Brain-specific deletion of histone variant H2A.z results in cortical neurogenesis defects and neurodevelopmental disorder

Abstract

Defects in neurogenesis alter brain circuit formations and may lead to neurodevelopmental disorders such as autism and schizophrenia. Histone H2A.z, a variant of histone H2A, plays critical roles in chromatin structure and epigenetic regulation, but its function and mechanism in brain development remain largely unknown. Here, we find that the deletion of H2A.z results in enhanced proliferation of neural progenitors but reduced neuronal differentiation. In addition, neurons in H2A.z knockout mice exhibit abnormal dendrites during brain development. Furthermore, H2A.zcKO mice exhibit serial behavioral deficits, such as decreased exploratory activity and impaired learning and memory. Mechanistically, H2A.z regulates embryonic neurogenesis by targeting Nkx2-4 through interaction with Setd2, thereby promoting H3K36me3 modification to activate the transcription of Nkx2-4. Furthermore, enforced expression of Nkx2-4 can rescue the defective neurogenesis in the H2A.z-knockdown embryonic brain. Together, our findings implicate the epigenetic regulation by H2A.z in embryonic neurogenesis and provide a framework for understanding how disruption in the H2A.z gene may contribute to neurological disorders.

Figure 1.

H2A.z is expressed in the embryonic cerebral cortex. (A) Western blot analysis of H2A.z in the development of the cerebral cortex. Brains in different developmental stages were isolated and lysed for western blot analysis with the antibodies indicated. (B) An overview of E13 brain sections labeled for H2A.z and DAPI. (C) Images of brain sections of the developing cerebral cortex labeled for H2A.z and DAPI. (D) H2A.z is abundantly expressed in Nestin positive neural progenitors and co labeled with Tuj1-positive neurons in the embryonic cortex. The E15 brain sections were immunostained with anti-H2A.z and anti-Nestin or anti-Tuj1 antibodies. (E) Western blot analysis reveals that the expression of H2A.z is effectively suppressed by H2A.z-shRNAs in primary neural progenitors. (F) Graph shows that the amount of H2A.z is obviously decreased in H2A.z-shRNA lentivirus-infected primary neural progenitors. n = 5 for all samples. (G) Electroporation of H2A.z-shRNAs results in abnormal cell distribution in the developing neocortex. The electroporation was performed at E13, and the mouse was sacrificed at E16. (H) Graph shows the percentage of GFP-positive cells distributed in the VZ/SVZ, IZ and CP. n = 5 for all samples. Representative images from at least three independent experiments. Error bars represent means ± S.E.M.; Two-tailed unpaired t-test, P < 0.05 (*), P < 0.01(**) or P < 0.001(***). n.s., not significant. CP, cortical plate; IZ, intermediate zone; VZ/SVZ, ventricular/subventricular zone. Scale bar represents 100 μm (B); 20 μm (C and D); 50 μm (G).

Figure 2.

H2A.z regulates the proliferation of NPCs during brain development. (A) Brain sections of H2A.z^fl/fl mice and H2A.z^cKO mice at E16 were immunostained with mitotic marker pH3 and DAPI. (B) Statistics of pH3-positive cells in rostral, central, and caudal areas of the cortex. n = 5 for all samples. (C) Representative images of E16 coronal brain sections were immunostained for Tbr2 and BrdU. BrdU was injected intraperitoneally to pregnant mice at E16 for 2 h of pulse labeling. (D) Statistic of BrdU-positive cells per 100 μm² surface of VZ/SVZ. n = 7 for all samples. (E) Statistics of Tbr2-positive cells per 100 μm² surface of VZ/SVZ. n = 6 for all samples. (F) Percentage of Tbr2⁺BrdU⁺ cells among all Tbr2⁺ cells. n = 7 for all samples. (G) Neural progenitor cells isolated from H2A.z^cKO mice form larger neurospheres. Primary neural progenitors were dissociated from E12 H2A.z^cKOor H2A.z^fl/fl mice and seeded at low density (50 cells/ml), then cultured in proliferated medium (suspension culture) for 3 days. (H) Statistic of the diameters of neurospheres. n = 8–9 for all samples. (I) Abnormal cell distribution is observed in H2A.z ablated neocortex. The GFP plasmid was electroporated into E13 mouse brains of H2A.z^fl/fl mice and H2A.z^cKO mice, and the mice were sacrificed at E16. (J) Graphs of the percentage of GFP-positive cells in the VZ/SVZ, IZ and CP. n = 8–10 for all samples. Representative images from at least three independent experiments. H2A.z^fl/fl, WT; H2A.z^cKO, cKO. CP, cortical plate; IZ, intermediate zone; VZ/SVZ, ventricular/subventricular zone. Error bars represent means ± S.E.M.; Two-tailed unpaired t-test, P < 0.05 (*), P < 0.01(**) or P < 0.001(***). n.s., not significant. Scale bar represents 15 μm (A); 20 μm (C); 80 μm (G); 50 μm (I).

Figure 3.

H2A.z regulates the differentiation of NPCs. (A) The number of neurons is decreased in H2A.z-shRNA-electroporated brains. Neurons were labeled with anti-Tuj1 in E16 brain sections. (B) Percentage of Tuj1⁺GFP⁺ cells among GFP⁺ cells. n = 4 for all samples. (C) Coronal brain slices of E16 H2A.z^fl/fl and H2A.z^cKO mice were immunostained using anti-Tuj1. (D) Relative Tuj1⁺ cells in the VZ of H2A.z^fl/fl and H2A.z^cKO mice. n = 8 for all samples. (E) The mRNA level of NeuroD1 and Tuj1 is decreased in H2A.z-deleted neurospheres. E12 primary neural progenitor cells (NPCs) were isolated from H2A.z^cKO and H2A.z^fl/fl mice, and cultured in proliferation medium (suspension culture) for 3 days, and then collected the neurospheres for further analysis (E and F). Actin mRNA served as an internal control. n = 4 for all samples. (F) Western blot analysis reveals that the expression levels of the neuron markers Tuj1 and NeuN are downregulated in H2A.z-loss neurospheres. (G) Statistics of the normalized density of H2A.z, Tuj1 and NeuN. n = 6 for all samples. (H) Cell cycle exit is decreased in H2A.z^cKO mice. E16 brain sections were stained with anti-BrdU and anti-Ki67. H2A.z^fl/fl mice and H2A.z^cKO mice were administered BrdU (100 mg/kg) for 24 h and were euthanized at E16. (I) Percentage of cell cycle exit (BrdU⁺Ki67⁻/Ki67⁺) in H2A.z^fl/fl or H2A.z^cKO mice. n = 6 for all samples. Representative images from at least three independent experiments. H2A.z^fl/fl, WT; H2A.z^cKO, cKO. VZ, ventricular zone. Error bars represent means ± S.E.M.; two-tailed unpaired t-test, P < 0.05 (*), P < 0.01(**) or P < 0.001(***). Scale bar represents 50 μm (A); 20 μm (C and H).

Figure 4.

The loss of H2A.z results in abnormal morphology of neurons. (A) Representative images of neurons after 4 days of culture in vitro. GFP-expressing plasmid was electroporated into the E13 cerebral cortices of H2A.z^fl/fl or H2A.z^cKO mice to mark neural progenitor cells. After 24 h, the GFP-positive cells were isolated and cultured in differentiation medium for 4 days. (B) Graph shows that the total dendritic length of neurons is decreased when H2A.z is deleted. n = 8 cells from three samples. (C) Statistics show that the number of primary dendrites is reduced upon H2A.z deletion. n = 8 cells from three samples. (D) Images of neurons cultured in differentiation medium for 6 days. (E) Sholl graphs of dendrites of neurons electroporated with GFP alone in H2A.z^fl/fl or H2A.z^cKO mice. The GFP-positive cells were isolated and cultured in differentiation medium for 6 days. H2A.z^fl/fl, n = 6 cells; H2A.z^cKO, n = 6 cells. (F) High-magnification confocal images of upper layer show that H2A.z knockout results in abnormally branched processes compared to the control. GFP plasmid was electroporated into the H2A.z^fl/fl or H2A.z^cKO embryonic brain at E13 and harvested at P0. (G) Quantification of dendritic numbers in H2A.z^fl/fl and H2A.z^cKO cortices. n = 6 for all samples. Representative images from at least three independent experiments. H2A.z^fl/fl, WT; H2A.z^cKO, cKO. Error bars represent means ± S.E.M., Kolmogorov–Smirnov test or two-tailed unpaired t-test, P < 0.05 (*), P < 0.01(**). Scale bar represents 15 μm (A and D); 10 μm (F).

Figure 5.

The loss of H2A.z results in behavioral deficits. (A) Representative tracing pathway of H2A.z^fl/fl and H2A.z^cKO mice in the open field test. (B) The total traveling distance was not different between H2A.z^fl/fl and H2A.z^cKO mice. (C) Time spent in the center is reduced in H2A.z^cKO mice. (D) Representative tracing pathway of H2A.z^fl/fl and H2A.z^cKO mice in the elevated-plus maze test. (E) H2A.z^cKO mice spent less time in the open arms. (F) H2A.z^cKO mice spent more time in the closed arms. (G) Social interaction behavior is impaired in H2A.z-null mice. H2A.z^cKO mice spent similar time in the chamber with the stranger 1 or stranger 2 mouse, while H2A.z^fl/fl mice spent more time in the chamber with stranger 2 than with the stranger 1 mouse. (H) Close social interaction is affected by H2A.z deletion. H2A.z^cKO mice did not spend more time in close interaction with stranger 2 than with stranger 1. (I) The immobile time of H2A.z^cKO mice was more than that of H2A.z^fl/fl mice during the forced swim test. (J) The percentage of freezing in H2A.z^cKO mice was less than in H2A.z^fl/fl mice, although they were explored in the same background. (K) Freezing response was also reduced in H2A.z^cKO mice in the cued test. (L) Compared with H2A.z^fl/fl mice, the entries of H2A.z^cKO mice into new arms were decreased, while the entries into old arms were increased. (M) Compared with H2A.z^fl/fl mice, H2A.z^cKO mice spent less time in the new arms and more time in the old arms. All mice were male, 8–10 weeks old. H2A.z^fl/fl, WT, n = 11–14; H2A.z^cKO, cKO, n = 9–16. Error bars represent means ± S.E.M.; one-way ANOVA, P < 0.05 (*), P < 0.01(**) or P < 0.001(***). n.s., not significant. Analysis was performed by an experimenter blinded to experimental conditions.

Figure 6.

H2A.z regulates neurogenesis by targeting Nkx2–4. (A) GO analysis for downregulated genes from E13 forebrain RNA-seq data sets in H2A.z-loss mice. (B) Analysis of gene ontology of biological functions for upregulated genes in H2A.z-depleted E13 forebrain. (C) Genome-wide distribution of H2A.z occupancies in E13 forebrain. (D) GO analyses of H2A.z binding genes. (E) Venn diagram showing the overlap between H2A.z bound genes by ChIP-seq with differentially expressed genes following H2A.z knockout (RNA-seq). (F) GO analyses for H2A.z-bound genes which are downregulated upon deletion of H2A.z. (G) GO analyses for H2A.z-bound genes which are upregulated upon deletion of H2A.z. (H) Volcano plots illustrate differentially expressed genes. Red dots are upregulated genes, and green dots are downregulated genes compared to the control. Nkx2–4 is one of the significantly differentially expressed genes. (I) H3K36me3 occupancy for all genes in E13 WT and H2A.z^cKO forebrain. (J) RPKM values for all genes in E13 WT and H2A.z^cKO forebrain.

Figure 7.

H2A.z and Setd2-mediated H3K36me3 on Nkx2–4 promoter regulate the expression of Nkx2–4. (A) The interaction between H2A.z and Setd2 is verified by co-immunoprecipitation experiments. HA-Setd2-expressing plasmids were transfected into N2A cells. Three days later, the transfected cells were subjected to immunoprecipitation using HA beads. Bound proteins were detected by western blotting with the antibodies indicated. (B) H2A.z promotes Setd2-mediated H3K36me3 modification. HA-Setd2 plasmids were co-transfected with Flag-H2A.z or control into N2A cells. The cell lysates were immunoprecipitated with anti-HA, and bound proteins were detected by western blotting with the antibodies indicated. (C) Higher H3K36me3 levels were pulled down in both the H2A.z and Setd2 overexpression groups. Flag-H2A.z plasmids were co-transfected with HA-Setd2 or control into N2A cells. The cell lysates were immunoprecipitated with anti-Flag and blotted with the antibodies indicated. (D) The levels of H3K36me3 and Nkx2–4 are decreased when both H2A.z and Setd2 are silenced. Primary neural progenitors were infected with lentivirus indicated (D, E and F). Three days later, the cell lysates were analyzed by western blotting with the antibodies indicated. Graph below depicts quantification for Nkx2–4 and H3K36me3, normalized to actin. (E) Higher levels of Nkx2–4 and H3K36me3 are detected when H2A.z and Setd2 are overexpressed, while no obvious increase is investigated when H2A is overexpressed. Graph below depicts quantification for Nkx2–4 and H3K36me3, normalized to actin. (F) Overexpression of H2A.z, but not H2A, can rescue the downregulation of Nkx2–4 and H3K36me3 caused by H2A.z-sh1. Graph below depicts quantification for Nkx2–4 and H3K36me3, normalized to actin. (G) ChIP in primary neural progenitors overexpressing Setd2 with or without H2A.z-sh1, as indicated, performed with antibodies indicated. (H) ChIP in primary neural progenitors overexpressing H2A.z with or without Setd2-shRNA, as indicated, performed with antibodies indicated. (I) ChIP in primary neural progenitors overexpressing H2A.z with or without Setd2, as indicated, performed with antibodies indicated. (J) ChIP analysis of enrichment of H3K36me3 on promoters of Nkx2–4 in primary neural progenitors infected with control, Setd2 or Setd2 together with H2A.z-sh1 lentivirus. A total of 5% of input was loaded for western blotting. Representative images from at least three independent experiments. ChIP enrichment is expressed as percentage of input. n = 3–4 for all samples. Error bars represent means ± S.E.M.; two-tailed unpaired t-test, P < 0.05 (*), P < 0.01(**) or P < 0.001(***). n.s., not significant.