Attenuation of Microglial Activation and Pyroptosis by Inhibition of P2X7 Pathway Promotes Photoreceptor Survival in Experimental Retinal Detachment

Abstract

Purpose: Photoreceptor (PR) death is the ultimate cause of irreversible vision loss in retinal detachment (RD). Although microglial infiltration in the subretinal space (SRS) was observed after RD, the molecular mechanism underlying microglial activation and the outcomes of infiltrating microglia remain unclear. We aimed to uncover the mechanism of initiation of microglial activation to help explore potential therapy to promote PR survival.

Methods: An RD model was conducted by injecting sodium hyaluronate into SRS of C57BL/6J wild type mice. Adenosine triphosphate (ATP) was measured by a ATP Microplate Assay Kit. Bioinformatics analysis was used to evaluate the upregulated receptor relating to ATP binding in human datasets and mouse transcriptomes of RD. Expression of P2X7, its downstream signaling pathways, and microglial pyroptosis were confirmed by qPCR, WB, and immunofluorescence in vivo and in vitro. The cell viability of PR was measured by cell counting kit-8. Brilliant Blue G, a P2X7 antagonist, was subretinally or intraperitoneally injected to inhibit microglial activation in vivo and was applied for microglia cell line treatment in vitro. The decrease in microglial activation and pyroptosis was detected by immunofluorescence and WB. The protective effect on PR was measured by hematoxylin and eosin staining, TUNEL assay, and electroretinogram analysis.

Results: The results showed that extracellular ATP released in the SRS after RD triggered P2X7 activation and attracted microglia. The downstream cascade of inflammasome activation induced by P2X7 activation contributed to microglial pyroptosis and then to PR death. ATP-activated microglia led to PR death in vitro. P2X7 blockade rescued PR morphologically and functionally by inhibiting microglial activation and pyroptosis.

Conclusions: These results elucidate that ATP-induced P2X7-mediated microglial activation leads to microglial pyroptosis, contributing to PR death. Appropriate inhibition of microglial pyroptosis might serve as a pharmacotherapeutic strategy for decreasing PR death in RD.

Retinal detachment (RD) is one of the most common causes of photoreceptor (PR) death and blindness in adults,¹ with severe vision loss more likely when RD follows other complicated retinal diseases.² Upon RD, the loss of oxygen and nutritional support is one reason for PR death.³ However, the role of inflammation in PR death cannot be neglected, because the subretinal fluid⁴ and vitreous humor⁵ contain cytokines, implying that RD is actually a sterile form of inflammation. Although surgery is an effective way to reattach the neurosensory and retinal pigment epithelium (RPE) layer, functional restoration requires more than physical recovery.^3,6–8 The degeneration of PR outer segments happens soon after detachment, but their recovery takes a long time after reattachment.⁹ Thus, an effective therapeutic intervention is needed to protect PR from death and to improve their functional recovery. 

Microglia, the main population of glial cells in the retina, have significant responsibilities for retinal health.¹⁰ Under normal circumstances, microglia extend their ramified processes to surveil neighboring cells.^11–13 The infiltration of microglia in the subretinal space (SRS) after RD has been reported as an important response of the innate immune system.^5,14,15 The activation of microglia is initiated 1 hour after RD¹⁴ and continues for at least 7 days after RD.¹⁵ However, the molecular mechanism underlying microglial activation and migration upon RD remains unclear. Moreover, the outcome of activated microglia is poorly studied previously. 

Under normal conditions, extracellular adenosine triphosphate (eATP) release from Müller cells is the main component to maintain microglial resting behavior.¹⁶ In the context of diseases, eATP released from cells in stress acts as a danger signal to alert neighboring cells,^17–20 and the chemical gradient of ATP generates a proinflammatory microenvironment to recruit immune cells to the damaged region.^18,21 The changes in morphology and location of microglia after RD give us a hint that eATP and its receptors could be involved in the induction of microglial activated behavior and migration. 

P2X7 receptors, one of the major cell surface receptors simulated by ATP,²² are known to distribute to microglia, and their activation by ATP is a key step in the release of inflammatory cytokines from microglia.^23–25 The consequences of P2X7 activation have been examined, yet its role involved in the pathological process of RD is unclear. ATP could enhance the nuclear factor-κB (NF-κB) activation via P2X7.^26,27 In addition, the crucial role of P2X7 in inflammation is the activation of P2X7/ nucleotide-binding domain, leucine-rich–containing family, pyrin domain-containing 3 (NLRP3) pathway, resulting in the release of IL-1 family cytokines^28,29 and induction of pyroptosis.^30,31 Pyroptosis, a highly inflammatory mode of regulated cell death, depends on the activation of inflammasomes followed by cleavage of Gasdermin D (GSDMD) and formation of the gasdermin channel on the cell membrane, resulting in cell explosion.^32,33 PRs have been reported to undergo pyroptosis.³⁴ However, pyroptosis is a consequence of inflammasome activation, and the infiltrating macrophages are considered as the major source of inflammasomes in RD.³⁵ Thus, the pyroptotic cell type in RD retina remains to be accurately described. 

In this study, we investigated the indispensable role of eATP-stimulated P2X7 activation in the mechanism underlying microglial activation and migration to the SRS. Our study demonstrated for the first time that pyroptosis was a likely outcome of subretinal microglia in RD and microglial pyroptosis played a detrimental role in PR death. The P2X7 antagonist Brilliant Blue G (BBG) effectively decreased microglial recruitment and pyroptosis, bringing about the morphological and functional recovery of PR. These findings suggest that antagonizing P2X7 receptor to block or alleviate ATP-induced microglial activation may mitigate inflammation in its acute phase and promote long-term PR survival. 

Materials and Methods

Animals

C57BL/6 mice (18–20 g) were used for the study. All animal experiments followed the guidelines of the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care and Use Committee of the Second Xiangya Hospital of Central South University. 

Experimental Model of RD and Drug Intervention

Briefly, mice were anesthetized with pentobarbital (50 mg/kg). Approximately 3 µL of 10 mg/mL sodium hyaluronate (Albomed, Schwarzenbruck, Germany) was injected into the SRS by a microsyringe (Hamilton, NV, USA) to make one-half of the retina detach from RPE (Fig. 1A). The edge of the detached retina was observed by microscopy to maintain consistency (Fig. 1B). A successful RD model is shown (Fig. 1C) in horizontal section with hematoxylin and eosin staining. Samples were collected at 1, 2, 3, 7, and 14 days after RD. BBG was treated by two methods: 10 or 100 µg/mL BBG were injected into the SRS immediately after RD as subretinal injections (SRi), or a serial of concentrations (15, 50, or 100 mM) of BBG were injected intraperitoneally 20 minutes before RD. 

Figure 1.

 

(A) Experimental RD model setup. A 33G needle installed on the microsyringe was inserted carefully through the nasal side of the retina, then sodium hyaluronate was injected into the SRS. The edge of the detached retina was observed under microscope (B) to keep consistency. The detached retina was marked by red *. (C) Representative horizontal section with hematoxylin and eosin staining of successful RD. Scale bars, 200 µm.

Human RD Retina Transcriptomic Analysis

Transcriptomic data of human RD (GSE28133) were obtained from the Gene Expression Omnibus database. Differentially expressed genes (DEGs) were identified based on fold change and P values were calculated using a t test (fold-change of >1.5 and P < 0.05). Volcano plots and heatmaps were generated using the OECloud tools at https://cloud.oebiotech.cn. Gene Ontology (GO) enrichment analyses were performed based on the hypergeometric distribution to explore the biological roles of these DEGs using R (v 3.2.0). 

RNA Sequencing, Sequencing Data Extraction, and Bioinformatics Analysis

Total RNA was extracted from the mouse retinas (naïve, n = 3; RD at 2 days, n = 5), then was purified, amplified, and labeled. Libraries were then constructed using VAHTS Universal V6 RNA sequencing Library Prep Kit (Vazyme, Nanjing, China) according to the manufacturer's instructions. The libraries were sequenced on an Illumina Novaseq 6000 platform and the read depth was around 40 million reads per sample. The clean reads were mapped to the GRCm39 reference genome (GCF_000001635.27) using hierarchical indexing for spliced alignment of transcripts. Fragments per kilobase million of each gene were calculated and the read counts of each gene were obtained by HTSeq-count. Principal component analysis was performed using R (v 3.2.0) to evaluate the biological duplication of samples. Differential expression analysis was performed using the R package DESeq2. The cutoff for significant DEGs were a Q value of <0.05 and a fold change of >1.5. Hierarchical cluster analysis, heatmaps, and GO enrichment analyses of DEGs were performed using R packages (v 3.2.0). 

Sample Preparation and ATP Detection

Subretinal fluid collection was based on a published interphotoreceptor matrix isolation protocol.³⁶ After separation, the retina and remaining eyecup were immersed in Hank's Balanced Salt Solution without Ca²⁺ and Mg²⁺ and shaken for 30 minutes on ice. The supernatant was centrifuged and collected as subretinal fluid. The retina was sonicated at 5% power for further tests. ATP level was detected using ATP Microplate Assay Kit (Absin, Shanghai, China) according to the manufacturer's instruction. The absorbance (optical density value) was measured at 635 nm. 

Immunohistochemistry

Eyes were fixed in 4% paraformaldehyde and dehydrated in gradient sucrose PBS. Samples were sectioned at 30-µm thicknesses. Slides were permeabilized using TBS supplemented with 0.3% Triton X-100 (BioFroxx, Guangzhou, China) for 40 minutes and then blocked in TBS-T with 1× BSA and 10% Donkey Serum for 1 hour at room temperature. The corresponding primary antibodies and secondary antibodies were listed in Table 1. The slides were mounted with ProLong Gold Antifade Mountant (ThermoFisher Scientific, Waltham, MA, USA), and were observed under a confocal microscope (Carl Zeiss, Oberkochen, Germany) with 20× Z-stack images captured. For quantification of subretinal cells, at least three areas from each slide were chosen at random. 

Table 1.

 

Antibodies

 

TUNEL Assay

Samples were sectioned at 12-µm thicknesses. Cell death was detected using an In Situ Cell Death Detection Kit, Fluorescein (Roche, Basel, Switzerland) following the supplier's instructions. The fluorescence was detected in the range of 515 to 565 nm under microscopy. 

Histology

After fixation in 4% paraformaldehyde, eyeballs were embedded in paraffin, sectioned into 12-µm slices, and stained with hematoxylin and eosin. For each section, the average thicknesses of the outer nuclear layer (ONL) and inner segment/outer segment (IS/OS) were calculated of 10 points of the detached region of retina.^37,38 The distances between each point were 150 µm. The thickness was measured using the CaseViewer application (3DHISTECH, Sysmex, Switzerland). 

Western Blot Analysis

Retinal samples were lysed in 1× RIPA lysis buffer (Solarbio, Beijing, China) supplemented with protease inhibitor (Roche) and phosphatase inhibitor (Roche). Proteins were separated by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to polyvinylidene difluoride membranes (Millipore, Burlington, MA). Membranes were then incubated with primary antibodies and secondary antibodies shown in Table 1. Membranes were treated with horseradish peroxidase substrates and visualized on a FluorChem FC3 system (ProteinSimple, San Jose, CA, USA). Image J (v1.8.0) software was used to analyze the grayscale values of the bands. 

Quantitative Real-Time PCR (qPCR) Analysis

Total RNA was extracted from retinal tissue using the TRIzol reagent (Invitrogen, San Deigo, CA, USA) according to the manufacturer's instructions. cDNA was synthesized using RevertAid Master Mix (ThermoFisher Scientific). qPCR was performed on an Applied Biosystems StepOne Plus Real-Time PCR System (ThermoFisher Scientific) using TB Green Premix Ex Taq II (Takara, RR820A, Shiga, Japan). GAPDH was used as the internal control. The primers used for qPCR are listed in Table 2. Gene expression was calculated via the 2^−ΔΔCt method. 

Table 2.

 

Primer Sequences for qPCR

 

Electroretinogram (ERG) Analysis

ERGs were recorded using RetiMINER System (AiErXi Medical Equipment Co., Ltd., Chongqing, China). After overnight dark adaption, mice were anesthetized and pupils were dilated. Round metal electrodes were attached to the center of the cornea and a needle electrode was inserted into the back head subcutaneously. Flash ERGs were recorded in response to a range of light intensities. The amplitude of the a-wave, which reflects PR function, was measured.³⁹ 

Cell Culture and Treatment

BV2 cells were purchased from the National Infrastructure of Cell Line Resource. BV2 cells were cultured in high glucose DMEM medium with 10% fetal bovine serum and 1% penicillin/streptomycin, and incubated at 37°C in a humidified atmosphere of 5% CO₂. For ATP administration, BV2 cells were seeded at 2 × 10⁵ per well into six-well plates, and treated with 1 mM or 3 mM ATP for 12 hours in the presence or absence of 0.5 µg/mL lipopolysaccharide (LPS). The supernatant of BV2 cells was collected for further assay. For BBG rescue, BV2 cells were seeded at 2 × 10⁵ per well into six-well plates, and exposed to 1 mM ATP for 12 hours in the presence or absence of 0.5 µg/mL LPS. The cells were pretreated with P2X7 inhibitor BBG (500 µM) for 30 minutes. The supernatant of BV2 cells was collected for further tests. 661W PR cells were grown in DMEM containing 10% fetal bovine serum, penicillin, and streptomycin in the presence of 5% CO₂ at 37°C. 

Cell Viability Assay

The 661W PR cells were seeded at 1 × 10⁴ per well into 96-well plates. For supernatant administration, BV2 cells were seeded at 2 × 10⁵ per well into six-well plates, and treated with 1 mM or 3 mM ATP for 12 hours in the presence or absence of 0.5 µg/mL LPS. The supernatant of BV2 cells was collected for further assay. BV2 cells supernatant was added in 661W PR cells into 96-well plates for 12 hours or 24 hours. For BBG rescue, BV2 cells were seeded at 2 × 10⁵ per well into six-well plates and exposed to 1 mM ATP for 12 hours in the presence or absence of 0.5 µg/mL LPS. The cells were pretreated with P2X7 inhibitor BBG (500 µM) for 30 minutes. The supernatant of BV2 cells was collected for further tests. BV2 cells supernatant was added in 661W PR cells into 96-well plates for 12 hours or 24 hours. For hydrogen peroxide administration, 661W cells were seeded in 96-well plates, and 3000 cells were grown for each well; then, 400 µM, 600 µM, 800 µM, and 1,000 µM hydrogen peroxide were added into cells in 96-well plates for 12 hours. For BBG rescue, 661W were grown in culture medium with 50 µM or 100 µM BBG along with 400 µM, 600 µM, 800 µM, and 1000 µM hydrogen peroxide for 12 hours. The same amount of doble-distilled H₂O were added in untreated cells as control. For cell viability assay, 100 µL fresh medium and 10 µL cell counting kit-8 (CCK8) reagent (Dojindo, Kumamoto, Japan) were added in each well, and incubate the plate for 1.5 hours in the incubator. Then measure the absorbance at 450 nm using a microplate reader (Varioskan Flash, ThermoFisher Scientific). 

Statistical Analyses

The statistical analyses were performed using Prism 9 software (GraphPad Software, Inc., San Diego, CA, USA). Data are presented as the mean ± SD. Experiments with only two treatments were analyzed using the unpaired Student t test. One-way ANOVA analysis was employed to calculate the difference between three or more groups. Significance was accepted at a P value of <0.05. 

Results

ATP-Induced P2X7 Upregulation Drives Microglia Activation in RD Retina

Microglia (Iba-1⁺ cells) started to react and migrate to the SRS within 24 hours as reported previously,¹⁴ via the ONL (Fig. 2A). The number of subretinal microglia accumulated gradually as RD continued (Fig. 2B). Thirty minutes after RD, the ATP level significantly increased in subretinal fluid compared with retina (Fig. 2C), indicating that eATP may act as a “find me” signal to recruit microglia. 

Figure 2.

 

ATP-induced upregulation of P2X7 initiated microglia activation. (A) Microglia were identified by Iba-1, and representative images of microglia attracted and migrated to SRS 1, 2, and 3 days after RD. (B) Quantification of intraretinal, subretinal and total Iba-1⁺ cells after RD. Consequently, two images from one retina were used to calculate the number of Iba-1⁺ cells. (1 and 2 days after RD, n = 4; 3 days after RD, n = 3.) (C) The optical density value reflecting ATP concentration in retina versus subretinal fluid after 30 minutes of RD (n = 5). (D) Heatmap showing DEGs in ATP binding (GO: 0005524) in human transcriptomic analysis, the color scale represents high expression values (red) to low expression values (blue). (E, F) The DEGs of P2X receptors in human (E) and mouse (F) transcriptomic analysis are shown in the heatmap (fold change > 1.5; P < 0.05). Fold changes and P values are listed in the table. (G) qPCR analysis for P2X7 expression in RD and naïve retinas at each observed time point (n = 3). (H) Western blot analysis for P2X7 expression in RD and naïve retinas at each observed time point (n = 3). (I) Representative immunofluorescence images of P2X7 and Iba-1 colocalization in cross-sections of retinas (n = 3). Scale bars, 50 µm. For immunofluorescence, Z-stack images of the entire thickness of the retina were created (n = 4). Nuclei staining: DAPI. Statistical analysis: (B) One-way ANOVA followed by Sidak's multiple comparison. (C) Unpaired t test, (G, H) one-way ANOVA followed by Dunnett's multiple comparison. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

The transcriptomic data of human and mouse RD retinas were analyzed to uncover the P2X receptor involved in the process. We focused on the DEGs in GO enrichment of ATP binding (GO: 0005524), as illustrated by the heatmap in Figure 2D, which uncovered 77 upregulated genes and 59 downregulated genes. Four significant DEGs in the P2X family were identified, with P2X4, P2X5, P2X7, and P2X5-TAX1BP3 significantly increased in the retinas of RD patients (Fig. 2E). Meanwhile, two P2X receptor genes P2X1 and P2X7 were significantly increased in the transcriptomic data of mouse retinas from 2 days after RD (Fig. 2F). P2X7 is considered to be most highly expressed by microglia in the brain,⁴⁰ whereas the P2X1/P2X5 heteromer is expressed predominantly on astrocytes in the mouse cortex.^41,42 P2X7 was the only elevated gene among P2X receptors in both datasets; thus, our emphasis was P2X7 on microglia in retina. The P2X7 expression level in the retina at 1 day, 2 days, and 3 days after RD was assessed. Both transcriptomic and protein levels of P2X7 increased in RD retinas compared with naïve (Figs. 2G, H). Because microglia comprise only 0.2% of the total retinal cells,⁴³ the gene and protein levels may vary when experiments were conducted using the whole retina. Immunofluorescent staining revealed that subretinal microglia showed clear P2X7 immunoreactivity 3 days after RD (Fig. 2I, Supplementary Fig. S1). Taken together, our results suggest that P2X7 was upregulated in the activated microglia after RD. 

P2X7-NLRP3 Pathway Activation Contributes to Microglial Pyroptosis and Inflammatory Factor Release In Vivo

The activation of P2X7/NLRP3 and P2X7/ NF-κB pathways was assessed in SRS microglia. mRNA levels of NLRP3, caspase-1, and GSDMD were elevated in RD retinas in the acute phase (Fig. 3A). Expression of NF-κB p-p65 was measured as an indication of increasing transcription of NF-κB. The protein levels of NLRP3 and NF-κB p-p65 (Fig. 3B) increased in whole retina lysate at 1 day, 2 days, and 3 days after RD. Elevated expression of the full length of GSDMD and cleaved GSDMD was also detected (Fig. 3C), suggesting the occurrence of pyroptosis. The expressions of related mRNA were also upregulated, as confirmed by transcriptomic data from mouse retinas (Fig. 3D). Immunofluorescence in Figure 3E further confirmed subretinal microglia as the major cell population undergoing pyroptosis after RD, which released harmful proteins, including caspase-1 and IL-1β, especially at 3 days after RD. No colocalization was found in the naïve retinas (Supplementary Fig. S2) between Iba-1 and these pyroptosis-related factors. Thus, these data suggest that subretinal microglia undergo irreversible pyroptotic cell death. We also observed acute PR death at 3 days after RD (Fig. 3E). Overall, these results indicate that ATP-induced microglial pyroptosis leads to inflammation and in turn to acute PR death. 

Figure 3.

 

P2X7 activation on microglia led to microglial pyroptosis. (A) qPCR analysis for inflammatory-related genes (GSDMD, caspase-1, and NLRP3) in RD retinas at each observed time point (n = 4). (B) Western blot analysis for NLRP3 and NF-κB p-p65 protein expression (NLRP3, n = 3; NF-κB p-p65, n = 5). (C) Western blot analysis for GSDMD-FL and cleaved-GSDMD at each observed time point (GSDMD-FL, cleaved-GSDMD: n = 6). (D) Heatmap of GSDMD, IL-1β, and caspase-1 from RNA sequencing of naïve and detached retina. (Naïve, n = 3; 2 days after RD, n = 5.) (E) Representative images from cross-sections stained with Iba-1 and caspase-1/IL-1β/GSDMD/NF-κB at 3 days after RD (n = 4). Representative cells of colocalization were marked by white arrows (caspase-1), yellow arrows (IL-1β), blue arrows (GSDMD), and red arrows (NF-κB p-p65), respectively. (F) Representative immunofluorescent images of TUNEL+ cells in ONL of naïve and retinas 4 days after RD (naïve, n = 3; 3 days after RD, n = 3). (E, F) Scale bars, 50 µm. Z-stack images of the entire thickness of the retina were created. Nuclei staining: DAPI. Statistical analysis: (A, B, C) one-way ANOVA followed by Dunnett's multiple comparison. *P < 0.05, **P < 0.01, ****P < 0.0001.

Microglial Pyroptosis Accelerates PR Death In Vitro

To understand the effect of ATP on microglia, we first examined inflammatory-related proteins in the BV2 cell line after different stimuli. Protein levels of NLRP3, NF-κB p-p65, and GSDMD were increased in BV2 cells when treated with ATP or ATP combined LPS (Fig. 4A), which was consistent with results in vivo. The results showed that ATP exerted an inflammatory inducement in the microglia and triggered the assembly of the NLRP3 inflammasome. To further validate whether ATP-mediated microglial activation and pyroptosis are vital to PR survival, PR cell line 661W was cultured with or without BV2 supernatant collected after different stimuli. We then use CCK8 to test PR cell viability under various treatments at 12 hours (Fig. 4B) and 24 hours after stimulation (Fig. 4C). We found that, by adding 1 mM and 3 mM ATP with or without LPS to the culture media of 661W cells, 661W showed worse viability, which may result from the drug cytotoxicity. In addition, different doses of ATP-treated BV2 cell supernatant promoted PR death significantly both at 12 hours (Fig. 4B) and 24 hours (Fig. 4C). The combination of ATP and LPS-induced microglial supernatant worsened the damage (Figs. 4B and 4C). Thus, our data confirmed that the ATP-stimulated microglial inflammatory effect led to PR death. 

Figure 4.

 

ATP-stimulated microglial inflammation and pyroptosis accelerate PR death in vitro. (A) Protein expression of NLRP3, NF-κB p-p65, and GSDMD in BV2 cells was measured through WB after treatment with 1 mM or 3 mM ATP with or without 0.5 µg/mL LPS for 12 hours. The graphs show the quantification of each protein level. (NLRP3, n = 6; NF-κB p-p65, n = 6; GSDMD-FL, cleaved-GSDMD, n = 6.) (B, C) After 1 mM or 3 mM ATP administration with or without LPS, supernatant of BV2 cells was added to the cell culture medium of 661W cells, then cell viability of PR was measured by CCK8 at 12 hours (B) and 24 hours (C) (mean ± SEM; n = 3). Statistical analyses: (A) unpaired t test, (B, C) one-way ANOVA followed by Dunnett's multiple comparison. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

P2X7 Blockade Reduces Microglial Migration and Pyroptosis

Given the data above, we hypothesized that inhibition of P2X7 activation on microglia may reduce PR death. To test this hypothesis, BBG, the P2X7 antagonist, was used for treatment through two different routes. CD68 served as an indicator of microglia activation and its colocalization with Iba-1 in the SRS revealing the activation of subretinal microglia. At all concentrations, SRi of BBG effectively reduced total microglia numbers in the retina, especially in the SRS (Figs. 5A, B) while massive microglia were observed in saline-treated retina at 3 days after RD (Fig. 5A). Similar trends were also observed in IP BBG pretreated RD retinas (Fig. 5C). IP BBG administration successfully reduced microglia in the SRS compared with saline administration (Figs. 5C, D). The effect of IP injection of BBG seemed to have a relatively mild effect because microglia counts were slightly higher in this group than the SRi group. The inhibition of microglial migration by BBG indicated that ATP-stimulated P2X7 activation played a crucial role in the initial response of microglia to RD. 

Figure 5.

 

Inhibition of P2X7 prevented microglial migration and inflammatory factors release. Representative cross-sectional images of microglia activation after SRi (A) or IP (C) administration of BBG. CD68 is used as an indicator for microglial activation. (B) SRi, saline (n = 6); BBG, 10 µg/mL (n = 8); 100 µg/mL (n = 3) and (D) IP, saline (n = 4); BBG, 15 mM (n = 3); 50 mM (n = 5); 100 mM (n = 4) are the quantifications of in-retina, subretinal and total microglia related to (A) and (C) after different concentrations of BBG treatment. (E) Western blot analysis for NLRP3 expression after 50 mM BBG administration (n = 3). (F) Representative immunofluorescent images from BBG-treated and saline-treated retina sections stained with Iba-1 and GSDMD or IL-1β (n ≥ 3). (A, C, E) Scale bars, 50 µm. Z-stack images of the entire thickness of the retina were created. Nuclei staining: DAPI. Statistical analysis: (B, D) one-way ANOVA followed by Sidak's multiple comparison, (E) unpaired t test. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

In accordance with the decreased number of microglia in the SRS, decreased protein levels of NLRP3 were found in retina after BBG administration at 3 days after RD (Fig. 5E). As the number of microglia decreased in SRS, the immunofluorescence of GSDMD or IL-1β also decreased in the BBG-treated retina and almost no colocalization is found (Fig. 5F). From the comparison of saline and BBG administration, only subretinal microglia revealed highly inflammatory phenotype, whereas intraretinal microglia do not seem to release inflammatory factors. Thus, the inhibition of microglial activation and translocation is a possible way to decrease inflammation. 

Inhibition of P2X7 Effectively Promotes PR Survival In Vivo

To investigate whether PR may benefit from inhibition of P2X7 on microglia, we focused on 3 days after RD, when PR death peaked,^3,44 and most subretinal microglia were present. Fewer TUNEL⁺ cells in ONL were counted in BBG-treated retinas than saline at 3 days after RD (Figs. 6A–D). Progressive PR degeneration in RD was observed from 7 days to 14 days after RD.⁴⁵ Thus, we measured the thickness of ONL and IS/OS at 14 days after RD to investigate the effect of BBG on PR survival in sustained RD (Figs. 6E, H). Recovery of ONL and IS/OS thickness were found in both SRi (Figs. 6E, F, G) and IP (Figs. 6H, I, J) BBG administration. Of note, even though the SRi BBG treatment showed better suppression of microglia activation (Figs. 5B, D), the recovery of ONL and IS/OS exhibited less significance (Figs. 6F, G) than IP treatment (Figs. 6 I, J). Additionally, ERG revealed amplitude recovery in response to all stimuli from IP BBG-treated retinas compared with saline-treated retinas (Fig. 6K). The increased amplitude of a-wave reflected the recovered PR function in BBG-treated retinas (Fig. 6L). 

Figure 6.

 

P2X7 receptor inhibition decreased cell death and promoted PR recovery. Representative immunofluorescent images of TUNEL+ cells in ONL from samples at 3 days after RD after SRi (A) and IP (C) injection of saline or BBG, respectively. (B, D) Quantifications relating to (A) and (C) (SRi: saline, n = 5; BBG, n = 4; IP: saline, n = 4; BBG, n = 6). Representative hematoxylin and eosin staining images of retinas 14 days after RD after SRi 10 µg/mL (E) and IP 50 mM injection (H) of saline or BBG, respectively. (F, I) Quantifications of ONL thickness related to (E) and (H), and (G) and (J) are related quantifications of IS/OS thickness (SRi, n = 4; IP, n = 6). (K) Representative dark-adapted ERG evoked at 0.01 cd·s/m², 1.0 cd·s/m², 3.0 cd·s/m², and 10 cd·s/m² at 14 days after RD after IP injection of saline or 50 mM BBG (saline, n = 6; BBG, n = 5). (L) Quantification of a-wave (micrometers) at different stimuli related to (K) (saline, n = 6; BBG, n = 5). (A, C) Scale bars, 50 µm. Z-stack images of the entire thickness of the retina were created. Nuclei staining: DAPI. (E, H) Scale bars, 20 µm. Statistical analyses: (B, D, L) unpaired t test, (F, G, I, J) one-way ANOVA followed by Sidak's multiple comparison. *P < 0.05, **P < 0.01, ***P < 0.001.

Interestingly, the protective effect of BBG was not dose dependent. ONL and IS/OS showed lower thickness when higher doses of BBG were administrated (Supplementary Figs. S3A–D). A sustained influence of BBG on the microglia was observed at 7 days after RD, when higher concentrations of BBG were administered, whereas microglia increased to saline-treated level under an effective dose of BBG treatment (Supplementary Figs. S3E, F). This finding suggests that a high dose of BBG decreases the beneficial effect on PR by oversuppression of microglia. Taken together, these results indicate that the appropriate dose of BBG treatment protects PR function in mice effectively. Microglia may play a dual role in regulating cell viability. 

Inhibition of P2X7 Attenuates Microglial Inflammation and Prevents PR Death In Vitro

The protective effect of BBG was also examined in vitro. Western Blot confirmed that BBG effectively decreased the protein levels of NLRP3, P2X7, and GSDMD of ATP-activated BV2 cells (Fig. 7A). Interestingly, NF-κB p-p65 showed no decrease after BBG treatment, indicating that BBG did not affect the downstream transcriptomic process of NF-κB pathway. BV2 supernatants under different conditions were added to the cell culture media of PR and PR cell viability was tested using CCK8 (Fig. 7B). BBG alone decreased the cell viability of PR, yet supernatant of BV2 pretreated with BBG still rescued PR (Fig. 7B). To assess the protective effect of BBG on apoptosis of PR directly, we used hydrogen peroxide as an apoptosis inducer to treat PR. BBG failed to reverse the lethal effect of hydrogen peroxide on PR (Fig. 7C) and a high dose of BBG can even cause cell death (Fig. 7B). Thus, our data proved that BBG protected PR through its role in inhibition of microglial inflammation and pyroptosis to mitigate the inflammatory damage instead of direct protection. 

Figure 7.

 

BBG administration rescued PR through decreased microglial inflammation in vitro. (A) Western blot analysis for NLRP3, P2X7, NF-κB p-p65, and GSDMD protein expression. Quantifications are shown on the right. (NLRP3, n = 6; NF-κB p-p65, n = 6; GSDMD-FL, cleaved-GSDMD, n = 6; P2X7, n = 3.) (B) Supernatants of BV2 under various conditions were added to the culture medium of PR and CCK8 was used to test PR cell viability after 500 µM BBG rescue (mean ± SEM, n = 3). (C) Hydrogen peroxide was used to induce PR apoptosis and CCK8 was used to test PR viability under different concentrations of BBG (mean ± SEM; n = 3). Statistical analyses: (A, B) unpaired t test, (C) one-way ANOVA followed by Tukey's multiple comparison. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001.

Discussion

This study provides a comprehensive understanding of microglial participation in the RD process. Our results showed that acute separation of PR and RPE caused the accumulation of eATP in the SRS, recruiting microglia to SRS via P2X7 stimulation. Pyroptosis through P2X7/NLRP3 pathway activation was shown to be a possible outcome of microglia. Pharmacological blockade of P2X7 by BBG effectively prevented microglial migration and pyroptosis, further preventing PR death in the acute phase of RD and in the long term (summarized in Fig. 8). 

Figure 8.

 

Schematic representation of ATP-induced P2X7 activation leading to microglial migration, inflammatory factor release and microglial pyroptosis. The inhibition of P2X7 using BBG effectively alleviated microglial inflammatory reaction by preventing microglial response to ATP initially. This figure was created with Biorender.com.

Microglia have been found to be involved in the pathological process of RD, yet the molecular factors initiating their activation and migration to the SRS were unknown. eATP release acts as a chemoattractant to regulate microglial branch dynamics in central nerve disease,⁴⁶ but little is known about its function in RD. In healthy retinas, ATP released by Müller cells is thought to be the direct source to maintain the morphology and motility of resting microglia in the inner retina.^47,48 Based on this concept, we hypothesized and confirmed that elevated concentrations of ATP released in the SRS could be a signal to recruit microglia. Among seven candidates of ATP-gated P2X receptors,^42,49 we then uncovered that P2X7 was the most possibly involved receptor, which was found upregulated in both transcriptomes of human and mouse RD samples. Thus, microglia activation and migration may be driven by eATP release in the SRS through the P2X7. 

P2X7 is the only member among P2 receptors associated with NLRP3 inflammasome activation⁵⁰ and interacts directly with NLRP3,⁵¹ which is the most related inflammasome response to sterile inflammation.⁵² The activation of NLRP3 relies on two signals. The first signal, also known as priming signal, induces transcription factor NF-κB moving from the cytoplasm to the nucleus, resulting in the upregulation of both NLRP3 and pro–IL-1β.⁵³ The second signal related to the activation pathway comprises molecules varying from pathogens to ATP.^54,55 In our model, the activation of NF-κB and NLRP3 pathways was observed in the subretinal microglia, yet BBG administration only downregulated NLRP3 expression, but not NF-κB in cell culture. These results imply that ATP-induced P2X7 activation acts as the second signal to induce NLRP3 pathway activation in subretinal microglia. Blockade of P2X7 was efficient in partially mitigating the inflammatory effect brought by NLRP3, but failed to inhibit the transcription of NF-κB pathway, which may be consistently activated by first signals through Toll-like receptors. 

The function of microglia in RD is not conclusive. Some studies suggest that microglia were the main source of inflammatory factors contributing to PR death,^35,37 whereas another study suggests that microglia decrease retinal damage during acute phase.¹⁴ Our data reinforced the former idea that microglia accelerate PR cell death. In our model, inflammatory factors released by microglia showed a cytotoxic effect on PR both in vivo and in vitro. Pyroptosis, as the inflammatory programmed cell death, was a likely outcome of microglia as a consequence of P2X7/NLRP3 pathway activation in RD model and ultimately contributes to PR death. We propose that pyroptosis of microglia is induced by eATP, then the ATP release from these dying microglia could accumulate in the SRS, creating more chemoattractant to monocytes^56–58 and in return contributing to their own death, creating a positive feedback loop accelerating the PR injury. Therefore, the regulation of microglial inflammation and pyroptosis is a crucial method to preserve PR cells. 

The regulation of P2X7 using BBG revealed successful inhibition of microglial migration in vivo and mitigation of inflammatory response of microglia in vitro. PR benefited from the significant decrease in microglial pyroptosis and inflammatory factors released in the early stage of RD. BBG treatment also rescued the outer retina thickness and recovered the PR function at the later phase. Because BBG has been proven to serve as a safe and effective dye in retinal surgery,^59–61 the application of BBG could be considered a promising pharmacological therapy or a preferred dye in surgery to regulate microglia in RD. 

It is worth noting that the protective effect of BBG was not dose dependent. SRi of BBG showed a better inhibitory effect on microglia than IP injection (perhaps owing to the higher local concentration of BBG), but the recovery of PR was more apparent with IP injection. Moreover, a higher concentration of BBG administration did not lead to better PR recovery. We propose two possible explanations. The drug toxicity of high doses of BBG may be the first reason: a higher concentration of BBG can cause PR death in vitro. Another assumption is that the appropriate suppression of microglial activation protects PRs, but oversuppression of microglia may negate this benefit. Previous studies identified that microglia can assume various phenotypes in some circumstances,^62,63 and that the timing of their activation or inhibition is an important element that determined disease outcome.⁶³ In a multiple sclerosis model, microglia are found to switch from harmful to beneficial during different phases of the disease.⁶⁴ More experiments are needed to explore the state and function change of microglia in different phases of RD. One limitation of our study is that Iba-1 cannot distinguish between microglia and monocytes because monocytic lineage cells can also express Iba-1 in the disease state.^65,66 The composition of subretinal Iba-1⁺ cells remains unclear. 

In conclusion, our study provides a deeper understanding of microglial participation in the RD process, from activation to death. The highly inflammatory mode leads to microglial pyroptosis and contributes to PR death. Our data suggest that appropriately induced microglial activation by P2X7 blockade could be considered a pharmacotherapeutic strategy to promote PR survival. 

Acknowledgments

M.C. designed the experiments and wrote the manuscript. M.C and X.H. performed the experiments and analyzed the data. L.Z. conceived experiments, reviewed and edited the manuscript. J.Z. was in charge of mouse husbandry. Y.W. and Y.P. contributed to the surgery procedure. X.Z. and L.T. reviewed the manuscript and helped with literature referencing. All authors read and approved the final version of the manuscript and agree to be held accountable for all aspects of the work. 

Supported by the National Natural Science Foundation of China (No. 81700837, No. 81903167) and the National Science Foundation of Hunan Province (No. 2020JJ5827). 

Disclosure: M. Cao, None; X. Huang, None; J. Zou, None; Y. Peng, None; Y. Wang, None; X. Zheng, None; L. Tang, None; L. Zhang, None