Evidence of a Maternally Acting Gene Allele (MAGA) for Autism in a Second Dataset in a Small Region of Chromosome 3p24.3

Background: Maternally-acting gene alleles, MAGAs, act in mothers prenatally to alter fetal environment and affect offspring phenotype, independently of any inheritance by the fetus. We have also used the terms “maternally acting alleles” and “teratogenic alleles” for these. There have been over 70 reports of MAGAs to date, mostly in neurodevelopmental disorders. Six of these reports were in autism. Evidence of MAGAs has previously been found by candidate gene approaches. Earlier, we reported the first, to our knowledge, genome-wide association study (GWAS) for MAGAs that implicated a small region of chromosome 3p24.3 for autism using existing data from 825 families in the Autism Genetic Resource Exchange (AGRE) studied on the Illumina Hap550 GWAS array. SNP rs12487874 (intronic in the RTFN1 gene) showed genome-wide significance (p=8.61E-11) with no evidence of child effect (presented at IMFAR 2012).

Objectives:  To replicate our preliminary findings with a second dataset.

Methods: We used the Weinberg log-linear method to analyze existing GWAS data from 1366 families in the Autism Genome Project (AGP) genotyped on the Illumina 1M array.

Results: The AGP dataset showed genome-wide significance (p = 1.83E-19) for SNP rs12636481, located near rs12487874 and not in the earlier dataset. SNP rs12487874 was not part of the AGP dataset. Both SNPs showed a marked asymmetry in parents in both studies, where the mothers identified were homozygotes for the minor allele.

Conclusions: Since this study confirmed a peak for autism in a small region of chromosome 3p24.3, we plan to analyze data from a much larger dataset, the Simons Simplex Collection (SSC), to confirm and extend these findings. We will also re-genotype the SNPs identified using a more stringent method to confirm these findings and to exclude possible artifacts. If these findings are confirmed by analysis in the SSC and re-genotyping, we will then test the hypothesis that a structural DNA variant in this region is responsible for the observed effect. These studies could lead to identification of DNA variation in this region whose action may contribute to autism. We hope these studies will lead to better understanding of the pathogenesis of autism, to approaches to identifying risk of autism prenatally or even before the onset of pregnancy, and perhaps to methods of preventing or treating autism at a very early stage.