Sprague-Dawley rats (male, weight 270–320 g) were acquired from the Animal Center of the First Affiliated Hospital of Harbin Medical University and raised in a room with appropriate humidity and a controlled temperature of 25-28°C. All rats were allowed to eat and drink freely prior to modeling. Intravascular perforation was performed as previously described (22). All animal experiments were carried out in accordance with the guidelines of the National Institutes of Health (NIH) and were approved by the Institutional Animal Care and Use Committee of the First Affiliated Hospital of Harbin Medical University (NO:2020047).

Two researchers independently graded the severity of SAH, as described previously (23). In brief, SD rats were sacrificed after the induction of SAH. Next, wo measured the volume of blood clots in different areas of the basal surface of the brain, as follows: no blood clots = 0; a small number of blood clots = 1; a moderate number of blood clots discernible within the arteries = 2 points, and blood clots covering all of the arteries = 3 points. Rats with a SAH grading score < 7 were excluded from subsequent experiments.

Am80 (Tocris, Ellisville, USA) was dissolved in carboxymethyl cellulose solution and administered by intraperitoneal (i.p.) injection 1 h after SAH. AGN196996 (Allergan Inc, USA) was diluted in Dimethyl sulfoxide (DMSO) and administered by intracerebroventricular (i.c.v.) injection 1 h prior to SAH. 500 pmol of Msr1 siRNA and scramble RNA (Thermo Fisher Scientific, Waltham, USA), in a total volume of 5 μL was injected (i.c.v.) 48 h prior to SAH. LY294002 (Selleck Chemicals, Houston, USA) was injected (i.c.v.) 2 h prior to SAH. The dose and time points of administration were based on previous research (24, 25).

Neurological outcomes were assessed 24 h after SAH by modified Garcia scores and the beam balance test. The modified Garcia score consists of six aspects, including autonomous activity, autonomous movement of the four limbs, forepaw outstretching, whisker proprioception, body proprioception, and climbing. The score for each part ranges from 0-3 points, and the score used for evaluation is the sum of the scores for each part. Beam balance tests were used to evaluate the ability of rats to balance. In brief, the rats were allowed to walk on a narrow wooden beam for 1 minute; the score ranged from 0 to 4 according to walking distance and the number of falls or slips. As previously described (4), the average scores of three consecutive trials was calculated for analysis, and blindly evaluated by two independent investigators.

The permeability of the BBB was evaluated at 24h after SAH, as reported previously (22). In brief, Evan’s blue (2%, 5 mL/kg, Sigma, USA) was injected into the left ventricle 1 h prior to sacrifice. Brain tissues were weighed after perfusion with normal saline, homogenized in 50% trichloroacetic acid (TCA), and then centrifuged. The supernatant was then collected, mixed with ethanol and TCA, and then maintained overnight at 4°C. Evan’s blue concentration determined by absorbance at 630 nm represented BBB permeability.

To evaluate water content in the brain. Rats were sacrificed 24 h after surgery and the brains were divided into four parts: the left cerebral hemispheres, the right cerebral hemispheres, the brain stem, and the cerebellum. Each brain segment was weighed to provide a wet weight. Then, each segment was placed in an oven for 24 h at 100°C to determine the dry weight. The formula used to calculate brain water content was as follows: [(wet weight-dry weight)/wet weight]×100%.

At set time points after SAH, The rats were sacrificed and perfused with cold phosphate buffer solution (PBS) and 4% paraformaldehyde (PFA). Then, fixed the brain in PFA solution at 4°C for 24 h. After embedding the tissue in paraffin, cut into 5um slices for further use. Nissl staining (14 days after SAH) was performed using 5% cresyl blue and observed by light microscopy for neuronal morphology and cell count (24, 27).

SH-SY5Y cells (Shanghai, China) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, NY, USA) containing Ham’s F-12 K Nutrient Mixture with 10% fetal bovine serum (HyClone, PA, USA) and 1% penicillin-streptomycin (Gibco), as reported previously (28). BV-2 microglia cells provided generously by Dr. Mou (Harbin Medicine University, HMU, China), were cultured in DMEM containing 10% FBS and 1% penicillin-streptomycin. BV-2 cells and SY5Y were seeded into 6-well plates (Coring Transwell, USA) and incubated at 37°C and 5% CO2 in a humid environment for the follow-up experiment.

Co-cultured cells were assigned into three groups randomly as previously reported (13, 29): control group (incubated with complete medium); Oxyhemoglobin (Hb; Sigma-Aldrich, MO, USA) +Vehicle group (10 μM Hb and 1% DMSO in complete medium); Hb+Am80 group (incubated with Am80 5 μM after 1 hour of Hb induction).

Extraction of total RNA from basal brain tissue by Trizol reagent (Takara, Kyoto, Japan). The sample is transcribed and amplified with the kit (cDNA Synthesis Kit, Takara, Kyoto, Japan; SYBR Premix Ex Taq kit Takara, Kyoto, Japan). All procedures were performed in accordance with the manufacturer’s instructions. The primer sequences used for amplification of IL-6, IL-1β, TNF-α and GAPDH have been previously reported (30, 31). All reactions were run in triplicate and quantification of the qRT-PCR results was performed by the 2−^△△CT method. The primer sequences are shown in Supplementary Table 1 .

Flow cytometry and a specific assay kit (Solarbio, China) were used to assess cell death, as reported previously (4). Data were analyzed by NovoCyte Express (ACEA, Biosciences, USA). Dead cell counts were calculated with the following formula: [Annexin V+/PI+ cells/total cells]×100%.

The Cell Counting Kit-8 (CCK-8) (Dojindo, Shanghai, China) was used to assessed the viability of SH-SY5Y cells. In brief, a total of 1000 cells in complete medium were cultured in a 96-well plate. Then, CCK-8 reagent (10 μL) was mixed with 90 μL of DMEM to generate a working solution. 100μL of working solution per hole, incubated for 1.5 h. Cell viability was detected 24 h after Hb induction. There were three replicate wells in each group and repeated three times.

Western blotting was performed as described previously (4, 30). Total or nuclear proteins from cells were extracted with RIPA lysis solution (Beyotime, China) or nucleoprotein extraction kit (Wanleibio, China), the protocols were performed in accordance with the manufacturer’s instructions. The protein sample loaded in an equal amount, and transferred to the polyvinylidene fluoride (PVDF) membrane after separation by SDS-PAGE electrophoresis. The membrane was incubated overnight at 4°C with the following primary antibodies: anti-RARα (1: 500, bs-0251R, Bioss), anti-Mafb (1:200, HPA005653, Sigma), anti-Msr1 (1:1000, ab123946, Abcam), anti-Iba-1 (1:2000, ab178846, Abcam), anti-phospho-Akt (1:1000, #9271, Cell Signaling Tech), anti-Akt (1:1000, #9272, Cell Signaling Tech), anti-p-NF-kB p65 (1:1000, #3033, Cell Signaling Tech), anti-NF-κB-p6 5 (1:1000, #8242, Cell Signaling Tech), anti-Arg-1 (1:100, ab96183, Abcam), anti-iNOS (1:100, ab178945, Abcam), anti-Bcl-xl (1:2000, ab32370, Abcam), anti-Bcl-2 (1:500, WL01556, Wanleibio), anti-Bax (1:2000, ab32503, Abcam), anti-cleaved caspase-3 (1:500, ab214430, Abcam), anti-IKBα (1:1000, ab32518, Abcam), anti-p-IKB-α (1:5000, ab133462, Abcam), anti-IL-6 (1:1000, ab259341, Abcam), anti-IL-1β (1:1000, ab254360, Abcam), anti-TNFα (1:1000, ab255275, Abcam), anti-β-actin (1:2000, ab8226, Abcam) and anti-Histone 3 (1:2000, ab176842, Abcam). The membranes were incubated in appropriate secondary antibodies (2 h, 37°C). Membranes detected with an enhanced chemiluminescence (ECL) reagent kit (NCM biotech, China) after washed three times. Images were analyzed by Image J software (Image J 1.4, NIH, USA).

Rats were transcardially perfused with ice-cold PBS and 4% PFA 24 h after SAH. Brain samples were then fixed in PFA for 24 h, followed by dehydration in 30% sucrose for 3 days. Next, the brains were treated with OCT embedding agent and frozen at -80°C. A cryostat was then used to cut the brain tissue into 7 μm sections for double immunofluorescence and TUNEL staining. The cells were seeded in a 6-well transwell plate and treated with oxyhemoglobin or Am80 for co-cultivation. Subsequently, cells were fixed with 4% PFA, permeabilized with 0.2% Triton-100 for 15 minutes, and then washed three times with PBS.

For immunofluorescence staining, the slides were blocked with 5% BSA for 1 hour and incubated overnight at 4°C with the following primary antibodies: anti-RARα (1:50, bs-0251R, Bioss), anti-MAFB (1:200, HPA005653, Sigma), anti-MSR1 (1:200, ab123946, Abcam), anti-NeuN (1:2000, ab177487, Abcam), anti-Iba-1 (1:1000, ab5076, Abcam; 1:1000, ab178846, Abcam; 1:1000, ab283319, Abcam), anti-Arg-1 (1:200, ab96183, Abcam), and anti-iNOS (1:200, ab178945, Abcam). After washing with PBS, the sections were incubated with an appropriate secondary antibody for 2 h at room temperature. TUNEL staining was performed to detect apoptotic cell death with Apoptosis Detection Kit (Roche, USA). Neuronal apoptosis was then evaluated by counting the number of TUNEL-positive neurons in the cortex. Six fields per brain were randomly selected, observed, and imaged on a fluorescence microscope (Lecia Microsystems, Germany).

All data are presented as mean ± standard deviation (SD) or median with interquartile range. Comparisons between two groups were performed using the student’s t-test One-way analysis of variance (ANOVA) was performed if comparisons involved more than two groups. The Mann Whitney U test or Kruskal-Wallis test, followed by the Dunn post hoc test, were used to test variables that were not normally distributed. Statistical analyses were performed using GraphPad Prism software 8.0 (GraphPad Software, US). P < 0.05 was defined as statistically significant.

34 rats were assigned to the Sham group, 221 rats were used to establish the SAH model, and 17 rats were ruled out (SAH grade score <7). The overall mortality rate of SAH modelling was 23.53% (52/221), ( Figure 1A ). Schematic diagram of successful modeling of subarachnoid hemorrhage ( Figure 1B ). There was no significant difference in the SAH grade score of all the SAH groups ( Figure 1C ).

Western blotting to detect the expression of RARα, Mafb and Msr1 ( Figure 2A ). The expression of RARα, Mafb and Msr1 began to increase at 3 h and reached a peak at 24 hours after SAH (P < 0.05, Figures 2B–D ). Double immunofluorescence staining further confirmed that RARα, Mafb, and Msr1, co-localized with microglia in the cerebral cortex ( Figure 2E ).

The neurological function of rats was scored by modified Garcia score and beam balance tests at 24 hours after SAH. The neurobehavioral scores were significantly reduced after SAH surgery when compared with the Sham group (P < 0.05, Figures 3A, B ). Am80 administration reversed the decline in neurological score (the modified Garcia score and beam balance tests) (P < 0.05, Figures 3A, B ). After SAH surgery, the extravasation of Evan’s blue was significantly increased, whereas, Am80 treatment reversed the extravasation. (P < 0.05, Figure 3C ). The brain water content weighing result displayed the extent of brain edema after SAH surgery on both hemispheres was significantly higher than Sham group 24h after SAH (P < 0.05, Supplementary Figure 1A ). Treating with Am80 administration, the brain edema of the both hemispheres was significantly alleviated after SAH (P < 0.05, Supplementary Figure 1A ). These results suggested that optimal therapeutic dosage of Am80 was 5 mg/kg after SAH in rats.

We used TUNEL staining to evaluate neuronal apoptosis. The results showed that 24 hours after SAH induction, the number of apoptotic neurons increased significantly in the SAH+Vehicle group (P < 0.05, Figures 3D, E ). In contrast, the administration of Am80 inhibited neuronal apoptosis (P < 0.01, Figures 3D, E ). Nissl staining showed that 14 days after SAH, the loss of neurons in the Cornu Ammonis area (CA) 1, CA3, and the dentate gyrus (DG), had increased and the morphology of the neurons exhibited shrinkage (P < 0.05, Figures 3F, G ). Am80 significantly reduced the loss of neurons in these parts of the hippocampus (P < 0.05, Figures 3F, G ).

BV2 cells and SH-SY5Y cells were co-cultured and induced by oxyhemoglobin to establish an in vitro model of SAH ( Figure 4A ). The results of CCK8 cell viability measurement indicated that the cell viability of the Hb+Vehicle group decreased compared to control group (P <0.05, Figure 4B ). Conversely, Am80 administration reversed this change (P <0.05, Figure 4B ).

Compared with the control group, the expression of Bcl-xl and Bcl-2 decreased after Hb induction, while the expression of Bax and Cleaved-Caspase 3 increased (P <0.05, Figure 4C–G ). AM80 administration reversed the changes in the expression of apoptosis-related proteins induced by Hb (P <0.05, Figure 4C–G ). The TUNEL staining results were consistent with the protein results. AM80 administration reduced the number of apoptotic neurons induced by Hb (P <0.05, Figure 4H ). The results of flow cytometry on apoptotic neurons are consistent with the above results (P <0.05, Figure 4I ).

The results of Iba-1 immunofluorescence suggested that compared with the Sham group, the number of activated microglia increased after SAH (P <0.01, Figures 5A, B ). Am80 treatment significantly reduced the number of Iba-1 positive microglia in the cerebral cortex (P <0.01, Figures 5A, B ). The western blot results of Iba-1 are consistent with the above results (P <0.05, Figures 5C, D ). The area of activated-microglia was increased after SAH, but the lacunarity, fractal dimension, and perimeter were reduced than the Sham group (P <0.05, Figures 5E–I ); Am80 treatment reversed this alteration (P <0.05, Figures 5E–I ).

In the in vitro SAH model, the expression of Arg-1 and iNOS in BV2 cells increased after Hb induction compared with the control group (P < 0.05, Figures 6A–C ). After Am80 treatment, the expression of Arg-1 further increased, on the contrary, the expression of iNOS decreased (P < 0.05, Figures 6A–C ). Immunofluorescence staining further confirmed that the number of microglia that were positive for Arg-1 and iNOS increased significantly after induced with Hb (P < 0.05, Figures 6D–F ). After Am80 treatment, the number of M2 type microglia (Arg-1-positive) increased, conversely, the number of M1 type microglia (iNOS positive) decreased (P < 0.05, Figures 6D–F ).

To confirm the role of RARα in the protective effects against neuroinflammation, the specific antagonist AGN196996 was used to inhibit the RARα ( Figure 7A ). After administration of the AGN196996, the neurological score of rats was significantly reduced (P < 0.05, Figures 7B, C ). Compared with the Sham group, the expression of Mafb, Msr1, p-NF-κB p65 were significantly increased after SAH; there was no change in level of the RARα (P < 0.05, Figures 7D–I ). Conversely, p-Akt level was reduced in the SAH+vehicle groups (P < 0.05, Figures 7D–I ). The expression of Mafb, Msr1 and p-Akt increased significantly after Am80 treatment, but the expression of p-NF-κB p65, IL-6, IL-1β and TNF-α decreased significantly under the same conditions. (P < 0.05, Figures 7D–I ).Immunofluorescence revealed that the area of TNF-α-positive cells in brain cortex after SAH was significantly increased when compared with the Sham group (P < 0.05, Figure 7M ). Am80 treatment reduced the fluorescence area of TNF-α-positive cells (P < 0.05, Figure 7M ). Conversely, the protective effects of Am80 were reversed by AGN196996 (P < 0.05, Figure 7M ). The mRNA expression of IL-6, IL-1β, and TNF-α were consistent with the protein result (P < 0.05, Supplementary Figures 1B–D ).

The rat neurological function scores significantly decreased after administration of Msr1 siRNA and LY294002 when compared with the SAH+Am80 group (P < 0.05, Figures 8A, B ).

In the SAH+Msr1 siRNA+Am80 group, the Msr1siRNA abolished the anti-inflammation effects of Am80, thus leading to a reduction in the expression of Msr1 and p-Akt. Meanwhile, the expression of p-NF-κB-p65 and inflammatory cytokines were upregulation when compared with the SAH+Am80 +Msr1 scrRNA group (P < 0.05, Figures 8C–J ). Compared with the SAH+DMSO+Am80 group, pretreatment with LY294002 downregulated the expression of p-Akt but increased the expression of p-NF-κB-p65, IL-6, IL-1β, and TNF-α (P < 0.05, Figures 8C–J ). Immunofluorescence staining further revealed that the fluorescence area of TNF-α-positive cells in the SAH+Am80 group was significantly decreased when compared with the SAH+Vehicle group (P < 0.05, Figure 8K ). The administration of Msr1 siRNA and LY294002 reversed these changes (P < 0.05, Figure 8K ). Furthermore, the mRNA expression of IL-6, IL-1β and TNF-α were decreased by treatment with Am80, whereas the effects were abolished by Msr1 siRNA and LY294002. (P < 0.05, Supplementary Figures 1E–G ).

To investigate whether the mechanism of Am80 inhibiting inflammation involves nuclear transcription of NF-κB. Western blotting results showed that compared with the control group, the expression of phosphorylated IKBα in BV2 cells increased after Hb induction, which was consistent with the nuclear transcription trend of NF-κB-p65 (P < 0.01, Figures 9A–D ). Compared with the Hb+Vehicle group, the expression of p65 in the cytoplasm increased and the expression of p65 in the nucleus decreased after Am80 administration, which means that the nuclear transcription of p65 was inhibited. Meanwhile, the phosphorylation of IκBα was significantly inhibited (P < 0.05, Figures 9A–D ). Western blot results of inflammatory cytokines showed that the expression of IL-6, IL-1β and TNF-α increased significantly after Hb induction when compared with the Control group (P <0.05, Figures 9E–H ). In contrast, Am80 treatment significantly inhibited the expression of inflammatory cytokines after Hb induction (P < 0.05, Figures 9E–H ).