Abstract
Heterozygous mutations in the histone lysine acetyltransferase gene KAT6B (MYST4/MORF/QKF) underlie neurodevelopmental disorders, but the mechanistic roles of KAT6B remain poorly understood. Here, we show that loss of KAT6B in embryonic neural stem and progenitor cells (NSPCs) impaired cell proliferation, neuronal differentiation, and neurite outgrowth. Mechanistically, loss of KAT6B resulted in reduced acetylation at histone H3 lysine 9 and reduced expression of key nervous system development genes in NSPCs and the developing cortex, including the SOX gene family, in particular Sox2, which is a key driver of neural progenitor proliferation, multipotency and brain development. In the fetal cortex, KAT6B occupied the Sox2 locus. Loss of KAT6B caused a reduction in Sox2 promoter activity in NSPCs. Sox2 overexpression partially rescued the proliferative defect of Kat6b−/− NSPCs. Collectively, these results elucidate molecular requirements for KAT6B in brain development and identify key KAT6B targets in neural precursor cells and the developing brain.
Graphical Abstract
Introduction
Genetic mutations in epigenetic regulators commonly underly intellectual disability disorders (Kleefstra et al, 2014; Kochinke et al, 2016). One of the most abundant epigenetic modifications is histone lysine acetylation, catalyzed by lysine acetyltransferases (KATs) and associated with transcriptional activation. Heterozygous mutations in the MYST family histone lysine acetyltransferase gene, KAT6B (MYST4/QKF/MORF) cause the Say-Barber-Biesecker-Young-Simpson variant of Ohdo syndrome, Genitopatellar syndrome, and similar disorders (Clayton-Smith et al, 2011; Campeau et al, 2012; Simpson et al, 2012), defined by a global development delay and cognitive impairment.
The Kat6b mRNA expression pattern has been extensively studied throughout development in mice. Unlike many chromatin regulators, Kat6b is subject to significant regulation at the mRNA level. Kat6b mRNA levels are low at embryonic day 9.5 (E9.5) but become up-regulated from E10.5, with substantial increases observed during cerebral cortex development from E11.5 and peaking in the E15.5 cortex (Thomas et al, 2000). Despite declining during neuronal differentiation, mRNA levels remain elevated in the subventricular neurogenic proliferation zone at E17.5, as well as on postnatal days 0, 1, 7, 14, 21, and in adulthood (Thomas et al, 2000; Merson et al, 2006). Congruently, Kat6b promoter activity is high during adult neurogenesis in the neural stem cell (NSC) population and gradually decreases as cells differentiate (Sheikh et al, 2012). Beyond the developing and adult brain, Kat6b mRNA levels are also elevated in developing facial structures, including the E11.5 and E12.5 frontal nasal and maxillary processes, the E15.5 eyelids, and in tooth primordia (Thomas et al, 2000; Clayton-Smith et al, 2011; Kraft et al, 2011). Elevated Kat6b mRNA or promoter activity were also observed in the anterior aspects of the developing limb bud at E10.5 and E11.5, limb skeletal elements at E12.5, E14.5, and E15.5 (Clayton-Smith et al, 2011; Kraft et al, 2011) and in rib/intercostal structures (Kraft et al, 2011).
Consistent with strong Kat6b expression in the developing cerebral cortex, embryos, fetuses and mice deficient in Kat6b mRNA displayed defects in cortex development. A reduction in the number of proliferating cells in the dorsal telencephalon at E11.5 and reduced numbers of differentiating cells in the developing fetal cerebral cortex were observed, but the rate of cell death was not affected by KAT6B status (Thomas et al, 2000). Consequently, Kat6b deficient mice have a small cortical plate during development and a small cerebral cortex in adulthood with fewer Otx1-positive large cortical pyramidal neurons in cortical layer V and fewer GAD67-positive interneurons in the cortex (Thomas et al, 2000), as well as fewer NSCs in the subventricular zone, a reduced number of migrating neuroblasts in the rostral migratory stream and smaller olfactory bulbs (Merson et al, 2006). Furthermore, Kat6b deficient NSCs isolated from the adult mouse brain display impaired self-renewal and neuronal differentiation (Merson et al, 2006). Despite this clear requirement for KAT6B in the developing and adult brain and in NSCs, how KAT6B controls neural development and progenitor activity remains poorly understood at the molecular level.
NSCs are multipotent cells that can be isolated from the developing and adult brain. In vitro, these cells form a heterogeneous population comprising NSCs and partially or completely lineage restricted progenitor cells, collectively referred to as neural stem and progenitor cells (NSPCs). NSCs and neural precursor cells give rise to the cellular diversity of the developing and adult nervous system (reviewed in Kriegstein and Alvarez-Buylla [2009]). NSCs must sustain adequate self-renewing divisions whilst also committing to neuronal or glial lineages as required. During differentiation, NSCs undergo extensive changes in gene expression, activating lineage-specific gene expression programs and silencing stem cell-associated genes. Various epigenetic mechanisms have been documented to control these processes (Yoon et al, 2018). For example, the histone methyltransferases EZH2, EHMT1, EHMT2, and DOT1L inhibit premature NSPC differentiation in a cell-based model (Ciceri et al, 2024). More specifically, DOTL1 restricts neural progenitor differentiation by ensuring the access of the stem cell transcription factor SOX2 to its target genes, thereby promoting the stem cell transcription program (Ferrari et al, 2020). The histone methyltransferase MLL1 maintains positional information of NSCs (Delgado et al, 2020), and the histone acetyltransferase KAT7 is essential for de novo gene activation during neuronal differentiation (Kueh et al, 2023). These examples underpin the importance of epigenetic control in brain development and NSCs function.
Despite the well-documented importance of KAT6B in brain development, the molecular function of KAT6B in neural cell types, including histone acetylation and gene targets, have not been reported. Using Kat6b loss and gain of function mice, as well as epigenomic and transcriptomic profiling, we report here that KAT6B controls neural precursor cells through the activation of key developmental control genes, including Sox2.
Results
KAT6B is essential for histone H3 lysine 9 acetylation
To investigate molecular targets of KAT6B during mouse development, we used Kat6b mutant mice (Kat6b– [Bergamasco et al, 2024a, 2024b]), which lack exons 2–12 of the endogenous Kat6b gene, and Kat6b BAC transgenic mice (Tg(Kat6b) [Bergamasco et al, 2024a]) which overexpress Kat6b ∼4.5-fold above endogenous levels. The presence of the Tg(Kat6b) transgene rescues the hematopoietic defects of Kat6b−/− mice (Bergamasco et al, 2024a), showing that the Tg(Kat6b) transgene produces functional KAT6B protein.
At embryonic day 12.5 (E12.5), a stage at which KAT6B is highly expressed (Thomas et al, 2000), Kat6b+/− and Kat6b−/− mutant embryos were externally indistinguishable from wild-type controls (Kat6b+/+; Fig 1A). Both genotypes were observed at expected Mendelian ratios in utero; however, Kat6b−/− mice were not present at weaning (3 wk of age; P < 10−6; Fig S1A) and Kat6b+/− animals were 18% underrepresented at weaning compared with controls (P = 5 × 10−6; Fig S1B), indicating that haploinsufficiency for Kat6b impairs survival in early life.
(A) Representative images of E12.5 Kat6b+/+, Kat6b+/− and Kat6b−/− embryos. Scale bar 1 mm. (B, C) RT-qPCR detecting Kat6b mRNA levels in Kat6b+/+, Kat6b+/−, and Kat6b−/− E12.5 dorsal telencephalon (B) and cultured neural stem and progenitor cells (NSPCs) (C), normalized to Gapdh mRNA. (D, E, F, G, H, I, J, K) Fluorescent Western immunoblots and quantification of H3K9ac (D, E, F, G) and H3K23ac (H, I, J, K) with pan H3 loading control in passage 4–5 Kat6b+/+ and Kat6b−/− E12.5 NSPCs (D, E, H, I) or E12.5 dorsal telencephalon (F, G, J, K). Each lane (D, F, H, J) contains protein from an individual embryo or NSPCs isolated from an individual embryo. Each circle (E, G, I, K) represents one lane of the associated immunoblot. 500 ng (D, F) and 250 ng (H, J) of acid extracted protein were loaded per lane. (L, M, N, O, P, Q, R, S) CUT&Tag sequencing results detecting the genomic distribution of H3K9ac, H3K23ac and RNA polymerase II, subunit A (POLR2A) in passage 5 Kat6b−/− versus Kat6b+/+ E12.5 NSPCs. N = 4 embryos per genotype. CUT&Tag data were analyzed as described in the methods section. FDR < 0.05 was considered significant. (L) Coverage plot of H3K9ac reads across the transcription start site ± 1 kb in Kat6b+/+ and Kat6b−/− NSPCs. (M, N, O) Log2-fold change versus average log2 read count (CPM) of H3K9ac in Kat6b−/− versus Kat6b+/+ NSPCs at the promoters (M) and gene bodies of protein coding genes (N) or active enhancers (O) defined as H3K4me1+H3K27ac+ in GSM2406793 and GSM2406791 (Bertolini et al, 2019). Regions with significantly reduced H3K9ac shown in blue and increased H3K9ac in red. (P) H3K9ac read counts per kilobase per one million reads (RPKM) at promoters and gene bodies of all protein coding genes and active enhancers in Kat6b+/+ and Kat6b−/− NSPCs. (Q) Coverage plot of H3K23ac reads across the transcription start site ± 1 kb in Kat6b−/− versus Kat6b+/+ NSPCs. (R) H3K23ac read counts per kilobase per one million reads (RPKM) at promoters and gene bodies of all protein coding genes and active enhancers in Kat6b+/+ and Kat6b−/− NSPCs. (S) Coverage plot of POLR2A reads across the transcription start site ± 1 kb in Kat6b−/− versus Kat6b+/+ NSPCs. (T) POLR2A read counts across the transcription start site ± 1 kb of all protein coding genes and active enhancers in Kat6b+/+ and Kat6b−/− NSPCs. N = NSPC cultures or E12.5 dorsal telencephalon samples from three to five embryos per genotype (A, B, C, D, E, F, G, H, I, J, K) and from four embryos per genotype (L, M, N, O, P, Q, R, S, T). (B, C, E, G, I, K) Each circle represents tissue or cells isolated from an individual embryo (B, C, E, G, I, K). Data are displayed as mean ± SEM, mean (L, Q, S), log2 of the fold-change (M, N, O) or 1–99 percentile plus outliers (P, R, T). Data were analyzed using a one-way ANOVA with Dunnett correction (B, C), an unpaired t test (E, G, I, K), as described in the methods section (L, M, N, O, Q, S) or by Kruskal-Wallis test (P, R, T).
Source data are available online for this figure.
Source Data for Figure 1.1[LSA-2024-02969_SdataF1.1.pdf]
Source Data for Figure 1.2[LSA-2024-02969_SdataF1.2.xlsx]
(A) Number of Kat6b+/+, Kat6b+/− and Kat6b−/− embryos and fetuses recovered from Kat6b+/− x Kat6b+/− matings at embryonic days 12.5–18.5 (E12.5–E18.5) and at weaning (3 wk of age). P-value shown below each time point. (B) Number of Kat6b+/+ and Kat6b+/− animals produced by Kat6b+/+ x Kat6b+/− matings at weaning. P-value shown below each time point. (C, D) RNA sequencing reads observed over Kat6b exons in Kat6b+/+ and Kat6b−/− NSPCs (C) and dorsal telencephalon tissue (D) demonstrating the effect of deleting exons 2–12 in Kat6b−/− samples. (E) RNA-sequencing read counts of Kat6b mRNA in Kat6b+/+ control and Tg(Kat6b) E12.5 dorsal telencephalon and E15.5 cortex confirming overexpression of Kat6b mRNA. N = 21–1,129 embryos, fetuses and mice as detailed in (A, B) and NSPCs, E12.5 dorsal telencephalon or E15.5 developing cortex isolated from 3–4 embryos or fetuses per genotype (C, D, E). Data are shown as number of embryos (A, B) and read count per kilobase per one million reads (RPKM; (C, D, E)). Data were analyzed by Chi-squared test (A, B) or as described for RNA-sequencing analysis in the methods section (C, D, E).
Compared with WT controls, Kat6b heterozygous (Kat6b+/−) and null (Kat6b−/−) samples displayed a 38–48% and 100% reduction in Kat6b mRNA, respectively, in E12.5 dorsal telencephalon and NSPCs (P < 10−6 to 0.001; Figs 1B and C and S1C and D). Conversely, the E12.5 dorsal telencephalon of Tg(Kat6b) mice overexpressed Kat6b approximately fourfold above endogenous levels (Fig S1E).
To investigate the histone lysine targets of KAT6B during mouse development, we assessed histone acetylation at seven lysines on histone H3 and four lysines on histone H4 by western immunoblotting in whole E12.5 embryos either lacking Kat6b (Kat6b−/−) or overexpressing Kat6b [Tg(Kat6b)]. Of the 11 histone residues examined, H3K9 was the only histone residue at which acetylation levels were reduced in Kat6b−/− versus control embryos (47% reduction, P = 8 × 10−6) and increased in Tg(Kat6b) versus control embryos (43% increase, P = 0.03; Fig S2A and B), as would be expected if KAT6B acetylated this residue. H3K14ac was increased 1.7-fold in Kat6b−/− versus control embryos (P = 2 × 10−6) but was unaffected in Tg(Kat6b) embryos. Acetylation levels at the other nine histone H3 and H4 lysine residues (H3K4, H3K18, H3K23, H3K27, H3K56, H4K5, H4K8, H4K12, H4K16) were not significantly affected by loss or gain of KAT6B (Fig S2A and B).
(A) Western immunoblots detecting acetylated histone H3 and H4 lysine residues as indicated in acid extracted protein from whole E12.5 Kat6b+/+ and Kat6b−/− embryos, as well as Kat6b+/+ and Tg(Kat6b) embryos. In the absence of non-interfering pan H3 and pan H4 antibodies, Ponceau S staining is shown as a loading control below each blot. Some immunoblots were probed for an acetylated H3 lysine residue and re-probed for an acetylated H4 lysine residue. These use the same Ponceau S loading control. (B) Densitometry quantitation of the blots shown in (A). N = 3 E12.5 embryos per genotype. Each immunoblot lane represents protein derived from an individual E12.5 embryo. 20 μg of acid extracted protein was loaded per lane. Each circle graphs represents one lane of the associated immunoblot (B). Data are presented as mean ± SEM and were analyzed using a two-way ANOVA with Šídák multiple comparison test.
Source data are available online for this figure.
Source Data for Figure S2[LSA-2024-02969_SdataFS2.pdf]
Congruent with the findings in whole E12.5 embryos, we found a 36% and 12% reduction in H3K9ac in Kat6b−/− NSPCs and E12.5 dorsal telencephalon, respectively, compared with control samples (P = 0.003 and 0.01; Fig 1D–G). In addition, H3K23ac was reduced by 21% in the Kat6b−/− dorsal telencephalon compared with control samples (P = 0.02), but not significantly changed in Kat6b−/− NSPCs (Fig 1H–K). Our data suggest a role for KAT6B in H3K9 acetylation and, in addition, a possible role in H3K23 acetylation in some tissues.
To investigate the locus-specific effects of KAT6B, we assessed H3K9ac, H3K14ac, and H3K23ac levels and RNA polymerase II, subunit A (POLR2A) occupancy by CUT&Tag and DNA accessibility by ATAC-sequencing in Kat6b+/+ and Kat6b−/− NSPCs. In WT samples, H3K9, H3K14 and H3K23 acetylation levels correlated strongly with POLR2A occupancy at all protein coding genes (R2 = 0.5–0.6; Fig S3A) and at brain development genes (GO:0007420∼brain development; R2 = 0.5–0.7; Fig S3B).
CUT&Tag sequencing results detecting the level of H3K9ac, H3K14ac and H3K23ac and RNA polymerase II, subunit A (POLR2A) and DNA accessibility assessed by ATAC-sequencing in NSPCs isolated from E12.5 Kat6b+/+ versus Kat6b−/− embryos at passage 3–5. N = 4 embryos per genotype. Data were analyzed as described in the methods section. FDR < 0.05 was considered significant. (A, B) Correlation of read counts (log2 CPM) between H3K9ac, H3K14ac and H3K23ac levels, ATAC-seq signal and POLR2A occupancy in the gene body (H3K9ac, H3K14ac, H3K23ac) or near the TSS (±1 kb) (POLR2A, ATAC-seq) of all protein coding genes (A) and genes annotated with gene ontology term (GO): 0007420∼brain development (B). R squared values of linear regressions are shown. (C, D, E) Multidimensional scaling plots showing the distance between all pairs of samples calculated using the root-mean-square of the log2-fold changes of the top 500 most variable genes between a given two samples for H3K9ac (C), H3K23ac (D) and POLR2A (E) at the genomic features indicated. (F) Log2-fold change versus average log2 read count (CPM) of POLR2A in Kat6b−/− versus Kat6b+/+ NSPCs at active enhancers defined as H3K4me1+H3K27ac+ in GSM2406793 and GSM2406791 (Bertolini et al, 2019). Numbers of enhancers with significantly reduced POLR2A read counts are shown in blue and those with increased counts in red.
H3K9ac was reduced at the transcription start site (TSS ± 1 kb) in Kat6b−/− versus Kat6b+/+ NSPCs (P < 10−6; Fig 1L). Significant loss of H3K9ac was observed individually at 112 promoters, 3,452 gene bodies and 44 active enhancers (false discovery rate [FDR] < 0.05; Figs 1M–O and S3C; Table S1) and collectively at promoters, gene bodies, and active enhancers (P < 10−6; Fig 1P). While loss of KAT6B did not appear to cause genome-wide effects on H3K23ac in NSPCs (Fig 1H and I), we investigated a locus-specific effect. H3K23ac coverage was reduced at the TSS (±1 kb) in Kat6b−/− versus Kat6b+/+ NSPCs (P < 10−6; Figs 1Q and S3D; Table S1) and was collectively slightly reduced at promoters, gene bodies, and active enhancers (P < 10−6; Fig 1R). POLR2A coverage over the TSS ± 1 kb and at active enhancers was reduced in Kat6b−/− versus Kat6b+/+ NSPCs (P < 10−6; Figs 1S and T and S3E). Significant loss of POLR2A occupancy was observed at 48 individual enhancers (Fig S3F; Table S1).
Table S1. Effects of loss of KAT6B on histone acetylation levels, POLR2A occupancy and DNA accessibility.
Similar to the unexpected increase in H3K14ac in whole Kat6b−/− versus Kat6b+/+ E12.5 embryos (Fig S2A and B), we also observed an increase in H3K14ac by Western blotting in Kat6b−/− versus Kat6b+/+ NSPCs (P = 0.01), but not in the dorsal telencephalon (Fig S4A–D). Assessed by CUT&Tag, H3K14ac was elevated in Kat6b−/− versus Kat6b+/+ NSPCs at 12,180 promoters, 15,943 gene bodies and 5,193 active enhancers (Fig S4E–L; Table S1). Similarly, DNA accessibility assessed by ATAC-sequencing was increased in Kat6b−/− versus Kat6b+/+ NSPCs at 14,000 promoters (P < 10−6; Fig S4M–O; Table S1).
(A, B, C, D) Fluorescent Western immunoblots and quantification of H3K14ac with pan H3 loading in Kat6b+/+ and Kat6b−/− E12.5 NSPCs (A, B) or E12.5 dorsal telencephalon (C, D). Each lane (A, C) contains protein from an individual embryo. Each circle (B, D) represents one lane of the associated immunoblot. 2 μg of acid extracted protein was loaded per lane. (E, F, G, H, I, J, K, L) CUT&Tag sequencing results detecting the level of H3K14ac in passage 5 NSPCs isolated from E12.5 Kat6b+/+ versus Kat6b−/− embryos. N = 4 embryos per genotype. (M, N, O) DNA accessibility assessed by ATAC-sequencing in passage 5 NSPCs isolated from E12.5 Kat6b+/+ versus Kat6b−/− embryos. N = 4 embryos per genotype. (E, F, G, H, I, J, K, L, M, N, O) Data were analyzed as described in the methods section. FDR < 0.05 was considered significant. Active enhancers were defined as H3K4me1+H3K27ac+ in GSM2406793 and GSM2406791, respectively (Bertolini et al, 2019). (E, F, G) Multidimensional scaling plots showing the distance between all pairs of samples calculated using the root-mean-square of the log2-fold changes of the top 500 most variable genes between a given two samples for H3K14ac at the genomic features indicated. (H, I, J) Log2-fold change versus average log2 read count (CPM) of H3K14ac in Kat6b−/− versus Kat6b+/+ NSPCs at the genomic features indicated. Numbers of features with significantly reduced H3K14ac read counts are shown in blue and those with increased counts in red. (K) Mean coverage plot of H3K14ac reads across the transcription start site ± 1 kb in Kat6b+/+ versus Kat6b−/− NSPCs. (L) H3K14ac read counts at promoters and gene bodies of all protein coding genes and active enhancers in Kat6b+/+ and Kat6b−/− NSPCs. (M) Multidimensional scaling plot for ATAC-seq results at promoters of protein coding genes. (N) Log2-fold change versus average log2 ATAC-seq read count (CPM) in Kat6b−/− versus Kat6b+/+ NSPCs at promoters of protein coding genes. Numbers of promoters with significantly reduced accessibility are shown in blue and those with increased accessibility in red. (O) Mean coverage plot of ATAC-seq reads across the transcription start site ± 1 kb in Kat6b+/+ versus Kat6b−/− NSPCs.
Source data are available online for this figure.
Source Data for Figure S4[LSA-2024-02969_SdataFS4.pdf]
Taken together, our data suggest that KAT6B is essential for normal H3K9 acetylation levels at a large number of genes. In contrast, the requirements for H3K23ac were more restricted to a smaller number of genes. Curiously, we found that the presence of KAT6B limited H3K14ac. Our peculiar observations of increased H3K14ac and DNA accessibility in Kat6b−/− versus Kat6b+/+ may suggest increased activity of another histone acetyltransferase.
Loss of KAT6B caused a reduction in expression of genes required for brain development
KAT6B is highly expressed in the neurogenic subventricular zone and NSCs (Merson et al, 2006; Sheikh et al, 2012). To assess the role of KAT6B in gene expression in NSPCs, we performed RNA-sequencing. RNA-sequencing profiles of Kat6b−/− versus Kat6b+/+ NSPCs segregated by genotype (Fig 2A). Kat6b deletion resulted in significant down-regulation of 3,861 genes and up-regulation of 3,833 genes, while Kat6b overexpression [Tg(Kat6b)] caused the up-regulation of 2,322 genes and the down-regulation of 2,561 genes (FDR< 0.05; Figs 2B and C and S5A–C; Table S2). Genes down-regulated in Kat6b−/− versus Kat6b+/+ NSPCs were generally up-regulated in Tg(Kat6b) NSPCs (P = 0.0001; Fig 2D), supporting a role of KAT6B in the direct or indirect regulation of these genes. Genes down-regulated in Kat6b−/− versus Kat6b+/+ NSPCs were enriched for nervous system development, gene transcription and metabolic processes, whereas up-regulated genes lacked such annotations (Figs 2E and S5D, Table S2). A subset of genes essential for NSC function was down-regulated (Fig S5E). The top 20 brain development genes (GO:0007420) down-regulated at the RNA level displayed a reduction in both H3K9ac and mRNA levels (Fig 2F; Tables S1 and S2) and included early regulators of neuroepithelium, brain development, and NSC function, e.g., Sox1, Sox2, Pax6, and Id2 (Schmahl et al, 1993; Malas et al, 2003; Ferri et al, 2004; Niola et al, 2012), as well as regulators of neurite elongation, Sall1, Hap1, and Rac3 (Orioli et al, 2006; Rong et al, 2006; Harrison et al, 2008), identifying these as putative KAT6B target genes. Curiously, these same down-regulated genes had an increase in H3K14ac and DNA accessibility (Fig S5F). Neural precursor proliferation genes (GO:0061351) were profoundly affected by loss of KAT6B (Fig 2G). At brain development genes and neural precursor proliferation genes, we observed a positive correlation between H3K9ac and POLR2A read counts per gene (log2 count per million [CPM]; Fig 2H and I). Correlations with RNA levels were strongest when the size of the gene was taken into consideration (log2 RPKM; Fig 2J–M). Positive correlations were observed between H3K9ac and mRNA levels (R2 = 0.6 and 0.5, respectively; both P < 10−6) and between H3K9ac and POLR2A levels (both R2 = 0.7 and P < 10−6; Fig 2H–M). Notably, the fold-changes in H3K9ac and mRNA levels due to Kat6b deletion also correlated significantly in brain development genes (R2 = 0.2; P = 0.002), and neural precursor proliferation genes (R2 = 0.3; P = 0.004; Fig 2N and O).
(A, B, C, D, E, G) RNA-sequencing data of N = 3 Kat6b−/− versus 3 Kat6b+/+ passage 5 NSPC isolates and 6 Tg(Kat6b) versus 2 Kat6b+/+ NSPC isolates. (F, H, I, J, K, L, M, N, O) Comparison of RNA-sequencing results to CUT&Tag and ATAC-seq results. For CUT&Tag and ATAC-seq experiments, N = 4 embryos per genotype. The CUT&Tag, ATAC- and RNA-sequencing data analyses are described in the methods section. FDR < 0.05 was considered significant. (A) Multidimensional scaling plots showing the distance between all pairs of samples calculated using the root-mean-square of the log2-fold changes of the top 500 most variable genes between any given two samples. (B, C) Volcano plot of −log10(P-value) versus log2 fold-change in RNA expression in Kat6b−/− versus Kat6b+/+ NSPCs (B) and Tg(Kat6b) versus Kat6b+/+ NSPCs (C). Significantly down-regulated genes shown in blue (B) or purple (C) and up-regulated genes in red (B) or orange (C) (FDR < 0.05). (D) Barcode plot showing the correlation between genes differentially expressed in Kat6b−/− versus Kat6b+/+NSPCs and Tg(Kat6b) versus Kat6b+/+ NSPCs. The block represents genes of the contrast Kat6b−/− (KO) versus Kat6b+/+ (WT) NSPCs ordered from down-regulated, left, to up-regulated, right, with the t-statistic on the x axis. The vertical lines represent genes in the contrast Tg(Kat6b) versus Kat6b+/+ NSPCs, their height represents the log2 fold-change. The worms indicate the enrichment, red for the genes up-regulated and blue for the genes down-regulated in Tg(Kat6b) versus Kat6b+/+ NSPCs. (E) Top 20 gene ontology (GO) terms, biological processes (BP), enriched in genes down-regulated in Kat6b−/− versus Kat6b+/+ NSPCs. (F) Log2 fold-changes in RNA and in H3K9ac of the top 20 differentially expressed brain development genes (GO:0007420) in Kat6b−/− versus Kat6b+/+ NSPCs. (G) Heatmap of neural precursor proliferation genes (GO:0061351) differentially expressed with FDR < 0.01 in Kat6b−/− versus Kat6b+/+ NSPCs. Each column represents a NSCs isolate from and individual E12.5 embryo. (H, I) Correlation of read counts (log2 CPM) levels between histone acetylation, DNA accessibility, POLR2A occupancy and RNA level in NSPCs for brain development genes (GO:0007420; (H)) and neural precursor proliferation genes (GO:0061351; (I)). (J, K, L, M) Correlation of read counts (log2 RPKM) between H3K9ac in gene bodies (J, K, L, M) and RNA (J, L) or POLR2A occupancy at the TSS (K, M) assessing brain development genes (GO:0007420; (J, K)) or neural precursor proliferation genes (GO:0061351; (L, M)). (N, O) Correlation between log2 fold changes in RNA and H3K9ac levels in gene bodies in Kat6b−/− versus Kat6b+/+ NSPCs, assessing brain development genes (GO:0007420; (N)) or neural precursor proliferation genes (GO:0061351; (O)) down-regulated in Kat6b−/− versus Kat6b+/+ NSPCs.
Source data are available online for this figure.
Source Data for Figure 2[LSA-2024-02969_SdataF2.xlsx]
(A, B, C, D, E, G, H, I, J, K, L) RNA-sequencing data of Kat6b+/+ and Kat6b−/− passage 5 NSPCs (A, B, C, D, E), Kat6b+/+ and Kat6b−/− NSPCs and Tg(Kat6b) versus Kat6b+/+ NSPCs (G), and of Kat6b+/+ and Kat6b−/− E12.5 dorsal telencephalon (H, I, J), E12.5 dorsal telencephalon c.f. NSCs (K, L, M, N), and E15.5 cortex (H, O, P). (A) Heatmap of the top 100 genes differentially expressed in Kat6b+/+ and Kat6b−/− NSPCs (FDR < 10−5). (B, C) Tissue annotation of genes down (B) and up-regulated (C) in Kat6b+/+ and Kat6b−/− NSPCs. (D) Top 20 gene ontology (GO) terms, biological processes (BP), enriched in genes up-regulated in Kat6b−/− versus Kat6b+/+ NSPCs. (E) Log2 fold change in RNA levels in Kat6b−/− versus Kat6b+/+ NSPCs for selected genes required for neural stem cell function. FDRs are indicated above the bars. (F) Log2 fold-change in H3K14ac and ATAC-sequencing reads at the top 20 brain development genes (GO:0007420) down-regulated in Kat6b−/− versus Kat6b+/+ NSPCs. Note that despite apparent increase in H3K14ac and DNA accessibility, mRNA levels in Kat6b−/− versus Kat6b+/+ NSPCs were reduced (RNA and H3K9ac shown in Fig 2F). (G) Effects of loss and gain of KAT6B on the expression of histone acetyltransferase, histone deacetylase and protein complex partner genes. Eight histone acetyltransferase genes (Crebbp, Ep300, Kat2a, Kat2b, Kat5, Kat6a, Kat7, Kat8), nine KAT6A, KAT6B and KAT7 protein complex members (Brpf1, 2, 3, Jade1, 2, 3, Ing4, 5, Meaf6) and 18 histone deacetylases (HDAC and sirtuin genes) in NSPCs were examined to detect possible compensatory changes. The four genes depicted were the genes with a change with an FDR < 0.05 and a direction of change opposite between Kat6b deletion and Kat6b overexpression. FDRs are indicated above and below the bars. (H) Multidimensional scaling plot for Kat6b+/+ and Kat6b−/− E12.5 dorsal telencephalon and E15.5 cortex RNA sequencing data. (I) Tissue annotation of genes down-regulated in Tg(Kat6b) versus Kat6b+/+ E12.5 dorsal telencephalon. (Note, only 14 genes were up-regulated in Kat6b−/− versus Kat6b+/+ E12.5 dorsal telencephalon, which did not attract a tissue annotation.). (J) Gene ontology terms (GO) terms, biological processes (BP) enriched in genes up-regulated (FDR < 0.05) in Kat6b−/− versus Kat6b+/+ E12.5 dorsal telencephalon. (K) Correlation of mRNA levels per gene (RNA-sequencing reads counts per protein-coding gene in counts per million reads within the library [log2 CPM]) between NSPCs and E12.5 dorsal telencephalon. A positive correlation was observed (R2 = 0.65, P < 10−6). (L) Venn diagram showing the overlap in genes down-regulated (DR) or up-regulated (UR) in NSPCs and E12.5 dorsal telencephalon, each with transcriptome-wide significance (FDR < 0.05) of the genes that were expressed in both tissue/cell sources. As expected, based on the difference in tissue/cell source, only few genes were regulated in an overlapping fashion. 16 genes were mutually up-regulated, and one gene was mutually down-regulated. (M) Numerical presentation of the genes down-regulated (DR) or up-regulated (UR) and overlappingly down-regulated (DR) or up-regulated (UR) in NSPCs and E12.5 dorsal telencephalon of the genes that were expressed in both tissue/cell sources. Over 5,600 of these genes were differentially expressed in the NSPCs with transcriptome-wide significance (FDR < 0.05), but only 50 in E12.5 dorsal telencephalon. Despite the difference in tissue/cell source, the number of mutually down-regulated genes (16) was significantly enriched (P = 5 × 10−5). (N) Log2-fold-change in mRNA levels of the genes mutually down-regulated in NSPCs and E12.5 dorsal telencephalon with FDRs indicated above the bars. (O, P) Heatmaps displaying genes differential expressed in Kat6b−/− versus Kat6b+/+ (O) and Tg(Kat6b) versus Kat6b+/+ E15.5 cerebral cortex (P). N = NSPCs isolated from three E12.5 embryos per genotype (A, B, C, D, E, K, L, M, N), four embryos per genotype (F), NSPC isolates from two to six E12.5 embryos per genotype (G), three to five E12.5 dorsal telencephalons or E15.5 cortices per genotype (H), three dorsal telencephalon from E12.5 embryos per genotype (I, J, K, L, M, N) and E15.5 cortex isolated from five (O) and four (P) fetuses per genotype. The CUT&Tag, ATAC-seq and RNA-seq data analyses are described in the methods section. FDR < 0.05 was considered significant.
We explored a possible compensatory regulation of other histone acetyltransferase or deacetylases. We postulated that any compensatory mechanism would operate in opposite directions in Kat6b−/− versus Kat6b+/+ and Tg(Kat6b) versus Kat6b+/+ NSPCs. Of the eight genes that, apart from Kat6b, encode nuclear histone acetyltransferases with defined acetyl-co-enzyme A binding sites, only Kat2a (Gcn5) mRNA was statistically significantly changed, namely up-regulated in Kat6b−/− versus Kat6b+/+ and down-regulated in Tg(Kat6b) versus Kat6b+/+ NSPCs (Fig S5G). In addition, two histone deacetylase genes, Hdac11 and Sirt2 were down-regulated in Kat6b−/− versus Kat6b+/+ and up-regulated in Tg(Kat6b) versus Kat6b+/+ NSPCs (Fig S5G).
Loss of KAT6B caused down-regulation of brain development genes in vivo
Examining the effects of loss and gain of KAT6B on the developing cortex, we found 64 genes were differentially expressed in Kat6b−/− versus Kat6b+/+ E12.5 dorsal telencephalon (FDR < 0.05; 50 down-regulated and 14 up-regulated) and 1,405 genes were differentially expressed in Tg(Kat6b) versus Kat6b+/+ E12.5 dorsal telencephalon (779 up-regulated and 626 down-regulated; Figs 3A–C and S5H; Table S3). Genes down-regulated in Kat6b−/− versus Kat6b+/+ or up-regulated in Tg(Kat6b) versus Kat6b+/+ dorsal telencephalon were strongly associated with brain tissue expression (Fig 3D and E), whereas genes up-regulated in Kat6b−/− versus Kat6b+/+ did not attract a specific tissue annotation and genes down-regulated in Tg(Kat6b) versus Kat6b+/+ dorsal telencephalon were not significantly associated with expression in the brain (Fig S5I). As expected, a negative correlation was observed between the effects of loss and gain of KAT6B on gene expression (P = 0.02; Fig 3F). Genes down-regulated in Kat6b−/− versus Kat6b+/+ E12.5 dorsal telencephalon were enriched for brain development and neuronal differentiation GO (BP) terms (Fig 3G; Table S3), whereas up-regulated genes did not prominently relate to brain development or function (Fig S5J). Loss of KAT6B resulted predominantly in the down-regulation of brain development genes (GO:0007420; FDR < 0.05; Fig 3H). Among the down-regulated genes were Lhx2 (Hsu et al, 2015) and Pou3f3 (Sugitani et al, 2002), genes individually required for neural precursor proliferation, and Bhlhe22 and Prdm8 (Ross et al, 2012), Emx1 (Qiu et al, 1996), Eomes (Sessa et al, 2008), Lhx2 (Porter et al, 1997), Neurod2 (Olson et al, 2001), Pou3f3 (Sugitani et al, 2002), Otx1 (Acampora et al, 1996), and other genes that are essential for normal brain and neuronal development. Genes associated with central nervous system neuron development (GO:0021954) were highly enriched among down-regulated genes in the absence of KAT6B and enriched among up-regulated genes when Kat6b was overexpressed (P = 0.0008 and 0.005, respectively; Fig 3I and J). Gene expression levels correlated positively between E12.5 dorsal telencephalon and NSPCs and 16 genes were mutually down-regulated with transcriptome-wide significance in E12.5 dorsal telencephalon and NSPCs (Fig S5K–N). Comparably few genes were detected as differentially expressed in E15.5 Kat6b−/− versus Kat6b+/+ or Tg(Kat6b) versus Kat6b+/+ cortex (Fig S5H, O, and P; Table S3). This may reflect the increased cell type complexity of the E15.5 cortex compared with the E12.5 dorsal telencephalon or NSPCs. Overall, our data suggest that loss and gain of KAT6B predominantly affects the expression of brain development genes in NSPC and the developing cortex.
(A, B, C, D, E, F, G, H, I, J) RNA-sequencing data of E12.5 dorsal telencephalon. N = 3 Kat6b−/− versus 3 Kat6b+/+ and 4 Tg(Kat6b) versus 4 Kat6b+/+ embryos. RNA sequencing data were analyzed as described in the methods section. FDR < 0.05 was considered significant. (A) Heatmap of the genes differentially expressed in Kat6b−/− versus Kat6b+/+ E12.5 dorsal telencephalon with transcriptome-wide significance (FDR < 0.05). (B, C) Volcano plot of -log10(P value) versus log2 fold-change in RNA expression in Kat6b−/− versus Kat6b+/+ (B) and Tg(Kat6b) versus Kat6b+/+ dorsal telencephalon (C). Down-regulated genes indicated in blue (B) or purple (C) and up-regulated genes in red (B) or orange (C) (FDR < 0.05). (D, E) Tissue annotation for genes down-regulated in Kat6b−/− versus Kat6b+/+ (D) and up-regulated in Tg(Kat6b) versus Kat6b+/+ dorsal telencephalon (E). (F) Barcode plot showing the correlation between genes differentially expressed in Kat6b−/− versus Kat6b+/+ NSPCs and Tg(Kat6b) versus Kat6b+/+ dorsal telencephalon. (G) Top 20 gene ontology (GO) terms, biological processes (BP), enriched in genes down-regulated in Kat6b−/− versus Kat6b+/+ dorsal telencephalon. (H) Brain development genes (GO:0007420) differentially expressed in Kat6b−/− versus Kat6b+/+ dorsal telencephalon. (I, J) Barcode plots depicting the results of gene set analyses for genes annotated with central nervous system development (GO:0021954) showing their enrichment in genes differentially expressed in Kat6b−/− versus Kat6b+/+ NSPCs (I) and Tg(Kat6b) versus Kat6b+/+ dorsal telencephalon (J).
Source data are available online for this figure.
Source Data for Figure 3[LSA-2024-02969_SdataF3.xlsx]
Impaired proliferation, self-renewal, and neuronal differentiation in embryonic Kat6b−/− NSPCs
Embryonic NSPC colonies (neurospheres) derived from the Kat6b−/− E12.5 dorsal telencephalon appeared smaller, proliferated slower (P < 10−6) and gave rise to fewer secondary neurospheres compared with WT control NSPCs (P = 0.03; Fig 4A–C). Kat6b−/− NSPCs showed a 1.3-fold greater proportion of cells in G0 of the cell cycle (P = 0.0003) and 8% decrease in the total percentage of proliferating (Ki67+) cells (P = 0.03) compared with Kat6b+/+ controls (Figs 4D–F and S6A and B). No difference in cell viability was observed between genotypes (Fig S6C–F).
(A) Cumulative cell counts of NSPC cultures derived from E12.5 dorsal telencephalon tissue of Kat6b+/+ or Kat6b−/− animals over 10 passages. (B) Representative images of Kat6b+/+ and Kat6b−/− neurosphere colonies at passage 4. Scale bar 100 μm. (C) Number of secondary neurospheres generated by passage 5 Kat6b+/+ and Kat6b−/− NSPC cultures in a dilution series. Circles represent averages of three technical replicates per three to four biological replicates per genotype and dilution. (D, E, F) Cell cycle analysis of Kat6b+/+ and Kat6b−/− NSPCs. Representative flow cytometry plot of Ki67 versus DAPI staining of Kat6b+/+ and Kat6b−/− NSPCs with the cell cycle stages G0, G1, S, G2 and M indicated (D). (E, F) Percentage of passage 3 NSPCs in each cell cycle stage (E) and percentage of proliferating (Ki67+) cells (F) in Kat6b+/+ and Kat6b−/− NSPC cultures. (G) Median fluorescence intensity of SOX2 protein per cell detected by intranuclear immuno-staining and flow cytometry in passage 3 NSPCs. (H, I) Representative images (H) and quantification (I) of passage 3 Kat6b+/+ and Kat6b−/− NSPC cultures grown for 6 d in differentiating conditions, stained to detect astrocytes (GFAP+, blue) and neurons (βIII tubulin+, red). Scale bar 100 μm. (J, K, L) Quantification of the percentage of SOX2–βIII tubulin+ (J) and SOX2–GFAP+ (K) cells in passage 3 Kat6b+/+ and Kat6b−/− NSPCs grown for 6 d in differentiating conditions and median fluorescence intensity of βIII tubulin per cell in βIII tubulin+ cells (L) assessed by flow cytometry. (M, N) Quantification of the average primary neurite length (M) and number of secondary neurites (N) in βIII tubulin and DAPI stained Kat6b+/+ and Kat6b−/− E16.5 cortical neurons after 5 d in culture. 100–200 neurons assessed per fetus. N = 3–4 (A, C), 3–5 (E, F, G, J, K, L), 3 (I) and 4 (M, N), embryos or fetuses per genotype. Each circle represents NSPCs (A, C, E, F, G, I, J, K, L) and E16.5 cortical neurons (M, N) derived from an individual embryo or fetus. Data are presented as mean ± SEM and were analyzed using a two-way ANOVA with Šidák correction (A, C, E, I) or unpaired t test (F, G, J, K, L, M, N).
Source data are available online for this figure.
Source Data for Figure 4[LSA-2024-02969_SdataF4.xlsx]
(A, B) Gating strategy for the assessment of the cell cycle by Ki67 and DAPI staining in NSPCs. (A) NSPCs were gated on single cells (doublet excluded) and size selected DAPI stained cells (A). (B) G0, G1, S, G2 and M cell cycle phase gates shown in a Ki67 stained sample and in a sample stained with an isotype control (B). (C, D) Gating strategy to assess cell death and apoptotic cell death. NSPCs were stained for annexin V and propidium iodide (PI). Positive control (cell death induced by incubation at 55°C for 20 min) (C) and uninduced sample (D). (E) Percentage live, early apoptotic, and late apoptotic cells in Kat6b+/+ and Kat6b−/− passage 3 NSPCs based on annexin V versus PI staining. (F) Viable cells based on trypan blue exclusion in cultures of Kat6b+/+ and Kat6b−/− NSPCs over 10 passages. (G) Gating strategy to assess the marker and regulator of NSPCs, SOX2, by flow cytometry. Cell doublet were excluded, and cells were size selected. Populations were gated on viable cells (negative for a LIVE/DEAD Fixable marker, an indicator of viability before fixation). (H) Percentage SOX2+ cells in Kat6b+/+ and Kat6b−/− passage 3 NSPCs. (I) Representative histogram of the median fluorescence intensity of Kat6b+/+ and Kat6b−/− NSPCs stained for SOX2 assessed by intranuclear flow cytometry. Each circle (E, F, H) represents NSPCs derived from the dorsal telencephalon of an individual embryo. N = 3–5 embryos per genotype. Data are presented as mean ± SEM and were analyzed using a two-way ANOVA with Šidák correction (E, F) and an unpaired t test (H).
Proliferating Kat6b−/− NSPCs contained a similar percentage of cells positive for the NSC marker and regulator SOX2 but showed a 30% reduction in the level of SOX2 protein levels per cell compared with Kat6b+/+ cells (P = 0.02; Figs 4G and S6G–I). Under differentiating conditions, Kat6b−/− NSPCs gave rise to 46% fewer βIII tubulin+ neurons (P = 0.01) and proportionally more GFAP+ astrocytes (P = 0.01) than Kat6b+/+ controls assessed by immunofluorescence (Fig 4H and I). Congruently, by flow cytometry, Kat6b−/− NSPCs gave rise to 68% fewer SOX2–βIII tubulin+ neurons, 37% more SOX2–GFAP + astrocytes (P = 0.01 and P = 0.04, respectively) and also showed a 58% reduction in βIII tubulin protein levels (P = 0.009; Figs 4J–L and S7A–C). Kat6b−/− cortical neurons displayed shorter primary neurites (P = 0.003) and fewer secondary neurites (P = 0.04) compared with Kat6b+/+ controls (Figs 4M and N and S7D–F).
(A) Gating strategy to assess differentiating NSPCs by flow cytometry. Cell doublets were excluded. Cells were size selected and gated on viable cells (negative for a LIVE/DEAD Fixable marker, as an indicator of viability before fixation). βIII tubulin+ neurons and GFAP+ astrocytes were assessed in the SOX2– population. (B, C) Representative flow cytometry plots for βIII tubulin+ (B) or GFAP+ (C) cells in passage 3 Kat6b+/+ and Kat6b−/− NSPCs grown for 6 d in differentiating conditions, gated on SOX2– cells with quantification of the median fluorescence intensity for GFAP (C) in GFAP+ Kat6b+/+ and Kat6b−/− cells. MFI for βIII tubulin is shown in Fig 4L. (D, E, F) Representative images of βIII tubulin (red) and DAPI (blue)-stained Kat6b+/+ and Kat6b−/− E16.5 cortical neurons plated at 10,000 cells/cm2 (D). Scale bar = 100 μm. Quantification of the average length (μm) of secondary and tertiary neurites ((E); primary neurite length displayed in Fig 4M) and average number of primary and tertiary neurites per cell ((F); number of secondary neurites displayed in Fig 4N). Average of 100–200 neurons per fetus. N = cell isolates from three to five embryos (C) and four fetuses (E, F) per genotype. Each circle represents a cell culture derived from the E12.5 dorsal telencephalon (C) or E16.5 cortex (E, F) of an individual embryo or fetus. Data are presented as mean ± SEM and were analyzed using an unpaired t test (C) or a two-way ANOVA with Šidák correction (E, F).
Overall, loss of KAT6B appeared to affect neurogenesis from embryonic NSPCs at several levels including stem cell self-renewal, proliferation, neuronal differentiation, and neurite outgrowth. KAT6B appeared to be required for normal protein levels of the stem cell transcription factor SOX2 and the neuronal marker βIII tubulin.
KAT6B promotes SOX family gene expression and directly targets the Sox2 and Pax6 genes
Among the brain development genes prominently affected by the absence of KAT6B was the SOX gene family. In NSPCs, SOX2 protein levels were reduced (Fig 4G), Sox1, 2, 4, 8, 9, and 21 were among the top 20 brain development genes and Sox2, 5, and 10 were among the top neural precursor proliferation genes that were down-regulated in Kat6b−/− versus Kat6b+/+ NSPCs (Fig 2F and G). We therefore examined the effects of KAT6B on the SOX gene family in further detail. Loss of KAT6B prominently down-regulated most SOX family genes in NSPCs, E12.5 dorsal telencephalon, and E15.5 cortex (FDR within gene family = 0.05 to 3 × 10−6; Figs 5A–C and S8A). The SOX gene family was enriched among genes down-regulated in Kat6b−/− versus Kat6b+/+ E12.5 dorsal telencephalon and E15.5 cortex (P = 0.003 and 0.001, respectively; Fig S8B and C). Congruently, some SOX genes were up-regulated when Kat6b was overexpressed (FDR within gene family = 0.05 to 3 × 10−5; Fig 5D–F). Consistent with KAT6B promoting Sox2 expression, we observed a strong positive correlation between genes differentially expressed in Kat6b−/− versus control NSPCs and those lacking Sox2 (Bertolini et al, 2019) (P = 0.0006; Figs 5G and S8D). Of 739 genes that were differentially expressed in both, Kat6b−/− versus Kat6b+/+ and Sox2−/− versus Sox2+/+ NSPCs, 91% were changed in the same direction in both comparisons, 640 down-regulated and 31 up-regulated (Table S4), indicating similar effects of loss of KAT6B and loss of SOX2 on gene transcription in NSPCs. Of the top 20 genes down-regulates in either Kat6b−/− versus Kat6b+/+ or Sox2−/− versus Sox2+/+ NSPCs, 16 were down-regulated in both (Fig S8). Genes down-regulated in both comparisons were enriched for nervous system development GO terms (Fig S8F) and included genes mutated in human nervous system disorders, including autism spectrum disorder genes, epilepsy genes, developmental disorders, and several neurodegenerative disorders (Fig S8G).
(A, B, C, D, E, F) Log2 fold change in mRNA for SOX family genes in Kat6b−/− versus Kat6b+/+ (A, B, C) and Tg(Kat6b) versus Kat6b+/+(D, E) passage 5 NSPCs (A, D), E12.5 dorsal telencephalon (B, E) and E15.5 cortex (C, F). An average expression cut-off of > 32 counts per one million reads was used. FDRs < 0.05 shown above or below each bar. (G) Barcode plot showing the correlation of genes differentially expressed in Kat6b−/− versus Kat6b+/+ NSPCs (this study) and Sox2−/− versus control NSPCs (Bertolini et al, 2019). (H, I, J, K, L, M, N) ChIP-qPCR results detecting: (H, L) KAT6B-V5 occupancy at the Sox2 and Pax6 promoters in in Kat6bV5/V5 versus Kat6b+/+ E15.5 cortex. (I, J, M) H3K9ac levels in Kat6b−/− versus Kat6b+/+ (I, M) and Tg(Kat6b) versus Kat6b+/+ E15.5 cortex (J) at the Sox2 (I, J) and Pax6 promoters (M). (K, N) H3K14ac levels in Kat6b−/− versus Kat6b+/+ E15.5 cortex at the Sox2 (K) and Pax6 promoters (N). (O) Read depth plots of CUT&Tag sequencing reads displaying H3K9ac and POLR2A occupancy at the Sox2 and Pax6 genes showing representative plots of N = 4 Kat6b+/+ and 4 Kat6b−/− NSPC isolates. (P) Representative flow cytometry plot showing the gating strategy for SSEA1 and CD133 double positive NSCs and histogram showing Sox2-GFP positive cells in Sox2Gfp/+ and GFP negative Sox+/+ cells isolated from E12.5 dorsal telencephalon tissue. (Q) Median fluorescence intensity of GFP in SSEA1+CD133+ cells from Kat6b+/+, Kat6b+/− and Kat6b−/− embryos with and without the Sox2Gfp allele. (R) Cumulative growth curve for NSPCs isolated from Kat6b−/− and Kat6b+/+ embryos and transfected with a Sox2 overexpression vector (Sox2-pMIG; pMSCV-Sox2-IRES-GFP II) or empty vector (pMIG; pMIG II) at passage 3. N = 2–6 NSPC cultures, 3–4 E12.5 dorsal telencephalon or 4–5 E15.5 developing cortex samples per genotype (A, B, C, D, E, F, G), 4–5 E15.5 cortex samples (H, I, J, K, L, M, N), NSPCs from four embryos per genotype (O, R) and 4–17 E12.5 dorsal telencephalon samples (Q). Each circle represents NSPCs or tissue from an individual embryo or fetus (H, I, J, K, L, M, N, Q). Data were analyzed as stated in the methods section (A, B, C, D, E, F, G), using an unpaired t test (H, I, J, K, L, M, N), one-way (Q) or two-way ANOVA with Dunnett correction (R).
Source data are available online for this figure.
Source Data for Figure 5[LSA-2024-02969_SdataF5.xlsx]
(A, B, C) RNA-sequencing results for the SOX gene family in Kat6b−/− versus Kat6b+/+ NSPCs (A), E12.5 dorsal telencephalon (B) and E15.5 cortex (C) displayed as a heatmap of RPKM (A) and barcode plots showing enrichment among down-regulated genes (B, C). RNA-sequencing data were analyzed as described in the methods section. (D) Correlation of the log2-fold-changes in expression of genes that were differentially expressed in both, Kat6b−/− versus Kat6b+/+ and Sox2−/− versus Sox2+/+ NSPCs. (E) Log2-fold-changes of the top 20 genes most strongly down-regulated in Sox2−/− versus Sox2+/+ NSPCs and also shown for Kat6b−/− versus Kat6b+/+ NSPCs. (F) GO term (biological process, BP) enriched in genes down-regulated in both, Kat6b−/− versus Kat6b+/+ and Sox2−/− versus Sox2+/+ NSPCs. (G) Human nervous system disorders associated with genes down-regulated in both, Kat6b−/− versus Kat6b+/+ and Sox2−/− versus Sox2+/+ NSPCs. (H, I) Schematic drawing of the construct used to target the Kat6b gene in ES cells to insert a triple tag (Flag-V5-biotinylation signal sequence) at the 3′end of the Kat6b coding sequence (H). ES cells were used to generate chimeric mice and a mouse strain. PCR genotyping gel (I). Mice homozygous for the triple tag insertion were viable and fertile, unlike Kat6b−/− mice, which died at birth, indicating that the KAT6B-Flag-V5-BIO protein was functional. (J, K, L, M) ChIP-qPCR results for acetylation levels at H3 and H4 lysine residues at the Sox2 promoter (J, K) and the Pax6 promoter (L, M) in Tg(Kat6b) versus Kat6b+/+ (J, L) and Kat6b−/− versus Kat6b+/+ E15.5 cortex (K, M). N = NSPCs, E12.5 dorsal telencephalon and E15.5 cortex isolated from two to five embryos or fetuses per genotype. Each circle represents a cortex from an individual E15.5 fetus (J, K, L, M). Data are displayed as mean and individual data points (J, K, L, M) and were analyzed by Welch t test. RNA-seq data analyses are described in the methods section.
To determine whether KAT6B could be a direct regulator of Sox2, we assessed if KAT6B bound the Sox2 locus. Given the absence of reliable antibodies against KAT6B, we generated mice in which we tagged the endogenous KAT6B with a V5-tag (Kat6bV5; Fig S8H and I). Unlike Kat6b−/− mice, which on a C57BL/6 genetic background die at birth, Kat6bV5/V5 mice were viable and fertile, indicating that the KAT6B-V5 protein was functional. We used anti-V5 antibodies to detect the tagged form of KAT6B in Kat6bV5/V5 mice by ChIP-qPCR. We observed a 3.2-fold enrichment of KAT6B-V5 binding to the Sox2 promoter in Kat6bV5/V5 compared with Kat6b+/+ E15.5 cortex (P = 0.002; Fig 5H). Furthermore, H3K9ac levels were reduced at the Sox2 promoter in Kat6b−/− versus Kat6b+/+ and increased in Tg(Kat6b) versus Kat6b+/+ E15.5 cortex (P = 0.01 and 0.006; Fig 5I and J). H3K14ac levels were increased in Kat6b−/− versus Kat6b+/+ in E15.5 cortex (P = 0.006; Fig 5K).
Apart from the Sox gene family, other neural development genes were also affected by loss or gain of KAT6B (Figs 2 and 3), including the paired homeodomain transcription factor gene Pax6, a NSC gene (Fig 2F and G). We found a 2-fold enrichment of KAT6B-V5 occupancy at the Pax6 promoter in Kat6bV5/V5 compared with Kat6b+/+ E15.5 cortex (P = 0.01; Fig 5L). H3K9ac levels were reduced (P = 0.004) and H3K14ac levels increased (P = 0.002) at the Pax6 promoter in Kat6b−/− versus Kat6b+/+ E15.5 cortex (Fig 5M and N). Acetylation at other histone residues was not affected by loss or gain of KAT6B at the Sox2 and Pax6 promoters (Fig S8J–M). Read depth plot examination suggested that loss of KAT6B was associated not only with a reduction in H3K9ac, but also a reduction in POLR2A occupancy at the Sox2 and Pax6 genes (Fig 5O) and at other Sox genes (Fig S9).
Read depth plots of CUT&Tag sequencing reads displaying H3K9ac and POLR2A occupancy at the Sox1, 3, 4, 8, 9, and 21 gene loci showing representative plots of N = 4 Kat6b+/+ and 4 Kat6b−/− NSPC isolates. Sox2 locus displayed in Fig 5O.
To further interrogate the effect of KAT6B on SOX2 expression, we assessed Sox2 promoter activity in NSCs isolated from E12.5 dorsal telencephalon of Kat6b+/+, Kat6b+/− and Kat6b−/− embryos that also expressed a Gfp reporter gene driven by the Sox2 promoter (Sox2GFP) (Arnold et al, 2011) by flow cytometry. Gating on SSEA1 and CD133 double positive cells, enriched for NSCs (Capela & Temple, 2002; Corti et al, 2007), we observed a gene-dose dependent reduction in Sox2-GFP levels in Kat6b+/− and Kat6b−/− compared with Kat6b+/+ cells (P = 0.04 and 0.0004; Fig 5P and Q), demonstrating a requirement for KAT6B for normal Sox2 promoter activity.
Given the role of SOX2 as a driver of NSC multipotency and self-renewal (Suh et al, 2007; Miyagi et al, 2008; Favaro et al, 2009; Arnold et al, 2011), we postulated that SOX2 overexpression might rescue the proliferation defect of Kat6b−/− NSPCs (Fig 4A). Consistent with previous results (Fig 4A), Kat6b−/− NSPC expressing the empty vector (Kat6b−/−pMIG) grew slower than Kat6b+/+pMIG NSPCs (P = 2 × 10−6; Fig 5R). Remarkably, overexpression of Sox2 in Kat6b−/−Sox2-pMIG NSPCs caused a partial rescue of proliferation. The proliferation capacity of Kat6b−/−Sox2-pMIG NSPCs was greater than Kat6b−/−pMIG+/+ NSPCs (P = 0.01) and only marginally different from WT control cultures (P = 0.05; Fig 5R). In addition, Kat6b+/+Sox2-pMIG NSPCs proliferated faster than Kat6b−/−Sox2-pMIG NSPCs (P = 0.0002), suggesting that SOX2 drove fast NSPC proliferation in the presence of KAT6B and that SOX2 was unable to exert this strong proliferative effect to its full extent in the absence of KAT6B (P < 10−6 for Kat6b+/+Sox2-pMIG vs. Kat6b−/− Sox2-pMIG NSPCs).
Discussion
In this study, we report the molecular effects of the histone acetyltransferase KAT6B during brain development and in embryonic NSPCs. We showed that KAT6B is essential for normal levels of H3K9ac at many genes, in particular brain development genes, including SOX family transcription factor genes. Notably, we show that KAT6B binds to and regulates the expression of a key regulator gene of neural precursor development, Sox2. In the absence of KAT6B, we observed a reduction in H3K9ac, Sox2 promoter activity, Sox2 mRNA, and SOX2 protein levels. We demonstrated direct binding of KAT6B to the Sox2 promoter. Consistent with these results, Sox2 overexpression partially rescued the proliferation defect of Kat6b−/− NSPCs in vitro.
Identification of KAT6B as promoting the expression of brain-specific genes in the developing cortex is consistent with the cognitive disorders caused by heterozygous mutations in the human KAT6B gene. The effects of KAT6B on the SOX gene family mirror the roles of its closely related paralogue KAT6A in promoting HOX, TBX and DLX family gene expression (Voss et al, 2009, 2012b; Sheikh et al, 2015a; Vanyai et al, 2019). These findings suggest that activation of important developmental control gene families has been divided between this pair of closely related proteins, KAT6A and KAT6B. Unique among chromatin regulators, Kat6b is strongly regulated at the mRNA level in the forebrain. Kat6b mRNA levels are low before the onset of cortex development but between E11.5 and E15.5 they are strongly up-regulated during neural precursor proliferation and differentiation (Thomas et al, 2000).
Heterozygous null mutation in the human KAT6B gene causes a reduction in global histone H3 acetylation (Kraft et al, 2011). Our data suggest that the specific lysine target of KAT6B is H3K9 (here and Bergamasco et al [2024a, 2024b]). Loss of KAT6B did not abolish H3K9ac completely, indicating that other histone acetyltransferases acetylate H3K9, too. This role has been ascribed to KAT2A (GCN5) and KAT2B (PCAF) (Jin et al, 2011) and to the KAT6B paralogue KAT6A (Voss et al, 2009, 2012b; Sheikh et al, 2015b; Vanyai et al, 2015, 2019; Lv et al, 2017; Yan et al, 2022). Interestingly, mRNA levels of the Kat2a gene were up-regulated in the absence of KAT6B and down-regulated when Kat6b was overexpressed, suggesting an attempt at compensatory up-regulation of this H3K9 acetyltransferase, which, however, did not achieve normal H3K9ac levels, gene expression or brain development. Similarly, the presence of Kat6a mRNA and the down-regulation of the histone deacetylase genes, Hdac11 and Sirt2, were insufficient for a rescue of H3K9ac levels.
In addition, we found H3K23 to be a potentially cell type-specific acetylation target of KAT6B. H3K23 was reported to be a KAT6B target in previous studies (Simo-Riudalbas et al, 2015; Klein et al, 2019). H3K23ac is also catalyzed by KAT6A (Lv et al, 2017; Yan et al, 2020; Sharma et al, 2023). KAT6A is required for H3K9ac at specific gene loci in mouse embryonic fibroblasts (Sheikh et al, 2015b), in E10.5 embryos and tissue (Voss et al, 2009, 2012b; Vanyai et al, 2015, 2019) and in AML cells (Yan et al, 2022), as well as for H3K23ac in glioblastoma cells (Lv et al, 2017) and breast cancer cells (Sharma et al, 2023). These findings indicate that the histone lysine target of KAT6B and KAT6A may be cell-type dependent. Alternatively, KAT6A may be better able to compensate for a reduction in H3K23ac caused by loss of KAT6B in NSPCs in vitro than in the developing cortex in vivo.
Surprisingly, we also observed an increase in H3K14ac in E12.5 Kat6b−/− embryos and NSPCs along with a genome-wide increase in DNA accessibility. H3K14ac is dependent on another MYST family histone acetyltransferase, KAT7, in tissues and cell types during mouse development and in several human cell types (Kueh et al, 2011, 2020, 2023; Mishima et al, 2011; Feng et al, 2016). Given that histone H3 lysine 9 and 14 residues are only five amino acids apart, steric hindrance may occur between the KAT7 and the KAT6B multiprotein complexes. Reduced KAT6B complex occupancy at H3K9 in Kat6b−/− samples may provide an opportunity for increased access by the KAT7 complex. Alternatively, or in addition, KAT7, KAT6A, and KAT6B share protein complex members (Doyon et al, 2006) and therefore loss of KAT6B may increase protein complex partner availability for KAT7 and/or KAT6A complexes, altering the stoichiometric balance of MYST family histone acetyltransferase complexes within the cell. Notably, the increase in H3K14ac did not rescue the effects of loss of KAT6B and reduction in H3K9ac at the Sox2 and Pax6 gene and reduced expression of over 3,000 genes, indicating the importance of KAT6B and normal H3K9ac levels for neural development and suggesting that a reduction in H3K9ac cannot be compensated for by an increase in H3K14ac.
KAT6B and SOX2 appear to form a positive feedback loop. Not only does KAT6B bind the Sox2 gene and promote Sox2 expression (shown here), it has also been demonstrated that SOX2 binds the Kat6b gene in ES cells (Whyte et al, 2013; Cosentino et al, 2019) and NSCs (Bertolini et al, 2019). Positive feedback loops involving SOX2 have been shown to regulate NSPC self-renewal (SOX2 and EGFR [Hu et al, 2010]) and inhibit premature differentiation (SOX2 and SOX6 [Lee et al, 2014]). Similarly, a positive feedback loop may also exist between KAT6B and PAX6, whereby KAT6B binds to and activates Pax6 gene expression (shown here), and PAX6 binds the KAT6B gene is ES cell-derived neural progenitors (Thakurela et al, 2016).
Consistent with SOX2 being a KAT6B gene target in the developing cerebral cortex, the Kat6b and Sox2 genes are both expressed strongly in the dorsal telencephalon at E10.5 to E12.5 (Thomas et al, 2000; Thomas & Voss, 2004; Iwafuchi-Doi et al, 2011; Roberts et al, 2014) and in the developing cortex at E14.5–E17.5 (Thomas et al, 2000; Diez-Roux et al, 2011; Cánovas et al, 2015). Both genes are more strongly expressed in the developing cortex than in neighboring tissues. Importantly, there is a notable overlap in the phenotypic consequences of SOX2 and KAT6B deficiency in neural cell types. While the complete loss of SOX2 and KAT6B cause dissimilar phenotypes, namely very early embryonic lethality before the egg cylinder stage (Avilion et al, 2003) versus death at birth (Thomas et al, 2000), Sox2 mRNA levels are not reduced to zero in the absence of KAT6B and therefore the homozygous null mutation of Sox2 is not a relevant comparison in this context. In contrast, hypomorphic alleles of Sox2 and Kat6b cause similar phenotypic anomalies. Mice carrying a Sox2 hypomorphic allele, lacking 70–75% of endogenous mRNA levels, have reduced GABAergic neurons in the newborn cortex and adult olfactory bulb (Cavallaro et al, 2008) and impaired neurogenesis in the adult hippocampus and subventricular zone niches (Ferri et al, 2004). Similarly, Kat6bgt/gt adult mice have fewer GAD67+ GABAergic neurons (Thomas et al, 2000) and fewer neuroblasts in the rostral migratory stream (Merson et al, 2006). Neurons derived from Sox2 deficient neuronal progenitors show impaired arborization and only weakly express markers of mature neurons (Cavallaro et al, 2008), consistent with impaired neurite outgrowth in Kat6b−/− E16.5 cortical neurons and reduced levels ßIII-tubulin protein in Kat6b−/− neurons shown here.
While our functional assessment focused on SOX2 as a key KAT6B gene target, other SOX family genes down-regulated in Kat6b−/− and up-regulated in Tg(Kat6b) samples also have important roles in brain development and NSC function. For example, Sox1 is expressed during neural plate formation and its overexpression promotes neuronal differentiation in P19 cells (Pevny et al, 1998) and mouse NSPCs (Kan et al, 2004). SOX3 inhibits premature astrocyte differentiation by repressing gene targets of SOX9 in the developing mouse spinal cord (Klum et al, 2018) and SOX21 promotes neurogenesis in mouse hippocampal neurons (Matsuda et al, 2012). While it is tempting to attribute each defect in Kat6b−/− NSPCs to an individual gene target or a small collection of genes, it is more likely that the molecular consequences of loss of KAT6B described in this study result from the cumulative disruption of multiple direct and indirect gene targets. To this point, we only see a partial rescue of the proliferative defect of Kat6b−/− NSPCs when Sox2 is overexpressed, indicating that the reduction in SOX2 is not responsible for this phenotype alone.
In conclusion, we have provided a characterization of the effects of loss of KAT6B on histone acetylation and gene expression in general and in particular on the SOX gene family. We identified Sox2 as a direct target of KAT6B, where KAT6B binds to the promoter of the Sox2 gene, promotes H3K9ac and Sox2 promoter activity, Sox2 mRNA levels, and ultimately SOX2 protein levels. Finally, we showed that overexpression of Sox2 can partially rescue a Kat6b loss of function phenotype, confirming a functional relationship between KAT6B, SOX2, and downstream events. Thus, our molecular analyses of KAT6B function reveal its importance for transcriptionally orchestrating neural development through H3K9ac and chromatin-based activation of development regulatory genes including the SOX2-driven molecular network.
Materials and Methods
Ethics statement
Animal experiments were conducted with approval of the WEHI Animal Ethics Committee and performed according to the Australian code for the care and use of animals for scientific purposes.
Mouse husbandry
Mice were housed four to six animals in ventilated cages (AirLaw) and provided with γ-irradiated feed (Ridley AgriProducts; Barastoc) and sterilized water. Mice were kept in a 14 h light/10 h dark cycle. Noon of the day a vaginal copulation plug was first observed was defined as embryonic day 0.5 (E0.5).
Mouse strains
We used Kat6b–mice lacking exons 2–12 of the Kat6b gene, which we reported previously (Bergamasco et al, 2024a, 2024b), and Tg(Kat6b) BAC transgenic mice, which overexpress Kat6b ∼4.5-fold, also reported previously (Bergamasco et al, 2024a). Sox2-GFP mice (B6; 129S-Sox2tm2Hoch/J [Arnold et al, 2011]) were obtained from the Jackson Laboratory. Mice were genotyped by PCR using primers displayed in Table S5.
Table S5. Genotyping primers.
Generation of Kat6bV5 mice
A FLAG-V5-biotinylation sequence triple tag with a neomycin phosphotransferase cassette (Soler et al, 2010; Schwickert et al, 2014) was inserted by homologous recombination in Escherichia coli into the endogenous Kat6b gene after the last codon of KAT6B in a Kat6b-containing BAC, followed by targeting of the Kat6b locus in embryonic stem cells and germline chimera production to generate Kat6bFag-V5-BIO mice (Kat6bV5). Kat6bV5 mice were crossed to Cre-deleter mice (Schwenk et al, 1995) to remove the neomycin phosphotransferase cassette. Kat6bV5 mice were genotyped using primers in Table S5.
NSPC culture
NSPCs were derived from E12.5 dorsal telencephalon and cultured as free-floating neurosphere colonies as described (Rietze et al, 2001) in neurosphere medium (DMEM/F12 [12500-062; Gibco], 5 mM HEPES [H-4034; Sigma-Aldrich], 13.4 mM NaHCO3 [G-7021; Sigma-Aldrich], 100 U/ml penicillin–streptomycin [15140-122; Gibco], 25 μg/ml insulin [I-6634; Sigma-Aldrich], 60 μM putrescine dihydrochloride [P-5780; Sigma-Aldrich], 100 μg/ml apo-transferrin [T-2252; Sigma-Aldrich], 30 nM selenium sodium salt [S-9133; Sigma-Aldrich], 20 nM progesterone [P-6149; Sigma-Aldrich], 0.2% BSA [A-3311; Sigma-Aldrich], 20 ng/ml EGF [78006.1; Neurocult] and 10 ng/ml FGF [78003; Neurocult]) at 37°C and 5% CO2. Neurospheres were dissociated using Accutase (A1110501; Gibco) in a 37°C water bath for 4 min, counted using an automated cell counter (Countess; Thermo Fisher Scientific) and replated at 10,000 cells/cm2. Neurospheres were imaged using an automated cell imager (ZOE, 1450031; Bio-Rad). Molecular and flow cytometry experiments were performed at passages 3–5. Secondary neurosphere formation assays were performed as described (Merson et al, 2006).
For NSPC differentiation, cells were dissociated using Accutase (A1110501; Gibco) and plated onto uncoated plates for flow cytometry or, for immunofluorescence, chamber slides (Nunc LabTek II, 154453; Thermo Fisher Scientific) pre-coated with 0.1 mg/ml poly-D-lysine (P6407; Sigma-Aldrich) followed by 10 μg/ml laminin (L2020; Sigma-Aldrich). Cells were grown in neurosphere medium, but without EGF or FGF and with 1% FCS for 6 d at 5% CO2 and 37°C. For flow cytometry, cells were processed as described (Kueh et al, 2023). For immunofluorescence, cells were fixed in 4% PFA (wt/vol) for 20 min at RT, permeabilized and blocked in 10% (vol/vol) FCS and 0.03% (vol/vol) Triton X-100 for 1 h at RT. Slides were incubated O/N at 4°C with anti-βIII tubulin (1:500, G7121; Promega) and anti-GFAP (1:500, Z0334; Dako) in 10% FCS, washed (2 × 5 min PBS) and incubated with goat anti-mouse Alexa Fluor 546 (A11003; Invitrogen) and goat anti-rabbit AMCA (711-155-152; Jackson Immunoresearch) secondary antibodies at 1:400 for 1 h at RT in the dark. Slides were washed (2 × 5 min PBS) and mounted in Dako fluorescent mounting medium (S3023; Agilent).
Flow cytometry
Intracellular flow cytometry analysis of NSPCs under proliferating and differentiating conditions was performed as described (Kueh et al, 2023) using antibodies listed in Table S6. For Ki67 versus DAPI cell cycle analysis, neurospheres were dissociated with Accutase (4 min, 37°C), fixed and permeabilized for 1 h on ice using the eBioscience Foxp3/Transcription factor staining buffer set (00-5523-00; Thermo Fisher Scientific), washed (2 × 5 min) in 1x permeabilization buffer from the same kit and resuspended in 100 μl/1 × 106 cells 2% FACS buffer (2% [vol/vol] FCS, 150 mM NaCl, 3.7 mM KCl, 2.5 mM CaCl2•2H2O, 1.2 mM MgSO4•7H2O, 0.8 mM K2HPO4, 1.2 mM KH2PO4, 11.5 mM HEPES, pH 7.4 in MQ-H2O) containing 20 μl/1 × 106 cells anti-Ki67 FITC-conjugated antibody or isotype control (AB_396302; BD biosciences). Samples were incubated O/N at 4°C on a rocker, washed (2 × 5 min) in permeabilization buffer and resuspended in 100 μl/1 × 106 cells with 10 μl/ml DAPI (62248; Thermo Fisher Scientific). Cells were incubated on ice for 30 min, washed (2 × 5 min) in permeabilization buffer and resuspended in 100 μl/1 × 106 cells 2% FACS buffer.
For annexin V versus propidium iodine (PI) cell death analysis, neurospheres were dissociated with Accutase (4 min, 37°C) and processed using the Dead Cell Apoptosis Kit (V13242; Invitrogen) according to the manufacturer’s instructions. For assessment of Sox2-GFP levels in E12.5 dorsal telencephalon, tissue was dissected and mechanically dissociated into a single cell suspension in 2% FACS buffer and passed through a 40 μm cell sieve (431750; Corning). Cells were centrifuged (200g, 5 min), supernatant removed, and cells resuspended in 2% FACS buffer containing anti-CD133-APC (17-1331-81; eBiosciences) and anti-SSEA1/CD15 (347420; BD Biosciences) antibodies at a 1:200 dilution. Samples were stained on ice for 1 h, washed (2 × 5 min) and resuspended in 2% FACS buffer containing 1:200 human anti-mouse PE-conjugated secondary antibody (made in house) and incubated on ice for 1 h. Samples were washed (2 × 5 min) and resuspended in 2% FACS buffer. Flow cytometry experiments were assessed on a LSRII or Fortessa flow cytometer (BD) at <7,500 events/sec and analyzed using FlowJo analysis software (version 10.7).
E16.5 cortical neuron culture
E16.5 fetal cortices were dissected under a dissecting microscope (Zeiss). Isolated cortices were washed once in PBS (14190144; Gibco) before dissociation in 200 μl trypsin/EDTA (10006132; Sigma-Aldrich) for 10 min at 37°C. Excess trypsin was removed and replaced with 1 ml cortical neuron medium (DMEM/F12 [12500-062; Gibco], 5 mM HEPES [H-4034; Sigma-Aldrich], 13.4 mM NaHCO3 [G-7021; Sigma-Aldrich], 100 U/ml penicillin–streptomycin [15140-122; Gibco], 25 μg/ml Insulin [I-6634; Sigma-Aldrich], 60 μM putrescine dihydrochloride [P-5780; Sigma-Aldrich], 100 μg/ml apo-transferrin [T-2252; Sigma-Aldrich], 30 nM selenium sodium salt [S-9133; Sigma-Aldrich], 20 nM progesterone [P-6149; Sigma-Aldrich], 0.2% BSA [A-3311; Sigma-Aldrich], and 1% FCS). Tissue was gently triturated, and cells passed through a 100 μm cell sieve (431751; Corning). Cells were plated onto chamber slides (C6932; Sigma-Aldrich) pre-coated with 0.1 mg/ml poly-D-lysine (P4832; Sigma-Aldrich) at 10,000 cells/cm2, determined using an automated cell counter (Countess; Invitrogen). Cells were cultured at 37°C in 5% CO2 for 5 d. On day 5, cells were fixed in 4% (wt/vol) paraformaldehyde (P6148; Sigma-Aldrich) in PBS, for 20 min at RT, nonspecific binding was blocked with 10% (vol/vol) FCS in PBS for 1 h at RT and cells incubated with anti-βIII tubulin antibody (G7121; Promega) at 1:500 dilution in 10% FCS and 0.3% (vol/vol) Triton-X 100 in H2O O/N at 4°C. Cells were washed (2 × 5 min) in PBS and incubated anti-mouse Alexa Fluro 546 antibody (A11003; Invitrogen) at 1:400 in 10% FCS for 1 h at RT. Cells were washed in PBS (2 × 5 min at RT) and incubated with 1 μg/ml DAPI (62248; Thermo Fisher Scientific) for 10 min at RT, washed in PBS (2 × 5 min) and mounted in Dako fluorescent mounting medium (S3023; Agilent).
SOX2 overexpression in cultured NSPCs
The Sox2 coding sequence (NM011443.4) was cloned by GeneArt Subcloning and Plasmid Services (Thermo Fisher Scientific) into a pMSCV-IRES-GFP II (pMIG II; Adgene [Holst et al, 2006]) vector using Xho1/EcoR1 restriction enzyme sites to generate a Sox2 expression construct (Sox2-pMIG; pMSCV-Sox2-IRES-GFP II). Sox2-pMIG or empty vector (pMIG; pMIG II) were virally packaged using human cells (Phoenix-AMPHO, ATCC CRL-3213). Phoenix cells were plated at a density of 3 × 106 cells/10 cm plate in DMEM + 10% FCS for 24 h before transfection. 5.0 μg Sox2-pMIG or pMIG were combined with 250 μl 0.5 M CaCl2, 250 μl MQ-H2O, and 500 μl 2x HBSS (H4385; Sigma-Aldrich), vortexed for 10 s and incubated for 10 min at RT. The transfection mixture was added dropwise over Phoenix cells, swirling the plate to ensure even distribution. The following morning, transfection media was replaced with 6 ml complete neurosphere medium. Viral supernatant was collected at 24 and 48 h, passed through a 0.45 μm filter syringe (16533-K; Sartorius Australia) and added to freshly passaged NSPCs. Transfection efficiency was confirmed by detecting GFP using a fluorescent microscope (Zeiss).
Western immunoblotting
Histones were extracted from whole E12.5 embryos, E12.5 dorsal telencephalon or cultured NSPCs by acid protein extraction. Cells or tissue were collected, washed in PBS (14190144; Gibco) containing 0.5 mM sodium butyrate (B5887; Sigma-Aldrich) and cOmplete EDTA-free protease inhibitor cocktail (11873580001; Roche) and collected by centrifugation (200g, 5 min). Samples were lysed in Histone acid lysis buffer (10 mM HEPES pH 7.9, 1.5 mM MgCl2, 10 mM KCl, and 0.5 mM DTT) for 30 min at 4°C on a roller, collected by centrifugation (10,000g, 10 min), resuspended in 0.2 M H2SO4, incubated on ice for 1–2 h and dialyzed in dialysis tubing (Spectrum Spectra/Por Dialysis Membrane Tubing, molecular weight cut-off 20 kD; 08-607-067; Thermo Fisher Scientific) against 0.1 M acetic acid (A6283; Sigma-Aldrich) for 1 h at 4°C and MQ-H2O overnight at 4°C. Protein concentrations were determined using a bicinchoninic acid assay (23225; Thermo Fisher Scientific). Acid extracted proteins were run on a 4–12% Bis–Tris gel (NP0322; Thermo Fisher Scientific) and transferred onto nitrocellulose membranes (926-31090; LI-COR Biosciences) for fluorescent detection or polyvinylidene fluoride membranes (45-3010040001; Kracheler scientific) for HPR detection. Membranes were blocked for 1 h at RT on a roller in blocking buffer (Intercept [PBS], 927-70001; LI-COR) for fluorescent detection or 5% skim milk in PBS + 0.1% (vol/vol) Tween-20 (P1379; Sigma-Aldrich) for HRP detection and probed with antibodies against acetylated histone lysines (Table S7) O/N at 4°C on a roller. Membranes were washed in PBS + 0.1% Tween-20 (P1379; Sigma-Aldrich) and incubated with secondary antibodies (Table S7) for 1 h at RT on a roller. For fluorescent detection, membranes were imaged and analyzed using an automated detection system (Odyssey, LI-COR). Acetylated histone H3 marks were normalized to pan H3 on the same membrane. For HRP detection, membranes were incubated with chemiluminescent HRP substrate (ECL, WBULS0500; Millipore) and exposed onto chemiluminescent film (GE healthcare). For detection of acetylated histone H4 lysine residues, membranes previously stained for an acetyl-histone H3 mark were incubated in 0.01% sodium azide (1 h RT), washed 3–4x in PBS and re-probed with an anti-acetyl H4 antibody. After exposure, membranes were washed in PBS and incubated with Ponceau S (A40000279; Thermo Fisher Scientific) for 5–10 min at RT. Exposed film and Ponceau S-stained membranes were scanned and staining intensity quantified with Fiji image analysis software. Acetylated histone H3 and H4 lysine bands were normalized to an abundance basic protein band at 50 kD on Ponceau S-stained membranes.
RNA isolation
Total RNA was collected from cultured NSPCs, E12.5 dorsal telencephalon or E15.5 cortex tissue using a kit (RNeasy mini kit, 74104; QIAGEN) according to the manufacturer’s instructions and including the optional DNase I digest step. RNA quality and quantity were assessed on an automated electrophoresis system (Tapestation 4200, G2991BA; Agilent).
RTqPCR
1 μg total RNA determined using a spectrophotometer (NanoDrop, Thermo Fisher Scientific) was used to generate cDNA using a cDNA synthesis kit (Superscript III, 18080051; Thermo Fisher Scientific) according to the manufacturer’s instructions. cDNA was amplified using sequence-specific primers (Table S8) on an automated real time PCR system (LightCycle 480; Roche).
Table S8. Primers used for RT-qPCR.
Chromatin immunoprecipitation
Chromatin immunoprecipitating (ChIP) was performed as described (Voss et al, 2012a) using antibodies in Table S7. qPCR was performed using a LightCycle 480 instrument (Roche) and sequence-specific primers in Table S9.
RNA sequencing and analysis
500 ng to 1 μg total RNA was used to generate sequencing libraries using a kit (TruSeq RNA prep kit v2, RS-122-2002; Illumina) according to the manufacturer’s instructions. Samples were sequenced (NextSeq500; Illumina) to give 66 bp paired end reads.
Analysis of E12.5 dorsal telencephalon and E15.5 cortex RNA sequencing data
All samples were aligned to the mm10 build of the mouse genome using Rsubread (v1.24.1) (Liao et al, 2019). In all cases, at least 85% of fragments were successfully mapped. Fragments overlapping genes were the summarized using Rsubread’s featureCounts function. Genes were identified using Rsubread’s inbuilt annotation for the mm10 genome. All sex-specific genes–Xist and those unique to the Y-chromosome – were removed to avoid sex biases. Genes with no official symbol were also discarded. Differential expression analyses were then carried out using the limma (v3.40.6) (Ritchie et al, 2015) and edgeR (v3.26.8) (Robinson et al, 2010) software packages. Independent analyses were performed for each mouse background–C57BL/6 (Kat6b−/− versus Kat6b+/+) and FVB x BALB/c (Tg(Kat6b) versus Kat6b+/+).
For each dataset, expression-based filtering was first performed. For those samples from the C57BL/6 background, all genes that did not achieve a CPM greater than 1 in at least three samples were filtered. For those samples from the FVB x BALB/c background, all genes that failed to achieve a CPM greater than 1 in at least four samples were removed. In both cases, sample composition was then normalized using the TMM method (Robinson & Oshlack, 2010).
Following filtering and normalization, the data in each analysis was transformed to log2-CPM. Differential expression between the genotype groups from each genetic background was then assessed using linear models and robust empirical Bayes moderated t-statistics with a trended prior variance (robust limma-trend pipeline) (Phipson et al, 2016). In the case of the C57BL/6 background samples, the linear models incorporated four surrogate variables to remove variation caused by a dissection error present in some samples as well as a litter batch effect. These were calculated using limma’s wsva function. For the analysis of the FVB x BALB/c samples, a factor representing sample litter was included in the linear models, and sample weights were calculated using limma’s array weights function with default parameters (Liu et al, 2015).
Analysis of NSPC RNA sequencing data
An index combining the mm10 build of the mouse genome and sacB genomic sequence was first built using Rsubread’s (v1.28.0) buildindex function. All samples were then aligned to this combined genome using Rsubread, achieving a mapping rate of at least 97% across all samples. Fragments overlapping genes were summarized using Rsubread’s featureCounts. Genes were identified using Rsubread’s inbuilt annotation to the mm10 genome. An additional line of annotation was added for the sacB sequence. Following fragment summarization, all genes with no symbol, ribosomal RNAs, non-protein coding immunoglobulin genes, predicted, unknown, and pseudo genes were all removed. Differential expression analyses were then carried out using limma (v3.40.6) and edgeR(v3.26.8). Independent analyses were conducted for each mouse genetic background - (Kat6b−/− versus Kat6b+/+) and FVB x BALB/c (Tg(Kat6b) versus Kat6b+/+).
For each analysis, expression-based filtering was first performed. In each case, edgeR’s filterByExpr function was used with default parameters. Following filtering TMM, normalization was applied to each data set.
The NSC from the C57BL/6 background was analyzed using a robust limma-voom pipeline. Thus, the data were first transformed to log2-CPM with associated precision weights using voom (Law et al, 2014). Differential expression between the genotype groups was then assessed using linear models and robust empirical Bayes moderated t-statistics.
The NSC from the FVB x BALB/c background was analyzed using a robust limma-voom with sample weights pipeline. Therefore, similarly to the C57BL/6 background analyses, the data was transformed to log2-CPM with associated precision weights using voom. Additional sample level weights were also calculated (limma voomWithQualityWeights function), and differential expression between the genotype groups was assessed using linear models and robust empirical Bayes moderated t-statistics.
For all analyses discussed here, the Benjamini and Hochberg method was applied to control the FDR below 5%. Pathway analyses were conducted using limma’s goana and kegga functions. Multidimensional scaling plots, mean-difference (MD) plots, and barcode plots were generated using limma’s plotMDS, plotMD, and barcodeplot functions respectively. Heatmaps were generated using ComplexHeatmap.
ATAC-sequencing and analysis
Using 50,000 NSPCs from Kat6b+/+ or Kat6b−/− E12.5 dorsal telencephalon tissue, combined with 50,000 Drosophila melanogaster S2 cells spike-in (automated cell counter [Countess; Thermo Fisher Scientific Scientific]), ATAC-sequencing was performed as described (Buenrostro et al, 2015). Briefly, combined cells were lysed in 100 μl lysis buffer (10 mM Tris–HCl, pH 7.4, 10 mM NaCl, 3 mM MgCl2, 0.1% IGEPAL, and EDTA-free Complete protease inhibitors, made up in DNase-free H2O). Nuclei were collected by centrifugation (500g, 10 min, 4°C) and resuspended in 50 μl TD buffer and TDE1 transposase (Nextera, #20034197; Illumina). Samples were digested (30 min at 37°C), purified using the QIAGEN MinElute PCR purification kit (28004; QIAGEN) and amplified using NEBNext High-Fidelity PCR master mix (M0541S) and P5/P7 indexing primer combinations (Table S10) under the following conditions: 72°C, 5 min, 98°C, 30 s, 10x cycles, 98°C, 10 s, 63°C, 30 s, 72°C, 1 min. Amplified samples were size-selected using 1.3x volumes AMPure beads (A63880; Beckman Coulter), eluted in 20 μl nuclease-free H2O and analyzed using a D1000 Tape and 4200 Tapestation (Agilent). Samples were processed on a high-throughput sequencing machine (NextSeq 2000; Illumina).
Following sequencing-read quality control, reads were aligned to the Mus musculus (mm39) and D. melanogaster (Dme1R6.32) genomes using Rsubread 2.12.3 (Liao et al, 2019). Mouse library sizes were normalized to Drosophila reads as controls, i.e., assuming that total Drosophila coverage should be equal across samples. Coverage was assessed across nonoverlapping 5 bp bins using the deepTools program (Ramírez et al, 2016). Read counts were obtained using the featureCounts function and the inbuilt mm39 annotation in Rsubread for gene promoters (TSS + 1 kb upstream) and transcription end sites (TES + 1 kb) of protein coding genes. Discontinued Entrez Gene IDs were excluded from analysis and genes with low counts were filtered using the filterByExpr function in edgeR. Counts were also obtained for active NSPC enhancers, defined as H3K4me1+/H3K27ac+, based on GSM2406793 and GSM2406791 (Bertolini et al, 2019). Enhancer regions overlapping with promoters or TES elements were removed. BEDTools (Quinlan & Hall, 2010) was used to identify overlapping H3K4me1/H3K27ac enriched regions. Differential accessibility upon gain or loss of Kat6b was assessed using quasi-likelihood generalized linear models in edgeR (Chen et al, 2016, 2024 Preprint).
CUT&Tag sequencing and analysis
Cleavage Under Targets and Tagmentation (CUT&Tag) sequencing was performed using 50,000 NSPCs combined with 50,000 D. melanogaster S2 cells, as described in Kaya-Okur et al (2019), with minor modifications as described (Wichmann et al, 2022), using antibodies against H3K9ac, H3K14ac, H3K23ac, or RNA polymerase II, subunit A (POLR2A) (Table S7) and indexing primers (Table S10). Libraries were sequenced (NextSeq 2000; Illumina).
Reads were aligned to the M. musculus (mm39) and D. melanogaster (Dme1R6.32) genomes using Rsubread. Mouse library sizes were normalized using the Drosophila read content and normalized read abundance assessed at the following genomic regions.
(i) Gene promoters, defined as up to 1 kb upstream of the TSS to the TSS of protein coding genes.
(ii) Active NSPC enhancers, defined as H3K4me1+H3K27ac+, based on GSM2406793 and GSM2406791 (Bertolini et al, 2019).
Enhancer regions overlapping with a promoter or TES downstream regions were removed. Overlapping and nonoverlapping H3K4me1 and H3K27ac enriched regions were identified using BEDTools. The number of reads pairs overlapping each genomic region was summarized using Rsubread’s featureCounts function (Liao et al, 2014). Coverages for non-overlapping 5 bp bins were computed using deepTools. Differential coverages upon gain or loss of Kat6b were assessed using quasi-likelihood generalized linear models in edgeR (Chen et al, 2016, 2024 Preprint).
Statistics
The specific statistical tests employed and the number of biological replicates for each experiment is detailed in the figure legends. Statistical analyses were performed in Prism Graphpad Version 8.3.1 for Mac (GraphPad Software), except the analyses of the RNA-sequencing, ATAC-seq and CUT&Tag sequencing data, which are provided above in the methods sections.
Data Availability
The RNA-sequencing, CUT&Tag sequencing, and ATAC-sequencing data are accessible at NCBI GEO under the accession numbers GSE280783 (NSPC RNA-seq data), GSE280784 (developing cortex RNA-seq data), GSE267672 (CUT&Tag data) and GSE267675 (ATAC-seq data).
Acknowledgements
The authors would like to thank R May, C Burström, and L Potenza for excellent technical support. S Wilcox for exceptional technical service. F Dabrowski, L Wilkins, N Blasch, S Bound, E Boyle, J Gilbert, S Oliver, and L Johnson for expert animal care. MI Bergamasco was supported by an Australian Government Postgraduate Award. The program of work was supported by the Pamela and Lorenzo Galli Charitable Trust and by the Australian National Health and Medical Research Council through Project Grants 1010851 to AK Voss and T Thomas; 1160517 to T Thomas; Ideas Grant 2010711 to T Thomas; Research Fellowships 1003435 to T Thomas, 575512 and 1081421 to AK Voss, and 1154970 to GK Smyth and Investigator Grants 1176789 to AK Voss and 2025645 to GK Smyth; through the Independent Research Institutes Infrastructure Support Scheme; by the Chan Zuckerberg Initiative through grant 2021-237445 to GK Smyth and by the Victorian Government through an Operational Infrastructure Support Grant.
Author Contributions
MI Bergamasco: formal analysis, validation, investigation, visualization, methodology, and writing—original draft, review, and editing.
W Abeysekera: formal analysis, visualization, methodology, and writing—original draft.
AL Garnham: formal analysis, visualization, methodology, and writing—original draft.
Y Hu: formal analysis.
CSN Li-Wai-Suen: formal analysis.
BN Sheikh: formal analysis, supervision, and methodology.
GK Smyth: formal analysis, supervision, funding acquisition, and methodology.
T Thomas: conceptualization, formal analysis, supervision, funding acquisition, methodology, and writing—original draft, review, and editing.
AK Voss: conceptualization, formal analysis, supervision, funding acquisition, visualization, methodology, project administration, and writing—original draft, review, and editing.
Conflict of Interest Statement
The T Thomas and AK Voss laboratories have received research funding from the Cooperative Research Centre for Cancer Therapeutics (CRC-CTx).
- Received July 31, 2024.
- Revision received November 1, 2024.
- Accepted November 4, 2024.
- © 2024 Bergamasco et al.


This article is available under a Creative Commons License (Attribution 4.0 International, as described at https://creativecommons.org/licenses/by/4.0/).