Adult male Sprague-Dawley rats (300–340 g) which were obtained from SLAC Laboratory Co., Ltd. (Shanghai, China) were used in this study. A total of 136 rats were used for this study. The Sprague-Dawley rats were raised in plastic cages with controlled temperature and humidity and a 12-hlight–dark cycle. All animal experiments were in accordance with the Guidelines for the Care and Use of Laboratory Animals of the National Institutes of Health and were approved by the Institutional Animal Care and Use Committee of Zhejiang University.

The in vivo study was divided into 2 parts. In the first part, rats were subjected to SAH with an endovascular perforation technique. Brains were harvested at 6, 12, 24, 48, and 72 h after SAH (n = 4 each group) for western blots and brain histology (n = 4 at 24 h after SAH). Brains of the sham group were harvested at 24 h after surgical procedures. In the second part, SAH groups were subjected to SAH and treated with vehicle or probenecid, and the sham group was subjected to a procedure similar to the SAH groups but without perforation and treatement with vehicle. All endpoints in this part were investigated 24 h after SAH based on the result of the first part. The experimental group design is shown in Figure 1A.

The rat SAH model was performed by the endovascular perforation technique as previously described (19). Rats were anesthetized with pentobarbital (40 mg/kg, intraperitoneal injection) to expose the left carotid artery and its branches. The external carotid artery was transected distally, and a blunt 4-0 nylon monofilament suture was inserted into the internal carotid artery, which was advanced through until resistance was felt and occurred at the junction of the anterior cerebral artery and the middle cerebral artery. The suture was further advanced ~3–5 mm to perforate the artery and create SAH with a stay of 10 s. The sham group underwent a similar procedure without perforation.

The SAH grades were quantified in accordance with the published grading system of Zhang et al. (20). The grade scores were based on the volume of a clot in six parts of the basal cistern. Each segment was scored 0–3, and the final scores were obtained by adding six partial scores.

The mouse hippocampal neuronal cell line HT22 cells were cultured (37°C, 5% CO₂) in DMEM/F12 (1:1) medium with 15% fetal bovine serum (FBS) and 100 U/ml penicillin. These cells were inoculated in 5 cm² × 5 cm²-cell culture flask at a density range of 4 × 10⁶/well−5 × 10⁶/well. Following SAH, the erythrocytes lyse and release their hemoglobin, which degrades to hemin. The toxicity of hemin occurs after it has been taken up into cells and thereby causes various harmful responses. We stimulated HT22 cells by administrating 75 μm hemin (Sigma-Aldrich, MO, USA, catalog number H9039) for 12 h to induce the SAH model in vitro. Meanwhile, cells were treated with probenecid (1 mM probenecid) or vehicles (Figure 1A).

In vivo: 50 mg probenecid (P8761, Sigma, USA) was dissolved in 2 mM NaOH PH was titrate to 7.0 with 0.2M HCl. Then, water was added to a final volume of 50 ml (1 mg/ml probenecid). Rats were intraperitoneally injected with probenecid at a dose of 1 mg/kg using two-injection protocol as described previously (21, 22). The injection of probenecid or vehicle was performed before SAH and again 2 h after SAH, for a total of two doses.

In vitro: HT22 cells were cultured in DMEM/F12 medium for 24 h. Cells were stimulated by hemin, hemin + probenecid, or probenecid for an additional 12 h. The control cells received the same volumes of vehicles.

The neurological deficits were evaluated 24 h after SAH according to the scoring system of Garcia et al. (23). The scoring system consisted of the following six tests: symmetry in the movement of four limbs, spontaneous activity, climbing, body proprioception, forepaw outstretching, and response to vibrissae touch. Each test was scored 0–3 or 1–3, and the total score ranged from 3 to 18. A lower score represents more serious neurological deficits. All the tests were assessed in a blind manner in a random sequence.

Rats were sacrificed 24 h after SAH. The brain was removed and immediately weighed to obtain the wet weight (WW). The samples were dried at 105°C for 24 h and weighed again to obtain the dry weight (DW). The brain water content was calculated as [(WW–DW)/WW] × 100% (24).

A reactive oxygen species (ROS) assay was performed as previously described with the ROS assay kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) following the manufacturer's instruction (25). The ROS levels in the brain tissue were expressed as fluorescence intensity/mg protein.

Ipsilateral basal cortical samples facing blood clots were extracted. Western blotting was performed as described previously (25). Cortical samples were homogenized and centrifuged (1,000 × g, 10 min, 4°C). The supernatant was further centrifuged, and then, the protein concentration was determined using the DC protein assay kit (Bio-Rad, Hercules, CA, USA). An equal amount of protein (50 μg) was suspended in loading buffer, denatured at 95°C for 5 min, and loaded on an SDS-PAGE gel. After being electrophoresed and transferred into polyvinylidene fluoride membranes, the membrane was blocked with nonfat dry milk buffer for 2 h and then incubated overnight at 4°C with the primary antibody for AIM2 (ab93015, 1:1,000, Abcam), pannexin-1 (ab124131, 1:1,000, Abcam), ASC (ab155970, 1:1,000, Abcam), caspase-1 (ab1872, 1:1,000, Abcam), P2X7R (1: 1,000, APR-004, Alomone, Jerusalem), IL-1β (SC-23460, 1:500; Santa Cruz), IL-18 (ab71495, 1:1,000; Abcam), and GAPDH (1:10,000, Fitzgerald, 10R-G109A). The membranes were incubated with horseradish peroxidase-conjugated secondary antibodies for 1 h at room temperature. The protein band densities were detected by x-ray film and quantified by ImageJ software (National Institutes of Health, Bethesda, MD, USA).

Propidium iodide (PI, 25535-16-4, sigma) was dissolved in 0.9% NaCl, and 10 mg/kg PI was administered intraperitoneally at 1 h prior to sacrifice. Brain coronal slices were obtained according to our previous protocol (4). Immunofluorescence staining was performed as previously described. The primary antibodies were AIM2 (ab93015, 1:1,000, Abcam), pannexin-1 (ab124131, 1:1,000, Abcam), and caspase-1 (ab1872, 1:1,000, Abcam). The secondary antibodies were rhodamine-conjugated antibody (1:200, Jackson Immuno Research) and fluorescein isothiocyanate-labeled antibody (1:200, Jackson Immuno Research). The sections were rinsed and stained with DAPI (1 μg/mL, Roche Inc, Basel, Switzerland) and then rinsed again and mounted with glycerol. The labeling was analyzed using a fluorescence microscope (Olympus).

The HT22 cells were inoculated in 96-well plates at a density range of 1 × 10⁴/well−3 × 10⁴/well, and then, these cells were used for immunofluorescence staining and measuring mitochondrial membrane potential. The cultures were washed two times with PBS and fixed with 4% paraformaldehyde in PBS for 15 min and then washed two times more with PBS at room temperature. The staining protocol and the primary and second antibodies were the same as described above. The mitochondrial membrane potential was measured by a JC-1 kit (Beyotime, Shanghai, China) following the manufacturer's instructions. HT22 cells were rinsed with PBS and incubated with JC-1 staining solution at 37°C for 20 min. Then, the inverted fluorescence microscope (Olympus, Tokyo, Japan) was used to capture the pictures and calculate the ratio of the red–green fluorescence.

In brief, the HT22 cells were inoculated in 96-well plates at a density range of 1 × 10⁴/well−3 × 10⁴/well. Culture medium (20 μl) was taken after 12 h of treatment, and cell death was quantified by lactate dehydrogenase release assay (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) following the manufacturer's instruction. The LDH release was determined by measuring the 450 nm absorbance using a microplate reader.

Continuous data were presented as mean ± standard error of the mean (SEM) or median (interquartile range) based on the normality and homogeneity of variance. For the data meeting normal distribution and homogeneity of variance, one-way analysis of variance (ANOVA) followed by a Tukey's multiple comparison test was used to evaluate significant differences among groups. For the non-normal distribution and unequal variance parameters, the Kruskal–Wallis test was used to compare the differences among the groups. The Chi-square test was used for mortality comparison. The p-value <0.05 indicated statistical significance. GraphPad Prism (version 9.0) and SPSS software (version 23.0) were used for statistical analyses.

Experimental SAH surgeries were performed in rats with the endovascular perforation technique. The pannexin-1 channel is the principal conduits of ATP release from dying cells in the brain, which release cytoplasmic contents that include dsDNA. AIM2 is a member of the HIN-200 family of proteins and was activated to form an inflammasome in the neuron by dsDNA. Western blot results showed that the pannexin-1 and AIM2 protein contents gradually increased with a marked accumulation at 24 h after SAH and still stay at a high level at 72 h post-SAH when compared to the sham group (p < 0.05, Figure 1B). PI staining was used for the detection of necrotic cells. Double immunostaining demonstrated that pannexin-1-positive and AIM2-positive cells were mainly expressed on PI-positive cells after SAH induction (Figure 1C).

The SAH grade of the sham group was 0, and the scores were not significantly different between SAH groups (Figure 2A). Probenecid is a pharmacological inhibitor of the pannexin-1 channel. Western blot showed that the treatment of probenecid reversed the increase of pannexin-1 protein expression after SAH (p < 0.05, Figure 2B). Brain water content was measured by wet–dry method and analyzed 24 h after SAH. The whole-brain water content of the SAH group was significantly higher than that of the sham group. The rats treated with probenecid showed reduced brain edema compared with the SAH group (p < 0.05, Figure 2C). The neurological score of the SAH group was significantly lower than that of the sham group. Probenecid treatment improved the neurological scores significantly compared with the SAH group (n = 21, p < 0.05, Figure 2D). When the mortality of SAH rats was estimated, we found that mortality was slightly lower in the SAH + pro group (20.59%) than in the SAH group (22.86%) with no statistical difference (Figure 1A). These results indicated that probenecid treatment provided a neuroprotective effect on EBI after SAH.

To assess the AIM2 inflammasome activation after SAH, we measured the expression levels of AIM2, ASC, and caspase-1. Compared with the sham group, the expression levels of AIM2, ASC, and caspase-1 were significantly increased in SAH group. The pannexin-1 inhibitor, probenecid, markedly reduced the protein expressions of AIM2, ASC, and caspase-1(p < 0.05, Figures 3A,B,D,E). Double immunostaining showed that caspase-1-positive cells were mainly expressed in PI-positive cells after SAH (Figure 3F). Pannexin-1 channel is associated with the ATP release of dying cells for large pore formation on the cell membrane, and this function is closely tied to that of ATP-activated P2X7 purinergic receptor (P2X7R) (12). Western blot analysis showed that the P2X7R level was significantly elevated 24 h after SAH, and probenecid treatment markedly reduced the expression of P2X7R compared with vehicle-treated rats (p < 0.05, Figures 3A,C), which might be due to the inhibition of the pannexin-1 channel.

The activation of inflammasome leads to the processing and secretion of the pro-inflammatory cytokines IL-1β and IL-18. To determine the expressions of IL-1β and IL-18, western blot was performed. The expressions of IL-1β and IL-18 were significantly increased 24 h after SAH, whereas the probenecid treatment markedly decreased the levels of IL-1β and IL-18 in SAH rats (p < 0.05, Figures 4A–C). The pro-inflammatory cytokines increased mitochondrial-generated ROS production, which indicated the cell damage after SAH. ROS levels were significantly higher in the SAH group than that in the sham group, and probenecid treatment markedly decreased the ROS content compared with SAH rats (p < 0.05, Figure 4D). These results suggested that treatment with probenecid inhibited the AIM2 inflammasome activation and resulted in less pro-inflammatory cytokines and ROS generation after SAH induction.

AIM2 inflammasome was primarily found in neurons. To investigate whether AIM2-mediated neuronal cell injury after SAH, we treated HT22 cells with hemin to mimic SAH injury for neurons. Immunostaining showed that AIM2-positive and ASC-positive cells were significantly elevated after hemin stimulation. The treatment of probenecid reduced the proportion of positive cells (p < 0.05, Figure 5). JC-1 staining was used to detect mitochondrial membrane potential, which was associated with ROS generation and cell damage. Normal membrane potential indicated red fluorescence intensity in the control group. Treatment of 75 μm hemin increased green fluorescence intensity, which represented the decline of membrane potential in HT22 cells. Co-incubation with probenecid attenuated hemin-induced collapse of membrane potential (Figure 6A). Quantification of JC-1 fluorescence intensity (red–green florescent area) is shown in Figure 6B. Moreover, the effect of probenecid treatment on protecting HT22 cells from hemin-induced cell injury was evaluated by LDH assay. Media LDH level was significantly increased after treating HT22 cells with 75 μm of hemin compared with the control group (p < 0.05, Figure 6C). This was attenuated by the treatment of probenecid (p < 0.05, Figure 6C). These results indicated that probenecid treatment alleviated hemin-induced AIM2 inflammasome activation and reduced neuronal cell death in vitro.