cGAS inhibition alleviates Alu RNA-induced immune responses and cytotoxicity in retinal pigmented epithelium

The degeneration of retinal pigmented epithelium (RPE) cells results in severe diseases, such as age-related macular degeneration (AMD) that causes blindness in millions of individuals. We report that targeting GMP-AMP (cGAMP) synthase (cGAS) alleviates Alu RNA-induced immune responses and cytotoxicity in RPE. We find that the deletion of cGAS in RPE inhibits the Alu RNA-stimulated interferon production. cGAS deficiency also protects RPE from cell death triggered by Alu RNA. Importantly, two natural chemicals, epigallocatechin gallate (EGCG) and resveratrol (RSVL), are effective in suppressing the immunogenic and cytotoxic effect of Alu RNA in RPE. Our findings further demonstrate the crucial role of cGAS in the Alu RNA-induced RPE damage and present EGCG and RSVL as potential therapies for AMD and other RPE degeneration-related conditions.


Introduction
Age-related macular degeneration (AMD) is a prevalent disease that causes blindness in aged individuals [1][2][3]. Millions of people worldwide are suffering from vision loss as a result of AMD [1][2][3]. There are two main forms of AMD, neovascular AMD and geographic atrophy (GA) [1,2,4,5]. The neovascular AMD has been effectively treated through anti-angiogenesis strategies, such as targeting the vascular endothelial growth factor A (VEGFA) [1,2,[5][6][7]. In contrast, GA, the advanced form of AMD, is still untreatable [2,4,5,[8][9][10][11][12]. RPE cells form a monolayer of tissue that play critical roles in supporting the homeostasis of retina [2][3][4]. Studies indicated that the RPE degeneration is not only a key characteristic manifestation of GA, but also a major pathological factor that triggers GA [1,4,9,10]. Therefore, understanding the mechanisms underlying the degeneration of RPE is critical for the development of therapies for GA.
Alu RNA is a type of noncoding RNAs transcribed from Alu elements, which are the most abundant repetitive elements in the genome of humans [13][14][15]. It is believed that there are about 1 million copies of Alu elements in each genome [13,15]. A growing number of evidence showed that these repetitive elements shape the genome both structurally and functionally [13,15]. Interestingly, recent studies illustrated the pathogenic role of Alu RNAs in GA and found that the accumulation of Alu RNAs in RPE induced cell death through the activation of inflammasome [14] and the cytosolic DNA sensor, cGAS [9].
As a primary intracellular DNA sensor, cGAS detects the cytosolic DNA to elicit the downstream immune responses, such as the production of type I interferons

Open Access
Cell & Bioscience † Jing Li and Feng Zhang contributed equally to this work *Correspondence: crystal_lijing@ccmu.edu.cn; wanxh@ccmu.edu.cn 1 Beijing Tongren Eye Center, Beijing Tongren Hospital of Capital Medical University, Beijing 100730, China Full list of author information is available at the end of the article (IFNs) [16][17][18]. The emergence of DNA in the cytoplasm can be a result of either cellular damage or microbial infections [19]. Alu RNAs were found to activate cGAS by inducing the release of mitochondrial DNA (mtDNA), which is a ligand for cGAS [9,20]. cGAS activation then drives the inflammasome activation and the RPE degeneration in AMD [9]. Thus, the activation of cGAS is an important upstream event during the Alu RNAs-induced RPE degeneration. It is therefore suggested that cGAS inhibition may be a potential mean to preserve RPE health and to treat GA. Several cGAS inhibitors were identified recently. For example, two natural chemicals, epigallocatechin gallate (EGCG) [16] and resveratrol (RSVL) [21], were showed to suppress cGAS activation efficiently. In the current study, we explored whether these cGAS-inhibiting reagents could be used to ameliorate the Alu RNAs-induced immune responses and cell death in RPE.

dsRNA induced the interferon production in RPE cells
Alu RNA accumulation is implicated in GA [14]. To mimic this condition, we synthesized Alu RNA transcripts and transfected ARPE-19, an RPE cell line, with Alu RNAs. As expected, the introducing of Alu RNAs into the cytoplasm of ARPE-19 cells triggered the robust expression of IFN (Fig. 1a). The intracellular nucleic acidstimulated expression of IFN is dependent on the transcriptional factor, interferon regulatory factor 3 (IRF3) and the phosphorylation level of IRF3 can be used to reflect its activation [22][23][24]. We then detected the phosphorylation of IRF3 using immunoblotting with the specific antibodies against the phosphorylated IRF3. We showed that the transfection of Alu RNAs strongly stimulated the activation of IRF3 (Fig. 1b). We also observed the Alu RNA-induced IFN expression at different time points post transfection and found that Alu RNAs triggered the expression of IFN in a time-dependent manner (Fig. 1c, d). We next confirmed these findings by measuring the production of IFN using enzyme-linked immunosorbent assay (ELISA) (Fig. 1e). Further, using the synthetic analog of double-stranded RNA (dsRNA), polyinosinic-polycytidylic acid [poly(I:C)], we obtained the consistent results ( Fig. 1f-j). Together, these data suggested that Alu RNAs and other dsRNAs induce the interferon production in RPE cells.

cGAS is required for Alu RNA-induced IFN expression
As previous publication indicated that the Alu RNA stimulates cGAS activation through inducing the release of mtDNA [9], we next tested this in cells that we studied. We first generated cGAS null RPE cells using CRISPR/  Cas9. As expected, cGAS deletion abolished different types of DNA-induced IFN expression (Fig. 2a-c). By detecting the HT-DNA-induced phosphorylation of IRF3, we obtained consistent results (Fig. 2d). We then transfected both wild-type (WT) and cGAS -/-ARPE-19 cells with Alu RNAs and confirmed that cGAS is required for Alu RNA-induced IFN expression (Fig. 2e, f ). We also used the WT and CGAS -/-U937 cells, the monocytic cell line that is widely used for cGAS study, to examine the role of cGAS in response to Alu RNA challenge. In U937 cells, cGAS deficiency disrupted the DNA-induced IFN expression (Fig. 2g). The cGAS-mediated response to Alu RNA seemed to be a universal mechanism, as the Alu RNA-stimulated IFN expression was also attenuated in cGAS null U937 cells (Fig. 2h). Thus, cGAS is a key mediator in the signaling pathway downstream of Alu RNA in RPE.

cGAS is critical for Alu RNA-induced RPE death
To assess the Alu RNA-induced cell death in ARPE-

cells, we transfected WT and cGAS -/cells with Alu
RNAs and harvested the cells 48 h post transfection. The cells were then stained with Annexin V and propidium iodide (PI), which respectively indicate the early apoptotic cell death and the late apoptotic or other forms of cell death [25]. Using flow cytometry, we analyzed the percentage of dead cells in the transfected ARPE-19. We showed that while cGAS deletion did not lead to detectable cell death, it significantly reduced the Alu RNA-induced death of ARPE-19 ( Fig. 3a, b). Thus, cGAS deficiency may prevent RPE from Alu RNA-induced cell death. Our data further suggested that inhibition of cGAS could be used to rescue the Alu RNA-associated RPE degeneration.

EGCG/RSVL inhibits cGAS activation
EGCG and RSVL were respectively showed to inhibit the activation of cGAS [16,21]. We therefore examined their effects in ARPE-19 cells. To do so, ARPE-19 cells were pretreated with either EGCG or RSVL, followed by the transfection of DNA, which specifically activates cGAS.
Our results showed that the pretreatment of EGCG significantly suppressed the DNA-induced IFN expression in RPE cells (Fig. 4a). We also showed that EGCG effectively blocked the DNA-triggered phosphorylation of IRF3 (Fig. 4b). Similarly, we further showed that the pretreatment of RSVL also led to the inhibition of DNAinduced cGAS activation (Fig. 4c, d). Because EGCG and RSVL were reported to inhibit cGAS activation through GTPase-activating protein SH3 domain-binding protein 1(G3BP1) [16,21], we confirmed the expression of G3BP1 in both ARPE-19 and hTERT RPE-1 cell lines (Fig. 4e). Further, using EGCG-and RSVL-conjugated Sepharose beads, we performed pull-down assays and showed that both EGCG and RSVL can selectively bind to G3BP1 protein (Fig. 4f, g).
We next examined the cytotoxicity of both EGCG and RSVL on ARPE-19 cells. To do so, we cultured ARPE-19 cells in the presence of EGCG or RSVL for 48 h and analyzed the cell death using CellTiter assays. Our data showed that EGCG did not induce obvious cell death at 200 μM (Fig. 4h), while it significantly suppressed cGAS at 20 μM (Fig. 4a). RSVL exhibited marginal toxic effect on ARPE-19 cells (Fig. 4i). Thus, both EGCG and RSVL can be tolerated by ARPE-19 cells, at the concentrations we used to inhibit cGAS activation.

EGCG/RSVL suppresses Alu RNA-induced IFN expression
We then tested the effect of EGCG on Alu RNA-transfected RPE cells. When the cells were pretreated with increasing amount of EGCG followed by the Alu RNA transfection, we found that EGCG effectively dampened Alu RNA-induced IFN expression at 5 μM. Strikingly, 40 μM of EGCG almost blocked IFN expression triggered by Alu RNA (Fig. 5a). Consistently, EGCG treatment inhibited the Alu RNA-induced phosphorylation of IRF3 (Fig. 5b). Although the inhibitory effect of RSVL was not as potent as EGCG, our data showed that RSVL can markedly reduce the Alu RNA-induced IFN expression and the activation of IRF3 (Fig. 5c, d). We also treated the cells with EGCG and RSVL together to explore whether there was a synergistic effect of these two chemicals. As shown in Fig. 5e, EGCG + RSVL did not obviously inhibit the Alu RNA-induced IFN expression further, probably because the effect of EGCG alone was efficient enough. Moreover, with hTERT RPE-1 cells, we confirmed the effect of EGCG and RSVL (Fig. 5f, g). Using poly(I:C), we obtained the similar data indicating that both EGCG and RSVL were effective in attenuating dsRNA-induced IFN expression (Fig. 5h-k). Taken together, both EGCG and RSVL can be used to inhibit the dsRNA-triggered IFN expression.

EGCG/RSVL restrained Alu RNA-induced cell death of RPE
As Alu RNA significantly induced cell death in ARPE-19 cells (Fig. 3a, b) and cGAS deficiency prevented such cell death (Fig. 3a, b). We reasoned that EGCG and RSVL may have the effect in restraining Alu RNA-induced cell death in RPE cells. We then verified our hypothesis by treating ARPE-19 cells with EGCG prior to the Alu RNA-transfection. Our data showed that EGCG significantly recured the cell death triggered by Alu RNA transfection (Fig. 6a, b). RSVL also showed a similar effected in preventing cell death of RPE in the condition of Alu RNA challenging (Fig. 6c, d).
Taken together, our data suggested that inhibition of cGAS by EGCG or RSVL could be a potential treatment for Alu RNA-induced RPE degeneration. Our study thereby presenting these natural chemicals as possible therapies for GA.

Discussion
AMD, especially the advanced form, GA, is a prevalent, severe, and currently untreatable disease that causes vision-loss in millions of individuals [3,8,14,26]. The degeneration of RPE cells has been known as a major player in the pathogenesis of GA [3,8,14,26]. However, the lack of detailed molecular mechanisms of the RPE degeneration has hampered the development of effective therapies for GA [1]. Recently, a series of exciting works highlighted the critical role of cGAS in the Alu RNA-induced RPE death [8,9,14]. In the current study, we showed that inhibition of cGAS with natural chemicals protected RPE from Alu RNA-triggered cell death. We first showed that the deletion of cGAS in RPE dampened the Alu RNA-stimulated interferon production. cGAS-deficient RPE cells were resistant to Alu RNA-induced cell death. Importantly, we found that two natural chemicals, EGCG and RSVL, were effective in suppressing the immunogenic and cytotoxic effect of Alu RNA on RPE. Thus, our findings further demonstrated the crucial role of cGAS in the Alu RNAinduced RPE damage and present EGCG and RSVL as potential treatments for RPE degeneration-related conditions, such as AMD. cGAS is a cytoplasmic DNA sensor that responsible for the detection of invading pathogens by sensing the emerging of DNAs in the cytosol [19]. The aberrant activation of cGAS by self-DNA can be a major cause for a type of human diseases [16,18]. For example, the insufficient clearance of self-DNAs derived from the transcription of endogenous retroviruses or retrotransposons led to the accumulation of self-DNAs in the cytoplasm, which chronically stimulate the activation of cGAS [18,27]. In RPE cells, elevated transcription of Alu element results in the release of mtDNAs, which activate cGAS and its downstream production of interferons, and the activation of cGAS is required for the further activation inflammasome [8,9,14]. These events finally caused the degeneration of RPE cells. Our data suggested that the Alu RNA-mtDNA release-cGAS activation could be a universal mechanism in different cells. Thus, cGAS is a key target for the treatment of many intracellular nucleic acid-related diseases.
Besides Alu RNA-mtDNA releasing pathway, Alu RNA may also trigger the intracellular RNA sensormediated immune responses. EGCG and RSVL were recently reported to block intracellular RNA-sensing signaling [21]. The inhibitory effect of these two chemicals were mainly attributed to the inhibition of a key factor, G3BP1 [16,21,52]. Interestingly, G3BP1 was also a core organizer for the assembly of stress granules (SG), which is an important molecular condensation assembled in response to stress signals, such as the emergence of irregular RNA molecules in the cytoplasm [16,[53][54][55]. Although we did not detect the formation of SGs upon Alu RNA challenges in our study, it is very likely that Alu RNA will trigger the assembly of SG. Thus, through targeting G3BP1, EGCG and RSVL could preserve RPE health by executing the inhibitory effects at multiple layers of the dysregulated immune responses during RPE degeneration. As natural chemicals, both EGCG and RSVL are abundant in nature and are easy to acquire from plants [16,21,56]. Our study therefore suggests these chemicals as tangible lead compounds to prevent the development of GA.

Conclusions
Our findings further demonstrate the crucial role of cGAS in the Alu RNA-induced RPE damage and present EGCG and RSVL as potential therapies for AMD and other RPE degeneration-related conditions.

Cell culture and transfection
ARPE-19 cells were cultured in Advanced DMEM/F12 medium containing 10% Fetal Bovine Serum, 2 mM glutamine, 100 mg mL −1 penicillin, 100 mg mL −1 streptomycin. Cells were grown in a 5% CO 2 incubator (Thermo Fisher Scientific) at 37 °C. All cell lines were tested to be mycoplasma free by PCR.

Cell viability assay
ARPE-19 cells were seeded into 96-well plates and incubated with EGCG or RSVL at indicated concentrations for 48 h. CellTiter 96 ® AQueous One Solution Cell Proliferation Assay (G3580, Promega) was performed to analyze the cell viability according to the manufacturer's instruction.

Annexin V and PI staining
WT and CGAS −/− ARPE-19 cells were transfected with Alu RNA (4 μg mL −1 ) for 48 h, with or without a 3-h pretreatment of EGCG or RSVL. The Annexin V-and PIpositive cells were then measured by flow cytometer (BD Accuri tm C6 Plus analyzer) using the Annexin V-FITC apoptosis detection kit (P04D03, Gene-Protein Link).

RNA extraction and quantitative PCR (qPCR)
Total RNAs were isolated from cells with TRIZOL reagent (93,289, Sigma-Aldrich) and reverse transcribed with PrimeScript RT Reagent Kit (TaKaRa, RR037A). qPCR was performed with Powerup SYBR Green Master Mix (A25742, Thermo Fisher Scientific) on an ABI StepOnePlus system according to the manufacturer's instructions. qPCR data was analyzed by StepOnePlus software. The sequences for qPCR primers are listed below. mRNA level of human GAPDH was used for normalization.

Immunoblotting
Cells were lysed with lysis buffer (20 mM Tris-HCl pH 7.5, 0.5% Nonidet P-40, 250 mM NaCl, 3 mM EDTA, 3 mM EGTA, 2 mM dithiothreitol) with protease inhibitor cocktail (Roche, 04,693,132,001). Cell lysates were separated by SDS-PAGE and proteins were transferred onto PVDF membranes. The transferred PVDF membranes were blocked by 5% milk for 1 h at room temperature and subjected to primary antibody incubation at 4℃ for overnight. Protein bands were visualized with enhanced chemiluminescence (ThermoFisher Scientific).

Alu RNA transcription
Alu RNA were transcribed using MEGAshortscript ™ Kit (AM1354) in vitro according to the manufacturer's instructions.

Enzyme-linked immunosorbent assay
ARPE-19 cells were seeded into 12-well plates at a density of 2 × 10 5 cells per well and treated as indicated.
The secreted interferon in cell culture medium was analyzed with enzyme-linked immunosorbent assay (ELISA) kits (EHC026b.96, Neobioscience, for human) according to the manufacturer's instruction.

Quantification and statistical analysis
A standard two-tailed unpaired Student's t-test was used for statistical analysis of two groups. Data are expressed as mean ± SEM. Graphs and statistical analysis were performed using GraphPad Prism (version 8.0). P values < 0.05 were considered as statistically significant. Flow cytometry data were analyzed by FlowJo (version 10).