H-151, a Selective STING Inhibitor, Has Potential as a Treatment for Neovascular Age-Related Macular Degeneration

Abstract

Purpose: The cyclic guanosine monophosphate–adenosine monophosphate synthase (cGAS) stimulator of interferon gene (STING) pathway is a crucial cascade in the inflammatory response initiated by the recognition of cytosolic double-stranded DNA (dsDNA). The aim of this study was to evaluate the effect of STING inhibitor in murine choroidal neovascularization (CNV).

Methods: To investigate whether the cGAS–STING pathway is activated during CNV, CNV was induced using laser photocoagulation in male C57BL/6J mice. The expression of change of cGAS and STING during CNV development was confirmed by Western-blotting. H-151, a potent STING palmitoylation antagonist, was used as a STING inhibitor. H-151 was administered intravitreally immediately after laser induction. To confirm the role of the cGAS–STING pathway in CNV formation, we evaluated CNV size and performed fundus fluorescein angiography.

Results: The expression levels of cGAS and STING were significantly upregulated in the RPE–choroid complex after CNV induction, and dsDNA merged with cGAS was observed in CNV lesions. Intravitreal administration of H-151 suppressed CNV development and fluorescent leakage from neovessels. In CNV lesions, the high expression of STING and cGAS was observed in infiltrating F4/80⁺ macrophages. H-151 administration attenuated downstream signals of the cGAS–STING pathway, including the phosphorylation of nuclear factor–κB, and downregulated the expression of interleukin 1β.

Conclusions: These findings support that the inhibition of cGAS–STING pathway treats abnormal ocular angiogenesis.

Age-related macular degeneration (AMD) is a leading cause of severe vision impairment in many countries worldwide.¹ Neovascular age-related macular degeneration (AMD), also known as wet or exudative AMD, is responsible for approximately 90% of AMD-induced blindness.¹ Neovascular AMD is characterized by the formation of new choroidal vessels in the retina that break through Bruch's membrane into the subretinal or sub-RPE space.² Aggressive choroidal neovascularization (CNV) into the retina causes fluid leakage and hemorrhage in the macula. In ocular fluids from patients with neovascular AMD, the levels of VEGF are markedly increased.^3,4 Currently, anti-VEGF agents, including ranibizumab, aflibercept, brolucizumab, and faricimab (a bispecific antibody against VEGF and angiopoietin 2), have been approved as therapeutic agents for neovascular AMD. Intravitreal injection of anti-VEGF agents improves visual acuity; however, some patients are resistant to anti-VEGF therapy.⁵ Importantly, the SEVEN-UP study showed that visual outcomes gradually decreased in patients receiving anti-VEGF therapy.⁶ To overcome resistance to anti-VEGF therapy, it is necessary to reveal the underlying pathophysiology and search for a novel target using animal models. 

Inflammation and oxidative stress play crucial roles in the development and progression of AMD. Epidemiologic surveys show that inflammation-related factors are increased in patients with AMD.^7,8 Crucial risk factors for AMD, such as smoking and excessive light exposure, enhance oxidative stress.^9,10 Oxidative stress damages the RPE cells to metabolize the substrate, leading to an increase in extracellular deposits. Dysfunction of RPE cells results in photoreceptor degeneration, as photoreceptor cells rely on nutrients from RPE cells. Extracellular deposits drive immune cell infiltration and inflammation, and activated immune cells upregulate AMD-related cytokine production. Macrophages are key immune cells that accumulate in the AMD region.^11,12 In an experimental CNV model, macrophage depletion suppressed CNV.¹³ Importantly, macrophages are responsible for the production of proinflammatory cytokines, including matrix metalloproteinase 9 (MMP-9),¹⁴ IL-6,¹⁵ and IL-1β.¹⁶ These data indicate that the reduction of oxidative stress might be a novel therapeutic approach for neovascular AMD. 

Stimulator of interferon genes (STING) was identified as an innate immune signaling response to intracellular DNA viruses.¹⁷ In 2013, cyclic guanosine monophosphate–adenosine monophosphate synthase (cGAS) was found to act as an important cytosolic DNA sensor, producing type Ⅰ interferon genes.^18,19 After binding to double-stranded DNA (dsDNA), cGAS catalyzes cyclic GMP-AMP (cGAMP) from ATP and GTP.¹⁹ cGAMP acts as a second messenger and activates the adaptor protein STING. After binding to cGAMP, the endoplasmic reticulum protein STING initiates the production of interferon genes via TANK-binding kinase 1 (TBK1)–interferon regulatory factor 3 (IRF3) activation against non-self-DNA pathogens.¹⁹ Recent studies provide the evidence that the cGAS–STING pathway also recognizes self-DNA, leading to inflammatory diseases^20–23 or cellular senescence.^24,25 Stress-induced DNA damage to nuclei or mitochondria increases dsDNA release into the cytoplasm. In patients with AMD, the levels of oxidative stress markers are increased in fluid samples.²⁶ These findings indicate that DNA damage-induced inflammatory induction might regulate AMD pathology. However, the role of the cGAS–STING pathway in the pathogenesis of neovascular AMD remains unknown. 

In this study, we investigated the effectiveness of potent STING inhibitor, H-151, on CNV development and the contribution of cGAS–STING to neovascular AMD pathogenesis using a murine experimental CNV model. 

Materials and Methods

Animals

C57BL/6J mice were purchased from Japan SLC (Shizuoka, Japan). The mice were housed under a 12-hour/12-hour light/dark cycle with free access to food (CE-2; CLEA Japan, Tokyo, Japan) and water. All procedures were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved and monitored by the Institutional Animal Care and Use Committee of Gifu Pharmaceutical University. 

Laser-Induced CNV Model

Male C57BL/6J mice aged 8 weeks were anesthetized via intraperitoneal injection of a mixture of 43.8 mg/kg ketamine (Daiichi Sankyo, Tokyo, Japan) and 2.5 mg/kg xylazine (Bayer Health Care, Leverkusen, Germany). The pupils were dilated with 0.5% tropicamide and 0.5% phenylephrine (Mydrin-P; Santen Pharmaceutical, Osaka, Japan), and the cornea was covered with 0.1% sodium hyaluronate ophthalmic solution (Hyalein; Santen Pharmaceutical). Laser photocoagulation was induced using a 647-nm laser with a spot size of 50 µm, power of 120 mW, and duration of 100 ms. Six lesions per eye were induced around the optic nerve head, and successful rupture of Bruch's membrane was confirmed by bubble formation. Eyes with hemorrhage after laser photocoagulation or intravitreal injection were excluded. Immediately after laser induction, 2 µL/eye of H-151 (0.1 or 1.0 mM; CAS No. 941987-60-6; Selleck Chemicals, Houston, TX, USA) diluted in Dulbecco's phosphate-buffered saline (D-PBS; FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) containing 1% dimethyl sulfoxide (DMSO) was injected intravitreally. In the vehicle group, D-PBS containing 1% DMSO was injected. 

Measurement of CNV Size

To visualize CNV, PBS (0.5 mL) containing 20 mg/mL FITC-conjugated dextran (2000 kDa; Sigma-Aldrich, St. Louis, MO, USA) was injected into the tail vein. Mice were euthanized by cervical dislocation 5 minutes after dextran perfusion. The enucleated eyes were fixed with 4% paraformaldehyde (PFA; pH 7.4) for 2 hours at 4°C. The cornea, lens, and neuroretina were carefully removed. The RPE–choroid complex was flat-mounted on a glass slide using aqueous mounting medium (Fluoromount; Diagnostic BioSystems, Pleasanton, CA, USA). CNV lesions were visualized using an all-in-one fluorescence microscope (BZ-X710; Keyence, Osaka, Japan). To evaluate the CNV size, the CNV lesion area was enclosed individually, and its inner area was measured using the BZ-X analyzer (Keyence). The mean of the CNV lesion area was used as one sample. The size of the CNV lesions was analyzed in a blinded manner. 

Fundus Fluorescein Angiography

Fourteen days after laser induction, the mice were treated with Mydrin-P eye drops. Fundus and fluorescein angiography images were captured using a retinal imaging microscope (Micron IV; Phoenix Research Laboratories, Pleasanton, CA, USA). To perform fundus fluorescein angiography (FFA), 0.1 mL saline containing 10 mg/mL fluorescein solution (Fluorescite; Alcon, Tokyo, Japan) was injected into the tail vein. Fluorescent images were taken 1 and 3 minutes after injecting Fluorescite, and the leakage grades were scored in a blinded manner as described previously.²⁷ The leakage grades were defined as follows: 1, no hyperfluorescence; 2, hyperfluorescence without leakage; 3, hyperfluorescence with moderate leakage; and 4, bright hyperfluorescence with excessive leakage. 

Western Blot

The RPE–choroid complexes were separated from the eyecups, and the samples were immediately frozen in liquid nitrogen. For protein extraction, samples were lysed in radioimmunoprecipitation buffer (150 mM NaCl [Kishida Chemical, Osaka, Japan], 50 mM Tris-HCl at pH 8.0 [Nacalai Tesque, Kyoto, Japan], 0.5% sodium deoxycholate [FUJIFILM Wako Pure Chemical Corporation], 0.1% sodium dodecyl sulfate [FUJIFILM Wako Pure Chemical Corporation], 1% Igepal CA-630 [Sigma-Aldrich]) containing a protease inhibitor cocktail (Sigma-Aldrich) and phosphatase inhibitor cocktails 2 and 3 (Sigma-Aldrich). The RPE–choroid complex was homogenized using a Handy micro homogenizer (NS-310E3; Microtec Co., Ltd., Chiba, Japan). The samples were then centrifuged for 20 minutes at 12,000 × g and 4°C. The supernatants were boiled in sample buffer solution containing 3-mercapto-1,2-propanediol (FUJIFILM Wako Pure Chemical Corporation) for 5 minutes. Samples were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis on a 5% to 15% gradient polyacrylamide gel (FUJIFILM Wako Pure Chemical Corporation) and transferred onto polyvinylidene difluoride membranes (Immobilon-P; Millipore, Billerica, MA, USA). After being blocked with Blocking One-P solution (Nacalai Tesque), the membrane was incubated with Can Get Signal Solution 1 (TOYOBO, Osaka, Japan) containing primary antibodies overnight at 4°C. The following antibodies were used: rabbit anti-cGAS (1:1000 dilution; Cell Signaling Technology, Danvers, MA, USA), rabbit anti-STING (1:1000 dilution; Cell Signaling Technology), rabbit anti-phospho-TBK1 (1:1000 dilution; Cell Signaling Technology), rabbit anti-TBK1 (1:1000 dilution; Cell Signaling Technology), rabbit anti–phospho-IRF3 (1:1000 dilution; Cell Signaling Technology), rabbit anti-IRF3 (1:1000 dilution; Cell Signaling Technology), rabbit anti–phospho-nuclear factor–κB (NF-κB) (1:1000 dilution; Cell Signaling Technology), rabbit anti–NF-κB (1:1000 dilution; Cell Signaling Technology), and mouse anti–β-actin (1:2000 dilution; Sigma-Aldrich). The membranes were washed and then incubated with secondary antibodies for 1 hour at room temperature. The following antibodies were used: horseradish peroxidase (HRP)–conjugated goat anti-rabbit (1:2000 dilution; Thermo Fisher Scientific, Waltham, MA, USA) and HRP-conjugated goat anti-mouse (1:2000 dilution; Thermo Fisher Scientific). Immunoreactive bands were visualized using ImmunoStar LD (FUJIFILM Wako Pure Chemical Corporation) and their images were captured using an Amersham Imager 680 (GE Healthcare, Chicago, IL, USA). 

Immunofluorescence Staining of Retinal Sections

After perfusion with saline, the enucleated eyes were fixed in 4% PFA (pH 7.4) for 48 hours at 4°C. Fixed samples were cryo-protected in 25% sucrose for 48 hours at 4°C, embedded in Tissue-Tek OCT compound (Sakura Finetek, Torrance, CA, USA), and frozen in liquid nitrogen. Cryo-sections (10 µm thickness) were prepared using a cryostat (Leica Microsystems, Wetzlar, Germany) and mounted on Matsunami adhesive silane–coated glass slides (Matsunami, Osaka, Japan). To detect ezrin, isolectin B4 (IB-4), and F4/80, the sections were subjected to antigen retrieval using HistoVT One (Nacalai Tesque) and blocked with Protein Block Serum-Free Ready-to-Use (Agilent Technologies, Santa Clara, CA, USA) according to the manufacturer's instructions. The sections were blocked with M.O.M Blocking Reagent (Vector Laboratories, Burlingame, CA, USA) or 10% horse serum (Vector Laboratories) containing 0.3% Triton X-100 for 1 hour. The samples were incubated overnight with the primary antibodies and Isolectin GS-IB4 from Griffonia simplicifolia–conjugated Alexa Fluor 647 at 4°C. The following primary antibodies were used: mouse anti-dsDNA (1:50 dilution; Santa Cruz Biotechnology, Dallas, TX, USA), rat anti-F4/80 (1:50 dilution; Bio-Rad, Hercules, CA, USA), rabbit anti-STING (1:50 dilution; Novus Biologicals, Littleton, CO, USA), and mouse anti-cGAS (1:50 dilution; Santa Cruz Biotechnology). The sections were then incubated with the secondary antibodies for 1 hour at 4°C. The following secondary antibodies were used: Alexa Fluor 647 donkey anti-mouse (1:1000 dilution; Thermo Fisher Scientific), Alexa Fluor 647 donkey anti-rat (1:1000 dilution; Thermo Fisher Scientific), Alexa Fluor 488 donkey anti-rabbit (1:1000 dilution; Thermo Fisher Scientific), Alexa Fluor 546 donkey anti-rabbit (1:1000 dilution; Thermo Fisher Scientific), and Alexa Fluor 647 donkey anti-rabbit (1:1000 dilution; Thermo Fisher Scientific). The nuclei were stained with Hoechst 33342 (1:1000; Thermo Fisher Scientific) for 10 minutes. Primary and secondary antibodies were diluted in 10% horse serum or M.O.M. protein concentrate (Vector Laboratories). Finally, the sections were mounted using ProLong Glass Antifade Mountant (Thermo Fisher Scientific). Images were captured using a confocal microscope (FV3000, Olympus, Tokyo, Japan), and Supplementary Figure S1A. For immunostaining analysis as shown in Figure 1D, Figures 4A–D, we prepared six and four cases of CNV model mice on days 3 and 14 of laser irradiation. In addition, we used four nontreated mice as normal. We stained one or three sections in each data and showed typical images. 

Figure 1.

 

The expression of cGAS and STING was increased in a laser-induced CNV model. (A) In a laser-induced CNV model, male C57BL/6J mice were subjected to laser induction, and their RPE–choroid complex was analyzed by Western blot. The representative immunoblots show the expression of cGAS, STING, and β-actin in the RPE–choroid complex. (B, C) Quantitative analysis of cGAS (B) and STING (C) expression. Data are shown as the mean ± SEM (n = 5). **P < 0.01, *P < 0.05 vs. normal group (Tukey's test). (D) Immunofluorescent images showing dsDNA expression (red) in the CNV lesion. The eyes were enucleated 3 days after laser irradiation. Nuclei were stained with Hoechst 33342 (blue). Highly magnified images of the boxed areas are shown in the lower panels. Scale bar: 100 (upper) or 40 µm (lower). (E) Representative images of immunofluorescence staining of dsDNA (red) and cGAS (green) in the choroidal sections collected 3 days after laser induction. Nuclei were stained with Hoechst 33342 (blue). Scale bars: 50 µm (upper) or 20 µm (lower). Ch, choroid; INL, inner nuclear layer; ONL, outer nuclear layer.

Immunofluorescence Staining of the RPE–Choroid Flat-Mount

Three days after laser induction, mice were injected with 0.5 mL of 20 mg/mL FITC-conjugated dextran (2000 kDa) via the tail vein. Enucleated eyes were fixed in 4% PFA (pH 7.4) for 2 hours at 4°C. The cornea, lens, and neuroretina were removed from the eyecups. The RPE–choroid complex was flat-mounted and blocked with 10% horse serum containing 0.3% Triton X-100 for 1 hour. The samples were incubated with rabbit anti–Iba-1 antibody (1:200 dilution; FUJIFILM Wako Pure Chemical Corporation) overnight at 4°C. Subsequently, the samples were then incubated with Alexa Fluor 546 donkey anti-rabbit antibody (1:1000 dilution; Thermo Fisher Scientific) for 1 hour. The nuclei were stained with Hoechst 33342 (1:1000 dilution) for 10 minutes. Both the primary and secondary antibodies were diluted in 10% horse serum. Samples were mounted with Fluoromount, and images were captured using an all-in-one fluorescence microscope (BZ-X710; Keyence). The intensity of Iba-1 signals was analyzed using the ImageJ software (National Institutes of Health, Bethesda, MD, USA). 

Quantitative Reverse Transcription PCR

Three days after laser photocoagulation, mouse RPE–choroid complexes were collected from the enucleated eyes. The samples were isolated with ice-cold D-PBS to prevent RNA degradation. To extract RNA, the samples were homogenized using a disposable homogenizer (BioMasher II; Nippi, Tokyo, Japan), followed by RNA extraction using a NucleoSpin RNA Kit (Takara, Shiga, Japan). Total RNA was reverse-transcribed into cDNA using a PrimeScript RT Reagent Kit (Takara). Total RNA and cDNA were prepared according to the manufacturer's instructions. Total RNAs was subjected to PCR using TB Green Premix Ex Taq II (Takara). PCR cycling was conducted as follows: 95°C for 30 seconds, followed by 40 cycles of 95°C for 5 seconds and 60°C for 30 seconds. The following primer sequences were used: IL-1β, forward 5′-AACCTGCTGGTGTGTGACGTTC-3′ and reverse 5′-CAGCACGAGGCTTTTTTGTTGT-3′; VEGF, forward 5′-ACATTGGCTCACTTCCAGAAACAC-3′ and reverse 5′-GGTTGGAACCGGCATCTTTATC-3′; and GAPDH, forward 5′-TGTGTCCGTCGTGGATCTGA-3′ and reverse 5′-TTGCTGTTGAAGTCGCAGGA-3′. Gene expression was normalized to GAPDH expression. 

Statistical Analysis

Data are presented as mean ± SEM. Statistical comparisons were made using the Student's t-test, Welch's t-test, Dunnett's test, Tukey's test, and Kruskal–Wallis test with post hoc Bonferroni test (SPSS Statistics software; IBM Corporation, Armonk, NY, USA). P < 0.05 was considered statistically significant. 

Results

Expression of Cytosolic DNA Sensors in a CNV Mouse Model

To investigate the role of the cGAS–STING pathway in the pathogenesis of neovascular AMD, we confirmed the expression levels in laser-induced CNV model mice. In the RPE–choroid complex, cGAS expression was increased 3 days after CNV induction (Figs. 1A, 1B). STING expression was also significantly upregulated in the RPE–choroid complex 3 and 5 days after CNV induction (Figs. 1A, 1C). In immunofluorescence staining of dsDNA, dsDNA that did not colocalize with the nucleus was very abundant in CNV lesions, which were mainly localized to the nuclei in normal mice (Fig. 1D). Moreover, the high expression of cGAS was observed on dsDNA in CNV lesions at 3 days after CNV induction (Fig. 1E), but its trend was not observed 14 days after CNV induction (Supplementary Fig. S1A). These data suggest that the cGAS–STING pathway may play a crucial role in recognizing cytosolic DNA or phagocyted extracellular DNA during CNV formation. 

Effects of STING Inhibitor on Vascular Permeability

Pathologic angiogenesis from choroidal vessels is characterized by leakage and associated with subretinal fluid. To evaluate the effects of STING inhibitors on vascular permeability, we assessed fluorescein leakage from CNV lesions. H-151 (0.1 and 1.0 mM) was administered into the vitreous body immediately after laser photocoagulation. Fourteen days after CNV induction, FFA revealed hyperpermeable CNVs were formed by laser photocoagulation. In contrast, H-151 (1.0 mM) significantly suppressed fluorescein leakage (Figs. 2A, 2B). The average FFA grade per eye also showed that injection of 1.0 mM H-151 inhibited the hyperpermeable vessels (Fig. 2C). 

Figure 2.

 

H-151 attenuated fluorescein leakage from CNV lesions. (A) Representative fundus images of mice immediately after laser induction (upper) and 14 days after laser induction (middle). The FFA images show the fluorescein leakage from CNV lesions (lower). H-151 (2 µL/eye) was intravitreally injected immediately after laser induction. (B) Graphs showing the FFA grades of the vehicle, 0.1 mM H-151, and 1.0 mM H-151 groups. *P < 0.05 vs. vehicle group (n = 8–15, Kruskal–Wallis test followed by post hoc Bonferroni test). (C) Graphs showing the average FFA grades. Data are shown as the mean ± SEM (n = 8–15). *P < 0.05 vs. vehicle group (Dunnett's test).

Effects of STING Inhibitor Against the CNV Formation

To elucidate the role of the cGAS–STING pathway in CNV, we evaluated the effects of a STING inhibitor on the CNV model. H-151 is a potent inhibitor of the STING signaling pathway. H-151 (0.1 and 1.0 mM) was injected into the vitreous chamber immediately after laser irradiation. Quantitative analysis showed that the CNV area was significantly decreased upon intravitreal injection of 1.0 mM H-151 (Figs. 3A, 3B). 

Figure 3.

 

A STING inhibitor, H-151, suppressed the CNV. (A) The representative images of CNV. Intravitreal injection of H-151 (2 µL/eye) was conducted immediately after the laser photocoagulation. Fourteen days after the laser induction, CNV was visualized by FITC-labeled dextran perfusion. Scale bar: 50 µm. (B) Quantitative analysis for the size of CNV. Data are shown as the mean ± SEM (n = 8–15). *P < 0.05 vs. vehicle group (Student's t-test).

Distribution of Cytosolic DNA Sensors in CNV Lesions

Immunofluorescence staining showed that high expression of CNV-related STING was not merged with the RPE marker (ezrin) (Fig. 4A). In addition, the immunofluorescence signal of the endothelial cell marker (IB-4) was not merged with that, indicating high expression of CNV-related STING (Fig. 4B). In contrast, cGAS and STING expression was colocalized with F4/80 macrophages in CNV lesions (Figs. 4C, 4D). These findings indicate that the cGAS–STING pathway might play pivotal roles in macrophages during CNV development. 

Figure 4.

 

The cGAS and STING expressions were observed in a CNV-related macrophage. (A) Representative images of immunofluorescence staining of ezrin (red) and STING (green) using the retinal sections collected 3 days after laser induction. Nuclei were stained with Hoechst 33342 (blue). The right panels show the bright field of the retinal sections. Highly magnified images of boxed areas are shown in the lower panels. Scale bar: 50 (upper) or 20 µm (lower). (B) Representative images of immunofluorescence staining of IB-4 (red) and STING (green) using the retinal sections collected 3 days after laser induction. Nuclei were stained with Hoechst 33342 (blue). Highly magnified images of boxed areas are shown in the lower panels. Scale bar: 50 (upper) or 20 µm (lower). (C) Representative images of immunofluorescence staining of F4/80 (red) and cGAS (green) in the choroidal sections collected 3 days after laser induction. Nuclei were stained with Hoechst 33342 (blue). The magnified images show the F4/80⁺ cGAS⁺ cells in the CNV lesion. Highly magnified images of boxed areas are shown in the lower panels. Scale bar: 50 (upper) or 20 µm (lower). (D) Representative images of immunofluorescence staining of F4/80 (red) and STING (green) in the choroidal sections collected 3 days after laser induction. Nuclei were stained with Hoechst 33342 (blue). Highly magnified images of boxed areas are shown in the lower panels. Scale bar: 50 (upper) or 20 µm (lower).

STING Inhibitor-Induced Changes in ProInflammatory Cytokines

To elucidate the inhibitory effects of H-151 on macrophages, we analyzed the immunofluorescence intensity of Iba-1 in CNV lesions (Fig. 5A). Representative immunostaining showed that Iba-1–positive cells accumulated around CNV lesions (Fig. 5A). Quantitative analysis showed that the fluorescence intensity of Iba-1 was not changed by H-151 administration (Figs. 5A, 5B). In macrophages, NF-κB signaling regulates the production of proinflammatory cytokines. Samples from the RPE–choroid complex showed that IL-1β mRNA level was decreased following H-151 administration (Fig. 5C). In contrast, VEGF mRNA level was not altered (Fig. 5D). To study the intracellular regulators of inflammatory cytokines, we analyzed the levels of cGAS–STING and its downstream pathway. Western blot analysis showed that CNV induction increased the expression levels of cGAS and STING, which were not changed by the STING palmitoylation inhibitor, H-151 (Figs. 5E–G). Phosphorylation of TBK1, IRF3, and NF-κB was significantly enhanced by CNV induction, and 1.0 mM H-151 injection significantly suppressed their phosphorylation in the RPE–choroid complex (Figs. 5E, 5H–J). These data indicate that H-151 suppressed CNV formation by inhibiting NF-κB activation. 

Figure 5.

 

H-151 suppressed CNV-related NF-κB phosphorylation in the RPE–choroid complex. (A) Representative images of immunofluorescence staining of Iba-1 (red) in the choroidal sections collected 3 days after laser induction. CNV was visualized by performing FITC-conjugated dextran perfusion (green). Nuclei were stained with Hoechst 33342 (blue). Scale bar: 100 µm. (B) Quantitative analysis of the fluorescence intensity of Iba-1 signals around the CNV lesion. Data are shown as the mean ± SEM (n = 7–8). N.S., not significant (Student's t-test). (C, D) Relative mRNA expressions levels of IL-1β (C) and VEGF (D). Data are shown as the mean ± SEM (n = 10–11). *P < 0.05 vs. CNV-vehicle group; N.S., not significant (Welch's t-test). (E–H) Western blot analysis was performed 3 days after laser induction. Representative immunoblots showing the expression of cGAS (F), STING (G), pTBK1 (H), pIRF3 (I), and pNF-κB (J) in the RPE–choroid complex. Data are shown as the mean ± SEM (n = 4). **P < 0.01, *P < 0.05 vs. normal group; ^##P < 0.01 vs. CNV-vehicle group (Student's t-test).

Discussion

Here, we showed that the activation of the cGAS–STING pathway occurred during CNV development, and the high expressions of cGAS and STING were observed in CNV-related macrophages. In addition, the intravitreal administration of STING inhibitor had an antiangiogenic effect on the experimental murine CNV model. To our knowledge, this is the first study to report that the cGAS–STING pathway regulates CNV pathogenesis. 

Under physiologic conditions, oxidative stress is regulated by the antioxidant system. However, patients with AMD cannot maintain a redox balance,^26,28 leading to oxidative stress and cellular damage. Key risk factors for AMD, including smoking and light exposure, increase oxidative stress.^9,10 It is speculated that accumulated stress is sufficient to trigger nuclear and/or mitochondrial DNA damage, followed by the release of DNA into the cytoplasm.²⁹ Cytosolic DNA is then recognized by the sensor protein cGAS, and its downstream signaling pathway promotes inflammatory cytokine production via the NF-κB pathway.³⁰ Therefore, increasing oxidative stress in patients with AMD would stimulate cGAS–STING signaling. Indeed, increased levels of 8-hydroxy-2′-deoxyguanosine (8-OHdG) were observed in the aqueous humor of patients with neovascular AMD,²⁶ and a free radical scavenger can reduce CNV development in mice.³¹ This study also supported evidence that cytosolic DNA might increase around CNV lesions and lead to the activation of cGAS–STING pathway. The origin and expression change of cytosolic DNA in CNV lesions is still unknown; however, our findings indicate the significance of the cGAS–STING pathway in the development of neovascular AMD. 

H-151 was discovered as a potent and selective STING antagonist that binds to STING Cys91, thereby inhibiting its palmitoylation and downstream activation.³² H-151 can be used as both a human and a mouse STING inhibitor, and expression of type I interferon- and NF-κB–regulated genes was suppressed by H-151. Notably, the cellular senescent phenotype is inhibited by H-151 treatment.^25,33 A recent study showed that systemic H-151 administration can alleviate inflammatory diseases, including acute kidney injury,³⁴ neuron loss, and neurologic disability in amyotrophic lateral sclerosis model mice.²¹ These data suggested that H-151 is a promising therapeutic agent for inflammatory disorders. The present study showed that intravitreal administration of H-151 reduced pathologic angiogenesis and vascular permeability in a laser-induced CNV model. In vitro studies have been conducted using 1 to 5 µM H-151,^21,25,32,33 while in in vivo studies, H-151 was injected intraperitoneally at 210 µg/mice.^21,32 Murine vitreous volume is estimated at 5.3 µL,³⁵ and we have injected 2 µL H-151 at 0.1 mM (55.868 ng) or 1.0 mM (558.68 ng) into each murine vitreous chamber in our study. Because small molecules were unstable and diluted in the vitreous chamber, it is reasonable to conclude that only 1.0 mM H-151 injection can suppress CNV pathogenesis. However, the concentration of administrated H-151 was higher than those in some previous reports; thus, further data about pharmacokinetics and specificity of H-151 in eyes would be needed to verify the existence of off-target effects. As one indicator, the specificity of H-151 confirmed that H-151 significantly reduced TBK1, IRF3, and NF-κB activation, which were downstream of the cGAS–STING pathway, without affecting the protein expression levels of cGAS and STING. Therefore, our results indicate that H-151 might suppress the activation of STING and its downstream, and it contributes to the decrease of pNF-κB and IL-1β. Since the essential action of H-151 is to inhibit the STING palmitoylation, it will need to investigate the detailed change of palmitoylation after H-151 administration. However, at least, our findings indicate that selective STING antagonists may be novel therapeutic agents for neovascular AMD. 

The importance of macrophages in CNV development has been demonstrated in human subjects and experimental animal models. In patients with AMD, macrophages are observed beneath the thinning Bruch's membrane,¹¹ suggesting that macrophages play pivotal roles in the pathogenesis of neovascular AMD. In the experimental CNV model, macrophage depletion can suppress CNV.¹³ The main sources of VEGF in the retina and choroid are macrophages,¹³ RPE,³⁶ and Müller cells.³⁷ Macrophages also produce proinflammatory cytokines, including IL-1β,¹⁶ MMP-9,¹⁴ and IL-6.¹⁵ This study showed that STING is markedly expressed by CNV-related macrophages but not by RPE or endothelial cells. The NF-κB pathway, a key regulator of proinflammatory cytokine production in macrophages, was attenuated by a STING inhibitor. Importantly, previous reports have shown that IL-1β is increased in patients with neovascular AMD,⁸ and IL-1β inhibition prevents CNV formation in a murine model.³⁸ In addition, CNV-induced upregulation of IL-1β expression was suppressed by H-151 injection; however, the expression levels of VEGF mRNA were not changed in CNV model mice. These data suggest that the STING pathway might play a pivotal role in the regulation of macrophages during CNV development. 

This study reveals the importance of the cGAS–STING pathway in neovascular AMD. The involvement of the cGAS–STING pathway in patients with neovascular AMD remains unknown, and this study supports the importance of the cGAS–STING pathway in CNV. In addition, since we focused on whether the activation of cGAS–STING pathway induced in neovascular AMD pathology and whether STING-specific inhibitor had an antiangiogenic effect on CNV formation, the essential function of the cGAS–STING pathway is still unclear. Therefore, further experiments using cGAS and STING knockout rodents would help to elucidate the physiologic function of cGAS–STING pathway in CNV formation in more detail. Taken together, our data suggest that the cGAS–STING pathway could regulate the pathophysiology of neovascular AMD and that STING inhibitors could be promising novel therapeutic agents for neovascular AMD. 

Acknowledgments

The authors thank Kota Aoshima for providing excellent technical assistance and Editage (www.editage.com) for English language editing. 

Supported by JST SPRING (grant JPMJSP2142) and Nagai Memorial Research Scholarship from the Pharmaceutical Society of Japan. 

Disclosure: M. Tanaka, None; H. Yasuda, None; S. Nakamura, None; M. Shimazawa, None