microRNAs and Gene-Environment Interactions in Autism: Prenatal Maternal Stress and the SERT Gene

Background: Both genetic and environmental factors may play critical roles in autism. However, there is less understanding of the potential environmental factors or gene X environment interactions (GxE). Recent evidence suggests that maternal stress exposure may be important in autism. By surveying for history and timing of prenatal psychosocial stressors, we found autism mothers to have a higher overall incidence of stressors as compared to the other groups, which was subsequently confirmed by larger epidemiological studies. In all of these studies, though, a significant proportion of stress-exposed mothers had unaffected children. To explain why prenatal stressors might result in autism in some cases but not others, we began to explore GxE. The serotonin transporter (SERT) gene is well studied for its role in stress reactivity. The most widely studied variation is an insertion/deletion within the promoter region, resulting in long (L) or short (S) alleles. The S-allele has been linked to autism in some but not all studies. These contradictory findings could, in part, be explained by the presence of a gene/stress interaction. We have subsequently demonstrated in two independent samples that the association between maternal stress exposure and autism is greatest with maternal presence of the S-allele. There is emerging evidence that microRNAs may play a regulatory role in the serotonergic pathway and in prenatal stress.

Objectives: To examine the microRNA profile in this GxE setting in maternal blood samples.

Methods: In the present study, we explored the role of epigenetic factors by profiling microRNA expression in blood samples (n=34) from mothers of children with autism, with known pregnancy stress history. The samples were divided into 5 groups based on SERT genotypes (LL, LS, and SS) and prenatal stress level (High and Low).

Results: Among the 2500 mature microRNAs examined, 5-way ANOVA showed significantly differential expression (DE) of 119 microRNAs, 90 (76%) of which showed a similar pattern of expression in High vs Low stress groups (i.e., stress-dependent microRNAs). Intriguingly, two of them, miR-1224-5p and miR-331-3p, were recently reported by our group to exhibit stress-dependent expression in rodent brain samples from embryos exposed to prenatal stress. Another stress-dependent microRNA found in our study, miR-145-5p, has been reported in association with maternal stress. To assess the role of SERT genotype, we conducted a 3-way ANOVA on three SERT genotype groups exposed to a high level of prenatal stress. This analysis showed a smaller number of significantly DE microRNAs (n=20), 5 (25%) of which were among the stress-dependent DE microRNAs. These 5 microRNAs may be candidates for stress X SERT genotype interactions. One of them, miR-663a, has been previously reported to be DE in response to the serotonin-specific reuptake inhibitor, fluoxetine. These findings are remarkable as these changes were detected in samples from mothers several years after stress-exposed pregnancies. Persistent microRNA changes have been observed previously in other conditions, such as after cessation of smoking.

Conclusions: Our study provides evidence for epigenetic alterations in relation to a promising GxE model (prenatal maternal stress X the SERT gene) in autism.