Objectives: The objectives of this study were to assess the role of epigenetic alterations in ASD in blood, a clinically accessible tissue. DNA methylation, defined as addition of methyl group to the cytosine residues at CpG dinucleotidesis is the best studied form of epigenetic regulation, and was chosen for the genome-wide analyses.
Methods: Genome-wide DNA methylation was assessed in 12 ASD samples and 12 controls at ~27,000 CpG sites using methylation microarray. Pyrosequencing was used for targeted validation and replication.
Results: We hypothesized that, similar to genetic variants, common and rare epigenetic variants contribute to the ASD phenotype. To test these possibilities we have undertook two statistical approaches. Using Mann-Whitney U test with correction for multiple testing we have identified only one ASD-associated epigenetic common variant, a 7 % gain of DNA methylation in the promoter of the natural cytotoxicity triggering receptor 1 (NCR1) gene. To identify rare variants in ASD cases, we selected CpG sites with DNA methylation changes of at least 10% compared to controls. Using this rare variant approach we identified 7 CpGs with loss and 5 CpGs with gain in ASD vs controls. Further, in order to assess the relevance of DNA methylation changes identified in blood to brain, we have compared DNA methylation levels at 13 CpG sites referred above to 150 brain samples of neurologically normal individuals run on the same microarray platform (Gibbs et al. 2010). To narrow the list of candidates we selected genes with: 1) similar levels of DNA methylation in blood and brain in controls, and 2) no outliers in the brain similar to the outliers levels in ASD blood. Only two rare variants with gain of DNA methylstion at CpG sites within promoters of arylsulfatase A (ARSA ) and phosphatidylcholine transfer protein (PCTP) genes survived the brain comparison. The ARSA gene is located within the region of 22q13.3 deletion syndrome characterized by developmental delay, and frequent autistic features, it is also adjacent to SHANK3, a known ASD susceptibility gene. Therefore, we have further assessed DNA methylation in six CpG sites of the ARSA promoter in 66 ASD cases and 70 controls of matched (European) ethnicity using targeted pyrosequencing. We analyzed data for each CpG site individually using a Chi Square test and set the cutoff at the maximum DNAm level in controls. We observed a statistically significant increase of DNA methylation in two CpG s within the assay (CpG#4 (4 ASD cases with gain of DNA methylation) and CpG#6 (5 ASD cases with gain of DNA methylation).
Conclusions: Our data support a role for epigenetic variants in ASD etiology and the feasibility of identifying the epigenetic variants in clinically accessible tissues such as blood