ACOD1, rather than itaconate, facilitates p62-mediated activation of Nrf2 in microglia post spinal cord contusion

Abstract

Background: Spinal cord injury (SCI)-induced neuroinflammation and oxidative stress (OS) are crucial events causing neurological dysfunction. Aconitate decarboxylase 1 (ACOD1) and its metabolite itaconate (Ita) inhibit inflammation and OS by promoting alkylation of Keap1 to induce Nrf2 expression; however, it is unclear whether there is another pathway regulating their effects in inflammation-activated microglia after SCI.

Methods: Adult male C57BL/6 ACOD1^-/- mice and their wild-type (WT) littermates were subjected to a moderate thoracic spinal cord contusion. The degree of neuroinflammation and OS in the injured spinal cord were assessed using qPCR, western blot, flow cytometry, immunofluorescence, and trans-well assay. We then employed immunoprecipitation-western blot, chromatin immunoprecipitation (ChIP)-PCR, dual-luciferase assay, and immunofluorescence-confocal imaging to examine the molecular mechanisms of ACOD1. Finally, the locomotor function was evaluated with the Basso Mouse Scale and footprint assay.

Results: Both in vitro and in vivo, microglia with transcriptional blockage of ACOD1 exhibited more severe levels of neuroinflammation and OS, in which the expression of p62/Keap1/Nrf2 was down-regulated. Furthermore, silencing ACOD1 exacerbated neurological dysfunction in SCI mice. Administration of exogenous Ita or 4-octyl itaconate reduced p62 phosphorylation. Besides, ACOD1 was capable of interacting with phosphorylated p62 to enhance Nrf2 activation, which in turn further promoted transcription of ACOD1.

Conclusions: Here, we identified an unreported ACOD1-p62-Nrf2-ACOD1 feedback loop exerting anti-inflammatory and anti-OS in inflammatory microglia, and demonstrated the neuroprotective role of ACOD1 after SCI, which was different from that of endogenous and exogenous Ita. The present study extends the functions of ACOD1 and uncovers marked property differences between endogenous and exogenous Ita.

Key points: ACOD1 attenuated neuroinflammation and oxidative stress after spinal cord injury. ACOD1, not itaconate, interacted with p-p62 to facilitate Nrf2 expression and nuclear translocation. Nrf2 was capable of promoting ACOD1 transcription in microglia.

Keywords: Nrf2; aconitate decarboxylase 1; itaconate; neuroinflammation; p62; spinal cord injury.

FIGURE 1

Differential gene expression and functional enrichment analysis. (A) Uniform manifold approximation and projection (UMAP) analysis plot of normalized gene chip GSE5296. (B) Volcano plot of the results of the differential gene expression analysis. (C) Volcano plot of differential expression of gene set variation analysis (GSVA); colors represent different biological process. (D) Heatmap of trend clustering results; the left part shows the trend clustering results; the middle part shows the heatmap of the expression of the genes in the four clusters (C1, C2, C3 and C4); and the right part is divided into three columns from left to right: the first column is the number of genes included in each cluster (C1, C2, C3 and C4); the second column is the biological process of each cluster; and the third column is the −log10 p value of the corresponding biological process term.

FIGURE 2

Neural debris and SCI induce increased expression of ACOD1 and Nrf2 during neuroinflammation. (A) Western blotting of iNOS and ACOD1 expression in microglia treated with LPS (1 μg/mL) and debris (1 mg/mL or 2 mg/mL) for 24 h. (B,C) Densitometric analysis of iNOS and ACOD1 expression. (D) Western blotting of iNOS and ACOD1 expression in microglia treated with debris (2 mg/mL) within 24 h. (E,F) Densitometric analysis of iNOS and ACOD1 expression. (G) Relative mRNA level of Nrf2 in the spinal cord within 24 h post‐SCI. (H) Relative mRNA level of ACOD1 in the spinal cord within 24 h post‐SCI. (I) Western blotting of Nrf2 and ACOD1 protein levels in the spinal cord within 24 h post‐SCI. (J,K) Densitometric analysis of Nrf2 and ACOD1 expressions. (L) The level of Ita in the spinal cord within 24 h post‐SCI. (M) Representative immunofluorescence labelling images of IBA‐1 (green) and Nrf2 (red) in the spinal cord within 24 h post‐SCI; scale bar = 50 μm. Data are representative of at least three biological replicates. Data are shown as mean ± SEM, and statistical significance was determined with one‐way ANOVA followed by the Tukey's post hoc test. *, p < .05, **, p < .01, ***, p < .001. SCI, Spinal cord injury; ACOD1, Aconitate decarboxylase 1; LPS, lipopolysaccharide.

FIGURE 3

ACOD1 loss results in cumulative dysfunction of the redox reaction in microglia. (A) Western blotting of ACOD1, NOX1 and NOX4 expression in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i. (B–D) Densitometric analysis of ACOD1, NOX1 and NOX4 expression. (E) Relative GSSG levels in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i. (F) Relative GSH levels in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i. (G) Flow cytometry and analysis of ROS levels in microglia. (H) Quantitative analysis of ROS expression. (I) Relative MDA content in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i. (J) Representative immunofluorescence labelling images of DCF (green) and DHE (red) in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i; Scale bar = 70 μm. (K) Representative immunofluorescence labelling images of IBA‐1 (green), NOX4 (red) and NOX1 (pink) obtained from longitudinal sections centred around the injured core 5 mm in WT and ACOD1^−/− mice at 3 days post‐SCI; scale bar a = 500 μm, b = 50 μm. Data are representative of at least three biological replicates. Data are shown as mean ± SEM, and statistical significance was determined with one‐way ANOVA followed by Tukey's post hoc test. *, p < .05, **, p < .01, ***, p < .001, n.s, No significance; ROI, region of interest; SCI, spinal cord injury; ACOD1, aconitate decarboxylase 1; ROS, reactive oxygen species.

FIGURE 4

ACOD1 loss increases neuroinflammation and acceleration of glial accumulation. (A) Western blotting of iNOS and COX2 in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i. (B,C) Densitometric analysis of iNOS and COX‐2 expression. (D,E) Relative mRNA levels of TNF‐α and IL‐1β in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i. (F) Representative immunofluorescence labelling images of CD68 (green) and iNOS (red) in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i; scale bar = 30 μm. (G) Representative immunofluorescence labelling images of IBA‐1 (green) and iNOS (red) obtained from longitudinal sections centred around the injured core 5 mm in WT and ACOD1^−/− mice at 3 days post‐SCI; scale bar a = 500 μm, b = 50 μm. (H) Representative images of astrocytes stained with crystal violet in trans‐well assay, which were treated with microglial medium for 72 h. (I) Quantitative analysis of the amounts of astrocytes under inserts. (J) Representative immunofluorescence labelling images for GFAP (green) and ACAN (red) obtained from longitudinal sections centred around the injured core 5 mm in WT and ACOD1^−/− mice at 7 days post‐SCI; scale bar = 50 μm. (K) Coculture of primary microglia and astrocytes. (L) Representative immunofluorescence labelling images of GFAP (green) and ACAN (red) in astrocytes co‐cultured with debris‐stimulated microglia for 24 h after transfection with ACOD1i; scale bar = 100 μm. Data are representative of at least three biological replicates. Data are shown as mean ± SEM, and statistical significance was determined with one‐way ANOVA followed by Tukey's post hoc test. *, p < .05, **, p < .01, ***, p < .001, ns, No significance; ROI, region of interest; SCI, spinal cord injury; ACOD1, aconitate decarboxylase 1; WT, wild‐type; ACAN, aggrecan.

FIGURE 5

Histological collapse and locomotor dysfunction become worse in ACOD1^−/− mice after SCI. (A) Representative immunofluorescence labelling images of GFAP (green), NF200 (red) and IBA‐1 (pink) obtained from longitudinal sections centred around the injured core 5 mm in WT and ACOD1^−/− mice at 7 days post‐SCI; scale bar a = 500 μm, b = 50 μm. (B) Quantitative analysis of the area of microglial scar at 7 days post‐SCI. (C) Quantitative analysis of the astroglial scar at 7 days post SCI. (D) Quantitative analysis of the axonal numbers at 7 days post‐SCI. (E) Representative immunofluorescence labelling images of GFAP (green), NF200 (red) and IBA‐1 (pink) obtained from longitudinal sections centred around the injured core 5 mm in WT and ACOD1^−/− mice at 28 days post SCI; scale bar a = 500 μm, b = 50 μm. (F) Quantitative analysis of the area of the microglial scar at 28 days post‐SCI. (G) Quantitative analysis of the area of the astroglial scar at 28 days post‐SCI. (H) Quantitative analysis of the axon area at 28 dpi. (I) A footprint analysis of WT and ACOD1^−/− mice performed at 7 days post‐SCI. (J,M) Quantification of the footprint analysis at 7 and 28 days post SCI in WT and ACOD1^−/− mice. (N) The BMS score within 28 days post‐SCI in WT and ACOD1^−/− mice. Data are representative of at least three biological replicates. Data are shown as mean ± SEM, and statistical significance was determined with one‐way ANOVA followed by Tukey's post hoc test. *, p < .05, **, p < .01, ***, p < .001, ns, no significance; ROI, region of interest; SCI, spinal cord injury; ACOD1, aconitate decarboxylase 1; WT, wild‐type.

FIGURE 6

ACOD1 promotes Keap1/Nrf2 complex disruption by activating p62 phosphorylation at Ser351. (A) Western blotting of p‐p62, p62, Keap1, Nrf2 and HO‐1 expression in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i. (B–F) Densitometric analysis of p‐p62, p62, Keap1, Nrf2 and HO‐1 expression. (G) The colocalisation of Nrf2 and Keap1 in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i was evaluated by immunofluorescence; scale bar a = 20 μm, b = 2 μm. (H,I) Quantitative analysis of Keap1–Nrf2 complex number. (J) The colocalisation of Nrf2 and DAPI in microglia treated with debris (2 mg/mL) for 24 h after transfection with ACOD1i was evaluated by immunofluorescence; scale bar a = 20 μm, b = 2 μm. (K) Western blotting of ACOD1, p‐p62, Nrf2 and HO‐1 expression in WT and ACOD1^−/− mice at 3 h post‐SCI. (L–O) Densitometric analysis of ACOD1, p‐p62, Nrf2 and HO‐1 expression. (P) Representative immunofluorescence labelling images of IBA‐1 (green) and p‐p62 (red) in the spinal cords of WT and ACOD1^−/− mice at 3 h post‐SCI; scale bar = 50 μm. (Q) Representative immunofluorescence labelling images of IBA‐1 (green) and Nrf2 (red) in the spinal cords of WT and ACOD1^−/− mice at 3 h post‐SCI; scale bar = 50 μm. Data are representative of at least three biological replicates. Data are shown as mean ± SEM, and statistical significance was determined with one‐way ANOVA followed by Tukey's post hoc test. *, p < .05, **, p < .01, ***, p < .001, ns, no significance; ROI, region of interest; SCI, spinal cord injury; ACOD1, aconitate decarboxylase 1; WT, wild‐type.

FIGURE 7

A positive feedback loop of ACOD1/p‐p62/Nrf2/ACOD1 was identified in microglia during neuroinflammation. (A) Co‐IP assay showing representative protein bands of p‐p62, Nrf2, Keap1 and ACOD1 in microglia after using an Ab against ACOD1. (B) Co‐IP assay showing representative protein bands of p‐p62 and ACOD1 in microglia treated with debris (2 mg/mL) or not using an Ab against ACOD1. (C) Co‐IP assay showing representative protein bands of p‐p62 and ACOD1 in WT or Mut ser using an Ab against ACOD1. (D) Representative protein bands of ubiquitin bound to Nrf2 using the Co‐IP assay. (E) Representative protein bands of ubiquitin linkage‐specific K63 bound to Nrf2 using the Co‐IP assay. (F) Representative protein bands of ubiquitin linkage‐specific K48 bound to Nrf2 using the Co‐IP assay. (G) Western blotting of Nrf2 and ACOD1 in microglia pretreated with NK252 or Nrf2‐IN‐1 and then treated with debris (2 mg/mL) for 24 h. (H,I) Densitometric analysis of Nrf2 and ACOD1 expression. (J) Co‐IP assay showing representative protein bands of Nrf2 and ACOD1 in microglia. (K) Seven binding sites between Nrf2 and the ACOD1 promoter using the Jaspar website. (l) ChIP assay shows binding at predicted binding sites between Nrf2 and ACOD1 promoters. (M) Densitometric analysis of ACOD1 expression. (N) The luciferase activity of ACOD1 treated with Nrf2 overexpression or not after mutation in the 6th promoter site 1st–7th promoter sites. Data are shown as mean ± SEM, and statistical significance was determined with one‐way ANOVA followed by Tukey's post hoc test. *, p < .05, **, p < .01, ***, p < .001. ACOD1, aconitate decarboxylase 1; WT, wild‐type; ChIP, chromatin immunoprecipitation.