Chromatin Remodelers in Autism: Deciphering Regulatory Networks That Contribute to Autism Risk

Background: Recent gene discovery efforts in autism spectrum disorder (ASD) have identified chromatin modifiers, such as the chromodomain helicase CHD8, as important new contributors to ASD pathogenesis.  These and additional ASD risk-associated genes are co-expressed in human midfetal cortex, suggesting that ASD risk genes may converge in regulatory networks that are perturbed in ASD. To investigate the factors shaping this regulatory network, we have undertaken studies to globally map regulatory targets of ASD risk-associated chromatin modifiers during human neurodevelopment. Additionally, these regulatory networks may aid in the identification and classification of non-coding regulatory regions, which is critical for interpreting whole genome sequencing (WGS) data.

Objectives: We have found that CHD8 regulates other ASD risk genes during human neurodevelopment. These studies provide an initial view of ASD-associated regulatory networks in the human brain. We will build on these results by identifying regulatory targets of other chromatin modifiers conferring risk for ASD.

Methods: We have mapped the binding sites of CHD8 in the developing human and mouse brain using ChIP-seq, and have characterized global dysregulation of CHD8 targets following CHD8 knockdown utilizing shRNAs followed by RNAseq. We will extend these analyses to the targets of ASD risk genes bound by CHD8. Integration of these ASD risk gene target maps with each other, and with maps of specific active and/or repressive histone modifications, will identify genes and regulatory elements targeted by other ASD risk genes within the CHD8 regulatory network. These studies will be further informed by cell model systems in which ASD risk gene expression will be decreased by shRNA and/or CRISPRi. Finally, we will integrate WGS data obtained from ASD cohorts in order to assess the impact of coding and regulatory variation on ASD-associated regulatory networks.

Results: We find that CHD8 gene targets in human and mouse developing brain are significantly enriched in ASD risk genes bearing one or more de novo loss of function mutations, and that genes found in ASD risk-associated spatiotemporal co-expression networks during human brain development are more likely to be targeted by CHD8. Knockdown of CHD8 expression levels by shRNA leads to significant dysregulation of ASD risk genes. CHD8 binds the ASD risk genes CHD2, SUV420H1, ARID1B, POGZ, MLL5, ASH1L, and SETD2, among others, and preliminary results suggest that CHD2 and CHD8 targets substantially overlap. In further studies, we will characterize the gene targets of CHD2 and other ASD-associated chromatin modifiers to elucidate the regulatory networks dependent on each gene.

Conclusions: Advances in next generation sequencing technologies have greatly advanced the field of autism genetics. However, obtaining biological insights from the ever-increasing numbers of variants from exome and genomic sequencing requires unbiased, functional genomic analysis of ASD-associated genes. Mapping regulatory targets of CHD8, CHD2 and other ASD-associated chromatin modifiers in the developing human brain will identify biological pathways underlying ASD etiology. This will provide the means to interpret ASD-associated noncoding variation, and offer potential avenues for drug discovery by revealing specific mechanisms contributing to ASD.