control vehicle (same solvent) was intravitreally injected using a microliter syringe (Fig. 3B). Retinas were collected for analysis at the designated time points (Fig. 3B). Mice were divided into GSK840-treated and untreated groups based on IR or non-IR control: a control + vehicle (Con) group, a control + GSK840 (Con + GSK840) group, an IR + vehicle (IR) group, and an IR + GSK840 (IR + GSK840) group. 

Hematoxylin & Eosin (HE) Staining and Histological Analysis

After fixation, the eyes were embedded in paraffin and sliced. Retinal sections were subsequently stained using HE, and images were acquired using a Panoramic MIDI scanner (3DHISTECH, Budapest, Hungary). For histological analysis, six eyeballs were included in each experimental group. Three discontinuous retinal sections through the optic nerve were randomly selected for each eyeball. Within each selected section, one area was chosen for analysis. Consequently, a total of three areas from three sections were used to measure the average values for each eyeball. The thickness of individual retinal layer, including nerve fiber layer/ganglion cell layer (NFL/GCL), inner plexiform layer (IPL), inner nuclear layer (INL), outer plexiform layer, and outer nuclear layer (ONL), was measured using ImageJ software (National Institutes of Health, Bethesda, MD, USA). Measurements were taken at three distinct regions, approximately 500 µm (central), 1100 µm (middle), and 1700 µm (peripheral) away from the optic nerve.²⁴ A reference line was placed perpendicular to the retinal layers at the measurement region to ensure accuracy. The segmentation of retinal layers was determined based on established literature,^25,26 as illustrated in Supplementary Figure S1. 

Spectral Domain Optical Coherence Tomography (SD-OCT)

Anesthetized mice with dilated pupils were positioned on the imaging platform. A carboxymethylcellulose sodium eyedrop was applied to maintain cornea moist. Living animals were imaged with a focus on the optic nerve head using SD-OCT imaging system (Envisu R4310; Bioptigen, Inc., Morrisville, NC, USA). The raw SD-OCT data was then exported and assessed through ImageJ software (National Institutes of Health) to obtain the thickness of the retinal nerve fiber layer (RNFL), ganglion cell complex (GCC, including RNFL, GCL, and IPL) and total retina. 

Electroretinography (ERG)

The mice underwent overnight dark adaptation prior to intraperitoneal anesthesia and pupil dilation. Two gold ring active electrodes were attached to the corneal surface, while the reference and ground electrodes were placed inside the cheek and tail skin, respectively. Flash stimuli were generated using the RetiPort system (Roland, Brandenburg, German). Scotopic ERG was elicited with three light intensities: 0.01, 3.0, and 10 cd s/m². Photopic negative response (PhNR) was elicited using a 20 cd s/m² stimulus. Amplitude measurements were obtained for the scotopic a-wave, b-wave, oscillatory potentials (OPs), and PhNR. 

Cell Culture and Treatment

To obtain purified primary mouse RGCs, a two-step immunopanning protocol was performed with minor modifications.^27,28 Retinas from newborn mice (postnatal day 0–3) were digested by trypsin-EDTA and then subjected to anti-macrophage antiserum incubation for macrophages removal. Next, nonadherent cells were shifted to plates coated with Thy-1.2 antibody and poly-D-lysine. After 30 minutes, adherent RGCs were rinsed with PBS and maintained in culture medium (neurobasal medium containing 1% GlutaMAX, 2% B27, and 1% penicillin-streptomycin). GSK840 (1 µM) or vehicle was administered to RGCs for two hours before the OGDR stimulation.^17,18 

OGDR Model

To mimic retinal IR in vitro, the OGDR model was constructed as described previously with minor changes.^29,30 The culture medium of primary mouse RGCs was changed to glucose-free DMEM after four-day cell culture. Primary RGCs were then put in a hypoxic chamber with 95% N₂ and 5% CO₂ at 37°C for two hours. The control group was exposed to culture medium in a normoxic incubator (95% air and 5% CO₂, 37°C) for the same duration. After the OGD treatment, the RGCs were returned to culture medium under normoxic conditions for zero to 12 hours. RGCs were harvested for further experiments. 

Immunofluorescence

Eyeball fixation at room temperature (RT) was performed by using 4% paraformaldehyde. For whole mounts, retinas were dissected and blocked with PBS containing 5% normal goat serum and 0.1% Triton-X-100 for one hour. After incubating throughout the night at 4°C with the Tuj1 primary antibody, secondary antibodies were subsequently applied and left to incubate at RT for two hours. Retinal sections were obtained by mounting the eyeballs in OCT compound at −80°C and cutting into 10 µm cryosections. Primary antibodies were applied onto the slides and allowed to incubate overnight at 4°C, after which secondary antibodies and DAPI were added. In vitro, incubation with primary and secondary antibodies was performed in the cultured cells after 15 minutes’ fixation with 4% paraformaldehyde. LSM 980 confocal microscope (Carl Zeiss, Jena, German) was used for image acquisition. 

For the quantification of immunostaining on retinal section, six eyeballs were used in each experimental group. From each eyeball, three discontinuous retinal sections passing through the optic nerve were randomly selected. Within each of these selected sections, one specific area was chosen for analysis, resulting in a total of three areas from three sections being used to calculate average values for each eyeball. Mean fluorescence intensity analysis of phosphorylated RIPK3 (p-RIPK3) expression was conducted by measuring the average pixel intensity of p-RIPK3 within the region of interest. Threshold settings for defining the region of interest remained consistent across all images. To correct for background intensity, the background fluorescence intensity was measured and subtracted. After the background subtraction, mean fluorescence intensity values were normalized to the control group. All labeled cells were manually counted using the Multi-point Tool of Image J software (National Institutes of Health). 

PI Staining

For in vivo experiments, mice were intravitreally injected with 1 µl of PI (5 µg/ml) and killed 1 hour post-injection. 10-µm frozen sections were obtained after eye enucleation and OCT embedding. For in vitro experiments, primary RGCs on slides were immersed in PI (5 µg/ml) at 37°C for 15 minutes, washed with PBS, and fixed with 4% paraformaldehyde at RT for 15 minutes. For double-staining of PI/Tuj1, PI staining was performed first, followed by immunostaining of Tuj1. DAPI was used to label nuclei. All samples were imaged with LSM 980 confocal microscope (Carl Zeiss, Jena, German). 

Western Blot (WB)

Protein extraction from retinas and cells was performed by homogenization in RIPA lysis buffer supplemented with protease and phosphatase inhibitors, followed by sonication. BCA Protein Assay Kit was used to measure protein concentration. Protein separation was achieved by SDS-PAGE and subsequently moved onto polyvinylidene fluoride (PVDF) membranes and blocked with nonfat milk at RT for two hours. Primary antibodies were added and allowed to react overnight at 4°C, subsequently treated with the respective secondary antibodies for two hours at RT. The gray intensity of protein bands was analyzed with ImageJ software (National Institutes of Health). 

Statistical Analysis

Each experiment was conducted a minimum of three times. Prism 8 (GraphPad, San Diego, CA, USA) was employed to perform the statistical analyses, including unpaired t test and one-way ANOVA with Tukey's post hoc test. The data were presented as mean ± SEM. P < 0.05 was considered as statistically significant. 

Results

IR Induced RIPK3-Dependent Necroptosis in RGCs

To elucidate the role of necroptosis in retinal IR, a murine model where retinal ischemia is triggered by rapid elevation of IOP was used.³¹ We assessed the expression of necroptosis-associated proteins. WB analysis revealed the upregulation of p-RIPK3, total RIPK3 (t-RIPK3), phosphorylated MLKL (p-MLKL), and total MLKL (t‐MLKL) in the retinas 12 hours after IR, then the protein levels were declined to normal one day after injury (Figs. 1A, 1B, Supplementary Fig. S3). Furthermore, p-RIPK3 immunoreactivity was significantly elevated in the cells of the INL, IPL, and GCL 12 hours after injury, as confirmed by immunostaining (Figs. 1C, 1D). To further specify the necroptotic cell types in the GCL, we conducted Tuj1/p-RIPK3 double staining. Co-immunostaining of p-RIPK3 and Tuj1⁺ RGCs revealed that necroptosis occurred in RGCs in response to IR (Fig. 1E). Our findings indicated that necroptosis was activated after IR, particularly in RGCs. 

Figure 1.

 

The RIPK3/MLKL pathway was activated in RGCs following IR. (A, B) Western blot and densitometry analysis of p-RIPK3, t-RIPK3, p-MLKL, and t-MLKL expression in control (Con) and injured retinas at six hours, 12 hours, one day, three days, and five days after IR. n = 3. *P < 0.05; **P < 0.01; ***P < 0.001; one-way ANOVA. (C, D) Representative p-RIPK3 (red) images and quantification of the fluorescence intensity in control and injured retinas at 12 hours after IR. n = 6. ***P < 0.001; Student's t test. (E) Representative images of Tuj1 (green) and p-RIPK3 (red) co-labeling in IR retinas. Magnified inset is shown on the right. Scale bar: 20 µm and 2 µm in the zoom picture. Data are shown as mean ± SEM. Blue, DAPI staining.

OGDR Induced RIPK3-Dependent Necroptosis in Primary RGCs

To further confirm whether RGCs underwent RIPK3-driven necroptosis in response to a pathological insult, primary mouse RGCs were challenged with OGDR as an in vitro model of IR.^32,33 The RGCs were cultured and identified by the markers Tuj1 and RBPMS (Fig. 2A).^34,35 WB analysis manifested that necroptosis-associated proteins expression in cultured RGCs were increased one hour after OGD, and subsequently declined to baseline levels six hours after stimulation (Figs. 2B, 2C). Immunostaining further confirmed the activation of RIPK3, as the OGDR group showed higher levels of p-RIPK3 than the control group (Figs. 2D, 2E). Together, our results demonstrated that OGDR induced necroptosis in primary RGCs via a RIPK3-dependent necroptosis pathway. 

Figure 2.

 

The RIPK3/MLKL pathway was activated in OGDR-induced primary RGCs. (A) Representative images of primary RGCs, identified by immunofluorescence of Tuj1 (green) and RBPMS (red). Scale bar: 50 µm. (B, C) Western blot and densitometry analysis of p-RIPK3, t-RIPK3, p-MLKL, and t-MLKL expression in control (Con) and treated RGCs at zero, one, three, six, and 12 hours after OGD. n = 3. *P < 0.05; **P < 0.01; ****P < 0.0001; one-way ANOVA. (D) Representative immunofluorescence images of Tuj1 (green) and p-RIPK3 (red) in control and treated-RGCs at one hour after OGD. Scale bar: 10 µm. (E) Quantification of the p-RIPK3 fluorescence intensity. n = 6. ***P < 0.001; Student's t test. Data are shown as mean ± SEM. Blue, DAPI staining.

GSK840 Promoted RGC Survival After IR

GSK840 is a newly identified small-molecule inhibitor that suppresses RIPK3 kinase activity.¹⁷ To assess its impact on RGC survival, we intravitreally injected GSK840 (Fig. 3A) at three concentrations (0.5, 1 and 5 mM) and quantified RGC numbers seven days following IR. The quantity of Tuj1⁺ RGCs of the retinal flat mounts did not increase with 0.5 mM of GSK840 compared to IR group, but it increased when the dose was elevated to 1 mM (Figs. 3C, 3D). However, a higher dose of GSK840 (5 mM) resulted in a decrease in cell counts compared to the value at 1 mM. These findings suggested that GSK840 at 1 mM could effectively promote RGC survival after IR. Therefore the concentration of 1 mM was selected for subsequent experiments. 

Figure 3.

 

GSK840 promoted RGC survival after IR. (A) Molecular weight, chemical formula and chemical structures of GSK840. (B) Outline of experimental design. GSK840 (2 µL; 0.5, 1, 5 mM) or control vehicle was intravitreally injected immediately after IR. Histological and functional evaluation were performed seven days after IR. (C) Representative images of Tuj1 (green) -labeled flat-mounted retinas from control and IR mice treated with or without GSK840. Scale bar: 50 µm. (D) Quantification of Tuj1⁺ RGCs. n = 6. ***P < 0.001; ****P < 0.0001; one-way ANOVA. Data are shown as mean ± SEM.

GSK840 Preserved Retinal Structure After IR

To evaluate the influence of GSK840 on IR-induced retinal structural defects, HE staining and SD-OCT were used. Histopathological analysis using HE staining revealed a marked reduction in the thickness of retinal sublayers (NFL/GCL, IPL, INL, outer plexiform layer, and ONL) after IR, which was partially restored by GSK840 treatment (Figs. 4A, 4B). Consistent with HE staining, SD-OCT analysis in living animals showed a considerable diminution in GCC, RNFL, and total retina thickness after IR. However, this reduction was also partially ameliorated by GSK840 treatment (Figs. 4C, 4D). These results suggested that GSK840 effectively preserved the retinal structure after IR. 

Figure 4.

 

GSK840 preserved the retinal structure after IR. (A) HE staining images from the control and IR groups treated with or without GSK840. Scale bar: 20 µm. (B) Quantification of retinal sublayers thickness. n = 6. (C) Representative SD-OCT images from the control and IR groups treated with or without GSK840. GCC: indicated as yellow line with straight ends. Scale bar: 100 µm. (D) Quantification of retinal thickness. n = 6. *P < 0.05; ***P < 0.001; ****P < 0.0001; ns: no significance; one-way ANOVA. Data are shown as mean ± SEM.

GSK840 Protected Inner Retinal Neurons After IR

To assess the effects of GSK840 on inner retinal neurons after IR, we conducted counting of cells seven days after IR using immunofluorescence with neuronal markers, involving calretinin (amacrine cells), PKC-α (bipolar cells), and calbindin (horizontal cells). The quantity of calretinin⁺ amacrine cells in the GCL and INL decreased after IR, which was partially restored by GSK840 treatment (Figs. 5A, 5B). However, the IR and control group displayed no discernible discrepancy in the quantity of PKC-α⁺ bipolar cells as well as calbindin⁺ horizontal cells (Figs. 5C, 5D, 5F, 5G). Although unchanged in number, bipolar cells exhibited a reduction in the length of axonal fibers, which was alleviated by GSK840 treatment (Figs. 5C, 5E). These findings suggested that GSK840 effectively rescued inner retinal neurons after IR, involving amacrine cells, as well as the axonal length of bipolar cells. 

Figure 5.

 

GSK840 protected inner retinal neurons after IR. (A, C, F) Representative immunofluorescence images of calretinin (green) , PKC-α (green) and calbindin (red) from the control and IR groups treated with or without GSK840. Scale bar: 20 µm. (B, D, E, G) Quantification of calretinin⁺ amacrine cells, PKC-α⁺ bipolar cells, axonal length of PKC-α⁺ bipolar cells, and calbindin⁺ horizontal cells. n = 6. *P < 0.05; **P <0.01; ***P < 0.001; ****P < 0.0001; ns: no significance; one-way ANOVA. Data are shown as mean ± SEM.

GSK840 Improved Visual Function After IR

To evaluate the impact of GSK840 on neural function in the retina, we conducted ERG to measure electrical potential changes, including PhNR, scotopic a-wave, b-wave, and OPs, which indicate the function of RGCs, photoreceptors, bipolar cells, and amacrine cells, respectively.^36,37 Our results showed a significant reduction in the amplitudes of the PhNR (Figs. 6A, 6F), a-wave (Figs. 6C, 6D, 6H, 6J), b-wave (Figs. 6 B–D, 6G, 6I, 6K), and OPs (Figs. 6E, 6L) after IR, indicating impaired visual function. However, treatment with GSK840 enhanced the amplitudes of PhNR (Figs. 6A, 6F), a-wave (Figs. 6C, 6D, 6H, 6J), b-wave (Figs. 6 B–D, 6G, 6I, 6K), and Ops (Figs. 6E, 6L), consistent with our histological findings. These results suggested that GSK840 could ameliorate visual dysfunction after IR. 

Figure 6.

 

GSK840 improved visual function after IR. (A–E) Representative waveforms of PhNR and scotopic ERG components from the control and IR groups treated with or without GSK840. The flashlight intensities used to elicit the responses are 20, 0.01, 3, 10 cd s/m², respectively. (F–L) Quantification of the amplitudes of PhNR, a-wave, b-wave and OPs. n = 6. *P < 0.05; **P < 0.01; ****P < 0.0001; ns: no significance; one-way ANOVA. Data are shown as mean ± SEM.

GSK840 Mitigated Necroptotic Cell Death in RGCs

To evaluate the impact of GSK840 on necroptotic cell death, we used PI, which permeates disrupted but not intact cell membranes.^10,28,38 There was a considerable elevation in PI⁺ cells in the GCL and INL after IR, whereas GSK840 application reduced the number of necroptotic cells (Figs. 7A, 7B). Nevertheless, the quantity of PI⁺ cells in the ONL did not exhibit any distinguishable variation across the four groups. Notably, PI labeling was present in Tuj1⁺ RGCs in the GCL (Fig. 7A). In vitro experiments also revealed that OGDR treatment increased the ratio of PI⁺ primary RGCs, which was alleviated by GSK840 treatment (Figs. 7C, 7D). Our results indicated that GSK840 effectively inhibited necroptotic cell death of RGCs. 

Figure 7.

 

GSK840 alleviated necroptotic cell death in RGCs. (A) Representative immunofluorescence images of Tuj1 (green) and PI (red) from the control and IR mice treated with or without GSK840. Scale bar: 20 µm. (B) Quantification of PI⁺ cells. n = 6. (C) Representative immunofluorescence images of Tuj1 (green) and PI (red) in control and OGDR-induced primary RGCs treated with or without GSK840. Scale bar: 10 µm. (D) Quantification of PI⁺ cells. n = 6. *P < 0.05; ****P < 0.0001; ns: no significance; one-way ANOVA. Data are shown as mean ± SEM.

GSK840 Inhibited RIPK3-Dependent Necroptosis in RGCs

Because the number of PI⁺ RGCs was reduced by GSK840, we further investigated the effect of GSK840 on the RIPK3-dependent necroptosis pathway of RGCs. WB results showed that GSK840 treatment reduced the protein levels of p-RIPK3, t-RIPK3, p-MLKL and t-MLKL in IR retinas (Figs. 8A, 8B, Supplementary Fig. S3). Furthermore, GSK840 treatment partially attenuated the increase in p-RIPK3 protein levels after IR, as shown by immunofluorescence analysis (Figs. 8C, 8D). Similarly, in vitro experiments showed that GSK840 treatment largely reversed the upregulated levels of necroptosis-related proteins in primary RGCs induced by OGDR (Figs. 8E, 8F, Supplementary Fig. S3). Immunofluorescence analysis revealed that GSK840 significantly reduced the upregulated p-RIPK3 immunoreactivity in OGDR-induced primary RGCs (Figs. 8G, 8H). Overall, these results showed that GSK840 effectively inhibited RIPK3-driven necroptosis signaling in RGCs. 

Figure 8.

 

GSK840 suppressed RIPK3/MLKL pathway in RGCs. (A, B) Western blot and densitometry analysis of p-RIPK3, t-RIPK3, p-MLKL and t-MLKL expression from the control and IR mice treated with or without GSK840. n = 3. (C, D) Representative p-RIPK3 (red) images and quantification of the fluorescence intensity from the control and IR mice treated with or without GSK840. Scale bar: 20 µm. n = 6. (E, F) Western blot and densitometry analysis of p-RIPK3, t-RIPK3, p-MLKL and t-MLKL expression from control and OGDR-induced primary RGCs treated with or without GSK840. n = 3. (G) Representative immunofluorescence images of Tuj1 (green) and p-RIPK3 (red) from control and OGDR-induced primary RGCs treated with or without GSK840. Scale bar: 10 µm. (H) Quantification of the p-RIPK3 fluorescence intensity. n = 6. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001; ns: no significance; one-way ANOVA. Data are shown as mean ± SEM. Blue, DAPI staining.

Discussion

Here, our results showed that necroptosis was rapidly activated in RGCs following IR or OGDR. By using in vivo and in vitro models of retinal IR, we observed that GSK840 reduced neuron loss and protected against retinal structural and functional impairments caused by IR. Mechanistically, GSK840 inhibited RIPK3-dependent necroptosis pathway in RGCs. Thus this study has shed light on the potential benefits of GSK840 for rescuing RGC loss after retinal IR. 

Although apoptosis has traditionally been associated with neuronal death,^39,40 recent studies have revealed that the pathogenesis of neurodegeneration is not exclusively caused by apoptosis.^4,7,9 Instead, an alternative programmed cell death pathway called necroptosis has emerged as a critical player in neuronal death.^{3,4,7,9,41,42} RIPK3 and MLKL are two core proteins that participate in the activation of necroptosis. Pathway-specific adaptor proteins including RIPK1, TIR-domain-containing adapter-inducing interferon-b (TRIF) and DNA activator of interferon are essential for transmitting the necroptotic signal to RIPK3.⁶ Once activated, RIPK3 recruits and phosphorylates MLKL, leading to plasma membrane destabilization, cellular swelling, and membrane disruption.⁶ RIPK3 has been detected in the RGCs of normal rat retinas, and its protein levels exhibit a marked elevation after retinal IR.⁴³ Additionally, Galina et al. found that necrostatin-1, a small molecule compounds that specifically targets necroptosis, could decrease necroptotic RGC death and promote RGC survival after retinal IR.⁴ These findings prompted us to postulate that RIPK3 might be involved in the pathophysiological mechanisms underlying IR by mediating RGC necroptosis. Thus we examined spatiotemporal expression profiling of necroptosis-related proteins and found an increase in the expression of p-RIPK3, t-RIPK3, p-MLKL, and t-MLKL in the retina at the early phase after IR. The pattern of RIPK3 expression was consistent with a previous study in a rat IR model.⁴³ Notably, our in vivo experiments revealed that p-RIPK3 was primarily located in Tuj1⁺ RGCs. In vitro experiments also showed an increase in p-RIPK3, t-RIPK3, p-MLKL, and t-MLKL expression in OGDR-stimulated RGCs. During necroptosis, swollen cells and organelles cause rupture of the plasma membranes,⁶ which can be detected by PI staining.³⁸ In agreement with the localization of p-RIPK3, we demonstrated that the PI-positive cells in the GCL were Tuj1⁺ RGCs. Overall, these results suggested that necroptosis was activated after IR and that the RIPK3/MLKL signaling contribute to IR-induced RGC necroptosis. 

After observing the activation of necroptosis following IR, we proceeded to examine the impact of inhibiting necroptosis on IR-induced neuronal injury. RIPK3 holds a crucial position in the downstream signaling cascade of necroptosis and is capable of triggering necroptosis in a RIPK1-independent manner.⁶ Hence, RIPK3 inhibitors have been discovered to provide protection against a wider range of pro-necrotic stimuli as compared to RIPK1 inhibitors.^44,45 The selective targeting of RIPK3 has garnered attention to develop inhibitors for various pathological conditions.¹⁷ In this study, we used GSK840, a selective small-molecule RIPK3 inhibitor that conjugates to the RIPK3 kinase domain and represses kinase activity in a highly specific manner.¹⁷ GSK840, initially identified by Mandal et al.¹⁷ in 2014, has primarily been applied in vitro, and its in vivo efficacy is not fully elucidated. First, we carefully considered the timing of GSK840 administration and its potential impact on efficacy. It is noteworthy that both pre- and post-insult administration of small-molecule necroptosis inhibitor have been investigated in studies, each with its own considerations and advantages.^21,46,47 However, post-insult administration more closely aligns with the clinical practice of intervening after injury occurs. Therefore we chose to administer GSK840 immediately after IR insult to mimic a clinically relevant scenario. Our results, in line with previous research,⁴³ showed that the expression levels of necroptotic proteins, including RIPK3 and p-RIPK3, increase significantly after IR and peak around 12 hours post-insult, with subsequent gradual reduction. This suggests the existence of a potential therapeutic window during which GSK840 could effectively inhibit necroptosis, thus exerting its neuroprotective effects. However, the exact duration of this time window requires further investigation. Beyond the timing of administration, we also considered the delivery method and concentration of GSK840. Wen et al.²² reported the potential teratogenic effects of high concentrations of GSK840 in zebrafish larvae, indicating possible dose-dependent toxicity when administered in vivo. Consequently, we chose intravitreal injection as the delivery route to minimize potential systemic effects. Furthermore, we designed a concentration gradient for intravitreal administration based on the study by He et al.²¹ Our results revealed an intriguing pattern: RGC counts did not significantly increase in the 0.5 mM group compared to the IR group. However, a significant increase in RGC survival was observed when the concentration was elevated to 1 mM. Intriguingly, further escalation to 5 mM did not yield a significant difference in RGC counts compared to the IR group. This prompted us to investigate potential retinal toxicity at the 5 mM. We conducted an additional experiment by injecting of 2 µL of 5 mM GSK840 or control vehicle into the vitreous of wild-type mice. The results revealed a significant reduction in RGC counts in mice injected with 5 mM GSK840 compared to those receiving the control vehicle (Supplementary Fig. S2). This observation aligns with the findings of Wen et al.²² and suggests that 5 mM GSK840 may indeed possess some retinal toxicity. Consequently, we selected the concentration of 1 mM for subsequent experiments. In addition to quantification of RGC survival, our HE and SD-OCT analyses further revealed that GSK840 treatment mitigated the reduction in NFL/GCL and GCC thickness caused by IR. In line with morphological results, our functional results showed GSK840 alleviated IR-induced reduction in PhNR amplitude, which reflects the function of RGCs.³⁷ In addition to its RGC rescuing effect, GSK840 treatment also prevented the loss of amacrine cells, as well as the reduction in INL thickness and OPs amplitude. Interestingly, GSK840 alleviated the reductions in b-wave amplitudes, which suggests damage to bipolar cells.³⁶ Nevertheless, no noteworthy difference in the population of PKC-α⁺ bipolar cells was observed among four groups, which is consistent with previous findings.^48,49 There is a possibility that bipolar cells continue to express protein kinase c after IR, but their functions are damaged, such as axonal neurotransmission.⁵⁰ In fact, we found IR induced a decrease in axonal length of bipolar cells, which was ameliorated by GSK840 treatment. Collectively, these findings support the neuroprotective potential of GSK840 in the context of retinal IR and highlight the critical considerations in the timing, method, and concentration of GSK840 administration for optimal therapeutic outcomes. 

The RIPK3-mediated necroptosis pathway has been recognized as a contributor to RGC death in ischemic retinopathy such as DR and glaucoma.^4,11 Inhibiting the RIPK3/MLKL pathway has been shown to be beneficial in reducing neuronal injury, specifically in RGCs. For instance, the RIPK3 inhibitor GSK872 has been found to substantially reduce glutamate excitotoxicity-induced RGC loss and retinal thinning via the blockade of RIPK1/RIPK3/MLKL signaling in RGCs.¹³ It is noteworthy that GSK840 can efficiently halt necroptotic process by selectively suppressing the phosphorylation of RIPK3 and its downstream effector MLKL.¹⁸ In the present investigation, to clarify the molecular mechanisms through which GSK840 mitigates IR-induced neuronal injury, the RIPK3/MLKL signaling was examined. Our WB and immunofluorescence results indicate that GSK840 administration notably attenuated the increased protein levels of p-RIPK3, t-RIPK3, p-MLKL and t-MLKL after IR. In vitro experiments also demonstrated that GSK840 reduced the activity of the RIPK3/MLKL pathway in OGDR-induced RGCs. Taken together, our findings suggested that GSK840 alleviate IR-induced neuronal injury possibly by inhibiting the RIPK3/MLKL signaling pathway in RGCs. 

In summary, our findings demonstrated that RIPK3-dependent necroptosis mediates RGC death after IR, highlighting the potential of RIPK3 as a target to ameliorate IR-induced neuronal injury. GSK840, with its ability to inhibit RGC necroptosis, is a hopeful candidate for such intervention. 

Acknowledgments

Supported by the National Natural Science Foundation of China (82271081), the Natural Science Foundation of Guangdong Province (2021A1515012142 and 2022A1515010302). 

Disclosure: Y. Feng, None; C. Hu, None; K. Cui, None; M. Fan, None; W. Xiang, None; D. Ye, None; Y. Shi, None; H. Ye, None; X. Bai, None; Y. Wei, None; Y. Xu, None; J. Huang, None