HDAC3 inhibition ameliorates ischemia/reperfusion-induced brain injury by regulating the microglial cGAS-STING pathway

Abstract

Rationale: It is known that neuroinflammation plays a critical and detrimental role in the development of cerebral ischemia/reperfusion (I/R), but the regulation of the cyclic GMP-AMP synthase (cGAS)-mediated innate immune response in I/R-induced neuroinflammation is largely unexplored. This study aimed to investigate the function and regulatory mechanism of cGAS in I/R-induced neuroinflammation and brain injury, and to identify possible strategies for the treatment of ischemic stroke. Methods: To demonstrate that microglial histone deacetylase 3 (HDAC3) regulates the microglial cGAS-stimulator of interferon genes (cGAS-STING) pathway and is involved in I/R-induced neuroinflammation and brain injury, a series of cell biological, molecular, and biochemical approaches were utilized. These approaches include transient middle cerebral artery occlusion (tMCAO), real-time polymerase chain reaction (PCR), RNA sequencing, western blot, co-immunoprecipitation, chromosome-immunoprecipitation, enzyme-linked immunosorbent assay (ELISA), dual-luciferase reporter assay, immunohistochemistry, and confocal imaging. Results: The microglial cGAS- STING pathway was activated by mitochondrial DNA, which promoted the formation of a pro-inflammatory microenvironment. In addition, we revealed that HDAC3 transcriptionally promoted the expression of cGAS and potentiated the activation of the cGAS-STING pathway by regulating the acetylation and nuclear localization of p65 in microglia. Our in vivo results indicated that deletion of cGAS or HDAC3 in microglia attenuated I/R-induced neuroinflammation and brain injury. Conclusion: Collectively, we elucidated that the HDAC3-p65-cGAS-STING pathway is involved in the development of I/R-induced neuroinflammation, identifying a new therapeutic avenue for the treatment of ischemic stroke.

Keywords: HDAC3; Ischemia/reperfusion; Microglia; Neuroinflammation; cGAS.

Figure 1

The cytosolic DNA-sensing pathway is upregulated in microglia after cerebral ischemia/reperfusion (I/R). (A) Kyoto encyclopedia of genes and genomes (KEGG) analysis of the upregulated genes from the RNA sequencing data of adult primary microglia isolated from the brains of mice that underwent tMCAO (10 mice for each sample). (B) Gene set enrichment analysis showed that the cytosolic DNA sensing pathway was upregulated in microglia isolated from mice subjected to tMCAO (Sham: the brain underwent sham operation; Contra: the contralateral side of the brain underwent I/R; Lateral: the side of the brain underwent I/R). (C) Genes involved in the cytosolic DNA sensing pathway were upregulated in microglia isolated from mice subjected to tMCAO. (D and E) Ischemic penumbra cortex was collected from mice that underwent tMCAO at different time points post-reperfusion, and the mRNA levels of IFN-β (D) and IL-6 (E) were quantified by real-time PCR (* indicates p < 0.05, ** indicates p < 0.01 by ANOVA or Student's t-test). Abbreviations: IFN-β, interferon-beta; IL-6, interleukin-6; tMCAO, transient middle cerebral artery occlusion.

Figure 2

The cytosolic DNA sensor cGAS is activated after cerebral ischemia/reperfusion. (A) The mRNA level of cGAS in primary microglia isolated from brain tissue 3 h post-tMCAO (9 mice per group). (B) The protein level of cGAS in primary microglia isolated from brain tissue at 1, 2, 3 and 6 h post-tMCAO, respectively. The gray values were analyzed by using ImageJ and were normalized to Iba1. (C) BV2 cells were transfected with siRNA against cGAS and harvested 48 h post-transfection to quantify the mRNA level of cGAS. (D) cGAS-silenced and control BV2 cells were transfected with poly(dA:dT) and harvested 3 h post-poly(dA:dT) treatment to analyze the protein levels of pIRF3, cGAS, and GAPDH. (E) cGAS-silenced, STING-silenced, and control BV2 cells were transfected with poly(dA:dT) and harvested 6 h post-poly(dA:dT) treatment to analyze the mRNA level of IFN-β. (F) BV2 cells were transfected with siRNA against STING and harvested 48 h post-transfection to quantify the mRNA level of STING. (G) cGAS-silenced, STING-silenced, and control BV2 cells were transfected with mtDNA and harvested 6 h post-transfection to analyze the mRNA level of IFN-β. (H) The cGAS-mtDNA complex was isolated from brain tissue extracted from mice subjected to tMCAO using an anti-cGAS antibody and quantified by real-time PCR. (I) The cGAS-mtDNA complex was isolated from OGD/R-treated primary microglia using an anti-cGAS antibody and quantified by real-time PCR with primer pairs against mtDNA. (J and K) The mRNA levels of IFN-β (J) and IL-6 (K) in control and cGAS-silenced BV2 cells after OGD/R treatment. (* indicates p < 0.05, ** indicates p < 0.01 by ANOVA or Student's t-test). Abbreviations: cGAS, cyclic GMP-AMP synthase; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; IFN-β, interferon-beta; IL-6, interleukin-6; mtDNA, mitochondrial DNA; OGD/R, oxygen-glucose deprivation/reperfusion; pIRF3, phosphorylated interferon regulatory factor 3; siRNA, small interfering RNA; STING, stimulator of interferon genes; tMCAO, transient middle cerebral artery occlusion.

Figure 3

Conditional knockout of cGAS in microglia attenuates ischemic/reperfusion-induced brain injury. (A) Adult primary microglia were isolated from WT (n = 6) and cGAS cKO (n = 6) mice 2 weeks after tamoxifen administration, and the mRNA level of cGAS was quantified by real-time PCR. (B) The neurological deficits of WT (n = 10) and cGAS cKO (n = 5) mice were assessed 24 h post-tMCAO. (C) Brain tissue isolated from WT (n = 19) and cGAS cKO (n = 11) mice 24 h post-reperfusion was stained with triphenyl tetrazolium chloride, and infarct volume was determined using ImageJ. (D) The mRNA level of IL-6 in the ischemic and contralateral cortices of WT (n = 6) and cGAS cKO (n = 6) mice 6 h post-tMCAO. (E) The protein level of IL-6 in serum from WT (n = 16) and cGAS cKO (n = 8) mice 6 h post-tMCAO. (F) The protein levels of cGAS and Iba1 in the ischemic penumbra brain tissue were detected by western blot and analyzed with the ImageJ. software (G to I) The number (G and I) and soma area (H and I) of Iba1⁺ cells in the contra and ischemic penumbra of cGAS cKO and WT mice subjected to tMCAO 6 h post-reperfusion. (* indicates p < 0.05, ** indicates p < 0.01 by ANOVA or Student's t-test). Abbreviations: cGAS, cyclic GMP-AMP synthase; cKO, conditional knockout; Iba1, ionized calcium-binding adapter molecule 1; IL-6, interleukin-6; tMCAO, transient middle cerebral artery occlusion; WT, wild-type.

Figure 4

HDAC3 promotes the transcription of cGAS in microglia (A) The mRNA levels of cGAS in TSA-, MS275-, and NAM-treated BV2 cells were quantified by real-time PCR. (B) The cGAS reporter plasmid and TK-Renilla reference reporter plasmid were co-transfected into HEK293T cells and the cells were treated with TSA, MS275, or vehicle for 24 h. (C and D) BV2 cells were pretreated with TSA, MS275, or vehicle for 12 h and poly(dA:dT) (0.5 μg/mL) was then transfected into the cells using Lipofectamine 2000. For real-time PCR, cells were harvested 6 h post-transfection (C). For western blotting, cells were harvested 3 h post-transfection (D). (E) BV2 cells stably expressing shRNA against HDAC1, HDAC2, and HDAC3 were collected and the mRNA levels of cGAS were analyzed by real-time PCR. (F) BV2 cells were treated with RGFP966 and the mRNA level of cGAS was determined by real-time PCR. (G) The enrichment of HDAC1, HDAC2, and HDAC3 in the promoter region of cGAS in BV2 cells was analyzed by chromatin immunoprecipitation followed by real-time PCR with antibodies against each HDAC. (H and I) Control and HDAC1-silenced, HDAC2-silenced, HDAC3-silenced, BV2 cell lines were transfected with poly(dA:dT) and harvested 0, 3, and 6 h post-transfection for real-time PCR (H) and western blot (I). (J) The mRNA level of IL-6 in control and HDAC3-silenced BV2 cells after OGD/R treatment. (K) The mRNA level of IL-6 in vehicle- and RGFP966-pretreated cells after OGD/R treatment. (* indicates p < 0.05, ** indicates p < 0.01 by ANOVA or Student's t-test). Abbreviations: NAM, nicotinamide; cGAS, cyclic GMP-AMP synthase; HDAC, histone deacetylase; IL-6, interleukin-6; OGD/R, oxygen-glucose deprivation/reperfusion; shRNA, short hairpin RNA; TSA, trichostatin A.

Figure 5

HDAC3 promotes cGAS transcription by deacetylating p65. (A) BV2 cells transfected with siRNA against p65 or scrambled siRNA were treated with poly(dA:dT)-Lipofectamine 2000 complex for 3 h and the expression of pIRF3, cGAS, p65, and GAPDH was analyzed. (B) The mRNA level of cGAS in p65-silenced and control cells was quantified by real-time PCR. (C) BV2 cells transfected with siRNA against p65 or scramble were treated with poly(dA:dT)-Lipofectamine 2000 complex for 6 h and the transcription levels of IFN-β and β-actin were analyzed by real-time PCR. (D) p65-silenced and control cells that underwent OGD/R were collected and the expression of IL-6 was analyzed by real-time PCR. (E) pFlag-HDAC3 and pMyc-p65 plasmids were co-transfected into HEK293T cells, and the cells were harvested for Co-IP with anti-Flag tag magnetic beads. (F) Co-IP was performed with a rabbit anti-p65 antibody or control rabbit IgG with BV2 cells. (G) The binding sites of p65 in the promoter region were analyzed by chromatin immunoprecipitation and real-time PCR with an antibody against p65. (H and I) The binding activity of p65 in the promoter region of cGAS in HDAC3-inhibited (H) and HDAC3-silenced (I) BV2 cells. (J) Cells were co-transfected with pMyc-p65, cGAS reporter plasmid and TK-Renilla plasmids, treated with RGFP966 or vehicle, and harvested 24 h post-transfection for the dual-luciferase assay. (K) Plasmids encoding WT p65, mutant p65_K122Q, mutant p65_K122/123Q, mutant p65_K122R, or mutant p65_K122/123R were transfected into HEK293T cells with the cGAS reporter and TK-Renilla reference plasmids, and the cells were harvested for the dual-luciferase assay 24 h post-transfection. (L) Plasmids encoding Myc-p65 and HA-p300 and empty vectors were co-transfected into HEK293T cells with plasmids encoding Flag-HDAC1, Flag-HDAC2, and Flag-HDAC3 and empty vectors, and the acetylation level of p65 was analyzed with anti-acetylated lysine (pan-Ac) after p65 was purified with anti-Myc tag magnetic beads. (M) The cytoplasm and nucleus of HDAC1-silenced, HDAC2-silenced, HDAC3-silenced, and control BV2 cells were separated and the expression of p65 was analyzed by western blot. (N and O) BV2 cell lines stably expressing Flag-tagged WT p65 or mutant p65_K122R or control cell lines were pretreated with RGFP966 for 12 h, transfected with poly(dA:dT), and harvested to analyze the protein levels of pIRF3, cGAS, p65, and GAPDH (N) and the mRNA level of IFN-β (O). (P) Schematic showing how HDAC3 regulates the transcription of cGAS by deacetylating p65. (* indicates p < 0.05, ** indicates p < 0.01 by ANOVA or Student's t-test). Abbreviations: cGAS, cyclic GMP-AMP synthase; Co-IP, co-immunoprecipitation; GADPH, glyceraldehyde 3-phsophate dehydrogenase; HDAC3, histone deacetylase 3; IFN-β, interferon-beta; OGD/R, oxygen-glucose deprivation/reperfusion; pIRF3, phosphorylated interferon regulatory factor 3; siRNA, small interfering RNA; WT, wild-type.

Figure 6

Deficiency of HDAC3 in microglia suppresses the cGAS-STING-mediated autoimmune response. (A and B) Adult primary microglial cells were isolated from HDAC3 cKO (n = 6) and WT (n = 6) mice, and the mRNA levels of HDAC3 (A) and cGAS (B) were analyzed by real-time PCR. (C) The neurological deficits of HDAC3 cKO (n = 7) and WT (n = 15) mice subjected to tMCAO were assessed 24 h post-reperfusion. (D and E) Brain tissue isolated from WT (n = 14) and HDAC3 cKO (n = 7) mice 24 h post-reperfusion was stained with triphenyl tetrazolium chloride, and infarct volume was determined using ImageJ (D) and infarct volume was determined using Image J (E). (F) RNA was extracted from the ischemic penumbra cortex and contra cortex of HDAC3 cKO (n = 12) and WT (n = 14) mice subjected to tMCAO 6 h post-reperfusion, and the mRNA level of IL-6 was quantified by real-time PCR. (G) Serum was isolated from HDAC3 cKO (n = 13) and WT (n = 21) mice subjected to tMCAO 6 h post-reperfusion, and the level of IL-6 was quantified by ELISA. (H to J) The number (H and J) and soma area (I and J) of Iba1⁺ cells in the ischemic penumbra and contralateral cortex of HDAC3 cKO and WT mice subjected to tMCAO 24 h post-reperfusion. (K) Proteins extracted from the ischemic penumbra and contralateral cortex of HDAC3 cKO and WT mice subjected to tMCAO 6 h post-reperfusion were analyzed with antibodies against cGAS, HDAC3, Iba1, and GAPDH. (L and M) Mice administered RGFP966 (n = 11) or vehicle (n = 5) for 2 days underwent tMCAO, and infarct volume was determined by TTC staining. (N) Serum was isolated from RGFP966-treated (n = 8) and vehicle-treated (n = 6) mice subjected to tMCAO 6 h post-reperfusion and the level of IL-6 was quantified by ELISA. (* indicates p < 0.05, ** indicates p < 0.01 by ANOVA or Student's t-test). Abbreviations: cGAS, cyclic GMP-AMP synthase; cKO, conditional knockout; ELISA, enzyme-linked immunosorbent assay; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; HDAC3, histone deacetylase 3; Iba1, ionized calcium-binding adapter molecule 1; IL-6, interleukin-6; STING, stimulator of interferon genes; tMCAO, transient middle cerebral artery occlusion; TTC, triphenyl tetrazolium chloride; WT, wild-type.

Figure 7

Proposed mechanism: HDAC3 regulates the transcription of cGAS by deacetylating p65 at K122, which plays a role in ischemic/reperfusion-induced neuroinflammation and brain injury. Briefly, mtDNA released from dead cells in the ischemic core is engulfed by microglia in the penumbra. Toxic mtDNA then activates the microglial cGAS-STING-IRF3 pathway and potentiates neuroinflammation, which in turn aggravates ischemia-induced neuronal cell death. Within microglia, HDAC deacetylates p65, which promotes its nuclear translocation and induces cGAS expression and cGAS-STING pathway activation. Abbreviations: cGAS, cyclic GMP-AMP synthase; HDAC, histone deacetylase; IRF3, interferon regulatory factor 3; mtDNA, mitochondrial DNA; STING, simulator of interferon genes.