Activating cGAS–STING axis contributes to neuroinflammation in CVST mouse model and induces inflammasome activation and microglia pyroptosis

A total of 201 male C57BL/6 J mice (8–10 weeks old, 22–25 g) were obtained from the Animal Experiment Center of Sun Yat-Sen University. Animals were housed in a specific pathogen-free laboratory with fixed 12-h light/dark cycles under controlled temperature and humidity conditions (22 ± 1 °C; 45–55%). Standard rodent food and water were available ad libitum throughout the whole experiment. All animal protocols were approved by the Sun Yat-sen University Ethics Committee in conformity to the guidelines for the Care and Use of Laboratory Animals by the National Institutes of Health (Approval No. L102012018040Z).

Modeling of SSS thrombosis was performed as previously described with some modifications [4, 33]. Briefly, each mouse was anesthetized by intraperitoneal injection of sodium pentobarbital (40 mg/kg) and fixed in a stereotactic frame. The skull was exposed at the dorsal aspect of the head by a median skin incision of 1.5 cm. After the bregma and lambda were identified, a high-speed drill was employed to create a longitudinal cranial window between them under a surgical microscope (Carl Zeiss, Inc., Wetzlar, Germany). The drill tip was cooled frequently with physiological saline to alleviate thermal damage. The SSS was carefully exposed to ensure that the sinus and dura remained intact. Thrombosis of the sinus was induced by topical placement of a 5.5 mm long 5–0 silk thread (orientated from the site of the lambda to the site of bregma) soaked with 20% ferric chloride (FeCl₃, 157740, Sigma-Aldrich) solution on the surface of SSS for 6 or 12 min. If a sufficient thrombus to cause complete blockage of SSS was not produced within 6 min, the FeCl₃ attachment was repeated for a further 6 min. After the attachment of the ferric chloride, the field was flushed several times with warm saline. Adjacently, the incision was sutured and the mouse was subcutaneously injected with 0.2 mL of saline to prevent dehydration. In sham treated animals, a saline-soaked silk thread was applied to the SSS. Throughout the surgical procedure, the body temperature was sustained at 37 °C using a thermostatic heating pad.

RU.521 (450 ug/kg, RU5-42-01, InvivoGen, USA) was dissolved in 1% DMSO + corn oil (DMSO, Sigma-Aldrich; corn oil, Cat.No.: HY-Y1888. MedchemExpress) and intranasally administrated at 2 h, 12 h, 24 h post-CVST, and then once a day until the seventh day. 2′3′-cGAMP (500 μg/kg, GA3-42-03, InvivoGen, USA) or vehicle (phosphate-buffered saline, PBS) was intranasally administrated at 2 h, 12 h, 24 h, and 48 h after CVST. STING siRNA (320 pmol, E-055528-00-0005, Dharmacon, USA) and scrambled siRNA (320 pmol, D-001910-01-05, Dharmacon, USA), dissolved in 2 ul of sterile DNase/RNase-free water, were slowly injected into the right lateral ventricle with the aid of a stereotaxic frame (the coordinates relative to bregma: 0.5 mm posterior, 1.0 mm lateral, 2.25 mm deep from the skull surface) at 48 h before CVST induction. The drug-delivery way and dosage mentioned were selected as previously described [12].

After deep anesthesia, mice were transcardially perfused with saline followed by 4% paraformaldehyde at specific time points after CVST. For paraffin slices, after fixation in 4% paraformaldehyde, dehydration, and paraffin-embedding, 4 μm-thickness coronary sections of the mouse brain (as shown in Fig. 1B) samples were prepared. For frozen sections, perfused brain tissues were immersed in 4% paraformaldehyde at 4 ℃ overnight and then incubated in 15 and 30% (wt/vol) sucrose in PBS for 24 h at 4 °C. Mouse brains were then embedded in the optimal cutting temperature (O.C.T.) compound and stored at − 80 °C. Consecutive 10-um-thick coronal sections were prepared using a freezing microtome (CM1950, Leica, Germany) at temperatures below − 20 °C.

Apoptosis was detected with a terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL) kit (C1090, Beyotime, China) according to the manufacturers instructions. Briefly, paraffin-embedded brain sections were dewaxed and incubated with Proteinase K (20 μg/mL) for 20 min at room temperature. After being washed with PBS, sections were incubated with the TUNEL reaction mixture for 60 min at 37 °C in the dark. Nuclei were counterstained with DAPI (C1005, Beyotime) and the sections were visualized and captured under a Leica fluorescent microscope (DM2500, Germany) and a Nikon microscope (TI2-E). TUNEL‐positive cells were expressed as a percentage of the total cell count.

Fluoro-Jade C (FJC) staining was performed using a commercial detection kit (Millipore, Darmstadt, Germany) to evaluate degenerating neurons as previously depicted [34]. Data are presented showing the average number of FJC-positive neurons in the fields as cells/mm². The number of stained cells was analyzed using Image J software.

DHE staining was employed to detect the ROS levels in mouse brains as previously reported [34]. Fresh frozen brain Sects. (10 um) were incubated with 2 μmol/L fluorescent dye dihydroethidium (D7008, Sigma-Aldrich, Germany) at 37 °C for 30 min in a dark humidified chamber. After counterstaining with DAPI for 10 min, images of the ROS staining were captured via a fluorescence microscope (Leica, DMI8, Germany). For the quantitative analysis, the average of the DHE-positive cells was calculated from two randomly selected areas with the use of ImageJ software (ImageJ, USA) and expressed as a percentage of the total cell count.

The post-CVST motor functions were investigated using the rotarod test as formerly described [35]. Briefly, mice were placed on a rotating drum with a speed accelerating from 4 to 40 rpm within 5 min. The times at which the mice fell off the drum (latency to fall) were recorded. Before surgery, 5 trials were carried out on each mouse and the mean time of trial numbers 3, 4, and 5 was used as the pre-surgery baseline. After the operation, mice were tested for 5 trials on the 1 d, 3 d, 5 d, 7 d, 10 d, 14 d post-CVST with intervals of 5 min at least. The data for trial numbers 3–5 were utilized to calculate the mean time of latency to fall.

Paraffin sections were prepared as aforementioned and antigen retrieval was performed with Citrate–EDTA Antigen Retrieval Solution (P0086, Beyotime, China) in a microwave for 25 min. The slides were then blocked with 5% goat serum (SL038, Solarbio life sciences, china) at room temperature for 30 min and then incubated overnight at 4 °C with primary antibodies as follows: rabbit anti-STING antibody (1:100, NBP2-24683SS, Novus), rabbit anti-GSDMD antibody (1:500, ab219800, Abcam), mouse anti-IBA-1 antibody (1:200, GB12105, Servicebio), mouse anti-NeuN antibody (1:500, ab104224, Abcam), mouse anti-GFAP antibody(1:200, GB12105, Servicebio), rabbit anti-TNF-α antibody (1:200, GB11188, Servicebio). Additionally, the IF staining of the frozen slices was conducted as previously described [9]. Briefly, after fixation, permeabilization, and blocking, the sections were incubated with the following primary antibodies at 4 °C overnight: mouse anti-dsDNA (sc-58749, 1:100, Santa Cruz Biotechnology), rabbit anti-cGAS antibody (GTX02874, 1:100, Gentex), mouse anti-8-OhdG antibody (1:300, ab62623, Abcam).

For immunofluorescent double-staining, the primary antibodies were mixed and incubated under the same conditions. The sections were washed with PBS and then incubated with the corresponding secondary antibodies as follows: goat anti-mouse IgG H&L (Alexa Fluor® 594), 1:400, Abcam, ab150116; goat anti-rabbit IgG H&L(Alexa Fluor® 488), 1:400, Abcam, ab150077; Goat Anti-Mouse IgG H&L (Alexa Fluor® 488), 1:400, Abcam, ab150113; Goat Anti-Rabbit IgG H&L (Alexa Fluor® 594),1:400, Abcam, ab150080) for 1 h at 37 ℃. After washing with PBS, the nuclei were counterstained with DAPI (C0065, Solarbio, China) for 10 min. Images were captured by fluorescence (Leica, DMI8, Germany) and confocal microscopy (Leica, TSCSP8, Germany).

IHC staining was performed according to the commercial kits (PV-6001 and PV-6002, ZSGB-Bio, Beijing, China). Antigen retrieval of the paraffin sections was performed using the same protocol described for IF staining. Subsequently, sections were incubated with 0.3% H₂O₂ for 10 min to inactivate endogenous peroxidase activity and then washed with PBS. After being blocked with 5% goat serum for 15 min, slices were incubated overnight at 4 °C with the primary antibodies as follows: rabbit anti-NLRP3 antibody (ab214185, 1:200, Abcam), rabbit anti-cleaved-caspase 1 p20 (1:100, AF4005, Affinity), IL-1β (1:100, GTX74034, Gentex), rabbit anti-GSDMD (1:800, ab219800, Abcam), mouse anti-IBA-1 (1:300, GB12105, Servicebio), rabbit anti-Ly6G (1:500, GB11229, Servicebio), rabbit anti-CD68 (1:100, DF7518, Affinity). The sections were incubated with enzyme-conjugated goat anti-mouse IgG or goat anti-rabbit IgG polymer for 30 min. Finally, immunoreactivity was visualized using 3,3-diaminobenzidine (DAB, ZLI-9017, ZSGB-Bio) followed by restaining with hematoxylin. Images were captured by a light microscope (Leica, DM2500, Germany).

After transcardial perfusion with pre-cooling saline, brain tissue samples from the injured cortex were immediately harvested and homogenized with a freezing grinding mill (70 Hz, 60 s, LUKA, Guangzhou). Total protein was isolated and extracted according to the protocol of a protein extraction kit (BC3711, Solarbio, China). Protein concentrations were measured using a BCA Protein Assay Kit (PC0020, Solarbio). Subsequently, 30 ug of protein sample was separated by 6–15% SDS-PAGE and then transferred to PVDF membranes. The PVDF membranes were blocked in 5% skimmed milk or BSA for 2 h at room temperature and incubated at 4 ℃ overnight with primary antibodies as follows: rabbit anti-cGAS (#31659, 1:1000, CST), rabbit anti-STING (1:1000, CST, #13647), rabbit anti-NLRP3 (1:1000, #15101, CST), rabbit anti-GSDMD (1:1000, #93709, CST), rabbit-anti-cleaved C-terminal GSDMD (ab255603,1:1000, Abcam), rabbit anti-TXNIP (1:1000, ab188865, Abcam), rabbit anti-p-NF-kb p65(Ser536) (1:1000, AF2006, Affinity), rabbit anti-NF-kb p65 (1:1000, ab16502, Abcam), mouse anti-MCP-1 (1:1500, 66272-1-Ig, Proteintech), caspase-1 p20 (1:500, sc-398715, Santa Cruz), IL-1β (1:100, GTX74034, Gentex) and HRP-conjugated β-actin mouse monoclonal antibody (1:5000, HRP-66009, Proteintech).

After being rinsed with PBS, the PVDF membranes were incubated with secondary antibodies (Goat Anti-Rabbit IgG(H + L), 1:10,000, CW0103, CWBIO, China, or Goat Anti-Mouse IgG(H + L), 1:10,000, CW0102, CWBIO) for 1 h at room temperature. Finally, the blots were visualized employing a Millipore ECL kit and a gel imaging system (Tanon, 5200SF, Shanghai, China). Protein expression was quantified using ImageJ 1.5 software (National Institutes of Health, USA).

The cerebral cortex tissues of mice were harvested and homogenized, consistent with the method of western blot. The protein concentrations were detected using a BCA Protein Assay Kit (MM-9227B). The levels of cerebral oxidative damage biomarkers including 8-hydroxy-2-deoxyguanosine (8-OHdG), 3-nitrotyrosine (3-NT) and malondialdehyde (MDA) were determined utilizing assay kits for 8-OHdG (MM-0221M2), 3-NT (MM-45185M2) and MDA (MM-0897M2) according to the manufacturer’s instructions. The levels of 2′3′-cGAMP were detected by a 2′3′-cGAMP assay Kit (MM-44862M2) in conformity to the manufacturer’s instructions. All the kits are purchased from Jiangsu Meimian Industrial Co., Ltd. China.

Total RNA was isolated from homogenized cerebral cortex tissues using Trizol reagent (R1100, Solarbio) and quantified using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Bremen, Germany). Reverse transcription was performed with a primescript RT reagent kit (AG11728, Accurate Biology) according to the manufacturer’s instructions. Real-time RT-PCR was performed in a total volume of 10 μL containing 1 μL of cDNA, 0.6 μL of primers, and 8.4 μL of SYBR® Green Pro Taq HS Premix (AG11701, Accurate Biology) employing a 7500 Real-Time PCR thermocycler (Applied Biosystem). The program steps for amplification were 95 °C for 30 s, 40 cycles of 95 °C for 3 s, and 60 °C for 30 s. Relative mRNA expression was calculated using the 2-∆∆CT method and β-actin as an endogenous control. The primers were synthesized by Takara Biomedical Technology and the primer sequences are listed in Table 1.

Flow cytometry was used to assess neuroinflammation after CVST as previously reported [36]. On day 3 after CVST and perfusion with pre-cooling PBS, the cerebral cortex tissues of mice were gently homogenized and passed through 70-um nylon cell strainers (15–1070, BIOLOGIX, China) in PBS. After centrifugation, the cell pellets were resuspended in 5 mL of 30% Percoll (P8370, Solarbio, China) and centrifuged at 700×g for 10 min. Cell pellets were harvested from the bottom of the tubes and washed once with 5 mL 1% BSA solution to block non-specific staining. Fresh single-cell suspensions were stained in the dark with directly labeled antibodies as follows: fluorescein isothiocyanate (FITC) anti-mouse CD45 (103107), phycoerythrin (PE) anti-mouse Ly6G (127607), allophycocyanin (APC) anti-mouse F4/80 (123115), or PerCP-Cy5.5 anti-mouse CD11b (101227). All fluorescent-labeled antibodies were purchased from BioLegend (San Diego, CA, USA). Finally, flow cytometry was performed using a FACS Aria III (BD CANTO, USA) and data were analyzed with FlowJo software.

The data were expressed as the mean ± SEM and all statistical analyses were carried out utilizing SPSS 20.0 software (SPSS, IBM, Armonk, NY, USA). Comparison between three groups was determined by one-way analysis of variance (ANOVA) and followed by LSD test or Dunnett’s T3 test for the two groups’ comparison within the multiple groups. The t-test was used for comparison between the independent two groups. Correlation analysis was performed by Pearson test (if the data do not conform to a normal distribution, the logarithm is taken). Statistical significance was accepted when P < 0.05. Data were analyzed by investigators blinded to the identity of groups during the whole experiment.

Mice were randomized to six groups containing a sham group and five CVST subgroups (at 6 h, 1 d, 3 d, 7 d, and 14 d post-CVST). The expression of dsDNA in the damaged cortex at indicated times was detected by IF staining. In sham controls, dsDNA expression was finite, mainly presenting a circular nuclear contour, as noted unanimously in the literature [9]; whereas, after the induction of CVST at 24 h and 3 days, the intensities of dsDNA notably elevated with increased distribution both in the cytoplasm and nucleus. Especially, arresting disintegration of the nucleus was observed at 3 days after CVST (Fig. 1A). The white frames in Fig. 1B denoted the selected brain cortex regions for the whole experiment. In line with the elevation of dsDNA accumulation post-CVST, by IF and qPCR, we demonstrated that enhanced expressions of cGAS and STING in the impaired cortex compared with the controls on day 3 following CVST and predominantly located in the cytoplasm (Fig. 1C and D). Simultaneously, the WB results suggested that compared to the sham controls, the protein expressions of cGAS and STING strikingly increased from 24 h and reached the summit at 3 days, and then maintained at a high level at 7 days post-CVST. Despite falling distinctly at 14 days, they were still higher than the sham groups (Fig. 1E and F). Importantly, by fluorescent double-labeling, we found that cGAS was mainly localized in cells abundant in aberrant or ectopic dsDNA, potently indicating that accumulation of self-derived or enthetic dsDNA in the cytosol may trigger the activation of the cGAS–STING pathway post-CVST (Fig. 1G).

Immunofluorescent double staining showed that the cGAS and STING are mainly expressed in the activated microglia/macrophages (IBA-1⁺) but rarely in neurons (NeuN⁺) and astrocytes (GFAP⁺) at 3 d after CVST (Fig. 2A and B).

The results of WB revealed that compared to the sham groups, the prominent elevation of NLRP3 started at 24 h post-CVST and peaked at 3 days, and then gradually declined until 14 days (Fig. 3A and B). Concurrently, the expressions of GSDMD and GSDMD-C have analogous tendencies as those of NLRP3 (Fig. 3A, C and D). Notably, GSDMD C-terminal is generated simultaneously and in equal proportion with GSDMD-N terminal from the total GSDMD, which may be used as a marker of GSDMD-N production as the prior report [37, 38]. Besides, IF staining exhibited that GSDMD was predominantly localized in IBA-1⁺ microglia/macrophages in the cerebral cortex but hardly in neurons and astrocytes at 3 d post-CVST (Fig. 3Ea).

Mice were randomly divided into three groups including sham + vehicle group, CVST + vehicle group and CVST + RU.521 treatment group. RU.521 is a small-molecule inhibitor of cGAS with potent selectivity, which can effectively restrain its dsDNA-dependent enzyme activity [12, 32]. At 3 days after CVST ictus, TUNEL-positive cells were markedly increased in the damaged cortex compared with the sham groups, nasal delivery of RU.521 significantly decreased the number of apoptotic cells (Fig. 4A and C). Additionally, by double fluorescence labeling, we found that TUNEL-positive cells co-localized mostly with NeuN, a neuronal marker, whereas hardly with IBA-1, a microglia/macrophage symbol (Additional file 2: Fig. S2). Synchronously, FJC staining revealed that the degenerated neurons were robustly augmented post-CVST, and RU.521 markedly reduced the number of dying neurons (Fig. 4B and D). Furthermore, compared with the sham groups, the CVST mice developed moderate motor deficits lasting for at least 10 days, as evidenced by the rotarod test (Fig. 4E). In contrast, the RU.521-treated mice exhibited significantly improved motor function relative to the CVST + vehicle group on day 7 and day 10 after CVST.

CVST induced eminently oxidative damage, as evidenced by DHE staining, IF, ELISA, and WB assays, demonstrating increases in the levels of ROS, 8-OHdG (a marker of DNA oxidative damage), MDA (an indicator of lipid peroxidation), 3-nitrotyrosine (an index of protein tyrosine nitration), and TXNIP when compared with those of the sham + vehicle groups (Fig. 5A–E). Contrarily, RU.521 administration partly counteracted the aforementioned variations and mitigated oxidative stress after CVST, which was basically in accord with the prior record [39]. Moreover, IF staining demonstrated that augmented 8-OhdG was noticeably co-localized with cGAS (Fig. 5F), implying that oxidized DNA may be strikingly involved in the activation of cGAS following CVST.

We employed the ELISA to determine the concentration of 2′3′-cGAMP in the injured cerebral cortex and revealed that endogenous level of 2′3′-cGAMP, which is naturally synthesized by activated cGAS, was markedly augmented at 3 days after CVST compared to the sham group, confirming that cGAS is activated following CVST. In contrast, treatment with RU.521 significantly reduced the increased level of 2′3′-cGAMP (Fig. 6A). Besides, data from WB and/or PCR showed that the levels of STING and NF-κb phosphorylation along with downstream inflammatory factors (MCP-1, IL-6 and TNF-a), were highly induced in the brain cortex of vehicle-treated CVST animals, whereas RU.521 treatment partly reversed the induction of STING and relevant mediators (Fig. 6B, C and E). In concert with the above results, parallel alterations of STING and TNF-a in diverse groups were also proved by IF staining (Fig. 6D and F).

Next, we attempted to determine the effects of RU.521 on brain inflammatory cells after CVST via flow cytometry (Additional file 3: Fig. S3). We discovered that compared to the vehicle-treated controls, RU.521 delivery notably dampened neutrophils (CD11b⁺CD45^highLy6G⁺) infiltration and decreased the number of microglia (CD11b⁺CD45^int) in the brain cortex on day 3 after CVST (Fig. 6G and H). A decrease of CVST-induced infiltration of monocyte/macrophages (CD11b⁺CD45^highF4/80⁺) by RU.521 treatment was also noticed, while that was not statistically significant. To further verify the impact of RU.521 on the above inflammatory cells, we performed IHC analysis demonstrating that the number of neutrophils and microglia/macrophages was reduced by RU.521 treatment as indicated by Ly6G and Iba-1 immunoreactivity (Additional file 4: Fig. S4). These results indicate that RU.521 can suppress neuroinflammation via regulation of both resident microglia and infiltrating immune cells after CVST.

By WB and IHC (Fig. 7A–D), we observed that protein levels of NLRP3 inflammasome-related components (NLRP3, caspase-1 p20, pro-IL-1β, cleaved-IL-1β, cleaved-IL-1β/Pro-IL-1β), and pyroptosis executor GSDMD along with its cleaved C-terminal (GSDMD-C) were strikingly increased in the ischemic cortex after CVST in comparison with the sham-operated groups, whereas RU.521 delivery markedly repressed the aforementioned alterations. Additionally, a notable co-localization link existed between IL-1β and IBA-1 (Fig. 7E).

To probe into the potential role of STING in the pathological process post-CVST, the specific STING agonist 2′3'-cGAMP and STING siRNA were utilized, separately. WB results indicated that in comparison with the CVST + vehicle group, the expression of STING was significantly increased after 2′3′-cGAMP delivery. In contrast, intraventricular injection of STING siRNA (The efficiency was confirmed by WB as shown in Additional file 5: Fig. S5) efficiently dropped the expression of STING when compared with the CVST + Scramble siRNA group (Fig. 8A and B). Analogously, the phosphorylation level of NF-κb was moderately upregulated by 2′3′-cGAMP compared to the CVST + vehicle group, while remarkably repressed by STING siRNA application. Additionally, we also demonstrated that the protein levels of NLRP3, caspase-1 p20, pro-IL-1β, cleaved-IL-1β, cleaved-IL-1β/Pro-IL-1β, GSDMD, and GSDMD-C were further enhanced after 2′3′-cGAMP delivery in comparison with the CVST + vehicle groups, whereas injection of STING siRNA markedly reduced the levels of aforementioned proteins when compared with the CVST + scramble siRNA groups (Fig. 8C and D). What’s more, delivery of STING siRNA also remarkably restrained cell apoptosis and decreased the counts of dying neurons in the CVST + scramble groups, as demonstrated by TUNEL and FJC staining (Fig. 8E and F).

All experimental procedures were approved by the Sun Yat-sen University Ethics Committee and were performed in conformity to the guidelines of the National Institutes of Health on the care and use of animals.

Not applicable.

The authors report no conflicts of interest in this work.

The datasets of the current study are available from the corresponding author on reasonable request.