miR-142-3p Is a Key Regulator of IL-1β-Dependent Synaptopathy in Neuroinflammation

Abstract

MicroRNAs (miRNA) play an important role in post-transcriptional gene regulation of several physiological and pathological processes. In multiple sclerosis (MS), a chronic inflammatory and degenerative disease of the CNS, and in its mouse model, the experimental autoimmune encephalomyelitis (EAE), miRNA dysregulation has been mainly related to immune system dysfunction and white matter (WM) pathology. However, little is known about their role in gray matter pathology. Here, we explored miRNA involvement in the inflammation-driven alterations of synaptic structure and function, collectively known as synaptopathy, a neuropathological process contributing to excitotoxic neurodegeneration in MS/EAE. Particularly, we observed that miR-142-3p is increased in the CSF of patients with active MS and in EAE brains. We propose miR-142-3p as a molecular mediator of the IL-1β-dependent downregulation of the glial glutamate-aspartate transporter (GLAST), which causes an enhancement of the glutamatergic transmission in the EAE cerebellum. The synaptic abnormalities mediated by IL-1β and the clinical and neuropathological manifestations of EAE disappeared in miR-142 knock-out mice. Furthermore, we observed that in vivo miR-142-3p inhibition, either by a preventive and local treatment or by a therapeutic and systemic strategy, abolished IL-1β- and GLAST-dependent synaptopathy in EAE wild-type mice. Consistently, miR-142-3p was responsible for the glutamatergic synaptic alterations caused by CSF of patients with MS, and CSF levels of miR-142-3p correlated with prospective MS disease progression. Our findings highlight miR-142-3p as key molecular player in IL-1β-mediated synaptic dysfunction, possibly leading to excitotoxic damage in both EAE and MS diseases. Inhibition of miR-142-3p could be neuroprotective in MS.

Significance statement: Current studies suggest the role of glutamate excitotoxicity in the development and progression of multiple sclerosis (MS) and of its mouse model experimental autoimmune encephalomyelitis (EAE). The molecular mechanisms linking inflammation and synaptic alterations in MS/EAE are still unknown. Here, we identified miR-142-3p as a determinant molecular actor in inflammation-dependent synaptopathy typical of both MS and EAE. miR-142-3p was upregulated in the CSF of MS patients and in EAE cerebellum. Inhibition of miR-142-3p, locally in EAE brain and in a MS chimeric ex vivo model, recovered glutamatergic synaptic enhancement typical of EAE/MS. We proved that miR-142-3p promoted the IL-1β-dependent glutamate dysfunction by targeting glutamate-aspartate transporter (GLAST), a crucial glial transporter involved in glutamate homeostasis. Finally, we suggest miR-142-3p as a negative prognostic factor in patients with relapsing-remitting multiple sclerosis.

Keywords: CSF; experimental autoimmune encephalomyelitis; glial glutamate transporter; glutamate excitotoxicity; microRNA; multiple sclerosis.

Copyright © 2017 the authors 0270-6474/17/370547-16$15.00/0.

Figures

Figure 1.

miR-142-3p is highly expressed in the EAE cerebellum, in particular in the inflammatory lesions of the WM. A, Two-color microarray analysis of miRNA differential expression in the EAE cerebellum (n = 2, 21 dpi). The plot represents relative change values, expressed as log2-ratio (EAE vs CFA), plotted against average log intensity ((log2Hy5 + log2Hy3)/2). Only miRNAs with 9 < log intensity < 15 (into the dotted gray lines) and with a log2-ratio > 0.5 or < −0.5 (out of the gray rectangle) were considered, to avoid nonlinear effects caused by the noise floor at low intensities or by saturation at high intensities. Black arrow indicates miR-142-3p in the plot. B, miR-142-3p quantification by qRT-PCR in the EAE cerebellum (CFA, n = 4; EAE, n = 5). Data were normalized to U6B by ΔΔCt calculation (mean ± SEM vs controls). *p < 0.05 (unpaired t test). C–E, ISH with miR-142-3p probe in sagittal cerebellar sections of both EAE and CFA mice. C, Low magnification of cerebellar slices derived from EAE and CFA showing ISH with miR-142-3p probe and with scramble (scr) probe. Of note, miR-142-3p staining was prominent in EAE relative to CFA slices, whereas scr staining on adjacent slices was almost undetectable. D, High magnification of cerebellar lobules showing higher ISH reactivity of miR-142-3p (blue) in WM lesions of EAE cerebellum compared with CFA (top left panels, respectively). The adjacent serial sections in each experimental group represent immunostaining (brown) of infiltrating lymphocytes (CD3, top right), microglial cells (Iba1, bottom left), and astroglial cells (GFAP, bottom right). A remarkable inflammatory reaction is evident in the WM lesions of EAE compared with CFA mice. Negative controls with scr probe have been included. E, Higher magnifications of D reveal that in EAE cerebellum miR-142-3p is enriched in proximity of positive immunostaining (brown) for CD3, Iba1, and GFAP, as indicated by arrows in the adjacent serial sections of EAE mice. Scale bars: C, D, 200 μm; E, 100 μm.

Figure 2.

miR-142-3p is overexpressed in immune cells following inflammatory stimulation. A, miR-142-3p quantification by qRT-PCR in the EAE CD3+ lymphocytes (CFA, n = 6; EAE, n = 5). B, qRT-PCR of miR-142-3p in primary cerebellar glial cells (ctr, TNF: n = 9; ctr, IL-1β: n = 6; ctr: n = 8, LPS: n = 6) after 24 h treatment with proinflammatory cytokines or LPS. Data were normalized to U6B by ΔΔCt calculation (mean ± SEM vs controls). *p < 0.05 (unpaired t test). **p < 0.01 (unpaired t test). C–E, ISH of miR-142-3p in primary cerebellar glial cells. C, Low magnification of glial cultures showing higher ISH staining of miR-142-3p (gray) in IL-1β-treated glial cells relative to control condition and scramble (scr) probe hybridization. D, E, Confocal images of ISH with miR-142-3p (gray) and scr probes coupled with immunofluorescence of Iba1 (red) and GFAP (green) in IL-1β-treated glial cells. Cyan represents DAPI counterstaining. Both Iba1-positive microglial cells in D and GFAP-positive astroglial cells in E are positive for miR-142-3p following IL-1β treatment (left panels, respectively). Control scr staining is shown in the relative right panels. Scale bars: C, 20 μm; D, E, 10 μm.

Figure 3.

Alteration of miR-142-3p levels in the CSF of MS patients with active brain lesions. A, B, Dot plots of miR-142-3p levels detected by qRT-PCR in CSF from healthy subjects (Ctr) and patients with RRMS (A, Ctr, n = 12; RRMS, n = 30; Mann–Whitney test, p > 0.05), later divided in patients with (active RRMS) and without (nonactive RRMS) active inflammatory brain lesions (B, Ctr, n = 12; active RRMS, n = 18; nonactive RRMS, n = 12; Kruskal–Wallis test followed by Mann–Whitney). *p < 0.05. Values were normalized to spiked-in cel-miR-39 using the ΔCt calculation. Error bars indicate mean ± SEM. C, D, Correlation plot between ranks of the miR-142-3p levels in the CSF and the age at the withdrawal (C) or disease duration (D) of patients with RRMS (p > 0.05, Spearman correlation).

Figure 4.

miR-142 KO mice are protected from both EAE motor symptoms and EAE glutamatergic alterations. A, Time course of clinical score in EAE WT littermates (black; WT) and EAE miR-142 KO (gray) mice (at least 8 mice per group). Data are from one representative immunization (n immunization = 2). B, Representative immunostaining of cerebellar sagittal sections from EAE WT, EAE miR-142 KO, and CFA miR-142 KO mice detecting Iba1 (red), CD3 (green), and IL-1β (cyan) as inflammatory markers upregulated in EAE (EAE-WT). EAE miR-142 KO mice do not show any sign of inflammation as in control condition. Gray represents DAPI counterstaining. Scale bar, 100 μm. C, Whole-cell patch-clamp recordings from PCs of sEPSCs in EAE WT and EAE miR-142 KO mice relative to controls. Decay time and half-width are represented in the graphs (CFA WT, n = 7; EAE WT, n = 9; CFA KO, n = 10; EAE KO, n = 9; one-way ANOVA followed by Tukey's HSD). **p < 0.01 versus CFA. ***p < 0.001 versus CFA. #p < 0.05 versus EAE WT. ##p < 0.01 versus EAE WT. Top, Examples of electrophysiological traces (sEPSCs) recorded from PCs in the different experimental conditions.

Figure 5.

Preventive and local inhibition of miR-142-3p in the CNS ameliorates glutamatergic synaptic alterations in EAE cerebellum. A, Clinical course of EAE in mice receiving continuous intracerebroventricular infusion of anti-miR-142-3p (gray) compared with mice receiving scramble (black). B, Whole-cell patch-clamp recordings from PCs of EAE mice (21–28 dpi) treated intracerebroventricularly with anti-miR-142-3p or scramble (score ≥ 2). Histograms represent glutamatergic sEPSC kinetic properties (decay time and half-width) of the two experimental groups after 28 d anti-miR-142-3p intracerebroventricular treatment to inhibit miR-142-3p or scramble administration (EAE scr, n = 20; EAE anti-miR-142-3p, n = 26). **p < 0.01 (unpaired Student's t test). Dotted lines indicate the mean values obtained in CFA-untreated mice. The electrophysiological traces on the right are examples of sEPSC mean peak obtained by group analysis in the different experimental conditions. C, Representative immunostaining of cerebellar sagittal sections from EAE WT mice treated with miR-142-3p inhibitor (anti-miR-142-3p) or scramble control (scr), detecting infiltrating CD3+-T cells (green), Iba1 (red), and IL-1β (cyan) as inflammatory markers upregulated in EAE. After 28 d-anti-miR-142-3p intracerebroventricular treatment, the local cerebellar inflammation was similar to control EAE conditions (EAE scr). Gray represents DAPI counterstaining. Scale bar, 50 μm.

Figure 6.

Therapeutic and systemic inhibition of miR-142-3p ameliorates glutamatergic synaptic alterations in EAE cerebellum. A, Clinical course of EAE mice receiving intravenous injection, repeated every 4 d, of anti-miR-142-3p (gray) and scramble (black). B, Whole-cell patch-clamp recordings from PCs of EAE mice (21–28 dpi) treated intravenously with anti-miR-142-3p or scramble (score ≥ 2). Histograms represent the sEPSC kinetic properties (decay time and half-width) of PCs in each experimental group. There is a recovery effect of the anti-miR-142-3p intravenous treatment (EAE scr, n = 11; EAE anti-miR-142-3p, n = 14). *p < 0.05 (unpaired Student's t test). Dotted lines indicate the mean values obtained in CFA-untreated mice. Bottom, Electrophysiological traces represent examples of sEPSC mean peak obtained by group analysis in the different experimental conditions. C, Representative immunostaining of cerebellar sagittal sections from EAE WT mice treated with intravenous injection of anti-miR-142-3p or scramble control (scr), detecting infiltrating CD3+-T cells (green), Iba1 (red), and IL-1β (cyan) as inflammatory markers upregulated in EAE. The local cerebellar inflammation was similar between control EAE conditions (EAE scr) and EAE-treated mice (EAE_anti-miR-142-3p). Gray represents DAPI counterstaining. Scale bar, 50 μm.

Figure 7.

miR-142-3p regulates the Slc1a3 mRNA coding for GLAST. A, Venn diagram of the mRNAs identified as candidate targets of miR-142-3p by at least 2 out of 3 prediction algorithms (microT version 3.0, TargetScan 5.2, and PicTar) and functionally annotated using DAVID 6.7. Slc1a3 is among the potential targets specifically expressed in the cerebellum. B, Alignment of the potential binding site for miR-142-3p in the 3′-UTR of Slc1a3 mRNA from different species (mmu, mouse; rno, rat; ocu, rabbit; mml, rhesus; ptr, chimpanzee; hsa, human), as predicted by TargetScan version 6.2 and DIANA micro-T version 3.0 algorithms. Bold indicates miR-142-3p “seed” region. Remarkable mRNA functional elements are indicated. C, Quantification of the Slc1a3 mRNA coding for the glial glutamate transporter GLAST in EAE cerebellum. β-actin was used as endogenous control. All reactions were analyzed by using the ΔΔCt calculation. Values are mean ± SEM versus CFA mice (CFA, n = 4; EAE, n = 5; p = 0.699, unpaired Student's t test). Data were pooled from two independent immunizations. D, Histograms show luciferase assay results validating Slc1a3 mRNA as direct target of miR-142-3p. The dual-luciferase control plasmid (Luc), the reporter construct containing full-length Slc1a3 3′-UTR (Luc- 3′-UTR WT), or the 3′-UTR bearing mutated (ACU→UAC) binding site for miR-142-3p (Luc-3′-UTR MUT) were cotransfected with miR-142 coding plasmid (pU1-miR-142/CMV-GFP, reported in figure as pU1-miR-142 for simplicity) in HEK293T cells, and the Firefly luciferase activity was measured 48 h after (normalized to Renilla luciferase). The empty vector was used as negative control. Data are mean ± SD and are normalized to empty vector (n = 6). ***p < 0.001 (unpaired t test).

Figure 8.

miR-142-3p is one of the effectors of the IL-1β-mediated synaptopathy in the cerebellum. A, Mature miR-142-3p quantification by qRT-PCR in cerebellar slices from healthy mice incubated with IL-1β (normalized to U6B using the ΔΔCt calculation). Values are mean ± SEM versus slices incubated only in ACSF (n = 15; IL-1β, n = 14). *p < 0.05 (unpaired Student's t test). B, Whole-cell patch-clamp recordings from PCs of WT and miR-142 KO slices incubated with IL-1β. Histograms represent decay time (left) and half-width (right) of PC sEPSC in WT and miR-142 KO slices upon 10 min-IL-1β incubation (miR-142 KO, n = 10; WT, n = 6). *p < 0.05 versus predrug value (paired Student's t test). In miR-142 KO slices, IL-1β does not reproduces the synaptic alterations observed in WT slices. Dotted lines indicate predrug sEPSC values. Right, Representative sEPSC mean peaks of the two experimental conditions. C, D, Representative WB images and quantification of GLAST protein levels (normalized to β-actin) in cerebellar slices from WT (C, ACSF, n = 3; IL-1β, n = 4). *p < 0.05 (unpaired Student's t test) or miR-142 KO mice (D, ACSF, n = 5; IL-1β, n = 7; p > 0.05, unpaired Student's t test) after 10 min IL-1β incubation. E, GLAST protein levels are unchanged in the cerebellum of miR-142 KO mice (n = 7) compared with WT mice (n = 7). *p < 0.05 (unpaired Student's t test). F, Whole-cell patch-clamp recordings from PCs of WT and miR-142 KO slices incubated with EAE-CD3+ lymphocytes isolated from the spleens of EAE mice. Histograms represent decay time (left) and half-width (right) of sEPSCs recorded in WT (n = 13) and miR-142 KO (n = 23) slices in the presence of CD3+ lymphocytes. *p < 0.05 (one-way ANOVA test). **p < 0.01 (one-way ANOVA test). EAE T cells fail to affect sEPSCs in miR-142 KO slices. Dotted lines indicate control mean values of sEPSC obtained from WT and miR-142 KO slices.

Figure 9.

miR-142-3p is potentially involved in MS synaptopathy and in MS disease severity. A, Electrophysiological recordings conducted in MS chimeric ex vivo model. Histograms represent decay time and half-width of sEPSCs after 1 h incubation of murine cerebellar slices with a pool of CSFs taken from patients with active lesions (active RRMS) and high level of miR-142-3p in the absence or in the presence of miR-142-3p inhibitor (anti-miR-142-3p) (CSF from active RRMS in both experimental groups, n = 17). *p < 0.05 (unpaired Student's t test). Top, Examples of electrophysiological traces (sEPSCs) recorded from PCs in the different experimental conditions. These results suggest that miR-142-3p is an indicator of MS-related synaptopathy. B, Correlation plot between ranks of miR-142-3p levels in the CSF and the Progression Index (PI) of the disease in patients with RRMS (n = 21, p < 0.05 Spearman correlation). miR-142-3p is a potential negative prognostic factor in patients with RRMS.