Reagents

Hyperoside (purity ≥ 98%, Lot. No. JOT-10,193) was ordered from Chengdu Pufei De Biotech Co., Ltd (China). 2-Aminoethyl diphenylborinate (2-APB) and lipopolysaccharides (LPS) were obtained from Sigma-Aldrich (USA). Calf thymus DNA (ctDNA), Lipofectamine 2000, Dulbecco’s Modified Eagle’s Medium (DMEM), and penicillin/streptomycin were products from Thermo Fisher Scientific (USA). Fetal bovine serum (FBS) was ordered from Nobimpex (Germany). 2′,3′-Cyclic guanosine monophosphate–adenosine monophosphate (cGAMP) sodium and 5,6-dimethyl-9-oxo-9 H-xanthene-4-acetic acid (DMXAA) were products from MedChemExpress (USA). Evans blue dye was purchased from Solarbio (China).

Cell culture

The BV-2 microglial cell line, acquired from the Cell Bank of the Chinese Academy of Sciences (China), was cultured using DMEM supplemented with 10% FBS and 1% penicillin/streptomycin at 37 °C in an atmosphere of 5% CO2. For LPS stimulation experiments, BV-2 cells were treated with hyperoside at the specified doses or vehicle for 1 h, followed by stimulation with LPS at 20 ng/ml for 12–24 h. For experiments involving ctDNA, after treating BV-2 cells with hyperoside at the specified doses or vehicle for 1 h, the cells were stimulated with ctDNA at 1 µg/ml for 3–6 h. For experiments involving DMXAA and 2′3′-cGAMP, BV-2 cells were treated with hyperoside or a vehicle for 1 h, followed by DMXAA stimulation at 10 µg/ml or 2′3′-cGAMP stimulation at 20 µM for 3–6 h. Transfection of ctDNA or 2′3′-cGAMP was performed using Lipofectamine 2000 following the manufacturer’s instructions. The doses and time points for the indicated treatments were determined by preliminary experiments.

Enzyme-linked immunosorbent assay (ELISA)

The supernatants from the BV-2 cells were collected after the indicated treatments. The levels of tumor necrosis factor (TNF), interleukin 6 (IL6), chemokine ligand 2 (CCL2) and interferon beta (IFNB) were then quantified using ELISA. Briefly, BV-2 cells were seeded at 6 × 104 cells per well in 24-well plates and stimulated with LPS, ctDNA, DMXAA or 2′3′-cGAMP in the absence or presence of hyperoside. The supernatants were then collected and subjected to the measurement using the Mouse TNF alpha Uncoated ELISA Kit (Thermo Fisher Scientific, USA), Mouse IL-6 Uncoated ELISA Kit (Thermo Fisher Scientific, USA), Mouse MCP-1/CCL2 Uncoated ELISA Kit (Thermo Fisher Scientific, USA), or Mouse IFN beta ELISA Kit (Multi Sciences, China) following the manufacturers’ instructions. For measuring the amount of 2′3′-cGAMP, BV-2 cells were lysed in RIPA buffer (Beyotime, China) containing proteinase inhibitors (Roche, Germany) on ice for 30 min, followed by centrifugation at 12,000 × g at 4°C for 30 min. The supernatant was then collected and analyzed using the 2′3′-cGAMP ELISA Kit (Cayman Chemical, USA) following the manufacturer’s instructions. For measuring the amount of 2′3′-cGAMP in cell-free system, hyperoside or CU-76, a cyclic GMP-AMP synthase (cGAS) inhibitor serving as the positive control, was co-incubated with cGAS in the presence of DNA, ATP, and GTP for 30 min, followed by measuring the amount of 2′3′-cGAMP using the cGAS Inhibitor Screening Assay Kit (Cayman Chemical, USA) according to the manufacturer’s instructions.

Scratch wound assay

The migratory capacity of BV-2 cells was evaluated by a scratch wound assay using the IncuCyte system (Essen Biosciences, German). In brief, BV-2 cells were seeded at 3 × 104 per well on 96-well ImageLock plates and were allowed to settle for 16 h. Subsequently, a scratch wound was created in each well using a 96-pin WoundMaker. The wells were then washed once with DMEM before adding fresh media containing either vehicle, LPS or hyperoside at the indicated concentrations. Images were automatically captured at the 1-h interval and recorded for the duration of 24 h using the IncuCyte Zoom with a 10× objective (Essen Biosciences, German). The data were then analyzed using an integrated metric of Incucyte, relative wound density. This algorithm measures cell density within the wound area relative to the cell density outside of the wound area at the specified time point. At 0 h post wounding, the relative wound density is set to be 0. When the cell density inside of the wound is the same as the cell density outside of the initial wound, the relative wound density is 100%.

Animals and treatments

Six to twelve-week-old female Balb/c mice were obtained from the Shanghai Laboratory Animal Research Center and maintained under the laboratory conditions including a 12/12-h light-dark cycle and the room temperature set at 20 ± 2 °C. Dark-adapted mice were subjected to 30-min experimental light exposure (10,000 lx) (Compact Fluorescence Lamp, 45 W, Chaoya Lighting, Shanghai, China), followed by intraperitoneal injection of hyperoside (prepared in 0.5% sodium carboxymethyl cellulose solution) at 100 mg/kg twice a day for 7 d. The first dose of hyperoside was delivered 3 h after the illumination. The volume for intraperitoneal injection of hyperoside was set at 100 µl per mouse. 2-APB (dissolved in DMSO) was administered intraperitoneally at 2.5 or 10 mg/kg 30 min before illumination for one-dose pretreatment and at 10 mg/kg twice a day for post-light damage treatment. The volume for 2-APB treatment was controlled at 50 µl per mouse. Normal controls (the mice unexposed to the experimental light) and the light-exposed mice without hyperoside or 2-APB treatment received vehicle treatment in the same manner. The animal handling protocols were reviewed and approved by the Institutional Animal Care and Use Committee at Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine (YYLAC-2023-199-1). The experiments were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research.

Evans blue dye uptake assay

Evans blue dye was injected via tail vein at 25 mg/kg. The mice were then subjected to illumination at 10,000 lx for 30 min. The eyes were enucleated at 1, 3 and 6 h post illumination. The eye cups free of the cornea and lens were then made and processed for cryosectioning. The autofluorescence of Evans blue dye was recorded using a fluorescent microscope (DM6000B, Leica, Germany). Quantification of the area of Evans blue autofluorescence in the outer nuclear layer (ONL) was performed using ImageJ.

Optical coherence tomography (OCT)

Image-guided OCT (OCT 2 with Micron IV, Phoenix Research Labs, USA) was performed to scan the retina 7 d post illumination as previously described [10]. In brief, after administering anesthetic cocktail of ketamine hydrochloride (82.5 mg/kg body weight) and xylazine (8.25 mg/kg body weight), pupil dilation was induced using 1% tropicamide (Santen Pharmaceutical, Japan) to prepare the mice for the OCT imaging. Five full-retinal scans were obtained and automatically averaged using the Phoenix Reveal OCT software (Phoenix Research Labs, USA). The thickness of the ONL was measured using the Insight Image Segmentation Software, an adjunct software to the Phoenix OCT and Retinal Imaging System (Version 2.0.5490, Voxeleron LLC, USA).

Electroretinography (ERG)

Seven days after the illumination, dark-adapted mice were subjected to ERG analysis under a safe light condition (5 lx) as previously described [10]. Flashes of green light (504 nm) were delivered at intensities of -2 (0.5 msec duration and 5 s inter-stimulus-interval), -0.8 (1 msec duration and 5 s inter-stimulus-interval), 0.4 (1 msec duration and 10 s inter-stimulus-interval), 1.6 (1 msec duration and 20 s inter-stimulus-interval), and 3.1 (1 msec duration and 60 s inter-stimulus-interval) log cd·s·m− 2. ERG responses were recorded using the Ganzfeld ERG system and analyzed by LabScribe software (ERG 2, Phoenix Research Labs, USA).

Histological examination and immunohistochemistry (IHC)

The enucleated eyes were fixed in 4% paraformaldehyde for 24 h prior to further processing for paraffin sectioning. Paraffin Sect. (4-µm-thick) were stained with hematoxylin and eosin (HE) or subjected to IHC examination using the primary antibodies including mouse anti-double stranded DNA (dsDNA) (1:500) (Santa Cruz, USA) and rabbit anti-high mobility group box 1 (HMGB1) (1:500) (Millipore, USA) as well as the secondary antibodies including Cy3-conjugated sheep anti-mouse (1:1000) or Cy3-sheep anti-rabbit secondary antibodies (1:1000) (Sigma-Aldrich, USA). Additionally, eye cups were fixed in 4% paraformaldehyde for 2 h at room temperature and processed for cryosectioning. Cryosections (12-µm-thick) were examined by IHC using primary antibodies including rabbit anti-ionized calcium binding adaptor molecule 1 (Iba-1) (1:500) (Wako Chemicals, USA), rabbit anti-CD68 (1:500) (Abcam, USA), rabbit anti-phospho-stimulator of interferon genes (STING) (Ser366) (1:250) (Thermo Fisher, USA), rat anti-F4/80 (1:250) (Thermo Fisher, USA), and the secondary antibodies including Cy3-conjugated sheep anti-rabbit (1:1000) (Sigma-Aldrich, USA), Cy3-conjugated goat anti-rat (1:1000) (Abcam, USA) or FITC-conjugated donkey anti-rabbit (1:1000) (Sigma-Aldrich, USA). Nuclei counterstaining was performed with 4-6-diamidino-2-phenylindole (DAPI) (Sigma Aldrich, USA). Microscopic imaging was conducted using a light microscope (DM2000, Leica, Germany) or a fluorescent microscope (DM6000B, Leica, Germany). Quantification of dsDNA positivity in the ONL was performed using ImageJ.

Real-time quantitative PCR (qPCR)

Total RNA from the mouse retinas was isolated using TRIzol reagent (Invitrogen, USA). Reverse transcription was then performed using PrimeScript RT Master Mix (TaKaRa, Japan). SYBR Green I Master (Roche, USA) was used to set up PCR reactions, which were run on a Roche Light Cycler 480 II (Roche, USA). The expression of Ccl5, Ccl6, Cd68, Cgas, Cxcl10, Ifi202b, Ifnb, Ikbke, Il6, Irf7, Sting1 and Tnf was analyzed along with the internal control 18 S rRNA. The primer sequences are included in Table 1. The fold change in gene expression was calculated based on 2−[Ct (candidate gene)−Ct (18s rRNA)].

Western blotting

Cell lysates were extracted from BV-2 cells using RIPA lysis buffer (Beyotime, China) supplemented with protease and phosphatase inhibitors (Roche, Germany). The lysates were then run on 10% SDS-PAGE gels and transferred onto polyvinylidene fluoride membranes (Millipore, USA). Primary antibodies used for Western blotting included monoclonal rabbit anti-cGAS (#31659S, Cell Signaling Technology, USA), monoclonal rabbit anti-phospho-tank binding kinase 1 (TBK1)/NAK (Ser172) (#5483S, Cell Signaling Technology, USA) and monoclonal rabbit anti-TBK1/NAK (#ab40676, Abcam, USA). Following incubation with the primary antibodies, horseradish peroxidase-conjugated goat anti-rabbit secondary antibody (Promega, USA) and the WesternBright ECL reagent (Advansta, USA) were sequentially applied. The blots were scanned using a ChemiScope 6000 imaging system (Clinx Science Instruments, China). Densitometry analysis was conducted using ImageJ.

Molecular docking and molecular dynamics (MD) simulations

The crystal structure of mouse cGAS was obtained from the Protein Data Bank (PDB ID: 4O6A) and the molecular structure of hyperoside was downloaded from PubChem (Compound CID: 5,281,643). Prior to the docking analysis, water molecules and original cGAS ligands such as DNA were removed from the file of cGAS. The polar hydrogen atoms and Gasteiger charges were added to hyperoside and cGAS, respectively. Non-polar hydrogen atoms were merged using AutoDock Tools1.5.6 software. A semi-flexible docking process between hyperoside and cGAS was performed using Autodock Vina software. The specific parameter is center_x=-31.129, center_y=-18.476 and center_z = 36.01. A total of 100 docking conformations were generated and ranked based on their docking energy values. MD simulation was then conducted using the GROMACS 19.5 package (https://manual.gromacs.org/) with the TIP3P explicit water model. The system was electrically neutralized by adding counterions. Subsequently, energy minimization (1000.0 kJ/mol/nm) was performed using the steepest descent method. Canonical ensemble simulation (NVT, 400 ps) and isothermal isobaric simulation (NPT, 400 ps) were used to guarantee that the system could endure constant temperature and pressure (300 K, 1.01325 bar). The annealing method was used to make sure that the system was slowly warmed from 0 to 300 K within 150 ps in the NVT simulation process. Lastly, a 50-ns MD simulation was started. The most populated cluster during the 30–50 ns MD simulation was selected for detailed analysis of the binding modes.

Surface plasmon resonance (SPR)

SPR was performed at 25 °C in the LMW multi-cycle mode on a BIAcore T200 instrument (Cytiva, USA). Briefly, the NTA biosensor chip (Cytiva, USA) was immobilized with human recombinant cGAS. Hyperoside were then flowed over the chip at the indicated concentrations (1.56 µM, 3.125 µM, 6.25 µM, 12.5 µM, 25 µM, 50 µM, 100 µM, and 200 µM) and the real-time responses were recorded. The binding affinity was assessed based on the equilibrium dissociation constant (KD) calculated using the BIAcore T200 evaluation software (Cytiva, USA).

Cellular thermal shift assay (CETSA)

BV-2 cells were treated with 200 µM hyperoside or DMSO for 3 h. The cells were then washed 3 times with PBS and resuspended in 1 ml PBS containing protease inhibitors (Roche, Germany). The cell suspensions were heated at 48 °C, 51 °C, 54 °C, 57 °C, and 60 °C. The cells were then processed for 3 cycles of snap-freezing in liquid nitrogen and thawing at 25 °C. Centrifugation was subsequently performed at 12,000 × g at 4 °C for 30 min to collect the supernatants. Equal amount of the supernatant from each treatment group was then subjected to Western blotting to assess the protein level of cGAS. The level of β-actin was simultaneously probed as a loading control. The protein bands were visualized using NBT/BCIP substrates (Solarbio, China).

Statistical analysis

Data were presented as mean ± standard error of the mean (SEM). Statistical analyses were conducted using one-way or two-way ANOVA, followed by Tukey’s multiple comparisons test (GraphPad Prism 9, USA). A P value less than 0.05 was considered statistically significant.