Origins and Proliferative States of Human Oligodendrocyte Precursor Cells

Abstract

Human cerebral cortex size and complexity has increased greatly during evolution. While increased progenitor diversity and enhanced proliferative potential play important roles in human neurogenesis and gray matter expansion, the mechanisms of human oligodendrogenesis and white matter expansion remain largely unknown. Here, we identify EGFR-expressing "Pre-OPCs" that originate from outer radial glial cells (oRGs) and undergo mitotic somal translocation (MST) during division. oRG-derived Pre-OPCs provide an additional source of human cortical oligodendrocyte precursor cells (OPCs) and define a lineage trajectory. We further show that human OPCs undergo consecutive symmetric divisions to exponentially increase the progenitor pool size. Additionally, we find that the OPC-enriched gene, PCDH15, mediates daughter cell repulsion and facilitates proliferation. These findings indicate properties of OPC derivation, proliferation, and dispersion important for human white matter expansion and myelination.

Keywords: EGFR; OPC; PCDH15; cortical expansion; oligodendrogenesis; self-repulsion.

Figure 1.. scRNA-seq of OPCs from the Developing Human Cortex

(A) Schematic of the workflow. Developing human cortex was dissociated and OPCs enriched by PDGFRA immunopanning. Single cells were isolated using the Fluidigim C1 microfluidic chip system, and pair-end single-cell RNA sequencing (scRNA-seq) was performed. Clustering was used to annotate cell types according to RNA profiles. (B) Scatterplot by t-stochastic neighbor embedding (tSNE) (left) and feature plot of PDGFRA-mRNA expression (right) from Nowakowski et al. (2017). (C) Validation of the immunopanning strategy by immunostaining or scRNA-seq. (D) Scatterplot by tSNE (left) and feature plot of PDGFRA-mRNA expression (right) of data generated by immunopanning in this study with Nowakowski et al. (2017). (E) Heatmap showing cluster annotations and marker gene expression. (F) Feature plots of marker genes of different cell types.

Figure 2.. Characterization of OPC Clusters from scRNA-seq

(A) Enlarged tSNE plot from Figure 1D focusing on the OPC clusters. Clusters are shown on the left, and metadata annotation of tissue source is represented on the right. (B) Violin plots of differentially expressed genes in OPC clusters. (C) Similarity matrix of OPC, NPC, and neuronal clusters. (D) Trajectory analysis of progenitor and OPC cell types using Monocle. Broad cell types are shown rather than individual clusters. (E) Feature plots of marker genes for NPCs, IPCs, neurons, and OPCs indicating lineage progression.

Figure 3.. EGFR-Expressing Progenitors Generate OPCs in the Developing Human Cortex

(A) Characterization of Pre-OPCs that co-express NPC and OPC markers. Images are from human OSVZ at GW20–24 (n = 12). (B) Characterization of EGFR-expressing progenitors by oRG or IPC marker co-staining. Images are from human OSVZ at GW20–24 (n = 12). (C) Outline of immunopanning workflow (left). Images of marker staining at DIV2 and DIV7 showing lineage progression (middle). Quantification of cell proportions (right). Mean ± SD (n = 5). (***p < 0.001; **p < 0.01; *p < 0.05; N.S., p > 0.05.) (D) Quantification of dividing OPCs in cultured human brain slices after pharmacologically activating [EGF(High) and EGF(Low)] or inhibiting (Iressa, AG1478, and PD153035) EGFR. Mean ± SD (n = 6). See also Figures S1 and S2.

Figure 4.. Characterization of Division Dynamics of EGFR-Expressing Progenitors in the Developing Human Cortex

(A) Outline of immunopanning workflow. Different types of progenitors in the OSVZ were positively selected with PDGFRA, EGFR, and LIFR antibodies by sequential immunopanning. (B) Schematic of the experimental strategy. Progenitors were enriched by immunopanning, infected with GFP adenovirus, injected into cultured cortical slices, and observed for time-lapse imaging. (C) Time-stamped still images of dividing progenitors enriched by LIFR, EGFR, or PDGFRA panning. Representative cells from cultured cortical slices at GW20–24 (n = 30). (D) Quantification of MST distance during division and daughter cell separation of different progenitors at 6 hr after division. Mean ± SD (n = 30). (***p < 0.001; N.S., p > 0.05.) (E) Quantification of OPCs from EGFR-panned cells in cultured human brain slices after pharmacologically activating (EGF) or inhibiting (Iressa) EGFR. Mean ± SD (n = 6). (***p < 0.001.) See also Figure S3 and Videos S1 and S2.

Figure 5.. Characterization of Division Properties of Human Embryonic OPCs

(A) Time-stamped still images of dividing OPCs. Two rounds of continuous division were observed and four granddaughter cells were generated. Representative cell from a cultured cortical slice at GW22 (n = 20). (B) Workflow to explore OPC division quantitatively (top left). OPCs were enriched with PDGFRA immunopanning, infected with GFP adenovirus before transplantation and thymidine analog labeling. Representative images show a GFP-positive immunopanned cell co-labeled with CldU and IdU and expressing OLIG2 (bottom left) (n = 20). Quantification of the proportion of GFP⁺OLIG2⁺ cells that express either CldU or IdU or both on cultured cortical slices (n = 10). Mean ± SD (right). See also Video S3.

Figure 6.. PCDH15 Mediates Daughter Cell Repulsion after OPC Division

(A) Volcano plots show differentially expressed genes between dividing OPCs and dividing RGs (left) or dividing IPCs (right). Two of the most highly differentially expressed genes are DSCAM and PCDH15. (B) Feature plots of DSCAM (left) shows expression in INs and OPCs, while PCDH15 is specific to OPCs. (C) Time-stamped still images of dividing OPCs treated with either scramble-shRNA control lentiviruses or PCDH15-shRNA lentiviruses. Representative cells are from cultured cortical slices at GW20–24 (n = 30). (D) Quantification of daughter cell distance after division. Mean ± SD (n = 30). (***p < 0.001.) (E) Quantification of migration speed before division. Mean ± SD (n = 30). (N.S., p > 0.05.) See also Figure S4 and Videos S4, S5, S6, and S7.

Figure 7.. Interfering with PCDH15 Function Affects OPC Proliferation

(A) Sample images of OPCs in the OSVZ of cultured brain slices treated with either IgG as control or PCDH15 antibody to interfere with PCDH15 function. Blocking to PCDH15 function did not cause obvious change in OPC morphology (O4 staining) and distribution (OLIG2 staining). (B) Time-stamped still images of dividing OPCs infected with PCDH15-shRNA lentiviruses. Two rounds of continuous division were observed, and only one of the daughter cells divided again. Representative cells are from cultured cortical slices at GW20–24 (n = 6). (C) Quantification of OPC division in the OSVZ of cultured brain slices treated with either IgG or PCDH15 antibody. Blocking to PCDH15 function inhibited OPC proliferation but did not change local OPC density. Mean ± SD (n = 6). (***p < 0.001; **p < 0.01; N.S., p > 0.05.) See also Videos S8 and S9.