Expression of Genes in the 16p11.2 Locus during Development of the Human Fetal Cerebral Cortex

Abstract

The 593 kbp 16p11.2 copy number variation (CNV) affects the gene dosage of 29 protein coding genes, with heterozygous 16p11.2 microduplication or microdeletion implicated in about 1% of autism spectrum disorder (ASD) cases. The 16p11.2 CNV is frequently associated with macrocephaly or microcephaly indicating early defects of neurogenesis may contribute to subsequent ASD symptoms, but it is unknown which 16p11.2 transcripts are expressed in progenitors and whose levels are likely, therefore, to influence neurogenesis. Analysis of human fetal gene expression data revealed that KIF22, ALDOA, HIRIP3, PAGR1, and MAZ transcripts are expressed in neural progenitors with ALDOA and KIF22 significantly enriched compared to post-mitotic cells. To investigate the possible roles of ALDOA and KIF22 proteins in human cerebral cortex development we used immunohistochemical staining to describe their expression in late first and early second trimester human cerebral cortex. KIF22 protein is restricted to proliferating cells with its levels increasing during the cell cycle and peaking at mitosis. ALDOA protein is expressed in all cell types and does not vary with cell-cycle phase. Our expression analysis suggests the hypothesis that altered neurogenesis in the cerebral cortex contributes to ASD in 16p11.2 CNV patients.

Keywords: ALDOA; CNV; KIF22; autism; cerebral cortex.

Figure 1

Bioinformatics analysis of scRNA-seq from the VZ and SVZ of the 16-18GW human fetal cortex. (a) 16p11.2 region and genes. (b) schematic adapted from Budday et al. 2015 shows the process of human fetal cortex development over time. Dotted box indicates approximately the time-period of interest for our study; 12-16PCW. (c) schematic of human cortical structure during development. (d) tSNE clustering of cell types. (e) changing mRNA expression levels of 16p11.2 genes as cells move from progenitors to neurons with KIF22 and ALDOA transcripts identified as changing significantly. Schematic of human cortex with dotted box indicates the compartments of the brain that were used to generate this dataset—the germinative VZ and SVZ. (f) Violin plots showing distribution of KIF22 in different cell types. (g) KIF22 gradient plot (arrow indicates progenitors with a subset expressing high levels of KIF22 mRNA). (h) Violin plots showing distribution of KIF22 at different cell cycle stages. (i) Violin plots showing distribution of ALDOA mRNA levels in different cell types. (j) ALDOA gradient plot. k) Violin plots showing distribution of ALDOA mRNA at different cell cycle stages.

Figure 2

In situ hybridisation of candidate genes. (a) PAX6 protein (green) at 12pcw. a-a’^*) show high magnification images of PAX6 protein expression in the (a’) SP/CP, a^*) IZ, a”) SVZ and a’^*) VZ. Low magnification scale bars = 50 μm, high magnification scale bars = 10 μm. b) Low magnification image of KIF22 mRNA in the 12pcw human fetal cortex, b’) High magnification showing KIF22 mRNA (blue) to be predominantly expressed in the germinative zones. c) Low magnification image of ALDOA mRNA in the 12pcw human fetal cortex, (b’) High magnification showing ALDOA mRNA (blue) to be predominantly expressed in the germinative zones but also some expression in the IZ and CP. (d) Low magnification image of HIRIP3 mRNA in the 12pcw human fetal cortex, (d’) High magnification showing HIRIP3 mRNA (blue) to be expressed throughout the telencephalic wall. e) Low magnification image of PAGR1 mRNA in the 12pcw human fetal cortex, e’) High magnification showing PAGR1 mRNA (blue) to be expressed throughout the telencephalic wall. (f) Low magnification image of MAZ mRNA in the 12pcw human fetal cortex, (f’) High magnification showing MAZ mRNA (blue) to be expressed throughout the telencephalic wall. For ISH, low magnification scale bars = 2 mm and high magnification scale bars = 100 μm.

Figure 3

KIF22 protein expression levels in the cerebral cortex at 12, 14, and 16 PCW (a, a’, a^*) schematic showing brain regions sectioned. (b, b’, b^*) images of whole brain section, scale bars =2 mm. c, c’, c^*) sections spanning the rostral-caudal axis showing KIF22 expression in the telencephalic wall, scale bars = 100 μm. (d, d’, d^*) high magnification images of different cortical zones rostral-caudal. KIF22+ cells in brown and examples indicated by green arrows, KIF22⁻ cells in pink, scale bars =25 μm. (e) Quantification of KIF22 expressing cells with all three ages combined. (f) Quantification of KIF22 expressing cells with rostral, middle, caudal values combined.

Figure 4

Immunofluorescence of KIF22 and KI67 proteins in the cortex. (a) KIF22 and KI67 at 12 PCW, low magnification scale bars = 4 mm, high magnification scale bars = 100 μm. (b) KIF22 and KI67 at 14 PCW, low magnification scale bars = 4 mm, high magnification scale bars = 100 μm. c) high magnification of KI67/KIF22 expressing cells. Scale bars = 10 μm. (d) Percentage of cells expressing KIF22, KI67, or both at 12 PCW. (e) Percentage of cells expressing KIF22, KI67, or both at 14 PCW. (f) Combined data of percentage of cells expressing KIF22, KI67, or both.

Figure 5

Quantification of KIF22 protein levels. (a, b, c) 12 PCW quantification of KIF22 fluorescence intensity in KI67+/KI67- cells (raw data transformation = +1(log), unpaired t-test with Welch’s correction,P = < 0.001). (d, e, f) intensity correlations of KIF22 and KI67 nuclear fluorescence intensity at 12 PCW. (g, h, i) 14 PCW quantification of KIF22 fluorescence intensity in KI67+/KI67- cells (raw data transformation = +1(log), unpaired t-test with Welch’s correction, P = < 0.001). j, k, l) intensity correlations of KIF22 and KI67 nuclear fluorescence intensity at 14 PCW. m) quantification of KIF22 fluorescence intensity in KI67+/KI67-cells 12 and 14 weeks combined (raw data transformation = (log), paired t-test, P = 0.0122). n) intensity correlations of KIF22 and KI67 nuclear fluorescence intensity for rostral-caudal points at 12 and 14 PCW with distance from apical surface indicated by dot color. o) diagram of KI67 protein levels throughout the cell cycle. p) model based on our results of KIF22 protein levels throughout the cell cycle.

Figure 6

ALDOA protein expression in the cortex. ALDOA protein expression across the telencephalic wall at a) 12, b) 14 and c) 16 pcw. Scale bars = 100 μm. White arrow indicates non-specific binding to blood vessels. d) high magnification immunofluorescence of ALDOA and KI67 proteins, scale bar = 10 μm. e) low power image showing how cells were randomly selected for analysis using the DAPI channel. e’) high power image showing how the nucleus and cytoplasm were delineated for analysis

Figure 7

ALDOA protein quantification. (a–d) Cell body ALDOA protein fluorescent intensity in KI67+ and KI67− cells at (a) 12 PCW (raw data transformation = +1(log), bimodal distribution, Mann–Whitney test, p = 0.3702), (b) 14 PCW (raw data transformation = +1(log), normal distribution, unpaired t-test with Welch’s correction, P = 0.2032), (c) 16 PCW (raw data transformation = +1(log), normal distribution, unpaired t-test with Welch’s correction, P = 0.3523). (d) ALDOA cell body protein fluorescent intensity in KI67+ and KI67- cells, 12, 14, and 16 PCW individual datasets averaged, (raw data transformation = +1(log)), paired t-test, P = 0.0836. (e–h) ALDOA cellular protein intensity levels ls correlated to nuclear KI67 protein intensity at (e)12, (f)14 and (g)16 PCW with distance from ventricular edge indicated. (h) ALDOA whole cell protein intensity levels correlated to nuclear KI67 protein intensity pooled 12, 14, 16 PCW. (i–l) Nuclear ALDOA protein fluorescent intensity in KI67+ and KI67− cells at (i) 12 PCW (raw data transformation = +1(log), normal distribution, unpaired t-test with Welch’s correction, P = 0.7543), (j) 14 PCW (raw data transformation = +1(log), normal distribution, unpaired t-test with Welch’s correction, P = 0.0694), (k) 16 PCW (raw data transformation = +1(log), normal distribution, unpaired t-test with Welch’s correction, P = 0.0772). (l) ALDOA nuclear protein fluorescent intensity in KI67+ and KI67− cells, 12, 14, and 16 PCW individual datasets averaged, (raw data transformation = +1(log)), paired t-test, P = 0.1330. (m–p) ALDOA nuclear protein intensity levels is correlated to nuclear KI67 protein intensity at (m)12, (n)14, and o)16 PCW with distance from ventricular edge indicated. (p) ALDOA nuclear protein intensity levels correlated to nuclear KI67 protein intensity pooled 12, 14, and 16 PCW. (q) schematic demonstrating ALDOA protein is predominantly in the cytoplasm and lower in the nucleus in both KI67+ proliferating cells and KI67- post mitotic cells. (r) model showing ALDOA levels do not change with the cell cycle.