Transcriptomes of germinal zones of human and mouse fetal neocortex suggest a role of extracellular matrix in progenitor self-renewal

Abstract

The expansion of the neocortex during mammalian brain evolution results primarily from an increase in neural progenitor cell divisions in its two principal germinal zones during development, the ventricular zone (VZ) and the subventricular zone (SVZ). Using mRNA sequencing, we analyzed the transcriptomes of fetal human and embryonic mouse VZ, SVZ, and cortical plate. In mouse, the transcriptome of the SVZ was more similar to that of the cortical plate than that of the VZ, whereas in human the opposite was the case, with the inner and outer SVZ being highly related to each other despite their cytoarchitectonic differences. We describe sets of genes that are up- or down-regulated in each germinal zone. These data suggest that cell adhesion and cell-extracellular matrix interactions promote the proliferation and self-renewal of neural progenitors in the developing human neocortex. Notably, relevant extracellular matrix-associated genes include distinct sets of collagens, laminins, proteoglycans, and integrins, along with specific sets of growth factors and morphogens. Our data establish a basis for identifying novel cell-type markers and open up avenues to unravel the molecular basis of neocortex expansion during evolution.

Fig. 1.

Correlation analysis of fetal human and embryonic mouse cortical zones. (A) Cartoon illustrating the major cell types (APs in blue, BPs in orange and red, neurons in green) in the different GZs and the CP of the developing mouse and human neocortex. (B and C) Heat-map showing pair-wise correlations of gene expression levels between the indicated zones of the fetal human neocortex at the indicated gestational weeks (B, 13–16 wpc) and of the E14.5 mouse neocortex (C). Spearman’s rank correlation coefficients (r_s) range from 0.7 (yellow) to 1.0 (blue). Numbers in the blank quadrants indicate the mean r_s values ± SD of the zone comparison located mirror-symmetrically to the diagonal line; asterisks indicate zone comparison mean r_s values that show a statistically significant difference to each other (P < 0.05).

Fig. 2.

Differential gene-expression analysis and functional annotation clustering of fetal human cortical zones. (A) DESeq scatter plot showing pair-wise comparisons of gene expression between the indicated zones of the 14–16 wpc human neocortex. Each dot represents the mean (five fetuses) expression level for a given gene, with differentially expressed genes (false-discovery rate <0.05) shown in green. Numbers above (below) the diagonal lines refer to the differentially expressed genes up-regulated in the zone indicated on top (on the right). These numbers include the genes expressed in only one of the two zones compared (green dots on vertical and horizontal 0.1 lines). For the corresponding analysis of genes differentially expressed in mouse zones and between human and mouse zones see Fig. S5. (B) Venn diagrams showing the numbers of differentially expressed genes (see A) that are up-regulated (Left, >) and down-regulated (Right, <) in a given zone compared with the other zones of the 14–16 wpc human neocortex. (C) Differentially expressed genes (see A) that are up-regulated (Left, >) and down-regulated (Right, <) in a given zone compared with the other zones indicated of the 14–16 wpc human neocortex were analyzed for significantly enriched FGA terms which were clustered, using DAVID. The five clusters with the highest enrichment scores are shown. Note that for certain FGA terms enriched in the various GZs, the same designation were chosen (e.g., secretory pathway), although the individual genes in these clusters were different.

Fig. 3.

Functional annotation clustering of genes differentially expressed among mouse cortical zones and between human and mouse cortical zones. Differentially expressed genes identified by DESeq (Fig. S5) that are up-regulated (Left, >) and down-regulated (Right, <) between the indicated zones of the E14.5 mouse neocortex (A) and between the 14–16 wpc human and E14.5 mouse neocortex (B) were analyzed for significantly enriched FGA terms, which were clustered, using DAVID. The five clusters with the highest enrichment scores are shown. See note on FGA terms in Fig. 2C.

Fig. 4.

Mean-shift clustering of genes related to ECM interaction in fetal human and embryonic mouse cortical zones. (A–J) Log₂-transformed FPKM fold-changes of ECM-related genes in the 14–16 wpc hVZ, hISVZ, hOSVZ, and hCP (five fetuses) and in the E14.5 mVZ, mSVZ, and mCP (five embryos) relative to the mean expression of each of the genes in all human or mouse zones were subjected to nonhierarchical K-nearest neighbor mean-shift (KNN-MS) cluster analysis, which yielded a total of nine clusters. (A) Heat map showing the nine clusters sorted hierarchically using an average-linkage algorithm. Each gene is represented by a single row and each developmental stage by a single column. Fold-changes range from 5.2 (red) to −6.6 (green). (B–J) Median log₂ fold-changes in expression of the genes (numbers in parentheses) contained in the nine clusters shown in A. (K–O) Occurrence of collagens (K), laminins (L), proteoglycans (M), integrins (N), and growth factors/morphogens (O) in the clusters indicated in B–J of the 14–16 wpc human (five fetuses) or E14.5 mouse (five embryos) neocortex. Numbers in parentheses refer to the total number of collagen, laminin, proteoglycan, integrin, or growth factor/morphogen genes expressed in the fetal human and embryonic mouse neocortex; numbers in the pie chart sectors refer to the number of the respective genes expressed in the indicated cluster. Black sectors, GZ clusters; white sectors, CP clusters. (P) Summary of the distribution of collagens (COL), laminins (LAM), proteoglycans (PG), integrins (INT), and growth factors/morphogens (GFM) between the various zones of fetal human and embryonic mouse neocortex (APs in blue, BPs in orange and red, neurons in green). Numbers in parentheses, total number of genes in respective class, as indicated in K–O; numbers in green/red/blue boxes, number of genes in respective class that are specifically overexpressed (red bars in B–J) only in the hCP or mCP (green), hISVZ, and hOSVZ or mSVZ (red), or hVZ or mVZ (blue); numbers in yellow boxes, number of genes in respective class that are overexpressed in more than one human or mouse zone; numbers in white boxes, number of genes in respective class that are overexpressed in both h/mVZ, hISVZ/hOSVZ/mSVZ, or h/mCP.