Adult male C57BL/6 J mice were purchased from Shanghai SLAC Laboratory Animal Co., Ltd. and raised to 12 weeks-old (25 to 30 g) for experiments. All mice used in this study were housed in a 12 h light/dark cycle room with water and food ad libitum. All animal procedures were performed in compliance with the guidelines and regulations approved by the Institutional Animal Care and Use Committee of Fudan University (ethical approval number: B2022-031R).

The Neuro-2a (N2a) cell line, obtained from the Cell Bank of Shanghai Institute of Biochemistry and Cell Biology (SIBCB), Chinese Academy of Sciences (CAS), was cultured in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin. The cells were maintained in a 37°C incubator with 5% CO₂.

Transient transfection was carried out using 4 μg vectors expressing U6-gRNA-EF1α-CasRx-eGFP according to the standard procedure with Lipofectamine 3,000 (Thermo Fisher Scientific). For the screening of Ripk1 gRNAs (Supplementary Table S1), the experimental plasmids included U6-Ripk1 (gRNA)-EF1α-CasRx-eGFP along with CAG-Ripk1-mCherry, while the control plasmid consisted of CAG-Ripk1-mCherry and U6-EF1α-CasRx-eGFP (no gRNA). Similarly, for the screening of Nsf gRNAs (Supplementary Table S1), the experimental plasmid was U6-Nsf (gRNA)-EF1α-CasRx-eGFP, and the control plasmid was U6-EF1α-CasRx-eGFP (no gRNA). N2a cells were seeded in 6-well plates and transfected with 4 mg vectors expressing U6-gRNA-EF1α-CasRx-eGFP using Lipofectamine 3,000. Two days post-transfection, approximately 50,000 cells per sample, positive for both eGFP and mCherry (for Ripk1 gRNAs) or positive for eGFP (for Nsf gRNAs), were collected by fluorescence activating cell sorting (FACS, MoFlo XDP, Beckman Coulter Ltd.) and lysed for qPCR analysis. RNA was extracted using Trizol (Ambion), and cDNA was synthesized using a reverse transcription kit (HiScript Q RT SuperMix for qPCR, Vazyme, Biotech). The qPCR Primers of Ripk1 and Nsf were shown in Supplementary Table S2. For RNA-seq, N2a cells were cultured in a 15 cm dish, and transient transfection was performed with 70 μg plasmids. The experimental plasmids included U6-Ripk1 (gRNA)-Nsf (gRNA)-EF1α-CasRx-eGFP and CAG-Ripk1-mCherry, and the control plasmid consisted of CAG-Ripk1-mCherry and U6-EF1α-CasRx-eGFP (no gRNA). Approximately 300,000 eGFP-positive and mCherry-positive (top 20%) N2a cells were collected by FACS. Subsequently, RNA was extracted and then converted to cDNA, which was used for transcriptome-wide RNA-seq analysis. High-throughput mRNA sequencing was performed using the Illumina Genome Analyzer, and adapter removal was carried out using Trimmomatic (v0.36) during the sequencing process. The hisat2 (v2.0.0) tool was employed to map qualified reads to the mouse reference genome (mm10) with default parameters. The expression levels of all mapped genes were estimated using stringtie (v2.0), and the gene expression abundances were indicated by FPKM (fragments per kilobase of transcript per million fragments mapped).

Due to the packaging capacity limitations of AAV, we reconstructed U6-Ripk1 (gRNA)-Nsf (gRNA)-EF1α-CasRx-eGFP by omitting eGFP. A gRNA for Ripk1 and a gRNA for Nsf were cloned into an adeno-associated viral backbone vector, where two gRNAs were driven by a U6 promoter and CasRx was driven by a EF1α promoter, forming the target plasmid EF1α-CasRx-Ripk1-Nsf. And the adeno-associated viral backbone vector itself can serve as a control for the target plasmid, which named as EF1α-CasRx. The adeno-associated viral backbone vector and the plasmid of EF1α-mCherry were a gift from Canbing Feng. The plasmids EF1α-CasRx-Ripk1-Nsf, EF1α-CasRx and EF1α-mCherry were custom-packaged using AAV9 by OBiO Technology (Shanghai) Corp., Ltd.

The mice were anesthetized using 3% isoflurane for induction and 1.5% for maintenance. Subsequently, they were secured in a stereotactic frame. AAVs containing either AAV-EF1α-CasRx-Ripk1-Nsf plus AAV-EF1α-mCherry or AAV-EF1α-CasRx plus AAV-EF1α-mCherry was injected into striatum (AP = 0 mm, ML = +2.2 mm, DV = −2.8 mm) and S1BF (AP = −0.5 mm, ML = +3.2 mm, DV = −1.3 mm) using a micromanipulator at a slow rate (approximately 6 min per microliter). Mice were injected in the striatum and S1BF with high-titer AAVs (>1 × 10¹³ vg/mL, 1 μL per site). The volume ratio between AAV-EF1α-mCherry and AAV-EF1α-CasRx-Ripk1-Nsf or AAV-EF1α-CasRx was 1:9.

The mice were anaesthetized using 3% isoflurane for induction and 1.5% for maintenance. Body temperature was carefully maintained at 37°C using a thermostatically controlled heating pad. Meloxicam (6 mg/kg) was administered subcutaneously 30 min before surgery for analgesia. A midline cervical incision was made, and under an operating microscope, the left common carotid artery (CCA), external artery (ECA), and internal carotid artery (ICA) were exposed. A 6-0 silicone-coated monofilament (602356PK5Re, Doccol Ltd.) was inserted into the CCA and advanced into the ICA approximately 9–11 mm from the common carotid bifurcation. The monofilament was kept in place for 60 min. After removing the monofilament and providing wound care, the animals were returned to their cages and closely monitored for 24 h. The same procedure was performed for sham-operated animals, except that the advancement distance of the monofilament was reduced to 5 mm from the common carotid bifurcation.

Coronal slices (1 mm thick) of murine brain were immersed in a 2% solution of TTC (298-96-4, Sangon Biotech Ltd., Shanghai, China) at 37°C for 10 min. At the 5th minute, the slices were flipped over. The stroke areas were quantified using ImageJ software. To compensate for cerebral swelling (oedema) and overestimation of the infarct volume, we used the following formula: percentage of corrected infarct volume = (contralateral hemisphere volume − (ipsilateral hemisphere volume − infarct volume))/total brain volume. Similarly, the percentage of corrected oedema volume = (ipsilateral hemisphere volume − contralateral hemisphere volume)/total brain volume.

Snap-frozen brain tissue was homogenized with ice-cold RIPA lysis buffer (P0013K, Beyotime Ltd.), and the supernatant was collected. 30 μg of total protein lysate was used for sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and transferred onto polyvinylidene fluoride (PVDF) membranes (IPVH00010, Merk Millipore). The membranes were incubated overnight at 4°C with the following primary antibodies: rabbit monoclonal anti-RIPK1 (1:1000, 3,493, Cell Signaling Technology), rabbit monoclonal anti-RIPK3 (1:1000, 10,188, Cell Signaling Technology), rabbit monoclonal anti-MLKL (1:1000, 37,705, Cell Signaling Technology), and rabbit monoclonal anti-β-actin (1:30000, AC308, Abclonal). Primary antibody dilution buffer (P0023A-500 mL, Beyotime Ltd.) and secondary antibody dilution buffer (P0023D-500 mL, Beyotime Ltd.) were used in this study. After incubation with the corresponding peroxidase-conjugated secondary antibody (anti-rabbit IgG, 1:5000, M21002, Abmart), the signal was visualized using the Omni-ECL^™ Enhanced Pico Light Chemiluminescence Kit (Epizyme Ltd., Shanghai, China) and quantified using Image J software (V1.48, National Institutes of Health, United States). β-actin was used for signal normalization primary antibody dilution buffer (P0023A-500 mL, Beyotime Ltd.) and secondary antibody dilution buffer (P0023D-500 mL, Beyotime Ltd.)

Snap-frozen brain tissue was homogenized with ice-cold immunoprecipitants lysis buffer (G2038-100ML, Servicebio Ltd.), and the supernatant was incubated on ice for 40 min, followed by centrifugation at 13,000 × g for 15 min. The supernatant was then incubated with 4 μg of the antibody at 4°C overnight. Protein G agarose beads (20 μL, Thermo Fisher Scientific) were added to the sample and incubated for 2 h at 4°C. The immunoprecipitants were washed three times with the lysis buffer, suspended with 2 × loading buffer, boiled, and run on SDS-PAGE. The proteins were transferred to PVDF membranes for immunoblotting. Additionally, aliquots of the original lysates were subjected to SDS-PAGE in parallel for immunoblotting to determine the input amount. The primary antibodies used were rabbit monoclonal anti-ASIC1a (1:100, ab284406, Abcam), mouse monoclonal anti-RIPK1 (1:1000, sc-133102, Santa Cruz), mouse monoclonal anti-NSF (1:1000, 515,043, Santa Cruz), and β-actin (1:5000, AC308, Abclonal). This was followed by incubation with horse radish peroxidase-conjugated secondary antibodies (Goat anti-rabbit IgG, 1:5000, GB23303; Goat anti-mouse IgG, 1:5000, GB23301; Donkey anti-goat IgG, 1:5000, GB23404, Servicebio, Shanghai, China) and visualization with enhanced chemiluminescence on an ImageQuant LAS 4000 mini digital imaging system (GE Healthcare Life Sciences). The primary and secondary antibodies were diluted in PBST with 1% BSA. The band intensities of the western blots were quantified using ImageJ with background subtraction.

The brains were perfused and fixed with 4% paraformaldehyde (PFA) overnight, followed by storage in 30% sucrose for 24 h. After embedding and freezing, the brains were sectioned with the thickness of 30 mm for immunofluorescence staining. The brain sections were thoroughly rinsed with 0.1 M phosphate-buffered saline (PBS). The following primary antibodies were used in this study: Guinea Pig anti-NeuN antibody (1:500, ABN90, Millipore), Mouse anti-Flag (1:1500, F3165, Sigma), Rabbit anti-RIPK1 (1:100, A7414, Abclonal), Rabbit anti-RIPK3 (1:100, A5431, Abclonal), Rabbit anti-MLKL (1:100, A5579, Abclonal), and Rabbit anti-NSF (1:100, A0926, Abclonal). The following secondary antibodies were used: Alexa Fluor^® 488 AffiniPure Donkey Anti-Rabbit IgG (H + L) (1:500, 711-545-152, Jackson ImmunoResearch), Cy5-AffiniPure Donkey Anti-Guinea Pig IgG (H + L) (1:500, 706-175-148, Jackson ImmunoResearch), and Goat anti-mouse IgG Alexa Fluor 594 (1:500, Abcam Cat# 150116). After antibody incubation, the slices were washed and covered with mountant (Life Technology). The images were visualized using an Olympus FV3000 microscope and image acquisition software is OLYMPUS FV31S-SW.

An experimenter blinded to the group allocation evaluated the neurological deficits after tMCAO. The neurological status was assessed at 2, 24, and 48 h after tMCAO using a five-point scale based on the Bederson score (Bieber et al., 2019). The scores were assigned as follows: score 0, no observable deficits; score 1, forelimb flexion; score 2, forelimb flexion and decreased resistance to lateral push; score 3, circling; score 4, circling and spinning around the cranial-caudal axis; score 5, no spontaneous movement. Neurological performance was also evaluated using the RotaRod (Cat. No. 47600, UGO BASILE S.R.L.) test and Adhesive removal test. In the RotaRod test, animals were placed on a rotating rod that gradually increased in speed (4–44 revolutions/min), and the latency to fall was recorded. The average staying time on the rod in three trials conducted one day before surgery was considered as the baseline value. Following surgery, the mean value of four trials was used as the time on the rod for the tested day. In the adhesive removal test, a 3 mm × 3 mm piece of adhesive tape was placed on the palmar surface of the contralateral forepaw. The time to contact and completely remove the tape from the forepaw was recorded, respectively. Three trials were conducted daily for each animal, starting from 3 days before surgery and continuing until specific time points after surgery. The mean value of three trials conducted one day before surgery was considered as the baseline value.

The figures and figure legends provided information about the number of mice used in each experiment. All values were analyzed using SPSS 24.0 software (SPSS Inc., Chicago, IL, United States) and were presented as mean ± SEM, except for Figure 1F, which was presented as median (interquartile range). The statistical significance was determined using unpaired two-tailed Student’s t test (p < 0.05), except for Figure 1F, where the Mann–Whitney U test was employed. Randomization was applied to all experiments, and sample sizes were not predetermined using statistical methods. The experimenters remained blinded to group allocation during data Research Topic to avoid experimenter bias.

To explore the specific pathophysiological relevance of RIPK1 and NSF binding to ASIC1a in the context of ischemic brain injury, we assessed the expression levels of RIPK1 and NSF in the sham-operation or the peri-infarct region of ipsilateral brain, 2 h following sham-operation/transient middle cerebral artery occlusion. Co-immunoprecipitation was carried out to isolate RIPK1 and NSF bound to ASIC1a. The result demonstrated an enhanced interaction between RIPK1 and ASIC1a (4.24 ± 1.29 vs. 1.00 ± 0.16, p = 0.003, Figure 2B), as well as between NSF and ASIC1a (4.33 ± 1.20 vs. 1.00 ± 0.31, p = 0.002, Figure 2B), in the peri-infarct brain tissue compared with the sham-operated brain tissue. Furthermore, the total levels of RIPK1 increased after brain ischemic injury (1.98 ± 0.31 vs. 1.00 ± 0.38, p = 0.007, Figure 2C), while NSF levels remained unchanged (1.02 ± 0.09 vs. 1.00 ± 0.08, p = 0.76, Figure 2C). These findings suggest an enhanced interaction of RIPK1 and NSF with ASIC1a during cerebral ischemia/reperfusion.

To evaluate the efficiency of CasRx-mediated knockdown of Ripk1 mRNA and Nsf mRNA, we initially screened fourteen gRNAs targeting Ripk1 and fifteen gRNAs targeting Nsf to assess their editing efficiency in N2a cells (Figures 3A,B). The gRNAs were designed to target exons, since introns are spliced out during mRNA processing. We found that the co-transfection of a vector containing the CasRx gene, a vector containing the Ripk1 gene, and a vector containing gRNA 8 targeting Ripk1 exon V, resulted in a significant reduction of Ripk1 mRNA by 93.54% ± 0.09% (n = 3 repeats) in N2a cells, as determined by qPCR (Figure 3C). Among the gRNAs targeting Nsf exons, gRNA 9, which targets Nsf exon IX exhibited the highest knockdown efficiency, resulting in a reduction of Nsf mRNA by 97.12% ± 0.48% (n = 3 repeats) in N2a cells (Figure 3D). Consequently, gRNA 8 for Ripk1 and the gRNA 9 for Nsf were co-loaded into a single plasmid along with CasRx gene, which was subsequently used in in vivo experiments. To evaluate the specificity of the CasRx-Ripk1-Nsf plasmid in knocking down Ripk1 mRNA and Nsf mRNA, we co-transfected a vector containing the Ripk1 gene and the CasRx-Ripk1-Nsf plasmid into N2a cells. Subsequently, we performed a transcriptome analysis to examine the expression levels of all detected genes in the RNA sequencing (RNA-seq) libraries. The analysis revealed a specific downregulation of Ripk1 and Nsf, while the transcriptional levels of all other detected genes in the RNA-seq libraries remained unchanged 48 h post co-transfection (Figures 3E,F).

Firstly, we examined the effect of Ripk1 and Nsf knockdown in vivo after tMCAO. The results showed a decreased association between RIPK1 and ASIC1a (0.21 ± 0.11 vs. 1.00 ± 0.06, p < 0.0001, n = 4, Supplementary Figures S1A,B), as well as between NSF and ASIC1a (0.29 ± 0.13 vs. 1.00 ± 0.20, p = 0.0019, n = 4, Supplementary Figures S1A,B), in mice treated with CasRx-Ripk1-Nsf compared with CasRx group after tMCAO. Furthermore, the total expression levels of RIPK1 and NSF also decreased in CasRx-Ripk1-Nsf group compared with CasRx group (RIPK1: 0.29 ± 0.11 vs. 1.00 ± 0.25, p = 0.0019, n = 4, Supplementary Figures S1A,C; NSF: 0.73 ± 0.09 vs. 1.00 ± 0.11, p = 0.0087, n = 4, Supplementary Figures S1A,C). Moreover, the colocalization of RIPK1 and Flag (Supplementary Figure S1D), as well as NSF and Flag (Supplementary Figure S1E), also shown a similar tendency after tMCAO in CasRx-Ripk1-Nsf group compared with CasRx group, evaluated by immunofluorescence staining. Given the high knockdown efficiency of Ripk1 and Nsf in vivo after tMCAO, we subsequently determined the therapeutic potential of Ripk1 and Nsf knockdown on stroke outcome through injecting wild-type mice with AAV-EF1α-CasRx-Ripk1-Nsf (containing gRNA 8 for Ripk1 and gRNA 9 for Nsf) into the S1BF and striatum 14 days prior to tMCAO/sham operation (Figure 1A), along with AAV-EF1α-mCherry, which served as a fluorescent marker indicating the infected area of the virus (Figure 1B). As a control, we used AAV- EF1α-CasRx that does not contain gRNA of Ripk1 and Nsf. Stroke size, edema and neurological deficits were assessed using the Bederson score, RotaRod test, and Adhesive removal test to determine the impact of Ripk1 and Nsf knockdown after tMCAO. Following 48 h of reperfusion, mice treated with AAV-EF1α-CasRx-Ripk1-Nsf exhibited a ~ 14% reduction in corrected infarct volume (21.17% ± 2.87% vs. 6.96% ± 0.81%, p < 0.0001, n = 6, Figures 1C,D), as determined by TTC staining. Moreover, knockdown of Ripk1 and Nsf also decreased the corrected edema volume (5.25% ± 1.35% vs. 0.91% ± 0.68%, p < 0.0001, n = 6, Figures 1C,E). The reduction in infarct and edema volume in the CasRx-Ripk1-Nsf group following tMCAO suggests that the knockdown of Ripk1 and Nsf provides a certain level of protection against ischemia brain injury. We observed a tendency towards improved neurological performance, as evaluated by the Bederson score, in the CasRx-Ripk1-Nsf group at 2 h, 24 h, and 48 h after tMCAO, compared with the vehicle-treated group (data presented as median [interquartile range]; 2 h: 4.00 (1.00) vs. 2.00 (0.25), p = 0.004; 24 h: 3.00 (2.00) vs. 1.00 (1.00), p = 0.009; 48 h: 2.00 (1.00) vs. 1.00 (0), p = 0.002; n = 6, Figure 1F). Additionally, the results of RotaRod test demonstrated a significantly prolonged latency to fall from the rotating rod in the CasRx-Ripk1-Nsf group at 24 h and 48 h after tMCAO, compared with the vehicle group (24 h: 65.88 ± 19.34 vs. 145.38 ± 25.68 s, p < 0.0001; 48 h: 104.04 ± 30.25 vs. 193.50 ± 30.55 s, p < 0.0001, n = 6, Figure 1G). Similarly, mice in CasRx-Ripk1-Nsf group exhibited faster response times to touch and to remove the adhesive tape at 24 h and 48 h after tMCAO, compared with the vehicle group(time to touch at 24 h: 51.39 ± 7.94 vs. 32.67 ± 3.39 s, p < 0.0001; time to touch at 48 h: 63.89 ± 5.87 vs. 24.06 ± 5.83 s, p < 0.0001, n = 6, Figure 1H; time to remove at 24 h: 101.5 ± 12.90 vs. 57.07 ± 6.86 s, p < 0.0001; time to remove at 48 h: 102.45 ± 16.01 vs. 56.00 ± 7.79 s, p < 0.0001, n = 6, Figure 1I). These findings suggest that mice in the CasRx-Ripk1-Nsf group exhibited significantly less sensorimotor deficits and better neurological recovery than vehicle-treated group. Overall, these results indicate that knockdown of Ripk1 and Nsf improves histological outcome and neurological recovery after tMCAO in mice.

To elucidate the process by which NSF and RIPK1 sequentially initiate necroptosis in neurons, we performed immunofluorescence staining to determine the colocalization of RIPK1 and NeuN, RIPK3 and NeuN, and MLKL and NeuN in cerebral frozen sections. Additionally, Western blot analysis was performed to assess the relative expression levels of RIPK1, RIPK3, and MLKL in the ipsilateral brain peri-infarct tissue and sham-operated brain tissue. Mice treated with CasRx-Ripk1-Nsf exhibited a lower degree of colocalization between RIPK1 and NeuN in NeuN⁺ cells 48 h after tMCAO, compared with vehicle group (30.70% ± 6.57% vs. 11.93% ± 2.67%, p < 0.0001, n = 6, Figures 4A,E). Furthermore, the ratio of RIPK3⁺ NeuN⁺/NeuN⁺ cells and MLKL⁺ NeuN⁺/NeuN⁺ cells was decreased in mice treated with CasRx-Ripk1-Nsf compared with vehicle group, 48 h after tMCAO (ratio of RIPK3⁺ NeuN⁺/NeuN⁺ cells: 70.04% ± 10.08% vs. 19.43% ± 4.30%, p < 0.0001, n = 6, Figures 4B,F; ratio of MLKL⁺ NeuN⁺/NeuN⁺ cells: 30.91% ± 5.39% vs. 16.81% ± 6.75%, p = 0.003, n = 6, Figures 4C,G). Consistent with the immunofluorescence staining results, Western Blot analysis also demonstrated a decrease in the expression levels of RIPK1, RIPK3, and MLKL in mice treated with CasRx-Ripk1-Nsf compared with vehicle group, 48 h after tMCAO (RIPK1: 1.97 ± 0.21 vs. 0.12 ± 0.30, p = 0.004, n = 4; RIPK3: 2.16 ± 0.19 vs. 1.22 ± 0.18, p = 0.004, n = 4; MLKL: 3.11 ± 0.59 vs. 1.31 ± 0.14, p < 0.001, n = 4, Figures 4D,H–J). Additionally, in the two sham-operated group, there were no statistically significant differences in the colocalization of RIPK1 and NeuN or the expression levels of RIPK1 in mice treated with CasRx-Ripk1-Nsf, compared with vehicle group (ratio of RIPK1⁺ NeuN⁺/NeuN⁺ cells: 4.12% ± 3.75% vs. 1.22% ± 1.19%, p = 0.12, n = 6; 1.00 ± 0.20 vs. 0.70 ± 0.16, p = 0.054, n = 4). This suggests that under physiological conditions, RIPK1 is barely expressed. The results of both immunofluorescence staining and Western blot analysis indicate that knockdown of Ripk1 and Nsf mediated by CasRx ameliorates necroptosis in neurons through the RIPK1/RIPK3/MLKL signaling pathway.