FGF21 attenuates neuroinflammation following subarachnoid hemorrhage through promoting mitophagy and inhibiting the cGAS-STING pathway

Male C57BL/6J mice, aged 8 weeks and weighing between 20 and 25 g, were provided by the Animal Center of the Air Force Medical University. STING-floxp mice were procured from the esteemed Shanghai Model Organisms Center, Inc., CX3CR1-CreERT2 mice were obtained from the reputable Gempharmatech Co., Ltd. These mice were housed in a specific pathogen-free (SPF) environment, maintained at a temperature of 23 °C with 60% humidity, and subjected to a 12-hour light/dark cycle. Prior to commencing the subsequent experiments, the mice underwent a 2-week period of acclimation, during which they were provided with enough food and water. In the first section, experiments involving immunofluorescence staining, western blot, brain edema measurement, and qPCR assay each utilized 18 wild-type C57BL/6J mice (6 mice each from the Sham group, SAH group, and SAH + rFGF21 group). Neurobehavioral tests utilized 36 mice (12 mice each from the Sham group, SAH group, and SAH + rFGF21 group). In the second section, experiments encompassed immunofluorescence staining, western blot, brain edema measurement, and qPCR each utilized 30 mice (6 mice each from Sting^fl/fl + Sham group, Sting^fl/fl + SAH group, Sting^CKO + Sham group, Sting^CKO + SAH group, and Sting^CKO + SAH + rFGF21 group). Neurobehavioral tests involved 60 mice (12 mice each from Sting^fl/fl + Sham group, Sting^fl/fl + SAH group, Sting^CKO + Sham group, Sting^CKO + SAH group, and Sting^CKO + SAH + rFGF21 group). All experimental procedures adhered strictly to the guidelines outlined in the National Institutes of Health Guide for the Care and Use of Laboratory Animals, and were approved by the Animal Experiment Center of Air Force Military Medical University (IACUC-20,220,555).

.

SAH mouse model was established using the endovascular perforation technique, as described previously [33]. Briefly, following intraperitoneal administration of ketamine/xylazine for anesthesia, the carotid artery was meticulously exposed. A fine and sharp 6–0 monofilament nylon suture was skillfully employed to puncture the bifurcation point of the anterior and middle cerebral arteries. Subsequently, mouse was carefully placed in a heated chamber to facilitate post-operative recovery. As for the sham-treated mice, they underwent identical procedures, albeit without the actual perforation of the arterial vessel.

Brain water content was assessed using the wet/dry method [34]. The brain tissues were removed and promptly weighed to determine its wet weight, subsequently subjected to desiccation in an oven maintained at 95 to 100 °C for a duration of 72 h. The brain tissues were then re-weighed to ascertain its dry weight. Brain water content = ([wet weight − dry weight]/wet weight) × 100%.

Neurological function was evaluated through the modified Garcia scale test [33]. This scoring system encompassed six measurements including spontaneous activity, spontaneous movement of four limbs, body proprioception, climbing, forepaw outstretching, and response to whisker stimulation, with a higher score indicating superior neurobehavioral outcomes. Neurobehavioral tests were performed in a blinded manner.

After a 72-hour post-SAH, mouse was sacrificed by an overdose of 2% pentobarbital sodium and underwent trans-cardiac perfusion with a 4% paraformaldehyde solution. Brain tissue was then removed and stored overnight in 4% paraformaldehyde. Dehydration was carried out using 30% sucrose solutions. Subsequently, brain sections measuring 15 μm were obtained and incubated in PBS solution containing 0.1% Triton X-100 for 30 min. To prevent non-specific binding, the sections were blocked with PBS solution containing 5% goat serum for an additional 30 min. Following this, the sections were incubated overnight at 4 °C with primary antibodies: IBA1 (ab289874, 1:300, Abcam) and CD68 (ab283654, 1:300, Abcam). Post-incubation, brain sections were thoroughly rinsed in PBS with three consecutive 5-minute washes. Subsequently, the sections were incubated at 25 °C for 1 h with corresponding secondary antibodies: Alexa Fluor 488 Anti-rabbit IgG (A11034, 1: 1000, Invitrogen) and Alexa Fluor 594 Anti-goat IgG (A11058, 1:1000, Invitrogen). Finally, the sections were stained with DAPI solution and observed using an A1 Si confocal microscope.

Cell death was assessed by an unbiased observer utilizing the In Situ Cell Death Detection Kit (Roche, Germany) following the manufacturer’s guidelines. Following treatment with 0.3% hydrogen peroxide (0.5 h) and incubation with 0.25% proteinase K (0.5 h, 37℃), the sections were stained using the TUNEL reaction solution in a lightless environment (1 h, 37℃). Finally, brain sections were stained with DAPI solution and observed using an A1 Si confocal microscope.

Total RNA was extracted using TRIzol Reagent (Invitrogen, Carlsbad, California, USA) in accordance with the manufacturer’s protocol. Subsequently, 1 mg of total RNA was reverse-transcribed using the HiScript II Q RT SuperMix for quantitative real-time PCR (qRT-PCR) (R223-01, Vazyme biotech). The qRT-PCR was performed using an iQTM 5 Optical Module Real-Time PCR Detection System (Bio-Rad, USA). ChamQTM SYBR qPCR master mix (Q311-02, Vazyme biotech) was utilized to quantify mRNA as per the manufacturer’s instructions. The mRNA levels were normalized to β-actin and calculated using the 2-ΔΔCt method.

Animal subjects were sacrificed by an overdose of 2% pentobarbital sodium and their brains were carefully extracted and dissected. To dissociate the tissues, a digestion process using Papain (2 mg.ml^{− 1}, Worthington) in RPMI 1640 medium (Gibico) was carried out at 37℃ for 1 h. The resulting dispersed cells were then filtered using the 70 mm nylon mesh. Subsequently, the resulting cells were further resuspended with a 30% Percoll density gradient (GE Healthcare) and centrifuged at 900 g for 25 min. The cells were isolated by collecting the bottom fraction in the 30% Percoll solution. The samples were then treated with FcR Blocking Reagent (Miltenyi Biotec) for blocking and subjected to flow cytometry analysis for the detection of surface antigens. For fluorescence acquisition, the cells were stained with FITC anti-mouse CD11b Antibody (BioLegend) and PerCp-CyTM5.5 anti-mouse CD45 Antibody (BD Pharmingen). Flow cytometry (FACS Aria II SORP, BD Biosciences, USA) was employed for fluorescence acquisition, and the samples were gated based on CD11b + CD45dim expression, which represents the microglia population.

Brain samples were homogenized using an ultrasonic homogenizer in a lysis buffer containing 1% protease inhibitor. After centrifugation of the lysate, the supernatant was collected and quantified using the BCA Protein Assay Kit (Thermo Scientific). Protein samples was then isolated by SDS-PAGE with 20 µg of protein loaded per well. Subsequently, protein was transferred onto PVDF membranes (IPVH-00010, Millipore Corporation, Billerica, MA, USA). The membranes were then incubated with 5% non-fat milk in TBST for 2 h. Primary antibodies, including anti-cGAS (NBP3-16666; Novus), anti-STING (D2P2F; Cell Signaling Technology), anti-phospho-STING (D8F4W; Cell Signaling Technology), anti-TBK1 (D1B4; Cell Signaling Technology), anti-phospho-TBK1 (D52C2; Cell Signaling Technology), anti-NF-κB (80979-1-RR; Proteintech), anti-p-NF-κB (93H1; Cell Signaling Technology), anti-IRF3 (11312-1-AP; Proteintech), anti-p-IRF3 (D6O1M; Cell Signaling Technology), anti-SQSTM1/p62 (D1Q5S; Cell Signaling Technology), anti-TIM50 (#62,317; Cell Signaling Technology), anti-Tom20 (D8T4N; Cell Signaling Technology) and anti-β-actin (13E5; Cell Signaling Technology), were then incubated with the membranes. Afterward, the membranes were washed three times with TBST for 5 min each and subsequently incubated with secondary antibodies. Lastly, protein bands were visualized using a Bio-Rad imaging system (Bio-Rad, USA).

All statistical analyses were carried out using SPSS software (version 21.0). Statistical comparisons between the two groups were conducted using an unpaired, two-tailed Student’s t-test. For comparisons among multiple groups, a one-way or two-way analysis of variance (ANOVA) followed by the Student-Newman-Keuls test was employed. The rating scale data, specifically neurological scores, were analyzed using Kruskal-Wallis one-way ANOVA on ranks, followed by Steel-Dwass multiple comparisons. The results were expressed as the median with the 25th and 75th percentiles. The data were presented as means ± standard deviation (SD). A p-value of less than 0.05 was considered statistically significant.

In our preliminary experimentation to determine the optimal FGF21 concentration, guided by prior research on the therapeutic application of FGF21 in CNS disorders [20, 35, 36], we administered three distinct doses (0.75 mg/kg, 1.5 mg/kg, and 3 mg/kg) of FGF21 to assess their efficacy in ameliorating cerebral edema and neurological impairment post-SAH. Evaluation of brain water content revealed that all three dosages of rFGF21 treatment effectively alleviated SAH-induced cerebral edema (Fig. 1A). Notably, the protective effect of 1.5 mg/kg was more pronounced than that of 0.75 mg/kg, with further dose escalation conferring no significant additional benefit. In line with these observations, rFGF21 treatment in a similar dosing regimen significantly mitigated SAH-induced neurological deficits (Fig. 1B). Consequently, 1.5 mg/kg rFGF21 was selected as the minimal effective dose for subsequent investigations. Moreover, TUNEL staining indicated a significantly increased number of apoptotic cells after SAH, which was markedly reduced upon additional administration of rFGF21 (Fig. 1C-D). Collectively, these results provide compelling evidence that rFGF21 mitigates apoptosis, cerebral edema, and neurological deficits following SAH.

To delineate the molecular mechanisms underlying the neuroprotective attributes of FGF21, we performed bulk RNA-sequencing analysis on cerebral tissue obtained from Sham, SAH, and SAH + rFGF21 mice. Out of the 1932 genes identified, a total of 199 were downregulated and 1328 were upregulated in the cerebral tissue post-SAH. A Venn diagram depicted the reversal of 1328 upregulated genes and 44 downregulated genes induced by SAH following rFGF21 administration (Fig. 2A). Gene ontology (GO) analysis of these reversed genes induced by rFGF21 treatment unveiled pivotal biological processes enriched in immune response and inflammatory response (Fig. 2B). Moreover, KEGG analysis revealed notable changes in the expression of genes linked to inflammation-related pathways, including the NF − kappa B signaling pathway, Toll − like receptor signaling pathway, Cytokine − cytokine receptor interaction, and NOD-like receptor signaling pathway, upon rFGF21 treatment (Fig. 2C). Additionally, GSEA indicated the activation of these inflammation-related pathways following SAH, which were subsequently suppressed after supplemental rFGF21 treatment (Fig. 2D-G). Collectively, these findings indicate that rFGF21 significantly suppresses neuroinflammation after SAH.

Microglia, the predominant immune cells in CNS, play a crucial role in the regulation of neuroinflammation following SAH. We found that the density of CD68, a marker of hyperactivated microglia, noticeably increased after SAH, but was subsequently reversed with rFGF21 treatment (Fig. 3A-B). Additionally, we employed the ImageJ software to delineate and skeletonize microglial morphology through Iba1 staining. The SAH group exhibited a significant decrease in branch numbers and branch length, while also showing an increase in microglial soma area compared to the sham group, suggesting the hyperactivated state of microglia. However, rFGF21 treatment effectively rescued the SAH-induced morphological alterations (Fig. 3C-D). Considering that different microglial phenotypes are associated with either cytotoxic or neuroprotective effects, we further examined the impact of rFGF21 treatment on microglial phenotype switching. Microglia were isolated by flow cytometry and subjected to qPCR analysis. Notably, rFGF21 treatment resulted in a significant downregulation of M1 microglia-specific transcripts (CD86, iNOS) and an upregulation of M2 microglia-specific transcripts (CD206, Arg1) following SAH (Fig. 3E-F). Collectively, these findings indicate that rFGF21 treatment suppresses microglial proinflammatory activities subsequent to SAH.

Notably, transcriptomic analysis unveiled a substantial attenuation of various immune signature genes upon rFGF21 treatment, including those associated with type I IFN signaling and cytosolic DNA-sensing pathways (Fig. 4A-B). The cytosolic DNA sensor, cGAS, is known to recognize cytosolic dsDNA, thereby initiating type I interferon and NF-κB pathways. Western blot analysis revealed an upregulation of cGAS protein levels and STING phosphorylation levels following SAH (Fig. 4C-E). Consistent with the elevated expression of cGAS-STING, phosphorylated levels of TBK1, NF-κB, and IRF3 were significantly increased post-SAH (Fig. 4F-H). Importantly, FGF21 treatment significantly decreased the SAH-induced activation of cGAS-STING signaling (Fig. 4F-H). Collectively, rFGF21 inhibits cGAS-STING pathway mediated inflammatory response following SAH.

In order to investigate the role of the cGAS-STING pathway in SAH-induced neuroinflammation and brain injury, we generated Sting^f/f:CX3CR1creER^/+ mice by crossing Sting^f/f mice with CX3CR1creER^/+ mice. Through the administration of tamoxifen, we induced the conditional knockout (cKO) of Sting specifically in microglia. Subsequently, we subjected both Sting^fl/fl and Sting^CKO mice to SAH injury, and observed that compared to the Sting^fl/fl mice, the density of activated microglia significantly decreased in Sting^CKO mice after SAH (Fig. 5A-B). And Sting cKO led to a substantial downregulation of phosphorylated levels of TBK1, NF-κB, and IRF3 (Fig. 5C-F) and a remarkable reduction in the IFN-β and TNF mRNA levels in mice following SAH (Fig. 5G-H). These findings underscore the critical role of the cGAS-STING pathway in SAH-induced neuroinflammation. Additionally, we observed that rFGF21 treatment failed to further alleviate neuroinflammation in Sting^CKO mice (Fig. 5A-H), suggesting that rFGF21 attenuates SAH-induced neuroinflammation through inhibition of cGAS-STING pathway.

Tunel staining revealed a significant reduction in SAH-induced neuronal apoptosis upon microglial deletion of Sting (Fig. 6A-B). Additionally, brain water measurement demonstrated a mitigated brain edema in Sting^CKO mice as compared to Sting^fl/fl mice (Fig. 6C). Furthermore, Sting^CKO SAH mice exhibited superior performance in modified Garcia tests when compared to Sting^fl/fl SAH mice (Fig. 6D). However, the administration of rFGF21 failed to further alleviate brain injuries in Sting^CKO mice (Fig. 6A-D). Collectively, these findings suggest that rFGF21 attenuates SAH-induced cerebral injuries through inhibition of cGAS-STING pathway.

cGAS-STING pathway can be triggered by self-DNA that is released from damaged mitochondria and nuclei (35). We verified that there was a significant upregulation of cytoplasmic mtDNA in isolated microglia following SAH, which was subsequently mitigated by the administration of rFGF21(Fig. 7A). Moreover, we did not observe any substantial alterations in cytosolic DNA originating from the nucleus following rFGF21 treatment (Fig. 7B). These results suggested that FGF21 inhibited the activation of cGAS-STING pathway by preventing the accumulation of mtDNA in cytosolic. Impairment of mitophagy has been demonstrated to be an important reason for the accumulation of mtDNA in the cytoplasm. We hypothesized that FGF21 may inhibit the activation of cGAS-STING pathway by inducing mitophagy. As expected, the level of mitophagy in microglia isolated from SAH mouse brains treated with rFGF21 was elevated, as evidenced by the decrease in SQSTM1, TOMM20 and TIM50 levels (Fig. 7C-F). (AMP-activated protein kinase) AMPK is a well-established downstream effector of FGF21 [17, 37]. Treatment with rFGF21 resulted in an elevation in the phosphorylated form of AMPK (Fig. 7C and G). Furthermore, the inhibition of AMPK diminished FGF21-mediated promotion of mitophagy (Fig. 7H-K), reduction of cytoplasmic DNA (Fig. 7L) and suppression of IFN-β and TNF mRNA expressions (Fig. 7M-N) in isolated microglia. These findings strongly suggest that FGF21 enhances the process of mitophagy, leading to a decrease in the leakage of mtDNA, and concurrently inhibits cGAS-STING-dependent neuroinflammation.