Chd7 is indispensable for mammalian brain development through activation of a neuronal differentiation programme

Feng, Weijun; Kawauchi, Daisuke; Körkel-Qu, Huiqin; Deng, Huan; Serger, Elisabeth; Sieber, Laura; Lieberman, Jenna Ariel; Jimeno-González, Silvia; Lambo, Sander; Hanna, Bola S.; Harim, Yassin; Jansen, Malin; Neuerburg, Anna; Friesen, Olga; Zuckermann, Marc; Rajendran, Vijayanad; Gronych, Jan; Ayrault, Olivier; Korshunov, Andrey; Jones, David T. W.; Kool, Marcel; Northcott, Paul A.; Lichter, Peter; Cortés-Ledesma, Felipe; Pfister, Stefan M.; Liu, Hai-Kun

doi:10.1038/ncomms14758

Download PDF

Article
Open access
Published: 20 March 2017

Chd7 is indispensable for mammalian brain development through activation of a neuronal differentiation programme

Weijun Feng¹^na1,
Daisuke Kawauchi²^na1,
Huiqin Körkel-Qu¹,
Huan Deng¹,
Elisabeth Serger¹,
Laura Sieber²,
Jenna Ariel Lieberman³,
Silvia Jimeno-González³,
Sander Lambo²,
Bola S. Hanna⁴,
Yassin Harim¹,
Malin Jansen¹,
Anna Neuerburg¹,
Olga Friesen¹,
Marc Zuckermann⁴,
Vijayanad Rajendran²,
Jan Gronych ORCID: orcid.org/0000-0003-0557-679X⁴,
Olivier Ayrault⁵,
Andrey Korshunov^6,7,
David T. W. Jones^2,7,
Marcel Kool²,
Paul A. Northcott⁸,
Peter Lichter^4,7,
Felipe Cortés-Ledesma³,
Stefan M. Pfister^2,7,9 &
…
Hai-Kun Liu¹

Nature Communications volume 8, Article number: 14758 (2017) Cite this article

9043 Accesses
95 Citations
6 Altmetric
Metrics details

Subjects

Abstract

Mutations in chromatin modifier genes are frequently associated with neurodevelopmental diseases. We herein demonstrate that the chromodomain helicase DNA-binding protein 7 (Chd7), frequently associated with CHARGE syndrome, is indispensable for normal cerebellar development. Genetic inactivation of Chd7 in cerebellar granule neuron progenitors leads to cerebellar hypoplasia in mice, due to the impairment of granule neuron differentiation, induction of apoptosis and abnormal localization of Purkinje cells, which closely recapitulates known clinical features in the cerebella of CHARGE patients. Combinatory molecular analyses reveal that Chd7 is required for the maintenance of open chromatin and thus activation of genes essential for granule neuron differentiation. We further demonstrate that both Chd7 and Top2b are necessary for the transcription of a set of long neuronal genes in cerebellar granule neurons. Altogether, our comprehensive analyses reveal a mechanism with chromatin remodellers governing brain development via controlling a core transcriptional programme for cell-specific differentiation.

CHARGE syndrome protein CHD7 regulates epigenomic activation of enhancers in granule cell precursors and gyrification of the cerebellum

Article Open access 29 September 2021

Naveen C. Reddy, Shahriyar P. Majidi, … Harrison W. Gabel

CHD7 promotes neural progenitor differentiation in embryonic stem cells via altered chromatin accessibility and nascent gene expression

Article Open access 15 October 2020

Hui Yao, Douglas F. Hannum, … Donna M. Martin

Regulation of human cortical interneuron development by the chromatin remodeling protein CHD2

Article Open access 17 September 2022

E. M. A. Lewis, G. Chapman, … K. L. Kroll

Introduction

The chromatin structure plays a central role in any DNA-dependent biological process such as transcription, DNA replication and DNA repair. Recurrent mutations of chromatin modifier genes have been identified by extensive high-throughput sequencing of a broad spectrum of human patient samples. These findings reveal perturbation of the epigenome as a new hallmark of cancer¹ and neurological diseases². This strong correlative evidence requires further functional validation to fully understand the role of chromatin modifiers in development and disease, however.

Chromatin remodellers are key players in the regulation of nucleosome positioning and thus control DNA accessibility in eukaryotic cells. Mutations in the ATP-dependent chromatin remodeller chromodomain, helicase, DNA binding (CHD) 7 are the major cause of CHARGE syndrome, which is characterized by multiple organ defects including Coloboma, Heart defect, Atresia choanae, Retarded growth and development, Genital abnormality and Ear abnormality³. Notably, many CHARGE syndrome patients have brain anomalies including hypoplasia of the olfactory bulb and cerebellum^4,5 and often display an autistic phenotype⁶. Moreover, CHD7 mutations have been identified in patients with Kallmann syndrome⁷, autism spectrum disorders (ASD)⁸ as well as subsets of patients with both congenital heart disease and neurodevelopmental disabilities⁹. These findings strongly suggest an important role of CHD7 in brain development. Heterozygous Chd7 gene trap mice have exhibited many CHARGE-relevant phenotypes^4,10,11, supporting the role of transgenic mice as a disease-relevant tool. Our studies and others have utilized Chd7 conditional knockout mice to specifically dissect the role of Chd7 in the brain, demonstrating that mutations of Chd7 could be a causal factor for olfactory defects as well as hearing loss and cognitive disabilities in CHARGE patients^12,13,14. However, the exact role of Chd7 in cerebellar development remains unclear.

In both human and mice, the cerebellum undergoes a rapid expansion and foliation during early postnatal age. This expansion of the cerebellum is primarily due to the fast proliferation of cerebellar granule neuron progenitors (GNPs) in the outer external granule layer (oEGL), which is stimulated by Sonic hedgehog (Shh) secreted from Purkinje cells¹⁵. On the other hand, GNPs secret a glycoprotein Reelin (Reln) to control the proper localization of Purkinje cells¹⁵. In a precisely and temporally controlled fashion, GNPs shift to the inner EGL (iEGL) coincident with exiting the cell cycle¹⁵. The postmitotic cerebellar granule neurons (thereafter called as CGNs) then migrate inwardly to reach the internal granule layer (IGL), where they become mature. Importantly, unbalance between GNP proliferation and differentiation can lead to cerebellar hypoplasia or tumour formation¹⁶. Emerging data demonstrate that chromatin landscape is dynamically changed during differentiation of cerebellar granule cells, implicating an important role of chromatin regulation in neuronal differentiation^17,18. Indeed, cerebellar anomalies including vermis hypoplasia and massive Purkinje cell heterotopia represent some of the most prominent features of CHARGE patients^4,19, strongly suggesting that CHD7 is essential for normal cerebellar development.

Here, we report in vivo evidence for an essential role of Chd7 during cerebellar development. Using genetic and biochemical approaches, we unravel molecular mechanisms of how Chd7 controls neuron differentiation to govern proper formation of the cerebellar architecture. These results provide not only new molecular insights into cellular differentiation, but also a better understanding of Chd7-associated human diseases.

Results

Chd7 is highly expressed in cerebellar granule cells

Many CHARGE patients carrying CHD7 mutations have defects in the cerebellum, implicating a functional role of CHD7 in cerebellar development. In support, expression of CHD7 is enriched in the cerebellum during human brain development when compared to the other brain regions (Supplementary Fig. 1a). To investigate the contribution of Chd7 to cerebellar development, we examined detailed spatial and temporal expression patterns of Chd7 in developing mouse cerebellum. Immunohistochemistry (IHC) with a Chd7 antibody at embryonic day (E) 14.5 revealed a higher expression of Chd7 in the external granule layer (EGL) as compared to cells in the ventricular zone (Fig. 1a). Chd7 is highly expressed in cells from EGL and IGL at neonates and persists in the IGL at P21 cerebella (Fig. 1a). A very similar expression pattern of GFP was observed in a bacterial artificial chromosome-based Chd7-GFP reporter mouse line carrying GFP fused to the Chd7 regulatory element (Fig. 1a, lower panels). Consistent with the expression of Chd7 protein, in situ hybridization data revealed that Chd7 mRNA was clearly detected in EGL at E14.5, and in the granule layer at P7 and adult cerebellum (Supplementary Fig. 1b). We next performed co-immunostaining assays to check the co-expression of Chd7 or GFP with cell type-specific markers in the cerebella of wild-type (WT) or Chd7-GFP reporter mice, respectively. Exampled as P7 cerebellum, over 90% of Pax6⁺ or Pax2⁺ cells, namely cerebellar granule cells or inhibitory interneuron progenitors, respectively, expressed GFP in Chd7-GFP mice (Fig. 1b). In contrast, Bergmann glia (Sox2⁺) and Purkinje cells (Calbindin⁺) only showed a very weak expression of Chd7 (Fig. 1c).

Intriguingly, higher levels of Chd7 were detected in iEGL, where cells have just exited cell cycle (p27⁺) as compared to proliferating GNPs (Ki67⁺) in oEGL (Fig. 1d). In support, quantitative RT–PCR (qRT-PCR) data of cultured granule cells validated a higher expression of Chd7 in postmitotic (Ccnd1-low) cells than proliferating (Ccnd1-high) cells (Fig. 1e). These data suggest that the expression of Chd7 is upregulated during the transition from proliferating into differentiated cerebellar granule cells.

Disruption of Chd7 in GNP leads to cerebellar hypoplasia

To examine roles of Chd7 during cerebellar development, we next genetically ablated Chd7 in vivo using Chd7 conditional knockout mice¹². We first utilized a Nestin-Cre driver to target neural stem cells in the brain. The homozygous mutant [Nestin-Cre::Chd7^f/f] mice exhibited largely perinatal lethality (Supplementary Fig. 2a). We observed conspicuous cerebellar hypoplasia in survived homozygous mutant mice (Supplementary Fig. 2b). To dissect the role of Chd7 in cerebellar development and overcome the lethality of Nestin-Cre-mediated Chd7 deletion, we decided to use cerebellar neuron-specific Cre drivers. The cerebellar neuroepithelium (Nestin⁺) is mainly divided into two compartments during embryogenesis: the upper rhombic lip (Atoh1⁺) and the cerebellar ventricular zone (Ptf1a⁺) that give rise to excitatory (for example, granule cells) and inhibitory neurons (for example, Purkinje cells), respectively¹⁵. Given that Chd7 is predominantly expressed in granule cells (Fig. 1), we first ablated Chd7 in Atoh1⁺ cells. IHC assays with a Chd7 antibody performed in P7 cerebella (Supplementary Fig. 3a) and in vitro cultured granule cells (Supplementary Fig. 3b) verified ablation of Chd7 in granule cells of [Atoh1-Cre::Chd7^f/f] homozygous mutant mice. While no structural anomalies of the cerebella were observed at E15.5, P0 and older [Atoh1-Cre::Chd7^f/f] homozygous mutant mice, but not [Atoh1-Cre::Chd7^f/+] heterozygous mutant cerebella showed severe defects in folia formation and cerebellar hypoplasia especially in the vermis (Fig. 2a, Supplementary Fig. 3c). The cerebellar defects in Chd7 mutant animals are more prominent in the anterior lobe (Fig. 2a). This is consistent with the previous report²⁰ and a remaining Chd7 protein in the posterior lobe (Supplementary Fig. 3d), which is probably due to the less recombination efficiency of the Atoh1-Cre mouse line.

**Figure 2: Genetic ablation of *Chd7* in GNPs leads to cerebellar hypoplasia and Purkinje cell ectopia.**

The crosstalk between GNPs and Purkinje cells is essential for cerebellar development, as revealed by severe cerebellum hypoplasia in Reln-deficient reeler mice²¹. Since massive Purkinje cell heterotopia has been reported in CHARGE patient brains^19,22, we next examined Purkinje cells in [Atoh1-Cre::Chd7^f/f] mutant mice. Although the Atoh1-Cre driver does not target Purkinje cells, abnormal distribution of Purkinje cells were clearly detected in cerebella from P0 onwards until adult, especially at the anterior lobe of the cerebellum (Fig. 2b, arrowheads). Importantly, we did not observe any abnormal phenotype of Purkinje cells, when directly targeting Chd7 in Purkinje cells using a Ptf1a-Cre driver²³ (Fig. 2c). Therefore, our findings suggest that the Purkinje cell phenotype in Chd7 mutant mice and very likely in CHARGE patients is not due to a Purkinje cell-autonomous effect, but rather results from Chd7 loss in granule cells.

Chd7 regulates neuronal differentiation and cell survival

To dissect cellular mechanisms responsible for the cerebellar hypoplasia phenotype of Chd7-deficient mice, we analysed cell proliferation, differentiation, cell death and migration of granule cells during cerebellar development. Two-hour in vivo pulse-labelling of BrdU or IdU revealed similar percentages of proliferating GNPs between Chd7 WT [Chd7^f/f] and Chd7-deficient [Atoh1-Cre::Chd7^f/f] cerebella at both P0 and P3 (Fig. 3a). As further validation, we freshly isolated GNPs at P7 using a percoll gradient centrifugation²⁴ after pulse-labelling of BrdU 2 h before the killing. The FAC-based cell cycle analysis revealed no significant change of the ratio of proliferating cells between WT and Chd7-deficient mice (Supplementary Fig. 3e). Together, these data revealed that the proliferation capacity of Chd7-deficient GNPs seems comparable to WT cells. However, 24-h BrdU labelling at P6 followed by co-staining of BrdU and Ki67 at P7 showed less Ki67-negative cells among BrdU-positive cells in the EGL of Chd7-deficient mice, indicating that Chd7 loss prevents appropriate cell cycle exit of GNPs in vivo (Fig. 3b). Failure of cell cycle exit of Chd7 mutant GNPs implicates less differentiated granule cells. Indeed, IHC analysis with a neuronal marker NeuN revealed significantly less granule neurons in the IGL of Chd7 mutant than WT cerebella (Fig. 3c). The decrease of granule neurons in Chd7 mutant cerebella was confirmed by the counting of BrdU⁺ cells in IGLs of P7 pups that were pulsed and traced with BrdU for 48 h (Supplementary Fig. 3f). Moreover, IHC for cleaved caspase3 detected more apoptotic cells in the Chd7-deficient EGL at postnatal stages (Fig. 3d). In line with this observation, Chd7 loss significantly increased cell death as detected by staining of cleaved caspase3 and Parp in cultured granule cells especially in absence of SHH (Fig. 3e), indicating that mutant cells are more prone to cell death upon differentiation. Meanwhile, we did not observe defects in the migration of granule cells in Chd7-deficient cerebella, as shown by BrdU labelling and tracing (for 48 h) assay (Supplementary Fig. 3f). Thus, our data suggest that cerebellar hypoplasia in [Atoh1-Cre::Chd7^f/f] mutant mice results from both granule cell-autonomous effect (that is, a failure of GNP differentiation and increased cell death) and non-autonomous effects from the abnormal distribution of Purkinje cells.

**Figure 3: Mutation of *Chd7* in GNPs impairs neuronal differentiation and leads to cell death.**

Chd7 loss alters expression of neuronal genes

To explore molecular consequences of Chd7 loss in GNPs, we next performed RNA sequencing (RNA-seq)-based transcriptome analysis using freshly isolated GNPs from P7 Chd7 WT, heterozygous [Atoh1-Cre::Chd7^f/+] (Chd7 Het.) and homozygous [Atoh1-Cre::Chd7^f/f] (Chd7 Hom.) mutant cerebella. Results from qRT–PCR showed a significant decrease of Chd7 transcripts according to their genotypes in Chd7 Het. and Chd7 Hom. mutant GNPs as compared to WT GNPs (Fig. 4a). At the significance of P<0.01, 151 genes were downregulated in Chd7 Hom. GNPs, with many of them (62.9%, 95/151, P<0.05 between WT and Chd7 Het.) significantly downregulated in Chd7 Het. GNPs (Fig. 4b), suggesting a dosage-dependent regulation of these genes by Chd7. Importantly, DAVID Gene Ontology (GO) term analysis of genes downregulated upon loss of Chd7 demonstrated neuronal functions including synaptic transmission, ion transport, neuron development and neuron differentiation as the most enriched biological processes (Fig. 4c), while no significantly enriched term (Benjamini adjusted P value<0.05) was identified in genes significantly upregulated upon Chd7 loss. For instance, the proper level of Chd7 is necessary for the expression of neural genes that are important for cerebellar development, including Cadps2, Gap43, Neurod1, Reln and Unc5c (refs 25, 26, 27, 28, 29) (Fig. 4d, Supplementary Fig. 4a). Consistent with downregulation of Reln, IHC analyses demonstrated that Dab1, the adaptor protein of Reln signalling pathway, was upregulated specifically in mislocalized Purkinje cells in Chd7 mutant cerebella (Supplementary Fig. 4b) in accordance with observation in the reeler mice²¹. To exclude the long-term or indirect effect of Chd7 loss on gene expression, we acutely ablated the expression of Chd7 by treating [Chd7^f/f] or WT granule cells with cell-permeable Cre for 48 h. Results from qRT–PCR showed a consistent downregulation of Chd7 regulated genes identified above (Supplementary Fig. 4c).

**Figure 4: Chd7 is required for activation of neuronal genes during GNP differentiation.**

To identify direct target genes of Chd7, we performed chromatin-immunoprecipitation coupled to sequencing (ChIP-seq) using freshly isolated P7 GNPs. In order to identify bona fide Chd7-binding sites, we in parallel performed Chd7 ChIP-seq on Chd7-deficient GNPs. Using the MACS (Model-based Analysis of ChIP-seq) peak calling method, 1,946 peaks were called from WT cells, in contrast to only 96 peaks called from Chd7-deficient cells. In consistency with reports showing that Chd7 preferentially binds to enhancers in ESCs and oligodendrocytes^30,31, Chd7-occupied regions in GNPs are enriched of H3K27ac, a histone mark for active promoters and enhancers (Fig. 4e). Furthermore, Chd7 was enriched in super-enhancers (SEs) defined by H3K27ac ChIP-seq on GNPs (ref. 31) (Fig. 4f,g, Supplementary Fig. 5b). GO analysis revealed a strong enrichment of Chd7-associated genes in biological processes like transcription, chromatin organization, neuron differentiation and neuron projection morphogenesis (Supplementary Fig. 5a). Among genes whose expression was significantly altered upon loss of Chd7 (P<0.05), 12.5% (93/746) downregulated and 9.9% (99/1002) upregulated genes showed association of Chd7. For instance, Chd7 binds Chd7-activated genes, including Neurod1, Cadps2, Gap43 and Reln (Fig. 4g, Supplementary Fig. 5b).

As Chd7 functions as an ATP-dependent nucleosome remodeller³², we hypothesized that Chd7 regulates transcription via controlling DNA accessibility of cis-regulatory elements. For this, we utilized an assay for Tn5 transposase-accessible chromatin with high-throughput sequencing (ATAC-seq) (ref. 33) and generated a genome-wide map of the open chromatin landscape in Chd7 WT and mutant GNPs at P7. As expected, most of H3K27ac ChIP-seq peaks (81.9%) colocalized with ATAC-seq peaks (Supplementary Fig. 6a). A small subset of ATAC-seq peaks showed significant alteration upon loss of Chd7 (2.1%), with Chd7 Het. cells exhibiting intermediate change between WT and Chd7 Hom. (Supplementary Fig. 6b). GO analysis of genes containing altered ATAC-seq peaks called transmission of nerve impulse, cell motion and cell adhesion as the most enriched biological processes (Supplementary Fig. 6c). Among genes whose expression was significantly changed upon loss of Chd7 (P<0.05), 14.2% (106/746) downregulated and 4.5% (45/1002) upregulated genes showed altered ATAC-seq peak. For instance, decreased ATAC-peaks upon Chd7 loss were detected in Chd7-activated genes including Cadps2, Nkain3, Reln, Bmp5 and Fstl5 (Fig. 4h, Supplementary Fig. 6d). Moreover, 22.8% (320/1402) of genes bound by Chd7 in GNPs showed altered ATAC-seq peaks upon Chd7 loss, revealing that the association of Chd7 is required for the maintenance open chromatin on these genes. Collectively, comprehensive analysis of ChIP-seq and ATAC-seq data (for example, Supplementary Table 1) demonstrate that Chd7 is involved in the maintenance of open chromatin at the regulatory elements of genes related with neuronal differentiation.

Chd7 regulates granule cell differentiation with Top2b

To get insight into molecular mechanisms underlying Chd7-dependent transcriptional regulation, we aimed at identifying CHD7-interacting proteins using immunoprecipitation (IP) coupled with mass spectrometry (MS). For efficient IP with a sufficient amount of exogenous CHD7 protein, a Flag-tagged, truncated CHD7 (1–1,899 amino acids) that preserves nucleosome remodelling activity in vitro³² was overexpressed and immunoprecipitated in CHD7-deficient HEK293T cells (Supplementary Fig. 7a,b). Besides previously published CHD7-interacting partners, such as BRG1, PARP1 and CHD8 (refs 34, 35), DNA topoisomerases TOP2A, TOP2B and TOP1 were identified by MS (Fig. 5a). Protein–protein interactions between CHD7 with these proteins were validated by co-IP assays using nucleic acids-free nuclear extracts (Supplementary Fig. 7c) from HEK293T cells expressing full-length Flag-tagged CHD7 (Fig. 5b). DNA topoisomerases are key enzymes that solve the topological problems occurring during virtually all DNA-related processes³⁶. Inhibition of enzymatic activities of DNA topoisomerases in postmitotic neurons selectively represses the expression of long genes^37,38. Top2b was shown to be important for neuronal migration in the developing neocortex by controlling the expression of Reln (ref. 39). Therefore, we decided to further explore the functional relevance of interaction between Chd7 and Top2b. Co-IP assays showed the interaction between endogenous Top2b and Chd7 proteins in not only HEK293T cells but also adult cerebellar cells, 85–90% of which consist of CGNs (ref. 40) (Fig. 5c). We also demonstrate that the N-terminal part (1–786 aa) of CHD7 is responsible for its interaction with TOP2B, whereas KU70 does not interact with the N-terminus of CHD7 and PARP interacts with several regions of CHD7 (Fig. 5d).

**Figure 5: Chd7 interacts with Top2b.**

Previous immunostaining analysis demonstrated that GNPs in the oEGL express Top2a, whereas differentiating cells in iEGL and differentiated CGNs in IGL only express Top2b (ref. 41). Consistently, the transcript level of Top2a in cultured granule cells drastically decreased upon differentiation of proliferating GNPs (cultured with Smoothened agonist SAG) into postmitotic (cultured without SAG) CGNs, as analysed using Affymetrix microarray (Supplementary Fig. 8a). As Top2b is the sole DNA topoisomerase II expressed in CGNs, we decided to use in vitro induced postmitotic CGN to dissect the functional interaction between Chd7 and Top2b. We performed immunostaining of granule cells at 1 and 48 h after plating in order to compare the percentage of proliferating and differentiated cells between WT and Chd7-deficient cells. At 1 h after plating, over 80% of cells were proliferating (Ki67⁺) and expressed Pax6 (Supplementary Fig. 8b). In contrast, after 48 h, most of cells exited cell cycle (Ki67⁻) and expressed p27, suggesting that these cells were newly differentiated in culture (Supplementary Fig. 8c). Importantly, no statistical difference of percentages of proliferating and differentiated cells between WT and Chd7 mutant cells at these two time points were observed (Supplementary Fig. 8b,c). To identify genes regulated by Top2b in CGNs, we performed microarray-based gene expression analysis using cultured CGNs treated with DMSO as a control or ICRF-193 (10 μM for 12 h), an enzymatic inhibitor of Topoisomerase II (ref. 42). Consistent with previous reports in cortical neurons^37,38, long genes (gene length>100 kb) were selectively downregulated in CGNs upon treatment of ICRF-193 (Fig. 6a, left panel). The average size of genes downregulated by ICRF-193 treatment (P<0.05, fold change>1.5) is approximately 300 kb, while the average sizes of all genes expressed in CGNs and the genes upregulated upon ICRF-193 treatment are both around 50 kb (Supplementary Fig. 9a). To directly compare with Top2b-dependent transcription, we next identified Chd7-regulated genes by microarray analysis using Chd7 WT and homozygous mutant CGNs. Interestingly, many long genes were significantly downregulated upon loss of Chd7 (Fig. 6a, right panel). We subsequently performed GSEA (Gene set enrichment analysis) to check the concordant regulation of gene expression by Top2b and Chd7. Genes that were significantly downregulated upon ICRF-193 treatment showed lower expression level in Chd7-deficient CGNs than WT cells (Fig. 6b, left panel). Also, genes that were significantly downregulated in Chd7-deficient CGNs were largely repressed upon ICRF-193 treatment (Fig. 6b, right panel). At the significance of P<0.05, 240 genes were commonly activated by Chd7 and Top2b in CGNs (Supplementary Fig. 9b). Intriguingly, the average gene length of Top2b and Chd7 co-activated genes was even longer than those genes activated by Top2b but not Chd7 (Supplementary Fig. 9b), implicating that transcription of these very long genes requires both enzymes. Next, we compared expression levels of these 240 genes between GNPs cultured with and without SAG. Most of these 240 genes (84.6%, 203/240) were significantly upregulated upon differentiation (Fig. 6c). Using qRT–PCR, we validated the downregulation of ASD-related genes involved in cerebellar development (Cadps2 and Reln) (refs 28, 43), and ASD- or epilepsy-related ion channels in neurons (Cacna2d1, Grik2 and Kcnd2) (refs 43, 44, 45) in ICRF-193 treated or Chd7 mutant CGNs (Fig. 6d,e). Of note, Chd7 expression in CGNs was significantly downregulated upon ICRF-193 treatment (Fig. 6d), indicating that Top2b is required for activation of Chd7 during differentiation.

**Figure 6: Chd7 recruits Top2b to activate long neuronal genes.**

We hypothesized that Chd7 may be required for proper targeting of Top2b to chromatin. To test this, we performed Top2b ChIP-seq in WT and Chd7-deficient CGNs. The expression level of Top2b in CGNs was not altered upon loss of Chd7, at the level of both transcript and protein (Supplementary Fig. 9c). The specificity of the Top2b antibody was verified with ChIP analysis on WT and Top2b^−/− mouse embryonic fibroblasts⁴⁶ using primer sets for Top2b-targeted regions at the Fos gene (Supplementary Fig. 9d)⁴⁷. We next aimed at comparing the genome-wide binding patterns of Top2b and Chd7. Since Chd7 ChIP-seq was done on GNPs, we performed ChIP-qPCR assays to validate the binding of Chd7 in CGNs. Using primers amplifying Chd7 binding regions identified by Chd7 ChIP-seq in GNPs, we detected the association of Chd7 with multiple regions in WT CGNs, and a clear decrease of its association in Chd7 mutant CGNs (Supplementary Fig. 9e). The comparison between the binding pattern of Chd7 in GNPs and that of Top2b in CGNs revealed that Top2b is associated with Chd7 binding sites (Fig.6f, Supplementary Fig. 9f). While the average length of genes bound by Chd7 was around 103 kb, the average gene length of genes co-occupied by Chd7 and Top2b was approximately 165 kb, revealing preferential binding of Chd7 and Top2b on long genes in cerebellar granule cells. Importantly, upon loss of Chd7, 23.2% of Top2b binding sites exhibited alteration of the association of Top2b (Fig. 6g). GO analysis of genes with decreased Top2b ChIP-seq peaks revealed cell morphogenesis, cell adhesion, ion transport and neuron differentiation as the most enriched biological process (Supplementary Fig. 9g), while no significantly (Benjamini adjusted P value<0.05) enriched term was detected in genes with increased peaks in Chd7-deficient CGNs. For instance, the binding of Top2b to promoters of two Chd7- and Top2b-activated genes Cadps2 and Cnksr3 was decreased in Chd7 mutant CGNs compared to WT, whereas the binding of Top2b to its known binding site at the Fos promoter was not altered (Fig. 6h, Supplementary Fig. 9h). Importantly, GSEA revealed that genes carrying decreased Top2b ChIP-seq peaks in Chd7-deficient CGNs tended to be downregulated upon ICRF-193 treatment (Fig. 6i), revealing that Chd7-dependent recruitment of Top2b is necessary for the expression of these genes. Furthermore, while the average length of genes bound by Top2b (6147 genes) in WT CGNs was around 95 kb, the average gene length of genes carrying decreased Top2b binding peak (1190 genes) or increased Top2b binding peak (718 genes) upon loss of Chd7 was around 153 or 115 kb, respectively. Thus, Chd7-dependent targeting of Top2b is preferentially observed in long genes. Moreover, we measured the amount of Top2b covalent complexes in Chd7 WT and mutant CGNs treated with etoposide, a DNA topoisomerase poison that traps DNA with topoisomerases as a complex. We observed reproducible decrease, although not statistically significant (P=0.2169), of Top2b covalent complexes in Chd7-deficient cells (Supplementary Fig. 9i), supporting that Chd7 could contribute to the targeting of Top2b.

Top2b regulates cerebellar development

While the reelin signalling is essential for the proper localization of Purkinje cells²⁷, the cerebellar phenotype of Top2b knockout mice has not been reported so far. To investigate this, we made use of the CRISPR-Cas9 system to ablate Top2b in the developing cerebellum. The cerebella of E13.5 Rosa26-CAG-LSL-Cas9-P2A-EGFP embryos were electroporated with plasmids encoding Cre recombinase and either sgRNA against Top2b or a negative-control sequence. The transfection of cerebellar cells was efficient as shown at P7 by the wide expression of GFP upon Cre-mediated recombination (Fig. 7a). In both control and Top2b-targeted cerebella, most of GFP⁺ cells in the IGL are Pax6⁺ (Fig. 7b), demonstrating efficient targeting of CGNs by electroporation. Around 80% of GFP⁺ cells in IGL of Top2b-targeted cerebella lost the expression of Top2b (Fig. 7b), demonstrating high knockout efficiency by the CRISPR-Cas9 system in vivo. Consistent with the data mentioned above (Fig. 6d), we observed that about 40% of cells lost Chd7 expression within transfected (GFP⁺) cells upon targeting of Top2b (Fig. 7b). Furthermore, Top2b loss resulted in the downregulation of Zfpm2, a common target gene of Chd7 and Top2b (see Supplementary Data 1) (Fig. 7b). These data provide in vivo evidence that Top2b is required for the expression of neuronal genes in granule cells. Importantly, clusters of mislocalized Purkinje cells were observed in all three Top2b-targeted, but not in the two control cerebella (Fig. 7c), as observed in Chd7-deficient [Atoh1-Cre::Chd7^f/f] mice. On the other hand, the percentage of proliferating (Ki67⁺) and differentiating (p27⁺) cells among GFP⁺ cells in EGL were not altered upon mutation of Top2b (Fig. 7d), similar to our previous finding that Chd7 loss does not affect cell proliferation of GNPs. Thus, the cerebella bearing Chd7- or Top2b-deficient granule cells partly share similar phenotype, which supports our conclusion that expression of a subset of neuronal genes are regulated by Chd7 and Top2b.

**Figure 7: Genetic ablation of *Top2b* leads to phenotypic changes analogous to *Chd7*-deficient cerebella.**

Discussion

Many CHARGE patients carrying CHD7 mutation exhibit cerebellar hypoplasia and massive Purkinje cell heterotopia^4,19. In both rodents and human, cerebellar development extends a long period from early embryonic stage until the first postnatal weeks (in mouse) or years (in human). One recent study revealed that Chd7 is necessary for cerebellum specification during early embryogenesis (around E8.5 in mouse) using Chd7 gene trap mice⁴. However, the contribution of Chd7 during later cerebellar development and its implication in CHARGE syndrome has not been studied, however. In this study, cell type-specific inactivation of the Chd7 gene using multiple Cre drivers identified failure of cerebellar granule cells differentiation as a contributing factor for cerebellar hypoplasia and Purkinje cell heterotopia. RNA-seq analysis revealed impaired downregulation of Reln in Chd7-deficient GNPs, by which mislocalization of Purkinje cells could be explained as observed in reeler mice^21,48,49. Mechanistically, Chd7 is specifically associated with and is required for the activation of a core neuronal differentiation programme through maintaining the open chromatin structure and recruiting DNA topoisomerase IIb to these genes. Our data provide a clear epigenetic mechanism of Chd7-mediated gene transcriptional regulation in vivo, thus shedding light on epigenetic mutation-associated human diseases.

Our previous study revealed that loss of Chd7 in adult neural stem cells impaired the transcriptional activation of two essential neurogenic transcription factors Sox4 and Sox11, leading to defects of both the quantity and quality of newborn neurons¹². In the current study, we made use of the abundance of primary cerebellar granule cells to perform comprehensive molecular analyses. Expression profiling data clearly revealed that Chd7 is required for the activation of genes involved in neural functions. ChIP-seq on freshly isolated GNPs demonstrated that Chd7 occupies active chromatin regions in genes involved in transcription and neuronal functions. Consistent with the result from human embryonic stem cells³¹, Chd7 was found to be associated with SEs, which exhibit densely spaced clusters of active histone marks H3K27ac and clustered occupancy of cell identity-specific transcription factors. The binding of these transcription factors forming self-reinforcing circuits within SEs shapes cell-type-specific transcriptional landscape and thus governs cell identity. Interestingly, two SEs were identified within the Chd7 gene, co-occupied by Chd7 protein and H3K27ac (Supplementary Fig. 5b), implicating that Chd7 could be a key chromatin regulator of cell identity in GNPs. In accordance with our study, a very recent investigation revealed an essential role of Chd7 in the onset of myelination via targeting the enhancers of key myelinogenic genes³⁰.

Both chromatin remodellers and DNA topoisomerases are crucial enzymes involved in DNA-related processes. The cooperative action of these two groups of enzymes was not appreciated until very recently⁵⁰. In this study, we revealed the collaboration between Chd7 and DNA topoisomerases Top2b to activate the expression of long neuronal genes. As an example, our molecular analyses identified Reln as a common target of both Chd7 and Top2b. Importantly, depletion of Top2b using Foxg1-Cre, a forebrain-specific Cre driver, leads to aberrant lamination of cerebral cortex due to the repression of Reln (ref. 39). Here, we demonstrated that the enzymatic activity of Top2b is crucial for the expression of Reln in cerebellar granule cells and genetic ablation of Top2b in cerebellum during development leads to mislocalization of Purkinje cells that is very likely due to lack of reelin signalling (Fig. 6c). In cortical neurons, the enzymatic activities of Top2b and Top1 are particularly necessary for transcription of long genes³⁸. It has been thought that transcription of very long genes causes more supercoiled DNA double helix structures on these genes, thus requiring Top enzymes to solve this topological issue during transcriptional elongation. Our results suggest that besides DNA topoisomerases, chromatin-remodelling activity is critical for transcription of very long genes. Two recent studies have revealed the cooperation between another chromatin remodeller the BAF complex with DNA topoisomerases. The first study demonstrated that in dividing cells, the BAF complex interacts with Top2a and regulates Top2a-mediated decantenation activity during mitosis⁵⁰. The second study revealed that the BAF complex-dependent chromatin remodelling activity is required for the Top1-mediated gene activation during inflammation⁵¹. Here, we observed Chd7 interacts with several topoisomerases including Top1, Top2a and Top2b (Fig. 5a). It will be of great interest to explore the potential functional interaction between Chd7 and Top1 in postmitotic neurons and Top2a in dividing cells. Furthermore, as reported in a previous study³⁴ and our current study (Fig. 5a), CHD7 protein interacts with the BAF complex. It could be interesting to investigate the cooperation action among CHD7, BAF complex and DNA topoisomerase.

The rapid progress of genomic sequencing technology has provided novel insights into genetic causes of many human diseases. On the other hand, the causal role of most of these newly identified mutations needs to be experimentally validated. It is striking that most of the patients carrying CHD7 mutations display neurological disabilities, suggesting that CHD7 is a master chromatin regulator for brain function. Using mouse genetics, we here discovered that Chd7 controls a core neuronal differentiation programme during cerebellar granule cell differentiation. Importantly, 16 out of the 151 genes (P=3.146E-15, Chi-square test) that were significantly downregulated in Chd7-deficient GNPs (Fig. 4b, Supplementary Data 1) were previously found as genetic mutation(s) in ASD patients (http://www.sfari.org). Furthermore, among 240 identified Chd7 and Top2b common activated genes in CGNs (Fig. 6c, Supplementary Data 1), mutations of 35 genes are associated with ASD (P<2.2E-16, Chi-square test). Emerging evidence has linked the cerebellum with higher cognitive functions. Indeed, cerebellar deficits have long been implicated in autism⁵². Our study provides supporting evidence showing Chd7-dependent regulation of transcription is involved in the pathogenesis of autism. Moreover, many Chd7-regulated genes identified in this study, such as Gap43, Cadps2 and Reln, function in other brain regions besides cerebellum^21,26,53. Thus, the molecular network controlled by Chd7 maybe a core network for brain development and function in human.

Methods

Animals

Chd7 conditional knockout mice were generated by breeding of Chd7^f/f (ref. 12) with Nestin-Cre (JAX, #003771), Atoh1-Cre (JAX, #011104) and Ptf1a-Cre mice (BCBC, RES184). Chd7-GFP (from the GENSAT project, http://www.gensat.org/index.html) and Rosa26-CAG-LSL-Cas9-P2A-EGFP (JAX, #024857) mice were purchased. Mice were housed in a vivarium with a 12-h light/dark cycle. All mice were bred in C57BL/6N background. Genotyping primers are listed in Supplementary Table 2. For in utero electroporation experiments, Rosa26-CAG-LSL-Cas9-P2A-EGFP homozygous knockin male mice were crossed with CD-1 female mice, yielding embryos carrying Rosa26-CAG-LSL-Cas9-P2A-EGFP heterozygous genotypes. For in vivo BrdU or IdU labelling, animals were subjected to intraperitoneal injection of BrdU or IdU (3 mg per 40 g body weight) and were killed at defined stages. All animal experiments were conducted according to animal welfare regulations and have been approved by the responsible authorities (Regierungspräsidium Karlsruhe, approval numbers: G32/14, G48/14 and G90/13).

Plasmids

The human CHD7 (1–1,899) cDNA with a C-terminal Flag tag was amplified from pFastbac-CHD7 (1-1,899) (a gift from Robert Kingston)³², and cloned into pEF-DEST51 using the Gateway technology (Thermo Fisher Scientific). The full-length CHD7 expression construct containing an N-terminal GFP tag was inserted into the Flag-tagged CHD7-encoding expression plasmid³⁴. The truncated CHD7 constructs were generated by the insertion of PCR products into pcDNA3.1 empty vector. The coding sequence of sgRNA against CHD7 (CTTTCTAGAGAAACCAGTGC) was cloned into pLenti-CRISPR v2 (Addgene #52961). For genetic ablation of Top2b, the coding sequence of sgRNA against Top2b (target sequence: #1: CTTCGTCCTGATACATACAT; #2: ACTGATCCAATGTATGTATC) or control region (target sequence: GCGACCAATACGCGAACGTC) was cloned into the pU6-sgRNA-Cbh-Cre plasmid, which was modified from pX330 (Addgene #42230) by replacing the hSpCas9 with Cre coding sequence.

In utero electroporation

In utero electroporation was performed as described previously⁵⁴ with some modifications. To ablate Top2b, pU6-sgRNA-Cbh-Cre (2 μg ml⁻¹) was injected into the fourth ventricle and electroporated (35 V, 50 ms-on, 950 ms-off, 5 pulses) into the cerebellar primordial of E13.5 Rosa26-CAG-LSL-Cas9-P2A-EGFP embryos. Only GFP-labelled cerebella were used for further analyses. No specific randomization or blinding was performed.

Primary cerebellar granule cell preparation

GNPs were purified from P7 C57BL/6N pups using percoll gradient centrifugation and plated on matrigel-coated plates as described previously²⁴. We did not discriminate the gender of pups. For BrdU analysis, freshly isolated cells were analysed by using the flow cytometry after staining with an anti-BrdU antibody (BD Pharmingen). In some experiments, the cells were cultured for defined days with or without either Shh protein or SAG (200 nM) and collected for analysis. To inhibit Top2b activity, cells were treated with ICRF-193 (Sigma-Aldrich) in DMSO at 10 μM for 12 h before collecting them. To delete Chd7 in vitro, Chd7 f/f GNPs were cultured without Shh, but treated with TAT-Cre (Merck Millipore) at 0.5 μM for 48 h before collecting them for RNA analysis.

Culture and transfection of HEK293T cells

Mycoplasma-free HEK293T cells from American Type Culture Collection were cultured in DMEM-Glutamax, (ThermoFisher Scientific) supplemented with 10% fetal bovine serum (ThermoFisher Scientific), 100 U ml⁻¹ of penicillin-streptomycin (ThermoFisher Scientific). For transient transfection, cells were transfected with a DNA-PEI (Sigma-Aldrich) mixture and collected 48 h later. To establish CHD7 knockout cell line with the CRISPR-Cas9 technology, HEK293T cells were transfected with pLentiCRISPR-sgCHD7 plasmid, selected with puromycin (1 μg ml⁻¹) for 7 days before single cell clones were picked up and analysed.

Preparation of cerebellar lysates

Cerebella from adult (2-month-old) mice were freshly dissected. After snap freezing in liquid nitrogen, tissue was thawed on ice. Cerebella were homogenized in buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl₂), and incubated on ice for 30 min. After centrifugation, nuclei were dissolved in IP buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, complete protease inhibitor cocktail (Roche)).

Immunoprecipitation and mass-spectrometry

We performed IP as described previously⁵⁵. Briefly, nuclei from transfected HEK293T were lysed in IP buffer, and treated with Benzonase (Sigma-Aldrich, 700 U ml⁻¹) for 1 h at 4 °C. Nuclear extracts were incubated at 4 °C overnight with M2 agarose (Sigma-Aldrich) or protein A/G sepharose (GE healthcare) with primary antibodies. After successive washing with IP buffer containing 200 mM NaCl, precipitated proteins were eluted in Laemmli buffer and analysed on western blots. Assays were repeated at least three times and the representative images were selected. For MS analysis, precipitated proteins were eluted with 400 ng μl⁻¹ 3 × Flag peptides (Sigma-Aldrich) in IP buffer containing 300 mM NaCl. Proteins were digested in solution with trypsin, and tryptic peptides were analysed with the Orbitrap XL mass spectrometer. The peptides were identified with the MASCOT searching engine against the SwissProt database. Uncropped immunoblots are shown in Supplementary Fig. 10.

Immunostaining

IHC was performed essentially described previously²⁴. The animals were killed at defined stages and fixed with 4% paraformaldehyde in PBS (PFA/PBS) overnight, followed by sectioning with cryostat (Leica) and microtome (Leica). For immunofluorescence on cultured GNPs, cells were fixed with 4% PFA/PBS for 10 min at room temperature, permeabilized for 5 min in PBS containing 0.2% Triton X-100. For BrdU staining, sections were treated with 2 N HCl for 30 min at 37 °C. Tissue sections and cultured cells were subsequently subjected to the primary antibody incubation for overnight at 4 °C after blocking with 10% normal donkey serum. For IHC, paraffin sections were incubated with biotinylated secondary antibodies (Vector) and signals were amplified by a horseradish peroxidase system (ABC kit, Vector) followed by DAB staining (Sigma-Aldrich). Fluorescent images were captured using a confocal laser-scanning microscopy (LSM780 and 800, Zeiss; and SP5, Leica). At least eight sections from one mouse were stained. Cell counting was done in a double-blind manner whenever applicable. Representative images were selected from sections from at least three independent mice.

Antibodies

Antibodies used in this study were the following: BrdU (1:500, AbDserotec, OBT0030CX), Brg1 (1:200, Santa Cruz Biotechnology, sc-374197), Cadps2 (1:500, Merck Millipore, ABN326), Calbindin (1:500, Swant, CB38), Chd7 (1:400, Cell Signaling, #6505), cleaved Caspase3 (1:200, Cell Signaling, #9661), Dab1 (1:200, Sigma-Aldrich, Ab232), Gap43 (1:500, Sigma-Aldrich, G9264), Ki67 (1:1,000, Abcam, ab15580), Ku70 (1:500, ThermoFisher Scientific, MA5-13110), GFP (1:500, Life technologies A11122; Abcam ab13970), H3K27Ac (2 μg per ChIP, Diagenode, pAb-174-050), NeuN (1:500, Merck Millipore, MAB377), p27 (1:500, BD bioscience, 610241), Parp (1:500, Cell Signaling,), Pax2 (1:1,000, Life technologies, 71-6000), Pax6 (1:500, Covance, PRB-278P), Reln (1:200, Merck Millipore, MAB5366), RPA116 (ref. 54), Sox2 (1:200, Santa Cruz Biotechnology, sc17320), Top2a (1:500, Santa Cruz Biotechnology, sc3659), Top2b (1:500, Santa Cruz Biotechnology, sc13059) and Zfpm2 (1:200, Santa Cruz Biotechnology, sc10755). These antibodies were quality-checked in previous publications.

Quantitative RT–PCR (qRT–PCR)

Total RNAs were extracted from cultured GNPs and tumour tissues with TRI-Reagent (Sigma-Aldrich) and purified with the RNeasy Mini kit (Qiagen). RNA was transcribed into cDNA using random primers (dN6, Roche) and M-MLV reverse transcriptase (Promega). cDNA were quantified by using SYBR gene expression assays (Qiagen) or Taqman Probe with Absolute Blue qPCR Rox mix (ThermoFisher Scientific), on the CFX96 Real-time System (Bio-rad). Primers and Probes used for qRT–PCR were listed in Supplementary Table 2.

Chromatin immunoprecipitation

We performed the IP as described previously⁵⁴. Freshly isolated P7 GNPs were fixed with 1% formaldehyde for 10 min at room temperature, and quenched with 125 mM glycine. Chromatin was extracted and sonicated for 20 min (30 s on, 30 s off) using a sonicator (Diagenode). After pre-clearing for 1 h by incubating with protein A/G beads, chromatin lysate was incubated overnight at 4 °C with antibodies. Beads were extensively washed. DNA–protein complex was eluted, reverse cross-linking and purified using Qiagen PCR purification column.

Topoisomerase covalent complex assay

Samples were processed according to the in vivo complex of enzyme (ICE) assay CGNs (ref. 56). CGNs were treated with etoposide (100 μM) for 1 h at 37 °C, and lysed in 1% sarkosyl in TE (10 mM Tris HCl, pH 7.5; 1 mM EDTA). DNA was sheared by passing the lysate ten times through a 2 ml syringe with a 25G/8 Gauge needle. Samples were centrifuged at 57,000 r.p.m. for 20 h at 25 °C using 3.3 ml 13 × 33 mm polyallomer Optiseal tubes (Beckman Coulter) in a TLN100 rotor (Beckman Coulter). Precipitated DNA (5 μg) was transferred onto Hybond ECL membranes (GE Healthcare) using a Bio-Dot SF Microfiltration Apparatus (Bio-Rad). Membranes were blocked 1 h with Odyssey Blocking Buffer (LI-COR Biosciences), incubated with anti-TOP2b antibody (1:1,000 dilution), in Odyssey Blocking Buffer-0.1% Tween20, washed (three times with TBS-0.1% Tween20), incubated with 1/15,000 IRDye 800CW Goat anti-Rabbit IgG (LI-COR) and finally washed (three times with TBS-0.1%Tween20 and once with TBS). Dried blots were analysed and quantified in Odyssey CLx using ImageStudio Odyssey CLx Software.

Gene expression analysis

Total RNA was extracted and subsequently profiled using the Affymetrix microarray chip (430V2). For RNA-sequencing, a sequencing library was prepared by using the Illumila TrueSeq Total RNA Sample Prep Kit and sequenced on the Illumina Hi-Seq 2000, single-end 50 bp reads. For each sample, about 20,000,000 single-end reads were obtained. RSEM-1.2.15 with Bowtie-1.0.0 is used to map the reads to the Ensemble mouse reference genome assembly mm10 transcriptome, and to estimate the expression levels for all the genes. For both Bowtie and RSEM, default parameters were used. Transcripts per million reads were adopted to quantify gene expression levels. RNA samples were sequenced in two batches, with the following samples as the first batch, Chd7 WT-1, 2, 3; Chd7 heterozygous mutant-1, 2, 3 and Chd7 homozygous mutant-1, 2, and the rest samples as the second batch. For this, we used ComBat with default parameters to remove the batch effect⁵⁷.

ATAC-seq analysis

ATAC-seq was performed with minor modification according to the original method³³. Briefly, 100,000 GNPs from P7 pups were isolated, permeabilized and incubated with Tn5 transposase (Illumina) for 30 min at 37 °C. DNA from Tn5-treated cells was purified and used as a template to amplify a DNA library by PCR. The Libraries were deep-sequenced with HiSeq 2000 v4 paired-end 125 bp. The ATAC-seq analysis was performed according to published papers^33,58. There are three technical replicates for both Chd7 WT and homozygous mutant GNPs, and two replicates for Chd7 heterozygous mutant GNPs. For each sample, approximately 30,000,000 paired-end reads were sequenced. Reads were first trimmed with trim-galore_0.4.1 to remove the adapters and trimmed reads are mapped with bowtie 1.0.0[1] to mm10 with parameters -X 2000 -m 1. Duplicated reads and reads from mitochondrial DNA were removed with picard_1.139. To call the peaks, macs2 2.1.0.20150731 (ref. 59) were applied with parameters -nomodel -shift -100 -extsize 200. In order to calculate total amount of peaks, peaks given by macs2 from each of above-mentioned three genotypes were merged when peak distance is less than 300 b.p. Differential peaks were identified with edgeR (ref. 60) according to the number of reads mapped to each peak counted by Rsubread (refs 61, 62) for each replicate. Both edgeR and Rsubread were run with default parameters. ATAC-seq peaks are annotated using PAVIS (ref. 61) with default parameters. Coverage density plots were done with deeptools.

Analysis of ChIP-seq

About 5 million GNPs collected from P7 pups were used for H3K27ac ChIP-seq according to standard ChIP protocol described above. DNA libraries generated from ChIP-DNA and input-DNA were deep-sequenced using HiSeq pair-end 125 bp. After trimming with trim_galore, more than 40 million pair-end reads were mapped to mm10 genome with bowtie 1.0.0 with parameter -X 1000 -m 1. Peaks were called with macs2 with default parameters for Chd7 and H3K27ac ChIP-seq on GNPs. For Top2b ChIP-seq (WT and Chd7 mutant CGNs), we use homer⁵⁸ findPeaks with parameter -F 2.0 -P 0.001 -L 2.0 -LP 0.001 -fdr 0.01 -style factor. For the differential peaks of Top2b Chip seq, MAnorm⁵⁹ was used with cutoff P<0.01 and fold change>2. SE was called using homer findPeaks with parameter -style super. ChIP-seq peaks are annotated using PAVIS (ref. 61) with default parameters. Coverage density plots were done with deeptools.

GO analysis and GSEA

Gene ontology biological process was performed with David6.7 (http://david.ncifcrf.gov). GO term with the Benjamini adjusted P value<0.05 is considered to be significantly enriched. For custom gene sets, GSEA (www.broadinstitute.org/gsea) was performed using default parameter. FDR<25%, P<0.05 was considered to be significantly associated.

Statistical methods

Two-tailed Student’s t-test with equal variance was performed to compare two groups. The Kruskal−Wallis test was performed for tests with three groups. Pearson test and Chi-squared test were performed as mentioned. For significance test, *means P<0.05. Error bar stands for mean±s.d.

Data availability

All sequencing and microarray data have been submitted to the US National Institute of Health GEO database (https://www.ncbi.nlm.nih.gov/gds) under the accession number GSE93741.

Additional information

How to cite this article: Feng, W. et al. Chd7 is indispensable for mammalian brain development through activation of a neuronal differentiation programme. Nat. Commun. 8, 14758 doi: 10.1038/ncomms14758 (2017).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

References

Plass, C. et al. Mutations in regulators of the epigenome and their connections to global chromatin patterns in cancer. Nat. Rev. Genet. 14, 765–780 (2013).
Article CAS PubMed Google Scholar
Jakovcevski, M. & Akbarian, S. Epigenetic mechanisms in neurological disease. Nat. Med. 18, 1194–1204 (2012).
Article CAS PubMed PubMed Central Google Scholar
Vissers, L. E. et al. Mutations in a new member of the chromodomain gene family cause CHARGE syndrome. Nat. Genet. 36, 955–957 (2004).
Article CAS PubMed Google Scholar
Yu, T. et al. Deregulated FGF and homeotic gene expression underlies cerebellar vermis hypoplasia in CHARGE syndrome. Elife 2, e01305 (2013).
Article PubMed PubMed Central Google Scholar
Bergman, J. E. et al. CHD7 mutations and CHARGE syndrome: the implications of an expanding phenotype. J. Med. Genet. 48, 334–342 (2011).
Article CAS PubMed Google Scholar
Hartshorne, T. S., Grialou, T. L. & Parker, K. R. Autistic-like behavior in CHARGE syndrome. Am. J. Med. Genet. A 133A, 257–261 (2005).
Article PubMed Google Scholar
Cadman, S. M., Kim, S. H., Hu, Y., Gonzalez-Martinez, D. & Bouloux, P. M. Molecular pathogenesis of Kallmann's syndrome. Horm. Res. 67, 231–242 (2007).
CAS PubMed Google Scholar
O'Roak, B. J. et al. Sporadic autism exomes reveal a highly interconnected protein network of de novo mutations. Nature 485, 246–250 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Homsy, J. et al. De novo mutations in congenital heart disease with neurodevelopmental and other congenital anomalies. Science 350, 1262–1266 (2015).
Article ADS CAS PubMed PubMed Central Google Scholar
Layman, W. S., Hurd, E. A. & Martin, D. M. Chromodomain proteins in development: lessons from CHARGE syndrome. Clin. Genet. 78, 11–20 (2010).
Article CAS PubMed PubMed Central Google Scholar
Hurd, E. A. et al. Loss of Chd7 function in gene-trapped reporter mice is embryonic lethal and associated with severe defects in multiple developing tissues. Mamm. Genome 18, 94–104 (2007).
Article CAS PubMed Google Scholar
Feng, W. et al. The chromatin remodeler CHD7 regulates adult neurogenesis via activation of SoxC transcription factors. Cell Stem Cell 13, 62–72 (2013).
Article CAS PubMed Google Scholar
Micucci, J. A. et al. CHD7 and retinoic acid signaling cooperate to regulate neural stem cell and inner ear development in mouse models of CHARGE syndrome. Hum. Mol. Genet. 23, 434–448 (2014).
Article CAS PubMed Google Scholar
Jones, K. M. et al. CHD7 maintains neural stem cell quiescence and prevents premature stem cell depletion in the adult hippocampus. Stem Cells 33, 196–210 (2015).
Article CAS PubMed Google Scholar
Leto, K. et al. Consensus paper: cerebellar development. Cerebellum 15, 789–828 (2015).
Article PubMed Central Google Scholar
Behesti, H. & Marino, S. Cerebellar granule cells: insights into proliferation, differentiation, and role in medulloblastoma pathogenesis. Int. J. Biochem. Cell Biol. 41, 435–445 (2009).
Article CAS PubMed Google Scholar
Zhu, X. et al. Role of Tet1/3 genes and chromatin remodeling genes in cerebellar circuit formation. Neuron 89, 100–112 (2016).
Article CAS PubMed Google Scholar
Frank, C. L. et al. Regulation of chromatin accessibility and Zic binding at enhancers in the developing cerebellum. Nat. Neurosci. 18, 647–656 (2015).
Article CAS PubMed PubMed Central Google Scholar
Legendre, M. et al. Antenatal spectrum of CHARGE syndrome in 40 fetuses with CHD7 mutations. J. Med. Genet. 49, 698–707 (2012).
Article CAS PubMed Google Scholar
Moreno, N. et al. Loss of Smarc proteins impairs cerebellar development. J. Neurosci. 34, 13486–13491 (2014).
Article PubMed PubMed Central Google Scholar
Rice, D. S. & Curran, T. Role of the reelin signaling pathway in central nervous system development. Annu. Rev. Neurosci. 24, 1005–1039 (2001).
Article CAS PubMed Google Scholar
Whitney, E. R., Kemper, T. L., Bauman, M. L., Rosene, D. L. & Blatt, G. J. Cerebellar Purkinje cells are reduced in a subpopulation of autistic brains: a stereological experiment using calbindin-D28k. Cerebellum 7, 406–416 (2008).
Article CAS PubMed Google Scholar
Hoshino, M. et al. Ptf1a, a bHLH transcriptional gene, defines GABAergic neuronal fates in cerebellum. Neuron 47, 201–213 (2005).
Article CAS PubMed Google Scholar
Kawauchi, D. et al. A mouse model of the most aggressive subgroup of human medulloblastoma. Cancer Cell 21, 168–180 (2012).
Article CAS PubMed PubMed Central Google Scholar
Ackerman, S. L. et al. The mouse rostral cerebellar malformation gene encodes an UNC-5-like protein. Nature 386, 838–842 (1997).
Article ADS CAS PubMed Google Scholar
Shen, Y., Mishra, R., Mani, S. & Meiri, K. F. Both cell-autonomous and cell non-autonomous functions of GAP-43 are required for normal patterning of the cerebellum in vivo. Cerebellum 7, 451–466 (2008).
Article CAS PubMed PubMed Central Google Scholar
Goldowitz, D. et al. Cerebellar disorganization characteristic of reeler in scrambler mutant mice despite presence of reelin. J. Neurosci. 17, 8767–8777 (1997).
Article CAS PubMed PubMed Central Google Scholar
Sadakata, T. et al. Autistic-like phenotypes in Cadps2-knockout mice and aberrant CADPS2 splicing in autistic patients. J. Clin. Invest. 117, 931–943 (2007).
Article CAS PubMed PubMed Central Google Scholar
Cho, J. H. & Tsai, M. J. Preferential posterior cerebellum defect in BETA2/NeuroD1 knockout mice is the result of differential expression of BETA2/NeuroD1 along anterior-posterior axis. Dev. Biol. 290, 125–138 (2006).
Article CAS PubMed Google Scholar
He, D. et al. Chd7 cooperates with Sox10 and regulates the onset of CNS myelination and remyelination. Nat. Neurosci. 19, 678–689 (2016).
Article CAS PubMed PubMed Central Google Scholar
Hnisz, D. et al. Super-enhancers in the control of cell identity and disease. Cell 155, 934–947 (2013).
Article CAS PubMed Google Scholar
Bouazoune, K. & Kingston, R. E. Chromatin remodeling by the CHD7 protein is impaired by mutations that cause human developmental disorders. Proc. Natl Acad. Sci. USA 109, 19238–19243 (2012).
Article ADS CAS PubMed PubMed Central Google Scholar
Buenrostro, J. D., Giresi, P. G., Zaba, L. C., Chang, H. Y. & Greenleaf, W. J. Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat. Methods 10, 1213–1218 (2013).
Article CAS PubMed PubMed Central Google Scholar
Bajpai, R. et al. CHD7 cooperates with PBAF to control multipotent neural crest formation. Nature 463, 958–962 (2010).
Article ADS CAS PubMed PubMed Central Google Scholar
Batsukh, T. et al. CHD8 interacts with CHD7, a protein which is mutated in CHARGE syndrome. Hum. Mol. Genet. 19, 2858–2866 (2010).
Article CAS PubMed Google Scholar
Nitiss, J. L. DNA topoisomerase II and its growing repertoire of biological functions. Nat. Rev. Cancer 9, 327–337 (2009).
Article CAS PubMed PubMed Central Google Scholar
Tiwari, V. K. et al. Target genes of Topoisomerase IIbeta regulate neuronal survival and are defined by their chromatin state. Proc. Natl Acad. Sci. USA 109, E934–E943 (2012).
Article CAS PubMed PubMed Central Google Scholar
King, I. F. et al. Topoisomerases facilitate transcription of long genes linked to autism. Nature 501, 58–62 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Lyu, Y. L. & Wang, J. C. Aberrant lamination in the cerebral cortex of mouse embryos lacking DNA topoisomerase IIbeta. Proc. Natl Acad. Sci. USA 100, 7123–7128 (2003).
Article ADS CAS PubMed PubMed Central Google Scholar
Lange, W. Cell number and cell density in the cerebellar cortex of man and some other mammals. Cell Tissue Res. 157, 115–124 (1975).
Article CAS PubMed Google Scholar
Tsutsui, K., Tsutsui, K., Hosoya, O., Sano, K. & Tokunaga, A. Immunohistochemical analyses of DNA topoisomerase II isoforms in developing rat cerebellum. J. Comp. Neurol. 431, 228–239 (2001).
Article CAS PubMed Google Scholar
Huang, K. C. et al. Topoisomerase II poisoning by ICRF-193. J. Biol. Chem. 276, 44488–44494 (2001).
Article CAS PubMed Google Scholar
Holt, R. et al. Linkage and candidate gene studies of autism spectrum disorders in European populations. Eur. J. Hum. Genet. 18, 1013–1019 (2010).
Article CAS PubMed PubMed Central Google Scholar
Vergult, S. et al. Genomic aberrations of the CACNA2D1 gene in three patients with epilepsy and intellectual disability. Eur. J. Hum. Genet. 23, 628–632 (2015).
Article CAS PubMed Google Scholar
Syrbe, S. et al. De novo loss- or gain-of-function mutations in KCNA2 cause epileptic encephalopathy. Nat. Genet. 47, 393–399 (2015).
Article CAS PubMed PubMed Central Google Scholar
Errington, F. et al. Murine transgenic cells lacking DNA topoisomerase IIbeta are resistant to acridines and mitoxantrone: analysis of cytotoxicity and cleavable complex formation. Mol. Pharmacol. 56, 1309–1316 (1999).
Article CAS PubMed Google Scholar
Madabhushi, R. et al. Activity-induced DNA breaks govern the expression of neuronal early-response genes. Cell 161, 1592–1605 (2015).
Article CAS PubMed PubMed Central Google Scholar
Miyata, T. et al. Migration, early axonogenesis, and Reelin-dependent layer-forming behavior of early/posterior-born Purkinje cells in the developing mouse lateral cerebellum. Neural Dev. 5, 23 (2010).
Article PubMed PubMed Central Google Scholar
Jensen, P., Zoghbi, H. Y. & Goldowitz, D. Dissection of the cellular and molecular events that position cerebellar Purkinje cells: a study of the math1 null-mutant mouse. J. Neurosci. 22, 8110–8116 (2002).
Article CAS PubMed PubMed Central Google Scholar
Dykhuizen, E. C. et al. BAF complexes facilitate decatenation of DNA by topoisomerase IIalpha. Nature 497, 624–627 (2013).
Article ADS CAS PubMed PubMed Central Google Scholar
Rialdi, A. et al. Topoisomerase 1 inhibition suppresses inflammatory genes and protects from death by inflammation. Science 352, aad7993 (2016).
Article PubMed PubMed Central Google Scholar
Wang, S. S., Kloth, A. D. & Badura, A. The cerebellum, sensitive periods, and autism. Neuron 83, 518–532 (2014).
Article CAS PubMed PubMed Central Google Scholar
Shinoda, Y. et al. Calcium-dependent activator protein for secretion 2 (CAPS2) promotes BDNF secretion and is critical for the development of GABAergic interneuron network. Proc. Natl Acad. Sci. USA 108, 373–378 (2011).
Article ADS CAS PubMed Google Scholar
Zuckermann, M. et al. Somatic CRISPR/Cas9-mediated tumour suppressor disruption enables versatile brain tumour modelling. Nat. Commun. 6, 7391 (2015).
Article CAS PubMed Google Scholar
Feng, W., Yonezawa, M., Ye, J., Jenuwein, T. & Grummt, I. PHF8 activates transcription of rRNA genes through H3K4me3 binding and H3K9me1/2 demethylation. Nat. Struct. Mol. Biol. 17, 445–450 (2010).
Article CAS PubMed Google Scholar
Nitiss, J. L., Soans, E., Rogojina, A., Seth, A. & Mishina, M. Topoisomerase assays. Curr. Protoc. Pharmacol. 57, 3.3.1–3.3.27 (2012).
Google Scholar
Johnson, W. E., Li, C. & Rabinovic, A. Adjusting batch effects in microarray expression data using empirical Bayes methods. Biostatistics 8, 118–127 (2007).
Article PubMed Google Scholar
Bao, X. et al. A novel ATAC-seq approach reveals lineage-specific reinforcement of the open chromatin landscape via cooperation between BAF and p63. Genome Biol. 16, 284 (2015).
Article PubMed PubMed Central Google Scholar
Zhang, Y. et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol. 9, R137 (2008).
Article PubMed PubMed Central Google Scholar
Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Article CAS PubMed Google Scholar
Dobin, A. et al. STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29, 15–21 (2013).
Article CAS PubMed Google Scholar
Liao, Y., Smyth, G. K. & Shi, W. The Subread aligner: fast, accurate and scalable read mapping by seed-and-vote. Nucleic Acids Res. 41, e108 (2013).
Article PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank Drs R. Kingston and H. van Attikum for providing us the CHD7 expression constructs; C. Shao, L. Linke and N. Mack for excellent technical assistance, Drs A. Forget, K. Reifenberg, N. Denk and K. Dell for helpful assistance for animal experiments; the Imaging and Cytometry, Genomics and Proteomics Core Facilities of the DKFZ and the Carl Zeiss Imaging Center in the DKFZ for their support. This work was supported by the Helmholtz Association (VH-NG-702), the Deutsche Forschungsgemeinschaft (KA 4472/1-1 for D.K., LI 2140/1-1 for H.-K.L.), the Deutsche Krebshilfe (110226 to H.-K.L.), the ERC (European Research Council) consolidator grant (647055) (to H.-K.L.), the Helmholtz Alliance ‘Preclinical Comprehensive Cancer Center’ Grant HA-305 (to J.G., P.L. S.M.P. and H.-K.L.), the DKFZ Intramural Grant (to W.F. and D.K.), the Spanish Government (SAF2010-21017, SAF2013-47343-P, SAF2014-55532-R and FEDER funds, to F.C.-L.), the Andalusian Regional Government (P11-CVI-7948 and FEDER funds, to F.C.-L.), the European Research Council (ERC-CoG-2014-647359, to F.C.-L.) and Predoctoral Studentships from the University of Seville to J.A.L. (PIF-2011). CABIMER is supported by the Andalusian Regional Government (Junta de Andalucía).

Author information

Weijun Feng and Daisuke Kawauchi: These authors contributed equally to this work

Authors and Affiliations

Division of Molecular Neurogenetics, German Cancer Research Center (DKFZ), DKFZ–ZMBH Alliance, Im Neuenheimer Feld 280, Heidelberg, 69120, Germany
Weijun Feng, Huiqin Körkel-Qu, Huan Deng, Elisabeth Serger, Yassin Harim, Malin Jansen, Anna Neuerburg, Olga Friesen & Hai-Kun Liu
Division of Pediatric Neuro-oncology, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280, Heidelberg, 69120, Germany
Daisuke Kawauchi, Laura Sieber, Sander Lambo, Vijayanad Rajendran, David T. W. Jones, Marcel Kool & Stefan M. Pfister
Centro Andaluz de Biología Molecular y Medicina Regenerativa (CABIMER), CSIC-Universidad de Sevilla-Universidad Pablo de Olavide, Sevilla, 41092, Spain
Jenna Ariel Lieberman, Silvia Jimeno-González & Felipe Cortés-Ledesma
Molecular Genetics, German Cancer Research Center (DKFZ), Im Neuenheimer Feld 280, Heidelberg, 69120, Germany
Bola S. Hanna, Marc Zuckermann, Jan Gronych & Peter Lichter
Institut Curie, CNRS UMR 3347, INSERM U1021, Centre Universitaire, Bâtiment 110, Orsay, 91405, France
Olivier Ayrault
Department of Neuropathology, Clinical Cooperation Unit Neuropathology, German Cancer Research Centre (DKFZ), University of Heidelberg, Heidelberg, 69120, Germany
Andrey Korshunov
German Cancer Consortium (DKTK), Core Center Heidelberg, Heidelberg, 69120, Germany
Andrey Korshunov, David T. W. Jones, Peter Lichter & Stefan M. Pfister
Department of Developmental Neurobiology, St. Jude Children’s Research Hospital, 262 Danny Thomas Place, Memphis, 38105, Tennessee, USA
Paul A. Northcott
Department of Pediatric Hematology and Oncology, Heidelberg University Hospital, Im Neuenheimer Feld 430, Heidelberg, 69120, Germany
Stefan M. Pfister

Authors

Weijun Feng
View author publications
You can also search for this author in PubMed Google Scholar
Daisuke Kawauchi
View author publications
You can also search for this author in PubMed Google Scholar
Huiqin Körkel-Qu
View author publications
You can also search for this author in PubMed Google Scholar
Huan Deng
View author publications
You can also search for this author in PubMed Google Scholar
Elisabeth Serger
View author publications
You can also search for this author in PubMed Google Scholar
Laura Sieber
View author publications
You can also search for this author in PubMed Google Scholar
Jenna Ariel Lieberman
View author publications
You can also search for this author in PubMed Google Scholar
Silvia Jimeno-González
View author publications
You can also search for this author in PubMed Google Scholar
Sander Lambo
View author publications
You can also search for this author in PubMed Google Scholar
Bola S. Hanna
View author publications
You can also search for this author in PubMed Google Scholar
Yassin Harim
View author publications
You can also search for this author in PubMed Google Scholar
Malin Jansen
View author publications
You can also search for this author in PubMed Google Scholar
Anna Neuerburg
View author publications
You can also search for this author in PubMed Google Scholar
Olga Friesen
View author publications
You can also search for this author in PubMed Google Scholar
Marc Zuckermann
View author publications
You can also search for this author in PubMed Google Scholar
Vijayanad Rajendran
View author publications
You can also search for this author in PubMed Google Scholar
Jan Gronych
View author publications
You can also search for this author in PubMed Google Scholar
Olivier Ayrault
View author publications
You can also search for this author in PubMed Google Scholar
Andrey Korshunov
View author publications
You can also search for this author in PubMed Google Scholar
David T. W. Jones
View author publications
You can also search for this author in PubMed Google Scholar
Marcel Kool
View author publications
You can also search for this author in PubMed Google Scholar
Paul A. Northcott
View author publications
You can also search for this author in PubMed Google Scholar
Peter Lichter
View author publications
You can also search for this author in PubMed Google Scholar
Felipe Cortés-Ledesma
View author publications
You can also search for this author in PubMed Google Scholar
Stefan M. Pfister
View author publications
You can also search for this author in PubMed Google Scholar
Hai-Kun Liu
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

W.F., D.K., S.M.P and H.-K.L. conceived the project. D.K., W.F., S.M.P. and H.-K.L. supervised the project. W.F., D.K., H.D., E.S., L.S., O.F., M.J., and A.N. analysed cerebellar tissues and primary granule cells. W.F. B.S.H. and E.S. performed experiments on cell lines. J.A.L. performed the Topoisomerase covalent complex assay. H.K-Q. analysed RNA-sequencing data, Affymetrix microarray, ATAC-seq data and ChIP-seq data. A.K. performed histopathological analysis. S.L., D.T.W.J., M.K. and P.A.N. analysed human data. Y.H., V.R., M.Z. and O.A. provided biomaterials. J.G., O.A. and P.L. provided intellectual inputs. W.F., D.K., J.A.L., J.G., D.T.W.J. M.K., F.C.-L., S.M.P. and H.-K.L. edited and wrote the manuscript.

Corresponding authors

Correspondence to Daisuke Kawauchi, Stefan M. Pfister or Hai-Kun Liu.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Supplementary information

Supplementary Information

Supplementary Figures and Supplementary Tables. (PDF 2695 kb)

Supplementary Data 1

Sheet 1: TPM values of genes downregulated upon loss of Chd7 in GNPs as identified by RNA-seq (corresponding to Figure 2b). Sheet 2: Microarray data showing the list of genes that are downregulated upon loss of Chd7 and inhibition of Top2b in CGNs (corresponding to Figure 6c). (XLSX 76 kb)

Peer Review File (PDF 230 kb)

Rights and permissions

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Reprints and permissions

About this article

Cite this article

Feng, W., Kawauchi, D., Körkel-Qu, H. et al. Chd7 is indispensable for mammalian brain development through activation of a neuronal differentiation programme. Nat Commun 8, 14758 (2017). https://doi.org/10.1038/ncomms14758

Download citation

Received: 23 August 2016
Accepted: 25 January 2017
Published: 20 March 2017
DOI: https://doi.org/10.1038/ncomms14758

This article is cited by

Small molecule modulators of chromatin remodeling: from neurodevelopment to neurodegeneration
- Dongfang Jiang
- Tingting Li
- Hongmei Liu
Cell & Bioscience (2023)
RENGE infers gene regulatory networks using time-series single-cell RNA-seq data with CRISPR perturbations
- Masato Ishikawa
- Seiichi Sugino
- Atsushi Mochizuki
Communications Biology (2023)
Preliminary studies on apparent mendelian psychotic disorders in consanguineous families
- Ambreen Kanwal
- Sohail A. Sheikh
- José V. Pardo
BMC Psychiatry (2022)
CHD7 regulates bone-fat balance by suppressing PPAR-γ signaling
- Caojie Liu
- Qiuchan Xiong
- Ning Kang
Nature Communications (2022)
Altered gene expression profiles impair the nervous system development in individuals with 15q13.3 microdeletion
- Marek B. Körner
- Akhil Velluva
- Diana Le Duc
Scientific Reports (2022)

Comments

By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.