Epigenetic Alterations in Autism Spectrum Disorder Following the Use of Fertility Treatments

Background:  Autism spectrum disorder (ASD) is a heterogeneous disorder both genetically and phenotypically. It has been estimated that ~25% of ASD cases can be accounted for by genetics, but the additional contributions of environment and epigenetics to the etiology of ASD have yet to be defined. One possible environmental exposure that may impact epigenetic regulation is the use of fertility treatments (FT). To date, FT, including assisted reproductive technologies (ART) and hormonal stimulation of ovulation, contribute to 6% of live-births in the United States annually. FT is associated with increased risks for multiple births, prematurity and low birth weight. Additionally, concerns have been raised about the potential for increased risks for birth defects and neurodevelopmental disorders. DNA methylation (DNAm) alterations are enhanced following FT in etiologically heterogeneous disorders such as Beckwith-Wiedemann and Angelman syndromes. The evidence for a similar effect in ASD is inconsistent. 

Objectives:  We hypothesize that epigenetic alterations contribute to the molecular etiology of ASD. This work also explores whether FT contributes to the risk of ASD and associated epigenetic modifications. 

Methods:  We examined DNAm variants (DMVs) in five groups of samples: whole blood from patients diagnosed with ASD and conceived by FT (ASD-FT) (n=27), patients diagnosed with ASD and conceived normally (ASD-NC) (n=21), and unaffected controls conceived normally (n=50). As an approximate control for FT alone, we also compared our results to a set of cord blood samples collected at birth from patients conceived by FT (n=22), with matched cord blood controls conceived normally (n=22). Whole blood was assessed for DNAm using sodium bisulfite converted DNA followed by the Illumina Infinium HumanMethylation450 BeadChip array, which interrogates >485,000 CpG sites at single nucleotide resolution. The top hyper- and hypomethylated regions will be validated independently by sodium bisulfite conversion followed by pyrosequencing. We employed a novel bioinformatics approach to analyze genome-wide DNAm data that is appreciably more successful in generating statistically significant epigenetic signals in subgroups of ASD cases. 

Results:  By initially grouping all ASD patients together (ASD-FT + ASD-NC), we identified >3000 DMVs when compared with controls (adjusted p < 0.05) with effect sizes of up to 14% difference. When subgroups based on DNAm and variance were identified, these signatures were even stronger at more stringent significance levels (adjusted p < 0.0001) and with greater effect sizes (8-28% difference). Within the subgroups, there appears to be a trend towards a higher frequency of use of FT vs. in the remaining ASD patients not placed into subgroups (“Other ASD”) (p > 0.05). We are now further delineating phenotypic and genomic associations between individuals in these subgroups, examining the functional relevance of the DMVs and analyzing cord blood DMVs. 

Conclusions:  We have identified DMVs that differentiate between ASD subgroups and controls. These findings support the role of epigenetics in the molecular etiology of ASD and we expect environmental and clinical patient stratification may ascertain specific risk associations. Such associations with epigenetic patterns could potentially be utilized as a diagnostic tool, allowing for earlier detection of ASD in subsets of patients.