Metabolic Control of Astrocyte Pathogenic Activity via cPLA2-MAVS

Abstract

Metabolism has been shown to control peripheral immunity, but little is known about its role in central nervous system (CNS) inflammation. Through a combination of proteomic, metabolomic, transcriptomic, and perturbation studies, we found that sphingolipid metabolism in astrocytes triggers the interaction of the C2 domain in cytosolic phospholipase A2 (cPLA2) with the CARD domain in mitochondrial antiviral signaling protein (MAVS), boosting NF-κB-driven transcriptional programs that promote CNS inflammation in experimental autoimmune encephalomyelitis (EAE) and, potentially, multiple sclerosis. cPLA2 recruitment to MAVS also disrupts MAVS-hexokinase 2 (HK2) interactions, decreasing HK enzymatic activity and the production of lactate involved in the metabolic support of neurons. Miglustat, a drug used to treat Gaucher and Niemann-Pick disease, suppresses astrocyte pathogenic activities and ameliorates EAE. Collectively, these findings define a novel immunometabolic mechanism that drives pro-inflammatory astrocyte activities, outlines a new role for MAVS in CNS inflammation, and identifies candidate targets for therapeutic intervention.

Keywords: MAVS; Miglustat; NF-κB; astrocytes; cPLA2; lactate; lactosylceramide; metabolism; multiple sclerosis; neuroinflammation.

Figure 1:. The LacCer biosynthetic pathway in astrocytes promotes NOD EAE progression.

(A) LacCer biosynthetic pathway. (B) Expression of sphingolipid metabolism-related genes in astrocytes during the progressive phase of NOD EAE. (n=5, N=2, unpaired T test to Naive) (C) Expression of sphingolipid metabolism-related genes in MS Lesions (n = 10) and MS NAWM (n = 5), or brain tissue from healthy individuals (n=6). (unpaired T test, compared to all) (D) EAE development in NOD mice treated with lentiviral constructs expressing shRNAs targeting B4galt4, Ugcg, Cers1, Cers2, Cers4, Cers5, Cers6 or control in astrocytes 37 and 44 days after EAE induction. (n=8, N=2, Regression slope T test, compared to shControl) (E-H) CNS samples were harvested 51 days after EAE induction from NOD EAE mice treated with lentivirus-delivered shRNAs as shown in Figure 1. (unpaired T test compared to shControl) (E) number of CNS-infiltrating inflammatory monocytes. (F) Axonal loss and demyelination in spinal cord. (G) Whole genome expression in astrocytes from control and knockdown mice. (n = 3 for each group) (H) mRNA expression determined by qPCR in astrocytes. (I-L) B4galt6, Ugcg, Cers2, Cers5 and Cers6 were knocked down in murine astrocytes in culture, which were activated with TNFα and IFNγ. (unpaired T test compared to shControl). mRNA expression was determined by qPCR in the astrocytes (I). Astrocyte-conditioned medium was tested using in vitro monocyte migration (J) and neurotoxicity assays (K). Alternatively, after washing astrocytes were co-cultured with microglia, and gene expression in microglia was determined by qPCR (L). Migrating monocytes and neuronal death in the shControl-treated group were set as 100%.

Figure 2:. LacCer boosts NF-κB driven inflammation via cPLA2 activation.

(A) Effect of LacCer on the enzymatic activity of recombinant human cPLA2 in a cell-free assay. (unpaired T test, compared to control) (B) Pla2g4a expression in astrocytes 51 days after NOD EAE induction, and PLA2G4A expression in CNS samples from MS patients and controls (The same set of samples as Figure 1C. (unpaired T test compared to Naïve or Control) (C) EAE development in NOD mice treated 39 and 47 days after EAE induction with lentiviral constructs expressing control or Pla2g4a targeting shRNAs in astrocytes. (n=7, N=2, Regression slope T test, compared to shControl) (D-G) CNS samples were harvested 51 days after EAE induction from mice treated with lentivirus-delivered shRNAs shown in Figure 2C. (unpaired T test, compared to shControl) (D) Axonal loss and demyelination in spinal cord. (E) the number CNS-infiltrating inflammatory monocytes. (F) mRNA expression in astrocytes determined by qPCR. (G) Whole genome expression in astrocytes from control or Pla2g4a knockdown mice (n=3) (H,I) Astrocyte conditioned medium was prepared for test in in vitro monocyte migration (unpaired T test compared to activated shControl group) (H) and neurotoxicity assays (I). Migrating monocytes and neuronal death in the resting shControl-treated group were set as 100%. (J) Pla2g4a was knocked down in astrocytes, which were then activated, and mRNA expression was determined by qPCR in microglia co-cultured with the astrocytes. (unpaired T test, compared to shControl) (K) The effect of LacCer on cPLA2 enzymatic activity in astrocytes. (unpaired T test, compared to all) (L) mRNA expression determined by qPCR in activated astrocytes in the presence of cPLA2i. (unpaired T test, compared to all) (M) mRNA expression determined by qPCR in human astrocytes activated in the presence of LacCer, cPLA2i or both. (unpaired T test, compared to all) (N) Effect of cPLA2 inhibition on NF-κB activation in astrocytes. The nucleus p65 level is analyzed by western blot. (unpaired T test, compared to all) (O) Effect of cPLA2 inhibition determined by ChIP assay on NF-κB recruitment to responsive elements in Ccl2, Csf2 and Nos2 promoters. (unpaired T test, compared to all)

Figure 3:. LacCer-induced cPLA2-MAVS signaling drives NF-κB dependent pro-inflammatory programs.

(A) Effect of LacCer and cPLA2 inhibition on MAVS oligomerization, analyzed by SDD-AGE and western blot. (unpaired T test, compared to all, intensity normalized to vehicle Naïve condition) (B) ROS levels in activated murine and data are shown relative to ROS levels in resting astrocytes. (unpaired T test, compared to all) (C) LacCer induces cPLA2 recruitment to mitochondria and co-localization with MAVS. Astrocytes were treated with 10 μM LacCer or vehicle for 4 hours then stained and analyzed by confocal microscopy. Bar plots depict the ratio of total cPLA2 co-localization with mitochondria, the ratio of total MAVS co-localization with mitochondria, and the ratio of cPLA2-MAVS co-localization in mitochondria. (n≥25 cells per group, unpaired T test, compared to vehicle) (D) MAVS-cPLA2 interaction in astrocyte is analyzed by CoIP assay with MAVS-specific antibody and western blot. (unpaired T test, compared to all, intensity normalized to vehicle naive condition) (E,F) Domain binding analysis for cPLA2-MAVS interaction. Full-length human cPLA2 together with flag-tagged human MAVS (Full length, W56A mutant, G67A/W68A/V69A mutant (AAA), M2–6L mutant or M2 isoform) (E); or human cPLA2 (full-length, phospholipid binding domain deficient or C2 domain deficient) together with full-length flag-tagged human MAVS (F) were co-expressed in HEK293 cells, protein complexes were pulled down with anti-Flag antibodies and analyzed by western blot. (G) The effect of cPLA2 overexpression in trigger MAVS oligomerization. cPLA2 (full-length and C2 domain only) and MAVS were co-expressed in HEK293 cells, mitochondria were isolated 24 hours later and analyzed by SDD-AGE and western blot. (H) The effect of cPLA2 overexpression in triggering MAVS-mediated NF-κB activation, analyzed by luciferase assay. (unpaired T test, compared to all) (I) Effect of MAVS on NF-κB activation in WT and Mavs knockout astrocytes. The level of nucleus p65 level is analyzed by western blot. (unpaired T test, compared to corresponding condition between WT and MAVS^−/− astrocytes) (J) mRNA expression determined by qPCR in activated WT and MAVS^−/− astrocytes. (unpaired T test, compared to corresponding condition between WT and MAVS^−/− astrocytes) (K) Recruitment of NF-κB to Ccl2, Csf2 and Nos2 promoters in murine WT and MAVS^−/− astrocytes in culture determined by ChIP assay as in Figure 2O. (unpaired T test, compared to all) (L) EAE development in NOD mice treated with lentiviral constructs expressing shRNAs targeting Mavs or control in astrocytes 37 and 43 days after EAE induction. (n≥7, N=3, Regression slope T test, compared to shControl). (M,N) CNS samples were harvested 50 days after EAE induction from mice treated with lentivirus-delivered shRNAs shown in Figure 3L. (unpaired T test, compared to shControl) (M) The number of CNS-infiltrating inflammatory monocytes. (N) Inflammatory genes expression in astrocytes from control and Mavs knockdown mice was detected by nCounter analysis system (n=4). (O, P) Astrocyte-conditioned medium was prepared from activated WT or MAVS^−/− astrocytes for test in in vitro monocyte migration. (O, unpaired T test, compared to all) and neurotoxicity assays (P, unpaired T test, compared to all). Migrating monocytes and neuronal death in the resting shControl-treated group were set as 100%.

Figure 4:. cPLA2-MAVS co-localize in astrocytes in MS.

(A) Immunostaining analysis of the co-localization (white arrowheads) of cPLA2 and mitochondria marker Tom20 in GFAP⁺ astrocytes in MS tissue. GFAP⁺ astrocytes in lesion, NAWM near lesion (20–50 μm from lesion) and NAWM far from lesion (>500 μm) (B) Immunostaining analysis of the co-localization (white arrowheads) of cPLA2 and MAVS in MS or HC tissue GFAP⁺ astrocytes. (C) Quantification of cPLA2⁺MAVS⁺ co-localization signal in GFAP⁺ and GFAP⁻ fields. (n = 30–54 cells per condition, unpaired T test, compared to all) HC = healthy control, WM = white matter, GM = gray matter, NAWM = normally appearing white matter, NAGM = normally appearing gray matter.

Figure 5:. cPLA2-MAVS signaling modulates astrocyte metabolism.

(A) Level change of glucose metabolism in activated astrocytes measured by metabolomic profiling. Green color indicates metabolites decreased by cPLA2i treatment, red color indicates increased metabolites, black color indicates metabolites for which no data were collected. (B, C) Effect of cPLA2 (cPLA2i) or MPCs (UK-5099) inhibition in astrocyte on lactate release (B, unpaired T test, compared to activated vehicle condition, normalized to naive vehicle condition) and pyruvate level in mitochondria (C, unpaired T test, compared to activated vehicle condition, normalized to naive vehicle condition). (D,E) Metabolomic profiling analysis of activated astrocytes in the presence of cPLA2i. (D) Effect of cPLA2 inhibition on saturated and unsaturated fatty acids level in activated astrocyte, shown relative to their levels on resting astrocytes. Each dot indicates a different fatty acid. (unpaired T test, compared to activated vehicle condition) (E) Metabolite categories altered by cPLA2i treatment. (F) Effect of methylglyoxal supply on lactate release by activated astrocytes. (unpaired T test, compared to activated control condition) (G) Effect of cPLA2 or MPCs inhibition in mitochondrial function of astrocyte analyzed by Mito Stress test assay. (unpaired T test, compared to all) (H,I) Effect of gene targeted siRNA knockdown in mitochondrial function of LacCer loaded astrocytes analyzed by Mito Stress test assay (H, unpaired T test, compared to all) and lactate release by astrocytes (I, unpaired T test, compared to naive siControl condition). (J) Effect of LacCer loading or cytokines stimulation on the interaction of MAVS with HK2 and cPLA2 in astrocyte, evaluated by pull-down assay and western blot. (K) Mitochondrial HK enzymatic activity in activated astrocytes. (unpaired T test, compared to vehicle condition) (L) MAVS binding domain analysis. Flag-tagged human MAVS (Full length, M2–6L, M2, AAA or W56A) was over-expressed in HEK293 cell and 24 hours later MAVS protein complexes were pulled down and analyzed by western blot. (M) Human cPLA2 (cPLA2-GFP or C2 domain-GFP respectively) were over-expressed in HEK293 cell and 24 hours later MAVS protein complexes were pulled down and analyzed by western blot. (N) Effect of cPLA2 or cPLA2 C2 domain overexpression in HEK293 cells on mitochondrial HK enzymatic activity, lactate release and mitochondrial pyruvate levels. (unpaired T test, compared to Control condition) (O) Effect of 2-DG treatment on lactate release and mitochondrial levels of pyruvate in astrocyte. (unpaired T test) (P) Lactate release by Bsg knockdown or control astrocytes. (unpaired T test) (Q) EAE development in C57BL/6 or NOD mice that received lentiviral constructs to knockdown Bsg in astrocytes at the time points indicated by the arrows. (n≥7, N = 2, Regression slope T test). (R) Axonal loss and demyelination in spinal cords from NOD EAE mice. (unpaired T test)

Figure 6.. Miglustat ameliorates chronic progressive NOD EAE.

(A) Miglustat levels in the CNS after oral administration of 600 mg/kg Miglustat. (unpaired T test) (B) EAE development in NOD mice treated with Miglustat (600 mg/kg administered orally before initiation of progressive phase, n≥6, N=3, Regression slope T test). (C-G) CNS samples were harvested 41 days after EAE induction from mice treated with Miglustat or vehicle as shown in Figure 6b. The number of CNS-infiltrating inflammatory monocytes (C, unpaired T test). Whole genome expression in astrocytes isolated from Miglustat treated NOD EAE mice (D; n = 6). Axonal loss and demyelination in spinal cord (E, unpaired T test). Immunoflourescence analysis of C3⁺GFAP⁺ astrocytes. Bar plots depict the number of C3⁺GFAP⁺ astrocyte within the observation field (F, unpaired T test, compared to EAE condition). Immunofluorescence analysis of NF-κB activation among MAVS⁺cPLA2⁺GFAP⁺ astrocytes (G, unpaired T test, compared to EAE condition). Bar plots depict the number of MAVS⁺cPLA2⁺acetyl-p65⁺GFAP⁺ astrocyte within the observation field. (H) mRNA expression determined by qPCR in activated human and mouse astrocytes in the presence of Miglustat. Data are shown relative to resting astrocytes. (unpaired T test) (I) Human and mouse astrocyte conditioned medium were analyzed using in vitro monocyte migration and neurotoxicity assays. Migrating monocytes and neuronal death in the resting vehicle-treated group were set as 100%. (unpaired T test) (J) mRNA expression in polarized microglia was determined by qPCR. Microglia was co-cultured with activated astrocytes. (unpaired T test) (K) Lactate release from activated astrocytes in the presence of Miglustat. (unpaired T test, compared to activated vehicle condition) (L) Effect of Miglustat treatment on cPLA2 enzymatic activity in astrocytes, evaluated by cPLA2 activity assay. (unpaired T test, compared to all)