Cspg4 sculpts oligodendrocyte precursor cell morphology

Abstract

The extracellular matrix (ECM) provides critical biochemical and structural cues that regulate neural development. Chondroitin sulfate proteoglycans (CSPGs), a major ECM component, have been implicated in modulating oligodendrocyte precursor cell (OPC) proliferation, migration, and maturation, but their specific roles in oligodendrocyte lineage cell (OLC) development and myelination in vivo remain poorly understood. Here, we use zebrafish as a model system to investigate the spatiotemporal dynamics of ECM deposition and CSPG localization during central nervous system (CNS) development, with a focus on their relationship to OLCs. We demonstrate that ECM components, including CSPGs, are dynamically expressed in distinct spatiotemporal patterns coinciding with OLC development and myelination. We found that zebrafish lacking cspg4 function produced normal numbers of OLCs, which appeared to undergo proper differentiation. However, OPC morphology in mutant larvae was aberrant. Nevertheless, the number and length of myelin sheaths produced by mature oligodendrocytes were unaffected. These data indicate that Cspg4 regulates OPC morphogenesis in vivo, supporting the role of the ECM in neural development.

Keywords: extracellular matrix; glia; myelination; neural cell differentiation; neural cell fate; zebrafish.

Figure 1.. Lectin staining intensity increases in the spinal cord during zebrafish development and is dynamically located.

Transverse slices of zebrafish larval spinal cord of larvae containing transgenes Tg(olig1:EGFP-CAAX) and Tg(mbpa:mCherry-CAAX) at 3, 5, and 7 days dpf. Arrowheads indicate specific fluorescence patterns of vicia villosa lectin: red arrowheads highlight puncta in soma-rich areas; orange arrowheads highlight puncta in axon-rich areas; white arrowheads highlight puncta near oligodendrocyte lineage cells (OLCs); and yellow arrowheads highlight strong deposition at the dorsal white matter tract. Scale bars 10 μm.

Figure 2.. OPCs and astrocytes have highest expression of structural ECM genes, and zebrafish OPCs express cspg4.

Heatmaps showing expression levels of structural extracellular matrix (ECM) genes – collagens (A), glycoproteins (B) – across cell populations in mouse cortex. Dataset from Zhang et al, 2014. Heatmaps generated using pluto.com. Transverse slices of wild-type zebrafish larval spinal cord at 2 (C), 3 (D), 5 (E), and 7 (F) dpf processed to detect sox10 (yellow), cspg4 (magenta), and mbpa (cyan) transcripts. Nuclei are labeled with DAPI (grey). White arrows highlight OPCs co-expressing sox10 and cspg4. Green arrows indicate OLs expressing sox10 and mbpa but not cspg4. The left column shows merged imaged of all 4 channels. The other columns show individual images of sox10, cspg4, and mbpa expression. Scale bars 10 μm.

Figure 3.. Schematic overview and confirmation of cspg4 mutant alleles.

(A) Locations of mutant alleles indicated by red arrowheads. Predicted proteins indicated below. compHet represents compound heterozygous cspg4^sa37999/cspg4sa39470. Abbreviations: extracellular domain (ECD), transmembrane (TM), intracellular domain (ICD). (B) Genotyping results for PCR-amplified regions displayed on an agarose gel. For cspg4^sa37999: 290 bp band indicates wildtype, ~150 bp band (split into 167 bp and 123 bp) indicates homozygous mutation (−/−), and dual bands indicate heterozygous (+/−). For cspg4^sa39470: 175 bp band signifies wildtype, 198 bp indicates homozygous mutation (−/−), and presence of both bands indicates heterozygosity (+/−).

Figure 4.. Loss of cspg4 function does not impair OPC formation or OL differentiation.

(A) Schematic representation of the experiment wherein larvae were fixed and sectioned at 2, 3, 5, and 7 dpf. Transverse sections were only analyzed if they showed the spine over yolk tube, indicated by crimson vertical lines. (B) Representative images of wild-type and cspg4^sa37999 homozygous mutant larvae showing sox10 (yellow), mbpa (blue), and cspg4 (magenta) RNAs detected using fluorescent in situ RNA hybridization; all cell nuclei labeled using DAPI (grey). White arrowheads highlight sox10+ mbpa− OPCs; green arrowheads highlight sox10+ mbpa+ OLs. Scale bars all 10 μm. (C) Box plot of the number of sox10+ mbpa− OPCs per larva at each timepoint. (D) Box plot of the number of sox10+ mbpa+ OLs at each timepoint. OPC and OL numbers were not statistically different between genotypes, although mutant embryos trended toward fewer OPCs at 2 dpf (p=0.054). Data analyzed by Kruskal-Wallis statistical test.

Figure 5.. OPCs from cspg4 mutant larvae have morphological differences in complexity compared to WT through development

(A) Schema of the imaging region in the spine above the yolk tube in larvae with example images of mosaically labeled fluorescent OPCs using olig1:EGFPcaax, imaged in live larvae. Color coding for genotypes is consistent across all figures. Scale bars all 10 μm. (B) The summed filament length of mutant larvae are not significantly different than WT at any timepoint using a two-way, non-repeated measures ANOVA. (C) Morphological complexity was assessed by Sholl analysis and Kruskal-Wallis test followed by a Dunn test when significant. This showed cspg4^sa37999 mutants have at least one significantly longer process on average than comHets (p = 0.001) at 3 dpf (i). cspg4^sa39470 mutants have significantly smaller cells with higher complexity closer to the soma compared to WT (p = 0.0001) at 5dpf (ii). At 7dpf (iii), cspg4^sa37999 mutants may be more complex close to the soma or have at least one longer process compared to WT (p=0.056).

Figure 6.. Loss of cspg4 does not alter myelin sheath number or length.

(A) Schema of the imaging region in the spine above the yolk tube in larvae with example images of mosaically labeled fluorescent OLs using mbpa:EGFP-CAAX. Using a Kruskal-Wallis statistical test, mutant larvae have oligodendrocytes with similar sheath number (B) and total summed length (C) compared to wildtype. Scale bars equal 10 μm.