Exploring Non-Coding Regulatory SNPs As Genetic Markers for Autism Spectrum Disorders

Background:  Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder of high prevalence that clearly involves genetic risk factors. However, the connection between ASD genetic risk factors and the development of ASD is poorly understood. Most current approaches to understand genetic risk factors in ASD focus on genes, although most human disease-associated single-nucleotide polymorphisms (SNPs) are located in regulatory regions which control gene expression. It remains challenging to pinpoint regulatory regions that are likely to be important for a specific disease. It is known, however, that functionally relevant regulatory regions are determined through epigenetic modifications which define where chromatin is accessible in the genome.

Objectives: The goal of this study was to evaluate whether regulatory regions obtained from epigenetic studies of mouse behavior could highlight non-coding regions with genetic links to ASD in human populations.

Methods: We used high-throughput sequencing (HTS) to uncover regulatory regions that show an increase in chromatin accessibility following contextual conditioning in the mouse hippocampus, which were disproportionally enriched around genes known to be linked to ASD according to the SFARI gene database (https://gene.sfari.org/autdb/Welcome.do). We then used genotyping to investigate whether SNPs in one of the candidate regions, located within and intron of Shank3, were significantly associated with ASD in in a cohort of 472 ASD (diagnosed by research reliable clinicians using the ADOS and ADI) and 209 typically developing children.

Results:  Using HTS we show that learning produces increases in chromatin accessibility at 3064 regulatory regions in the mouse hippocampus.  59 of the 420 autosomal ASD associated genes according to SFARI show an increase in chromatin accessibility following conditioning within their regulatory regions (promoters or introns), including an internal promoter within the Shank3 gene. Genotyping of common SNPs within this region reveals that there is a significant increase in frequency (Fisher exact test p=0.03) of the heterozygous allele in ASD cases versus controls for SNP rs6010065.

Conclusions:  Overall our results demonstrate that using epigenetic data obtained from mouse is a viable strategy to highlight functionally relevant regulatory regions for genetic association studies of ASD. Future work will include genotyping of the remaining 58 candidate regions as well as further investigation of the mechanism underlying the effect of mutations within the Shank3 internal promoter.