Demyelination causes adult CNS progenitors to revert to an immature state and express immune cues that support their migration

Abstract

The declining efficiency of myelin regeneration in individuals with multiple sclerosis has stimulated a search for ways by which it might be therapeutically enhanced. Here we have used gene expression profiling on purified murine oligodendrocyte progenitor cells (OPCs), the remyelinating cells of the adult CNS, to obtain a comprehensive picture of how they become activated after demyelination and how this enables them to contribute to remyelination. We find that adult OPCs have a transcriptome more similar to that of oligodendrocytes than to neonatal OPCs, but revert to a neonatal-like transcriptome when activated. Part of the activation response involves increased expression of two genes of the innate immune system, IL1β and CCL2, which enhance the mobilization of OPCs. Our results add a new dimension to the role of the innate immune system in CNS regeneration, revealing how OPCs themselves contribute to the postinjury inflammatory milieu by producing cytokines that directly enhance their repopulation of areas of demyelination and hence their ability to contribute to remyelination.

Keywords: Oligodendrocyte progenitor cells; cytokines; migration; multiple sclerosis; remyelination.

Figure 1.

Flow cytometry sorting of OPCs and OLs. a, Coronal section of a control adult PDGFαR::GFP brain showing homogeneously distributed GFP-positive aOPCs. Scale bar, 200 μm. b, GFP-positive cells are also expressing NG2. Scale bar, 50 μm. c, Coronal section of a demyelinated brain treated with cuprizone for 5 weeks, showing increased GFP-positive cell density within the demyelinated areas (lack of MBP staining). Scale bar, 200 μm. d, GFP-positive cells expressing NG2 on a cuprizone-treated brain section. Scale bar, 50 μm. e, f, h, Neonatal OPCs (e), and adult OPCs from control (f) and from demyelinated conditions (h) are sorted by flow cytometry from PDGFαR::GFP brains. g, Mature OLs are isolated from PLP-GFP brains. All sorted cells are GFP positive and PI negative. i–l, Flow cytometry analysis of O4 expression in neonatal OPCs (i), adult OPCs from control (j), demyelinated brains (l) and mature OLs (k). CC, Corpus callosum.

Figure 2.

In vitro characterization of the sorted cell populations. a, b, Immunolabeling on sorted GFP-positive cells isolated from PDGFαR::GFP brains, 60 min after cell platting (a) or after 2 d in culture (b). Sorted aOPCs express NG2 and O4, as well the mature marker MBP. Scale bar, 50 μm. c, Triple staining with GFP/NG2/MBP. d, Immunolabeling on GFP-positive cells isolated from PLP-GFP brains, after 2 d in culture. Only a percentage of sorted aOLs express O4, whereas almost all express MBP. NG2 or PDGFαR expressions are not detected on aOLs. Scale bar, 50 μm.

Figure 3.

mRNA profile of the sorted populations. a, Microarray analysis restricted to oligodendroglial genes showing the different profiles of each population. b, Hierarchical clustering of all genes (dendogram): each column represents the gene expression of one replicate (Pearson correlation, using Multi-Experiment Viewer). Volcano plot (x-axis = Log2 ratio activated vs nonactivated aOPCs; y-axis = −Log10 p value) shows the changes induced by demyelination in aOPCs, and colored dots point to some known oligodendroglial markers. c, On the right are genes overexpressed, and on the left genes underexpressed by activated aOPCs.

Figure 4.

In vitro functional changes in activated aOPCs. a, b, After 2 d in vitro, the proportions of activated and nonactivated aOPCs undergoing proliferation (a) and apoptosis (b) are similar. c, In vitro vertical migration assessed during a 2 d period, showing at 24 and 48 h a 1.3-fold increased migration of activated aOPCs compared with nonactivated aOPCs (n = 6; paired Student's t test). Differentiation is assessed using morphological classification (stages 1–5; adapted from Huang et al., 2011; schematic representation on the left). d, After 3 d, activated aOPCs are more differentiated than nonactivated aOPCs, with a higher proportion of cells in stages 4 and 5, a pattern ressembling nOPC differentiation (n = 3; paired Student's t test, ***p < 0.001, **p < 0.005, *p < 0.05).

Figure 5.

Biostatistical analysis to identify genes of interest. a, Gene ontology performed on GOrilla software for all genes differentially expressed between activated aOPCs and nonactivated aOPCs. b, Gene ontology for genes differentially expressed in activated aOPCs versus those in the nonactivated aOPC database and in the reported early repair database of caudal cerebellar peduncle (CCP) lesions comparing 14 versus 5 dpl. c, Representation of the 119 genes differentially expressed in both databases. d, Identification of a group of seven interacting genes, using Ariadne Genomics–Pathway Studio. e, qPCR detection of CCL2, CCR2, IL1β, and IL1R1 on activated aOPCs and nonactivated aOPCs. qPCR showing CCL2 and IL1β expression in activated aOPCs (n = 4; Student's t test, ***p < 0.001).

Figure 6.

In vivo Il1β and Ccl2 expression by aOPCs in control and cuprizone-treated PDGFαR:GFP adult mice. a, Immunostaining on coronal brain sections of control and cuprizone-treated PDGFαR::GFP adult mice showing increased numbers of aOPCs (GFP-positive cells) in the demyelinated area (corpus callosum) associated with increased expression of Il1β and Ccl2. Scale bar, 200 μm. b, Higher-magnification image showing GFP-positive aOPCs expressing Il1β and Ccl2 (white arrowheads). Scale bar, 50 μm. c, A GFP-positive cell expressing Ccl2. Scale bar, 10 μm. d, Quantification of the percentage of aOPCS expressing Ccl2 showing a 2.5-fold increase after demyelination (n = 5; Student's t test ***p < 0.001). e, ELISA performed on lysates and supernatants of purified aOPCs from control and demyelinated brains showing a twofold increase of Ccl2 secretion by activated aOPCs compared with nonactivated aOPCs (Student's t test, *p < 0.05).

Figure 7.

Influence of Il1β and Ccl2 on aOPC migration in vitro. a, b, Soluble Ccl2 (20 ng/ml) and Il1β (5 ng/ml) increase aOPC migration, compared with control in a vertical migration system. This effect is blocked by the respective antagonists INCB (for Ccl2, 8 μm) and Il1ra1 (for Il1β, 200 ng/ml; Student's t test, *p < 0.01, **p < 0.005, and ***p < 0.001). c, d, Horizontal migration assessed by video microscopy (each colored spot represents a single cell track, c) showing that nonactivated aOPCs treated by soluble Ccl2 or Il1β become as mobile as activated aOPCs (treated or not with Ccl2 or Il1β; d). e, f, A similar increase is induced by lentiviral-mediated overexpression of Ccl2 (aOPCs-Ccl2) compared with control vector (aOPCs-DsRed; n = 3; Student's t test, *p < 0.05, **p < 0.01, and ***p < 0.002).

Figure 8.

Influence of Ccl2 on aOPC migration in vivo. a, Four days after transduction, CG4 transduced with the control (CG4-DsRed) or with the Ccl2 lentivirus (CG4-Ccl2, tagged with mCherry) were sorted. b, CG4 cells were plated and immunostained 4 d after transduction. Scale bar, 50 μm. Nontransduced and transduced CG4 cells express immature markers PDGFαR, A2B5, and O4, but not the mature marker MBP. c, qPCR detection of CCL2 on nontransduced CG4 cells and transduced CG4 cells (CG4-DSRed or CG4-Ccl2). CCL2, expressed at a low level in nontransduced cells, is increased after transduction with the Ccl2 lentivirus (n = 3; Student's t test, ***p < 0.001, *p < 0.05). Sorted cells are injected in P2 PDGFαR:GFP brains. d, The circles represent distances. Scale bar, 200 μm. Red arrow represents the injection track, the inner circle represents the injection site. CC, Corpus callosum. e, No difference in the distance of migration [dorsoventral (DV), anteroposterior (AP), and lateral (Lateral)] was detected between CG4-DsRed and CG4-Ccl2 cells. f, g, Quantification of the percentage of transduced CG4 cells, which have migrated at different distances from the injection site: CG4-Ccl2 cells are more migratory compared to CG4-DsRed cells (f), without difference in the percentage of proliferating cells (g; n = 3; Student's t test, **p < 0.01, *p < 0.05).

Figure 9.

a, b, Ccl2 immunostaining on MS tissues. OPCs (nuclear Olig1 staining) expressing Ccl2 (arrowhead) in an active lesion (a) and active borders of a chronic lesion (b). c, d, Virtually no Ccl2-expressing OPCs in a chronic lesion (c) or NAWM area (d). Scale bar, 50 μm. e, High-magnification image of one OPCs, stained by nuclear Olig1 antibody (red), expressing Ccl2 (green). Scale bar, 10 μm. f, Ratio of the percentage of nuclear Olig1-positive cells expressing Ccl2 in multiple sclerosis lesions, compared to adjacent NAWM of the same size (n = 3–6, depending on the type of lesion; one-way ANOVA and Holm–Sidak multiple-comparisons test, *p < 0.05).