Inhibition of Gli1 mobilizes endogenous neural stem cells for remyelination

Abstract

Enhancing repair of myelin is an important but still elusive therapeutic goal in many neurological disorders. In multiple sclerosis, an inflammatory demyelinating disease, endogenous remyelination does occur but is frequently insufficient to restore function. Both parenchymal oligodendrocyte progenitor cells and endogenous adult neural stem cells resident within the subventricular zone are known sources of remyelinating cells. Here we characterize the contribution to remyelination of a subset of adult neural stem cells, identified by their expression of Gli1, a transcriptional effector of the sonic hedgehog pathway. We show that these cells are recruited from the subventricular zone to populate demyelinated lesions in the forebrain but never enter healthy, white matter tracts. Unexpectedly, recruitment of this pool of neural stem cells, and their differentiation into oligodendrocytes, is significantly enhanced by genetic or pharmacological inhibition of Gli1. Importantly, complete inhibition of canonical hedgehog signalling was ineffective, indicating that the role of Gli1 both in augmenting hedgehog signalling and in retarding myelination is specialized. Indeed, inhibition of Gli1 improves the functional outcome in a relapsing/remitting model of experimental autoimmune encephalomyelitis and is neuroprotective. Thus, endogenous neural stem cells can be mobilized for the repair of demyelinated lesions by inhibiting Gli1, identifying a new therapeutic avenue for the treatment of demyelinating disorders.

Extended data 1. Gli1-expressing cells generate oligodendrocytes following demyelination

Additional markers used to analyze fate mapped Shh-responsive NSCs are shown. a, 2 weeks after removal from cuprizone diet, GFP-labeled cells in the CC of Gli1^CE/+ mice co-expressed the oligodendroglial progenitor markers PDGFRα, NG2 and Sox10, the mature oligodendrocyte marker CC1, and the astrocytic markers GFAP and S100β. Scale bar, 10 μm. b, 4 weeks after removal from cuprizone diet, GFP-labeled processes co-localized with myelin proteins MBP and MOG but not with peripheral myelin protein P0. GFP-labeled processes also overlaid the axonal, paranodal marker Caspr. n=5 mice/group, Scale bar, 10 μm.

Extended data 2. Gli1-expressing neural stem cells in the SVZ egress and generate labeled cells in the corpus callosum

a, Expression of Gli1 in the forebrain was confirmed in Gli1^nLacZ/+ mice by immunofluorescence for LacZ. Labeled cells were observed in the SVZ, cortex and basal forebrain but not in the CC of mice. The right panel shows the magnified images for the corresponding boxes in the left panel. n=5 mice, Scale bar, 50μm. b, Double staining for PDGFRα and LacZ in the Gli1^nLacZ/+ forebrain does not show any co-labeled cells indicating that Gli1 is not expressed by OPCs. n=5 mice, Scale bar, 50μm. c, The ventral SVZ lining the body of the lateral ventricle from a human brain specimen shows colocalization of Gli1 with GFAP+ cells (yellow arrow) as well as a Gli1+ cell not expressing GFAP (arrow head). n= 1 brain, Scale bar, 50μm. d, e, f, Time course analysis of the SVZ and CC of Gli1^CE/+ mice stereotactically injected with saline (control, left panels) or Lysolecithin (LPC) (right panels) to induce demyelination; n=3 mice/group. No labeled cells were seen within the CC following saline injections; areas of ingress into the LPC-injected CC are boxed. At 1 day post lesion (dpl) (d) GFP-labeled cells diverted towards the CC; at 2 dpl (e), a few labeled cells were seen within the CC; at 6 dpl (f), many GFP⁺ cells had accumulated at the site of LPC injection (arrowhead). Scale bar, 50μm.

Extended data 3. Neural stem cells generate oligodendrocytes in various white matter tracts in Gli1-null brains upon demyelination

a,b Fate-mapped Gli1-null cells migrated into the anterior commissure (AC) (a) and striatum (b) following cuprizone-mediated demyelination (inset shows the area of the forebrain) in Gli1^Null (Gli1^CE/nLacZ) mice. Scale bar, 50μm c, 2 weeks after removal from cuprizone diet, GFP-labeled cells are present throughout the CC of Gli1^Null (Gli1^CE/nLacZ) mice. n=5 mice/group, CC= corpus callosum, Scale bar, 100 μm.

Extended data 4. Myelination commences earlier in developing Gli1 null mice

a, Gli1^nLacZ/nLacZ (Gli1^Null) mice (right panel) show increased Myelin Basic Protein (MBP) levels in the forebrain at post-natal day 9 (P9) compared to Gli1^nLacZ/+ (Gli1^Het) mice. b, Quantification of the extent of MBP expression at P9 in the CC of Gli1 nulls (47.12 ± 10.44%) vs. hets (29.71 ± 1.77%) corroborates myelination is accelerated, n=5 mice/genotype c, Analysis of healthy adult forebrain shows the intensity of Black-Gold myelin stain in Gli1 hets and nulls was comparable; d, Quantification of the sizes of the CC shows the CC in Gli1^Null (Gli1^nLacZ/nLacZ) were slightly larger on average than those in Gli1^Het (Gli1^nLacZ/+) mice, although the difference was not statistically significant n=5 mice/genotype. e, Quantification of G ratios (ratio of myelinated axon diameter to the axon diameter) from electron micrographs of healthy Gli1^Het (Gli1^nLacZ/+) and Gli1^Null (Gli1^nLacZ/nLacZ) mice revealed no difference in the thickness of myelin sheaths in the CC. n=3 mice/genotype, Scale bar, 50μm, CC= corpus callosum, Data are mean ± SEM. Student’s T test.

Extended data 5. Effects of gain of Smoothened function in Gli1-het vs. Gli1-null cells during remyelination

Gli1^Het (Gli1^CE/+), Gli1^Het;Smo^M2 (Gli1^CE/+;Smo^M2) and Gli1^Null;Smo^M2 (Gli1^CE/nLacZ;Smo^M2) mice were injected with tamoxifen, fed either a regular or a cuprizone-supplemented diet for 6 weeks and analyzed by immunofluorescence 2 weeks after removal of cuprizone. a, GFP⁺ cells are only seen in the CC of mice on cuprizone diet (right panels) and not in the control mice (left panels). b, Quantification of the GFP⁺ cells in the CC shows significantly higher numbers of cells in Gli1^Null;Smo^M2 mice compared to Gli1^Het and Gli1^Het;Smo^M2 mice. c, Quantification of the proportion of GFP-labeled co-expressing glial markers in the CC of cuprizone treated Gli1^Het, Gli1^Het;Smo^M2 and Gli1^Null;Smo^M2 mice shows an increase in percentage of GFP-labeled OPCs (PDGFRα⁺) in Gli1^Het;Smo^M2 mice and mature oligodendrocytes (CC1⁺) in Gli1^Null;Smo^M2 mice; n=3 mice/group/genotype, Data are mean ± SEM. Student’s T test.

Extended data 6. Proliferation of NSCs and expression of Shh in Gli1-null mice

a,b, At the start of demyelination (three weeks of cuprizone diet), Gli1^Null (Gli1^nLacZ/nLacZ) brains have a higher proportion of proliferating nLacZ⁺ neural stem cells indicated by the percentage of EdU incorporating cells co-expressing nLacZ in the subventricular zone (SVZ) compared to Gli1^Het (Gli1^nLacZ/+) brains, i.e. 31 ± 8.9% vs 8.22 ± 5.33%, respectively. n=3 mice/group/genotype; Scale bar, 50μm. c, The numbers of fate-mapped Gli1⁺ neural stem cells in the SVZ were quantified as the proportion of GFP⁺ cells co-expressing GFAP in Gli1^Het (Gli1^CE/+) and Gli1^Null (Gli1^CE/nLacZ) mice at two weeks of recovery from cuprizone diet. The percentage of GFAP⁺GFP⁺ cells in the SVZ of mice receiving cuprizone diet was comparable to those on a control diet suggesting that the stem cell pool is not depleted during remyelination. n=3 mice/group/genotype. Data are mean ± SEM, Student’s T test. d, Fate-mapping of Shh expressing cells using a membrane GFP reporter labels neurons in the basal forebrain (left panel) with their neurites reaching the ventral SVZ (right panel) in Gli1^Het and Gli1^Null brains e, Gli1^Null mice show expression of Sonic Hedgehog (Shh) in the SVZ and CC following demyelination which is mostly co-localized to GFAP-expressing cells. f, Quantification of the proportional area of the CC expressing Shh does not show any significant difference between Gli1^Het and Gli1^Null mice either on control or cuprizone diet. n=3 mice/group/genotype, CC= corpus callosum, SVZ=Subventricular Zone, Data are mean ± SEM, Student’s T test. Scale bar, 50μm.

Extended data 7. GANT61 reduces Gli1 levels but does not deplete neural stem cells in the SVZ

a, Relative expression of Gli1 and Gli2 mRNA in the forebrain of Gli1^CE/+ mice was examined by qPCR after administration of GANT61 (50mg/Kg/day) for 4 weeks. GANT61 decreases the mRNA levels of Gli1 significantly without changing Gli2 levels. n=3 mice/group, Data are mean ± SEM. Student’s T test. b, c, The numbers of fate-mapped Gli1⁺ neural stem cells in the SVZ were analyzed by immunofluorescence as the proportion of GFP⁺ cells co-expressing GFAP in Gli1^CE/+ mice treated with vehicle or GANT61 at two weeks of recovery from cuprizone diet (b). The percentage of GFAP⁺GFP⁺ cells in the SVZ of mice treated with GANT61 was comparable to those treated with vehicle suggesting that the stem cell pool is not depleted by GANT61 (c). n=3 mice/group, Data are mean ± SEM. Student’s T test.

Extended data 8. Pharmacological inhibition of Gli1 does not affect OPC recruitment or differentiation during remyelination

a, NG2^CE/+ mice were treated with two doses of i.p. tamoxifen to sparsely label OPCs and analyzed at 2 weeks of recovery from cuprizone; Scale bar, 50μm. b, The numbers (~40 GFP⁺ cells/field), and c proportions of GFP-labeled OPCs (PDGFRα) and mature oligodendrocytes (CC1) were similar in the GANT61 vs. the vehicle treated mice in the CC indicating GANT61 does not alter OPC remyelination. Scale bar, 50μm. n=3 mice/group, Data are mean ± SEM. Student’s T test.

Extended data 9. Gli1 is not expressed by immune cells of spleen, thymus and liver of healthy mice

Cells from the spleen, liver and thymus of tamoxifen treated Gli1^Het (Gli1^CE/+) and Gli1^Null (Gli1^CE/nLacZ) mice were analyzed by flow cytometry for GFP expression. Wt (wild type) mice did not express GFP and were used as controls. Representative flow cytometry scatterplots showing absence of GFP expression in (a) CD45⁺/CD3⁺/CD4⁺/CD19⁺ T and B cells; (b) CD45⁺/CD3⁻/CD19⁻/B220⁺/MPDCA⁺ plasmacytoid dendritic cells; (c) CD45⁺/CD3⁻/CD19⁻/B220⁻/CD11b⁺ macrophages, monocytes, dendritic and natural killer (NK) cells and (d) CD45⁺/CD3⁻/CD19⁻/B220⁻/CD11b⁻ cells in Gli1^CE/+ and Gli1^CE/nLacZ mice. n=3 mice/genotype.

Extended data 10. Effects of GANT61 on spinal cord axons in the PLP-induced EAE model

a, Scatterplot of G ratios with respect to axonal diameters (n=500 axons in 3 mice/group, exponential trend line). b, Analysis of electron microscopy images showing the relative proportion of axons binned by their diameters in the 4 groups. c, Analysis of electron microscopy images indicating the G ratios of axons relative to their diameters in the 4 groups (n=500 axons in 3 mice/group, Data are mean ± SEM. Student’s T test). d, Immunofluorescence image of a spinal cord section from Gli1^nLacZ/+ mice shows that LacZ is not expressed by NG2⁺ OPCs. The inset shows expression of LacZ in the germinal zone around the central canal. n=3 mice, Scale bar, 50μm.

Figure 1. Gli1-expressing cells are recruited to and generate myelinating oligodendrocytes at sites of demyelination

a, Serial sections from the brains of Gli1^CE/+ mice on a control or cuprizone diet (at the peak of demyelination) were stained with Black-Gold myelin (left panels) or immunostained for GFP (right panels). The inset in the upper left panel shows a coronal section of the forebrain; the box highlights the area of CC analyzed. The black arrow in the lower left panel indicates a demyelinated region of the CC. Labeled cells in the CC are restricted to the site of demyelination (white arrow). n=5 mice/group, Scale bar, 100μm; inset, 40 μm. b, Comparison of Gli1^CE/+ mice with Nestin^CE/+ mice shows GFP⁺ cells are present in the CC of Nestin^CE/+ mice but not of Gli1^CE/+ mice on control diet (top panels). During recovery from cuprizone, GFP⁺ cells appear in the CC of Gli1^CE/+ mice (bottom panels) and increase in the CC of Nestin^CE/+ mice. n=5 mice/group/genotype, Scale bars, 100 μm; inset, 40 μm. c,d Percentage of GFP⁺ oligodendrocyte progenitors (PDGFRα⁺), mature oligodendrocytes (CC1⁺), and astrocytes (GFAP⁺) in the CC of (c) Gli1^CE/+ and (d) Nestin^CE/+ mice. n=5 mice/group/genotype e, 10 weeks after cessation of cuprizone, GFP-labeled processes flank nodes of Ranvier, demarcated by β4 spectrin (top). n=3 mice, Scale bar, 10 μm. Immunoelectron microscopy shows membrane-targeted GFP (mGFP), indicated by immunogold particles, present within compact myelin sheaths that surround axons (bottom). CC= corpus callosum, n=3 mice, Scale, 100 nm. Data are mean ± SEM.

Figure 2. Loss of Gli1 enhances oligodendrogliogenesis during remyelination

a, Brains of Gli1^CE/+ (Gli1^Het), Gli1^CE/nLacZ (Gli1^Null) and Gli1^CE/+;Smo^fx/fx (Smo^Null) mice were analyzed two weeks after cessation of cuprizone by immunofluorescence. GFP⁺ cells are only observed in the CC of mice receiving cuprizone. n=10 mice/group/genotype, Scale bar, 50μm. b, Quantification of the numbers of GFP⁺ cells in the CC shows a significant increase in Gli1^Null mice compared to Gli1^Het and Smo^Null mice. n=3 mice/group/genotype. c, Gli1^Null mice have a greater proportion of labeled mature oligodendrocytes and reduced proportions of astrocytes than do Gli1^Het and Smo^Null mice. n=3 mice/group/genotype. d, 3 weeks after cessation of cuprizone, Gli1^Null mice have enhanced Myelin Basic Protein (MBP) expression in the CC. n=4 mice/group/genotype. Scale bar,50μm, CC= corpus callosum, Data are mean ± SEM, Student’s T test.

Figure 3. Pharmacological inhibition of Gli1 promotes NSC recruitment and differentiation during remyelination

a, Gli1⁺ cells were fate-mapped in cuprizone-fed Gli1^CE/+ (Gli1^Het) mice treated with vehicle or GANT61 for 4 weeks before analysis. Mice that received GANT61 had more GFP+ cells than those receiving vehicle. n=5 mice/group; Scale bar, 50μm b, c, Mice that received GANT61 had ~ 7 fold increase in the numbers of GFP-labeled cells in the CC (b) and a significant increase in the percentage of mature oligodendrocytes (c). d, Wildtype mice were fed a cuprizone diet and treated with vehicle or GANT61 for 9 weeks. The brains were examined 6 weeks after recovery from cuprizone diet. Mice that received GANT61 showed significantly higher MBP fluorescence intensity levels than vehicle treated mice. n=5 mice/group. Data are mean ± SEM. Student’s T test.

Figure 4. GANT61 improves functional outcomes and is neuroprotective in a RR-EAE model

a, EAE clinical scores following prophylactic or therapeutic treatment with GANT61 compared with vehicle administration. n=9 mice/group, Data are mean ± SEM, *p<0.05, Two-way Anova with Tukey’s multiple comparison test. b,c, Electron micrographs from the ventral lumbar spinal cords of control, GANT61- and vehicle-treated EAE mice shows axonal pathology (b) including axolysis (arrow) and dense axoplasm (arrowhead). Quantification of 500 axons/group from 3 mice/group (c) shows a higher proportion of pathological axons in vehicle-treated group compared to GANT61 treated EAE mice. Scale bar, 2nm, n=3 mice/group. d,e, Alpha motor neurons labeled for NeuN and CHAT in the lumbar spinal cords of control, vehicle- and GANT61-treated EAE mice (d). Quantification shows more motor neurons in the GANT61- vs. vehicle-treated mice (e). Scale bar, 50μm, n=3 mice/group, f, Quantification for the area of MBP expression shows lowest levels of myelin in vehicle-treated group. n=3 mice/group. Data are mean ± SEM, Student’s T test.