25013
Whole-Genome Methylation Screen Identifies Enriched Neuronal Pathways in a South African Autism Cohort
An increasing number of studies support a role for epigenetic factors in the aetiology of autism. Recent work on monozygotic twin pairs, discordant for autism spectrum disorder (ASD) phenotype, implicated DNA methylation as a contributor to autism. Despite intense research over the past decade, no definitive biological markers for this disorder have been identified given the highly heterogeneous nature of ASD. This highlights the importance of a clearly defined ASD phenotype in molecular research.
Objectives:
We investigate whole-genome methylation patterns in a cohort of South African children. Differentially methylated regions are used to identify biologically relevant pathways. We validate selected genes within these pathways using quantitative methylation assays in both the discovery and replication cohorts.
Methods:
South African boys aged between 6-11 years old (n=32 ASD and n=16 controls) were recruited. All participants were assessed using the Autism Diagnostic Observation Schedule-2, ADOS-2. DNA was extracted from study participants using buccal cells. We performed a whole genome DNA methylation screen using Illumina 450K Human Methylation Array. Differentially methylated loci associated with ASD were used in gene ontology pathway analysis to investigate significant biological pathways. Selected genes from these pathways, both novel and previously associated with ASD, were validated using a quantitative mass spectrophotometry assay(Epityper®).
Results:
We identify over 5000 differentially methylated CPG sites (corrected FDR<0.5) associated with ASD. These sites are distributed across ~4000 different genes and UTRs. Gene ontology analysis identified Nervous System Development and System Development biological processes as being significantly enriched. When comparing ASD endophenotypes, with find significant associations with smaller gene sets. For example, children with severe Repetitive and Restricted behaviours, RRB (ADOS-2 score >8) show significant association for a smaller set of CpG sites across less than 100 genes, and thus we recover more targeted biological processes. We validate differential methylation quantitatively using mass spectrophotometry for two genes: a gene previously associated with ASD (SLC6A4) and a novel signalling gene. The quantitative data supports a role for differential methylation in ASD individuals with high social affect, SA (ADOS-2 SA>8) at both loci.
Conclusions:
This study is the first to investigate whole-epigenome profiles of DNA methylation in a South African cohort of children with ASD. We find that methylation patterns differ significantly between ASD and typically developing children at a large number of genes involved in biologically relevant processes, particularly in nervous system development. When looking at ASD subtypes, we are able to isolate smaller groups of CpG sites that display differential methylation, thus reducing the number of genes and processes involved. We have validated that differential methylation in signaling pathways is associated with ASD subtypes, which is consistent with the central role of neuronal signaling in neuropsychiatric disorders. Our approach of defining ASD subtypes in the search for molecular epigenetic markers is central to the identification of potential ASD biomarkers and effective drug targets.